Search

Establishing a unified global height datum is one of the core objectives of the international geo‐ detic scientific community, and also a necessary infrastructure for geoscience research on a global scale and cross-border engineering applications. The international geodesy association (IAG) released the definition of international height reference system (IHRS) in 2015 and then in 2019 set out the goal of establishing the international height reference frame (IHRF). In this paper, the theory, methods and practical problems re‐ lated to the IHRS and IHRF are reviewed and studied on the basis of the theoretical foundation and defini‐ tion of the global height reference system. The main research contents include the determination of gravity potential W0 of the global geoid, the determination of gravity potential based on high degree gravity field model, and the determination of gravity potential based on regional gravity field modeling. Two typical in‐ vestigation cases are Colorado geoid model experiment organized by IAG and the realization of IHRS in 2020 Qomolangma height survey. The results of two typical cases demonstrate that the accuracy of gravi‐ metric geoid model can reach 1 cm (0.1 m² /s² in gravity potential) in flat areas and general mountainous areas, and it is expected to reach 2− 3 cm accuracy (0.2− 0.3 m² /s² in gravity potential) even in very complex mountainous areas such as Mount Qomolangma. Based on the results of two case studies, a preliminary strategy for establishing the IHRF is proposed, including layout plan of IHRF reference stations, determination methods of IHRF related gravity potential, data requirements, standards/conventions to be followed and expected accuracy. In addition, we prospect the potential contributions of optical atomic clocks and relativistic geodesy to the unification of global height datum. [ABSTRACT FROM AUTHOR]

Digital development has been vigorously promoted in China to drive a comprehensive and indepth change in production, life and governance through the construction of digital economy, digital society and digital government. It has put forward new and higher requirements for basic surveying and mapping products and services. Recently, the Ministry of Natural Resources of the People􀆳s Republic of China issued a notice to promote the construction of 3D realistic geospatial scene of China, aiming to develop new 3D geo-spatial information products which can describe and represent the real 3D spaces of industrial production, living environment and ecological development. The construction of such 3D realistic geospatial scene of China will lead to a 3D digital space which can be well connected with the real geo-spatial space through real-time inter-connection, and will provide a new generation of spatiotemporal information framework for the development of digital China. It is not only a major task for national and local basic surveying and map‐ ping agencies in the next few decades, but also a challenging scientific and technological initiative where several complex factors and technical problems remain to be solved. This paper analyzes the background and problems, discusses the development directions and general technical logic, and conductes the major technological path. [ABSTRACT FROM AUTHOR]

On September 16th, 2021, general secretary Xi Jinping pointed out in his congratulatory letter to the first international summit on BeiDou scale application: At present, the development of global digitalization is accelerating, and spatiotemporal information, location and navigation services have become an important new infrastructure. With the advent of the new infrastructure era, people engaged in surveying and mapping should think about what the country wants, be anxious about what the country is anxious about, and devote themselves to the wave of “the second 100 years” and “new infrastructure”. Firstly, this paper expounds the definition of new infrastructure, and analyzes the difference between new infrastructure and traditional infrastructure. Secondly, it discusses the mission of geo ‐ spatial information science in the three systems of new infrastructure, including information infrastructure, integration infrastructure and in‐ novation infrastructure. The authors believed that we have walked through the so called traditional surveying and mapping which mainly serves topographic maps, and developed into the so called ubiquitous surveying and mapping which mainly focuses in the services of geo ‐ spatial information. At present, we should seize the opportunity and expand the new mission of geo ‐ spatial informatics science in the new infrastructure era, provide spatiotemporal data with good integrity, strong reality and high accuracy for the new infra‐ structure, and realize digital industrialization, industry digitization and intelligence. [ABSTRACT FROM AUTHOR]

Objectives: The spatial delimitation of urban area is the foundation for monitoring the urbanization level and determining city scale from the perspective of geographical space.Methods: Based on the requirements of urban area defined in Document No. 51 of the State Council, this paper clarifies the connotations of urban area with further understandings and proposes a process for delimiting urban area in geographical space by connecting map polygons of governmental residents and neighborhood committee, which can reduce the subjectivity and fuzziness in the spatialization of process of urban area. Results: This paper develops the main idea and spatial delimitation about urban area and takes Wuhan as an example to illustrate the details about spatial delimitation of urban areas.Conclusions: This paper mainly discusses how to unify urban delineation under the current technical conditions, and does not mention how to use other auxiliary data, such as the application of multi⁃source big data in urban delineation, because of the unbalanced development of big data infrastructure in different cities. But it does not mean that the application of population big data and urban big data is not important.On the contrary, the application of big data in the work of urban area⁃ delimitation is the next research focus and the future development trend. [ABSTRACT FROM AUTHOR]

Objectives:In video⁃based multiple object tracking, the object detection and re⁃identification have a strong correlation. The existing methods generally train the object detection and re⁃identification net⁃ works separately, which makes the tracking speed fail to meet the real⁃time requirements. In this paper, we integrate the detection and re⁃identification into one network to accelerate the tracking process, and also solve the problems of identity switching and failure tracking. Methods:This paper develops a pedestrian mo⁃ tion model and obtains the optimal state estimation of pedestrians using center point detection. The person reidentification model with deep layer features uses the Mahalanobis distance and cosine distance to enhance the ability of person identification. And the Hungary algorithm is used for data online association, where the state estimation results become more accurate using Kalman filtering, and the unrelated lost objects are predicted by motion. Results: Experiments are conducted on MOT15, MOT16 and MOT17 data⁃ sets using the proposed algorithm and other multi⁃pedestrian tracking algorithms， and the multiple object tracking accuracy of tracking results using our proposed algorithm is 63.5, 72.4 and 70.9, respectively, and the identity F1⁃measure is optimal, with the real⁃time rate. Conclusions: The proposed algorithm can accelerate the tracking speed by network parameter sharing, and improve the recognition accuracy by per⁃ son re⁃identification training. [ABSTRACT FROM AUTHOR]

Objectives：With the acceleration of urbanization and the rapid expansion of urban population, the population data on the street scale plays an important role in urban economic, social, resource and environmental development. Therefore, estimating population on urban street scales with high‐resolution remote sensing images is of great theoretical and practical value for promoting sustainable urban development. Methods：This paper uses the building information in remote sensing images and census data to analyze the correlation between street building information and population data, and the city street scale population estimation models based on building information are proposed. The best estimation model for the number and geometric characteristics of buildings and population is established based on the multivariate stepwise regression method and the Akaike information criterion that determine the significant feature variables of the building.Results：The Experiments show that the population estimation model proposed in this paper can estimate the population of street scale with high accuracy. On the one hand, the floor data can be directly applied to the estimation model. On the other hand, the floor data is also the basis for classifying street buildings. Dividing street buildings into low‐rise, middle‐rise and high‐rise buildings according to the number of floors can significantly improve the population estimation effect of the model. Conclusions：The population estimation model proposed in this paper can estimate the population of street scale with high accuracy. This paper uses street‐scale data in administrative divisions, but the estimation range of the proposed model is theoretically also valid in population estimation across street areas, which needs to be verified in subsequent studies. [ABSTRACT FROM AUTHOR]

Objectives: The Chinese Mars probe has been successfully launched and is currently orbiting Mars. This article analyzes several different satellite orbit characteristics in detail to find out the suitable orbits with stable operation characteristics. Methods: This paper uses the MAGREAS （Mars gravity recovery and analysis software/system） independently developed by Wuhan University to analyze the evolution of satellite orbits under different initial orbit heights, different initial orbital inclinations, different initial orbital eccentricities, and different orders of Martian gravitational field models. Then we make orbit predictions for the four situations.Results: The results show that satellite can run for nearly two years without any orbital adjustment. When the initial orbital inclination is 90°, it does not contribute to the running time of the satellite. In the calculation of orbital integration, the Martian gravity field model can be intercepted to 100 degrees without affecting the calculation accuracy.Conclusions: This paper has done a lot of orbit predictions based on different initial orbit elements, and analyzed the evolution characteristics of different satellite orbits. This research can provide a certain reference for China's Mars exploration project. [ABSTRACT FROM AUTHOR]

：Objectives:With the development of urbanization, people's travel behaviors have diversified. An in‐depth understanding of human behavior and the modeling and prediction of individual travel behaviors are helpful in explaining several complex socio ‐ economic phenomena, and are important in offering location ‐ based services, transportation planning, and public safety. Individual travel behavior prediction is based on a deep understanding of human activity characteristics. In the era of mobile Internet, the online behavior of cyberspace is inseparable from the travel behavior of real space. Methods: This paper integrates individuals' mobile phone tracking data and Internet traffic data, and constructs a multi‐model fusion model of individual travel location prediction on Markov and machine learning methods. Considering the classification probability of prediction results, an adaptive fusion strategy based on frequency distribution graph is proposed. The prediction results of Markov model and machine learning multi‐classification model are merged together to obtain the final mobile phone user travel location prediction result. Results:This paper performs individual travel location prediction experiments on the basis of multi ‐ source data. And the experiments show that the correct rate of the first result and the top three results of the multi‐model fusion location prediction model based on histogram is respectively 74.59% and 94.19%, higher than the prediction accuracy of the basic model with the highest accuracy and the vote strategy.Conclusions: Under the prediction time granularity of 30 minutes, the individual travel location prediction is better. [ABSTRACT FROM AUTHOR]

Objectives: Aiming at the mixed observation data of multiple distribution forms, a expectation-maximum (EM) combined p-norm distributed model(EM_p) is established.Methods: Considering that the mixed number in the mixture model belongs to incomplete data, the EM algorithm is introduced to estimate the parameters of the mixture model and the p-model mixture model parameters are derived in detail. The estimated iteration formula and the corresponding iteration steps are given.The mixture Gaussian distribution data, Laplace distribution and Gaussian distribution mixture data, and the residual data of measured global positioning system(GPS) observations are used to verify the correctness and adaptability of the formula in this paper. Results and Conclusions: The results of the calculation examples show that, compared with the single probability distribution, the p-norm mixture model can accurately reflect the actual situation of the data distribution, and the model parameters estimated by the EM algorithm have higher accuracy. [ABSTRACT FROM AUTHOR]

Objectives: In surveying and cartography, solving geodetic latitude is a common reverse calcula⁃ tion by conformal latitude. Methods: Several different algorithms of reverse solution are studied, including iterative method, series expansion of solving geodetic latitude by difference of conformal latitude and direct algorithm of inverse solution of geodetic latitude with conformal latitude. The calculation formulas of the lat⁃ ter two algorithms are derived by Mathematica. The higher-order coefficient expansion formula are given. At the same time, problems in existing algorithms are analyzed. Results and Conclusions: Examples re⁃ sults of Lambert conic conformal projection show that the accuracy of the formula derived in this paper can reaches (1×10−7)″-(1×10−8)″. It fully meets the requirements of high precision in surveying and map projection. [ABSTRACT FROM AUTHOR]

Objectives: It is difficult to effectively evaluate and weaken the impact of observation environment change on data quality. To solve the problem, this paper studies a multi-point hemispherical grid model which can be applied to various satellite navigation systems.Methods: It integrates the observations of different global navigation satellite system (GNSS) satellite pairs in different observation periods to establish a global optimal multipath model. GNSS multipath error can be estimated effectively using multi-point hemispherical grid model (MHGM), image-based inverse of multipath effect and reflect the orientation of interference sources in the observation environment around stations. Results: To a certain extent the observation environment around stations can be modelled. The test under deliberate high multipath environment reveals that using MHGM significantly improved the root mean square error (RMSE) of the double-differenced carrier phase observation residuals in the period of fixed ambiguity, reducing the mean RMSE from 0.983 cm to 0.318 cm, a 67.7% improvement. In addition, the abnormal areas with large value of correction in the MHGM results is consistent with the direction of the baffle mounted at the corresponding stations. This model indicates multipath effects on GNSS carrier phase observations in different directions from the station. The test with International GNSS Service historical observations shows MHGM can effectively reflect the influence of changing multipath interference around stations on carrier phase observations, with an average improvement of 29.5% in the RMSE of residuals over the past 18 years. Conclusions: This model can be used to evaluate the changes in observation environments surrounding stations over time. It has potential application value to improve the accuracy and reliability of augmentation system products. [ABSTRACT FROM AUTHOR]

Objectives: Existing population spatialization methods mainly use administrative-unit-level data to train regression model, and transfer it to grid cell-level to achieve population allocation. However, the significant scale difference between the analytical units in training and estimation leads to the issues of crossscale model transfer. Meanwhile, only the attributes of current cell are considered in cell-level feature modeling, which causes the innate spatial association between cells to be eliminated and cells to be isolated. Methods: This paper proposes a novel population spatialization based on random forest by considering pixellevel attribute grading and spatial association (PAG-SA). In the cell-level feature modeling, we firstly construct the night light grading features embedded with building category constraints based on natural breaks, and count the grid proportion of each grading level at the administrative-unit-level as the training input to reduce the cross scale error; secondly, the influence and distance attenuation of neighborhood point of interests (POIs) upon the current cell is modelled by using kernel density estimation; thirdly, based on overlay analysis, the numbers of POIs in the contours of different building types are counted to improve the precision of feature modeling. Results: To verify the effectiveness of the proposed method, we selected Wuhan city as the experimental area and compared its spatialization accuracy with the datasets of WorldPop, GPW and PopulationGrid_China at street scale. The results show that the mean absolute error of PAG⁃SA is only 1/6-1/3 of the comparison datasets. In addition, the influence of feature composition, grid size and kernel density bandwidth on the accuracy is also discussed. Conclusions: By fusing multi source data and considering pixel⁃level attribute grading and spatial association, the proposed method PAG⁃SA is effective for achieving population spatialization in urban areas with finer grid sizes and higher accuracy. It can also provide references for spatialization applications of other geographic attributes that also face with scale mismatch issue in spatial regression modeling. [ABSTRACT FROM AUTHOR]

Objectives: Multi-source image matching is primarily disturbed by nonlinear intensity difference, contrast difference and inconspicuous regional structure features, while the significant differences of texture features result in lack of part structure seriously between light detection and ranging (LiDAR) depth map and aerial image, and this problem causes a mutation in the phase extremum, which further increases the difficulty of matching. Methods: In this paper, a method of efficient matching of LiDAR depth map and aerial image based on phase mean convolution is proposed. In the image feature matching stage, a histogram of phase mean energy convolution (HPMEC) is established, which extended the phase consistency model in order to solve a mean convolution sequence and phase maximum label map by constructing phase mean energy convolution equation. Then the nearest neighbor matching algorithm was completed the initial match and marginalizing sample consensus plus was used to remove outliers. Based on the thread pool parallel strategy, the images were matched by dividing the overlapping grid. Multiple sets of LiDAR depth map and aerial image with different types of ground coverage are used to as dataset to experiment with position scale orientation-scale invariant feature transform (PSO-SIFT), Log-Gabor histogram descriptor (LGHD), radiation-variation insensitive feature transform (RIFT) and histogram of absolute phase consistency gradients (HAPCG) methods respectively. Results: The results show that the performance of HPMEC method is superior to the other four methods in the matching of LiDAR depth map and aerial image, the average running time is 13.3 times of PSO-SIFT, 10.9 times of LGHD, 10.4 times of HAPCG and 7.0 times of RIFT, at the same time the average correct matching points are significantly higher than the other four methods, the root mean square error is lightly better than the other four methods within 1.9 pixels. Conclusions: The proposed HPMEC method could achieve efficient and robust matching between LiDAR depth map and aerial image. [ABSTRACT FROM AUTHOR]

Objectives: Mars is the main target object for deep space exploration. Mars rovers, or roving probes, are important tools for surface exploration and scientific research on Mars. For the growing amount of remote sensing data collected by Mars rovers, there is an urgent need for a method that can intelligently detect novel targets of scientific value from the massive amount of images, reduce the time cost of detection planning, and provide information for subsequent scientific analysis. The traditional novel detection methods mostly include distance-based metrics and image-based reconstruction methods, distance-based metrics calculate novel scores pixel by pixel without considering spatial contextual information, and image-based re‐ construction methods focus on reconstruction of typical landscape backgrounds, and novelty is manifested by image reconstruction errors, which is not effective in extracting small novel targets such as boreholes and dust removal points. Methods: To address the above problems of traditional novel detection methods in Mars rover novel target detection, this paper proposes an improved Mars rover multispectral image depth novel target detection method, uses full convolutional self-coding neural network to extract deep features for typical landscape reconstruction, and joints Mahalanobis distance for novel target and typical landscape background separation, fully exploits the spatial and spectral dimensional features to improve the accuracy of Mars rover novel target detection results. Results: The experiments use the multispectral image dataset of Curiosity rover released by NASA ( National Aeronautics and Space Administration ), and the proposed convolution auto-encoder combined Mahalanobis distance method(CAE-M) is compared with Reed-Xiaoli detector, principal component analysis, convolution auto-encoder convolution, and generative adversarial networks on the surface of Gale crater. The results show that CAE-M outperforms previous detection methods in terms of detection accuracy and visual interpretation, and has a balanced and stable performance in different classes of novel target detection. Conclusion: The proposed CAE-M method comprehensively utilizes spatial and spectral information of multispectral images, which can help Mars rover exploration missions to quickly and intelligently screen and sort novel data with scientific value in massive data, save the time and cost of route planning, improve scientific returns. [ABSTRACT FROM AUTHOR]

Objectives: Few corresponding points can easily affect the calculation of image pose information, increase the difficulty of constructing a regional network in an aerial triangulation solution, so that lead to problems such as image stitching misalignment, incorrect bundle adjustment results or even failure. In order to better complete the matching of unmanned aerial vehicle (UAV) images, this paper proposes a robust UAV image matching algorithm considering log-polar description and position scale distance. Methods: Firstly, a Gaussian multi-scale image collection is established and feature points are extracted. Secondly, the descriptors are constructed using log-polar coordinates, and a descriptor suitable for UAV image characteristics is established. Then, the feature matching is performed by the distance function of position and scale constraints. Finally, the mode seeking and fast sample consensus method are used to eliminate the outliner and complete the extraction of correspondence. Results: The image obtained by four-rotor UAV is used as the data source, and a comparison experiment of image matching with scale invariant feature transform (SIFT) algorithm and synthetic aperture radar-scale invariant feature transform (SARSIFT) algorithm is carried out. The experimental results show that a 210-dimensional log-polar coordinate descriptor is constructed through the gradient location and orientation histogram. The descriptor can better describe the feature points in 10 directions through the circular neighborhood, making the matching results more robust. The position scale Euclidean distance matching function established by integrating factors such as position and scale can better calculate the UAV image matching relationship, and match more correct corresponding points. In terms of the number of correct corresponding points extracted under the same parameter settings, the proposed algorithm is significantly more than the other two algorithms, and in terms of the root mean square error of the matching results, the algorithm in this article is also significantly better than the two compared algorithms. Conclusions: The proposed algorithm can better extract the corresponding points of UAV images. [ABSTRACT FROM AUTHOR]

Objectives: On 8th January 2022, a large earthquake (Mw 6.7) struck Menyuan County, Qinghai, China, causing serious damage to Lanzhou-Xinjiang high speed railway and forcing the closure of the railway for repairs, which has attracted highly domestic and international attention. Methods: We presented a technical framework to determine earthquake surface ruptures by integrating optical, synthetic aperture radar (SAR) and unmanned aerial vehicle (UAV) images as well as light detection and ranging (LiDAR) data, and evaluated its damage to traffic networks. Firstly, we acquired a range of datasets including GF-1, GF-7, Sentinel-2 optical images and Sentinel-1A SAR images. GF-1 and GF-7 images were used to determine the spatial distribution characteristic of the surface ruptures. Secondly, we employed to estimate 2D surface displacement fields using optical pixel offset technique. One in the east-west (EW) direction and the other in the south-north (SN) direction. SAR pixel offset technique was utilized to acquire surface displacements in the range and azimuth directions whilst interferometric SAR(InSAR) was mainly for surface displacements in the radar line of sight (i. e. the range direction). Structure from motion (SfM) was used to process UAV images to obtain high precision digital surface models (DSMs). Finally, all the abovementioned information was used to precisely determine the spatial distribution and surface displacement characteristics of the earthquake surface ruptures. Results: Our results show that the maximum surface displacement in the EW direction was about 2.0 m, the maximum in the range direction was approximately 1.5 m, and the total length of the surface ruptures was around 36.22 km. Furthermore, we performed an assessment of traffic inefficiency in Menyuan and its surrounding areas based on the distribution of historical geohazards as well as the earthquake surface ruptures using machine learning methods support vector ma‐ chine models. Conclusions: The Menyuan earthquake had the greatest impacts on highways, and the least impacts on rural roads. The southeast sections of the major highways G0611 and G338 had high risks. The technical framework demonstrated in this paper appears to be promising to precisely map surface ruptures, which in turn will directly benefit to earthquake disaster reduction. [ABSTRACT FROM AUTHOR]

Objectives: The rapid development of information and communication technology has facilitated the online tourism service and massive web text, which provides a new opportunity for tourism sector planning and personalized recommendation. However, owing to the characteristics of semantic vagueness and low signal‐to‐noise ratio, the web text is difficult to get utilized directly. Therefore, how to integrate the technologies of knowledge engineering, natural language processing and machine learning, so as to form a formalized domain knowledge graph from abundant tourism text, has attracted much attention. Methods: This paper proposes a tourism knowledge graph construction method based on tourism domain ontology and transfer learning. Firstly, the ontology of tourist attractions is defined based on the domain specifications and standards, which support a comprehensive and systematic description of the semantic characteristics of attractions. Secondly, a transfer learning method is adopted to transform the pre-training language model into a customized knowledge extractor to acquire knowledge triples accurately from web text, which is integrated with the scattered tourism-related information including tourist check-ins and POI (point of interest) attributes to build a systematic knowledge graph. Results: Experimental results show that the proposed knowledge extractor improves the accuracy (average area under the curve) and integrity (the number of sematic characteristics) of acquisition of sematic knowledge by 50.7% and 670%, respectively, compared with the common LDA (latent Dirichlet allocation) model. The constructed knowledge graph of tourist attractions contained 77 039 entities, 16 types of relationship, and total 10 971 810 triples. Conclusions: Through the unified organization paradigm of triplet knowledge, the study realizes the fusion and integration of multi-source heterogeneous tourism data, and addresses the potential systemic risk in the decision-making process based on a single data source. It is argued that the constructed knowledge graph can fully capture the real tourism scene, support in-depth analysis of tourist behaviors and demands at different scales and granularities, and provide decision support for sustainable developments of tourist destinations. [ABSTRACT FROM AUTHOR]

Objectives: Intelligent interpretation based on high-resolution remote sensing is an important way to realize the fine generation and rapid update of geographic information. Based on the background of fine-accurate geographical application using remote sensing, this paper analyzes the limitations of remote sensing information products in production and application. We explain the necessity of remote sensing serving geographical research and the key of coupling the fine shape of the map and the accurate content of the spectrum. Methods: Based on the basic understanding that geography guides the research of intelligent remote sensing, we first put forward the development direction and technical ideas of high-resolution remote sensing geoscience analysis based on fine geographical scenes. Then, we propose an intelligent computing mode via the space-time/satellite-ground collaboration. Furthermore, we take the evaluation of rocky desertification cultivated land in Guanling County, Guizhou Province, China as an application case. Results: Through this case study, three basic models, namely zoning-stratified perception, spatiotemporal synergistically inversion and multi-granular decision-making, are used to show how to carry out fine-accurate application by high-resolution remote sensing collaborative computing in complex mountain areas from a comprehensive perspective of refinement, quantification and topicalization. Conclusions: Combined with the previous work experience and practical cognition, several key scientific problems are discussed, such as spatial expression using irregular grids, multimodal reconstruction of temporal features, multi-source un‐ certainty analysis and iterative optimization guided by uncertainty. We give some potential study directions and research ideas in order to establish a more complete and feasible theoretical system and explore the development paths for the research on intelligent remote sensing under the guidance of geography and the fine-accurate geographical application with remote sensing data. [ABSTRACT FROM AUTHOR]

Objectives: The scale of buildings and their distribution is key indicators to measure the economic and social development of a region. Therefore, it is significant to study the extraction of buildings based on remote sensing images. Existing neural network methods still have shortcomings in the completeness of building extraction and the accuracy of building edges. To solve the above problems, this paper proposes a multi-level feature fusion network (MFFNet) based on high-resolution images. Methods: Firstly, we use edge detection operators to improve the ability of the network to recognize the boundaries of buildings. Secondly, we use a multi-path convolution fusion module to extract building features from multiple dimensions, and introduce a large receptive field convolution module to break through feature extraction. The process is limited by the size of the receptive field. After fusing the extracted features, the convolutional attention module is used to compress them, and the global features are further mined by pyramid pooling, so as to achieve high-precision extraction of buildings. Results: The current mainstream UNet, pyramid scene parsing network (PSPNet),multi attending path neural network (MAPNet) and multiscale-feature fusion deep neural networks with dilated convolution (MDNNet) are used as the comparison methods, and we use Wuhan University Aerial Image Dataset, Satellite Dataset II (East Asia) and Inria Aerial Image Dataset as experimental data for testing. Compared with the other four methods, MFFNet improves intersection over union, precision, recall, F1-score and mean average precision by 1.53%, 2.65%, 2.41%, 3.32% and 1.19% on average, achieves a better effect. Conclusions: MFFNet not only accurately captures the detail features of buildings, but also strengthens the extraction and utilization of global features. It has better ex‐ traction effect on large buildings and buildings in complex environment. [ABSTRACT FROM AUTHOR]

Objectives: In precise point positioning (PPP) technology, the correct fixing of ambiguity is al⁃ ways an essential part of achieving high-precision positioning. However, the wrong fixing of ambiguity will lead to serious positioning errors. Therefore, it is necessary to optimize the selection of the ambiguity sub⁃ set to ensure a more reliable fixing of PPP ambiguity. Methods: This paper proposes a PPP partial ambiguity resolution (PAR) method that combines quality control and Gram-Schmidt orthogonalization. By utilizing the multi-GNSS(global navigation satellite system) experiment (MGEX) observations in the un-differenced and un-combined PPP model, the PAR method based on Gram-Schmidt orthogonalization was com⁃ pared and analyzed with the selected reference satellite method with maximum elevation, and the performance of ambiguity resolution and positioning was assessed under the multi-GNSS conditions. Results: Experimental results show that the average epoch fixing rate of each day and each station based on the Gram Schmidt method is improved by 7.74% and 11.46% in the static mode, and 7.90% and 7.78% in the pseudo-kinematic mode, respectively, compared with that of the selected reference satellite method considering the maximum elevation. Meanwhile, time-to-first-fix (TTFF) of each day and each station based on the Gram-Schmidt orthogonalization can be significantly reduced by 22.30% and 25.42% in the static mode, corresponding to an improvement of 20.44% and 19.65% in the pseudo-kinematic mode, respectively, compared with that of the selected reference satellite method considering the maximum elevation. As far as multi-GNSS PPP AR is concerned, multi-GNSS PPP AR has a significant improvement on the convergence time of BDS (BeiDou navigation satellite system) only. The convergence time of horizontal components is shortened by 20.00 min on average at 95% confidence level, and the vertical components are shortened by 19.00 min on average. For the positioning accuracy aspects, the positioning accuracy of horizontal components is improved by 1.50 cm on average, while that of vertical components is improved by 1.12 cm on average. Conclusions: The PPP PAR algorithm based on the Gram-Schmidt method can significantly improve the performance of ambiguity resolution and positioning accuracy. [ABSTRACT FROM AUTHOR]

Objectives: On July 28, 1976, a strong Ms 7.8 earthquake sequence occurred in Tangshan, and the influence of this earthquake continues to now. Methods: In order to study the deformation characteristics and seismogenic fault activities of Tangshan Ms 7.8 earthquake in 1976, we calculate quantitative results and error distribution of three-dimensional deformation before and after the 1976 Tangshan earthquake by dynamic adjustment method. And the coseismic slip distribution on the fault plane is also inversed in this paper. Results: The deformation results show that before the 1976 Tangshan earthquake, an overall uplift area appeares near the epicenter and the peak area of uplift which is located near Tangshan, the horizontal deformation of the epicenter area is small, and the Tangshan rhombic block rotates counterclockwise relative to the surrounding block. In the coseismic period,the Tangshan fault is dominated by right-lateral strike slip, with the vertical deformation characteristics of rising in the west and falling in the east. The inversion results show that multi-segment faults participate in the energy release of Tangshan Ms 7.8 earthquake. The distribution of seismic slip is mainly distributed in 6 - 18 km depth and 50 km width in the southern segment, and 7-17 km depth and 50 km width in the northern segment.The fault of Luanxian earthquake also experiences little sliding. The maximum horizontal dislocation is 7.82 m, the maximum vertical dislocation is 2.04 m, and the equivalent magnitude is Mw 7.58.Conclusions: The results of three-dimensional deformation field before and after the Tangshan earthquake provide valuable data for the in-depth study of the preparation and occurrence mechanism of the Tangshan earthquake and its impact. [ABSTRACT FROM AUTHOR]

Objectives: Revealing the characteristics of deformation field related to earthquake preparation process is an important way to distinguish earthquake preparation stage and earthquake risk. In order to reveal the significance of crustal deformation observation, the deformation field related to the Wenchuan Ms 8.0 earthquake preparation and the relationship between the deformation field and the seismic activity are discussed from a large spatial scale. Methods: The original data of crustal tilt observed at Wenchuan station in the epicenter area of the Wenchuan earthquake since 2002 are collected, and the phased characteristics of crustal tilt change are extracted by using wavelet analysis method and piecewise fitting method. Based on the daily solution data of global navigation satellite system (GNSS) station coordinates since 2001, the baselines between Luzhou station and the surrounding stations are calculated. By comparing the baseline change with the Benioff creep, the synergistic characteristics between them are found. Results: The tilt rate in EW direction at Wenchuan station began to decrease significantly at the end of 2004, from 230 ms/a to 130 ms/a, and it decreased to 60 ms/a until the earthquake occurrence. There was no significant exponential acceleration change from the end of 2007 to the earthquake occurrence. The rapidly shortening anomaly of Guangzhou-Luzhou GNSS baseline in early 2005 has a high correlation with the earthquake creep loss anomaly, and the correlation coefficient is 0.89. The duration of rapidly shortening is about 2.5 years, and the magnitude of the corresponding earthquake is estimated to be 7.9, which is close to the magnitude of the Wenchuan earthquake. The Sumatra Ms 8.9 earthquake in 2004 has a significant impact on the stress-strain field in the eastern Tibet Plateau and accelerates the preparation process of the Wenchuan earthquake. Conclusions: The different influence of surrounding earthquakes should be considered in the analysis of crustal deformation and earthquake risk. The research results of this paper have reference value for understanding the significance of crustal deformation measurements and innovating earthquake prediction ideas. [ABSTRACT FROM AUTHOR]

Objectives: Manual seismic phase picking and earthquake catalog building become challenging because the number of seismic stations increases. With the rapid development of artificial intelligence, several deep learning models are developed for seismic phase picking, but their performance and applicability remain ambiguous. Methods: Phase-Net and EqT can extract relevant features at different levels from raw seismic waveforms and map them into probabilistic sequences. In this paper, the performance of these two widely-used deep learning models for earthquake detection and phase picking is investigated with the 2019 Ridgecrest earthquake (California, USA) sequence and the 2019 Weiyuan Ms 5.4 earthquake (Sichuan, China) sequence. Besides, the completeness and accuracy of the constructed earthquake catalogs are analyzed and evaluated by cross-validation. Results: The results show that the two models, though trained with datasets from different regions, perform well in cross-region applications and are comparable to manual analysis. EqT1 provides high detection precision with few false alarms, even under a low threshold. But EqT1 has a limited recall rate, implying that fewer earthquakes are detected compared with Phase-Net. The picking performance of Phase-Net is sensitive to the employed threshold. With the low and moderate thresholds, Phase-Net can detect more earthquakes than EqT1 and manual catalog, while registering a relatively low detection precision. With the high threshold, the precision of Phase-Net is comparable to that of EqT1. EqT2 is not recommended due to its inferior performance over EqT1 and Phase-Net. Conclusions: EqT1 provides a higher precision when being employed for seismic phase picking, leading to earthquake catalogs with better accuracy. Phase-Net has a higher recall rate and generates more complete catalogs. Both models can facilitate automatic and rapid construction of earthquake catalogs. [ABSTRACT FROM AUTHOR]

Objectives: North China is an important social and economic area in China but suffers from severe surface deformation. Monitoring the surface displacement effectively and timely can provide data basis for understanding the deformation mechanism and predicting disasters. However, the surface vertical displacement in North China shows significant seasonal fluctuations. Hence, this paper aims to analyze the spatiotemporal feature of seasonal vertical displacement and determine the driving sources. Methods: Firstly, the vertical time series of 20 global navigation satellite system (GNSS) stations from January 2011 to November 2019 are decomposed by independent component analysis. Secondly, the first six independent components (ICs) are derived and compared with the vertical displacements caused by environmental loadings. Results: The results show that GNSS-IC1 presents obvious annual variation, as well as a relative uniform spatial response, which is highly consistent with the displacement time series derived from atmospheric loading, and the average correlation coefficient at all stations is 0.60. GNSS-IC2 and GNSS-IC3 are both the mixed signals constrained by annual and intra-annual periods, while the spatial responses reflect locally clustered characteristics, which corresponds to the distribution of water resources in North China. After combining GNSS-IC2 and GNSS-IC3 as GNSS-IC2/3, the average correlation coefficient between GNSS-IC2/3 and the hydrological loading displacement time series at all stations is 0.50. Conclusions: The temporal component of each IC is generally less clustered by different time cycles and the spatial response shows distinct localized characteristics. Atmospheric and hydrological loadings are the primary driving sources of seasonal vertical displacement in North China. However, the seasonal displacements induced by other effects, including non-tide ocean loading and thermal expansion effect of bedrock and observation pillar, are not extracted and need further analysis. [ABSTRACT FROM AUTHOR]

This paper briefly reviews the temporal and spatial characteristics of crustal deformation and its relationship with strong earthquakes in Chinese mainland based on geodetic observations. It also introduces the progress and understanding of extracting multi-scale crustal deformation characteristics associated with seismogenic process and various deformation anomalies before earthquakes in Chinese mainland. However, it is still insufficient that the effective basis for the crustal deformation can give the prediction of strong earthquakes in a clear time scale of 10 years and shorter. In order to promote the progress of continental strong earthquake prediction in a short time scale, it is necessary to effectively identify the process of strong source approaching earthquake in late seismogenic stage. Based on the analysis that the crustal deformation information observed by global navigation satellite system(GNSS) before several strong earthquakes, the crustal deformation state in the focal region of the late pregnancy earthquake could be transferred to the inelastic deformation state. Considering the factors such as the progress of observation technology, the possibility and related problems of the effective identification of the late pregnancy earthquake approaching to the strong source are discussed. [ABSTRACT FROM AUTHOR]

In recent years, the rapid development of satellite geodesy provides unprecedented multidimensional observations for the accurate measurement of crustal deformation and fault behavior. Combined with seismology, the temporal and spatial resolution of earthquake cycle deformation monitoring is greatly improved, which provides a chance for further study of earthquake cycle process and mechanism. Seismogeodesy can quantitatively describe the crustal movement and accurately model the fault activity. It can also pro⁃ vide a scientific basis to understand the stress and strain evolution of the whole earthquake cycle process, and provide scientific guidance for the assessment of seismic hazard and earthquake prediction and early warning. This paper analyzes the kinematic characteristics of faults at different stages of the earthquake cycle, including interseismic, coseismic, and postseismic, and reviews some important findings of seismic geodesy in terms of source physics by using satellite geodetic observations to investigate fault deformation. The satellite geodetic data can be applied to determine the fault location in the stage of earthquake cycle. It is a feasible idea for earthquake prediction. [ABSTRACT FROM AUTHOR]

Objectives: A method of deep learning change detection with domain knowledge as an optimization strategy was proposed to improve the change detection precision of high-resolution remote sensing images. Methods: The improved change vector analysis algorithm and grey-level co-occurrence matrix algorithm were used to obtain the spectral and texture changes of images, and reasonable thresholds were set to divide the changed samples from the unchanged samples based on the spectral and texture change intensity maps. The pattern shape index and spectral knowledge in domain knowledge were introduced as an optimization strategy to filter the changed samples for obtaining high-quality training samples. The deep belief net⁃ work model was constructed and trained, and the results of deep learning change detection were optimized by the optimization strategy to reduce the influence of “salt and pepper noise” and false change zones on the detection accuracy. Results: The Results of change detection experiments show that the accuracies of Gaofen⁃2 and IKONOS images were increased by 7.58% and 14.69% and the recall by 17.08% and 23.87%, respectively, while the false alarms and were decreased by 30.22% and 23.30% and the missing alarms by 17.08% and 23.87%, respectively. Conclusions: Compared with the method before the optimization strategy was adopted, the proposed method in this paper can effectively improve the precision of change detection, and it provides a new way of using remote sensing images to improve the precision of deep learning change detection. [ABSTRACT FROM AUTHOR]

Objectives: For the reliability of the geomagnetic navigation of an autonomous underwater vehicle (AUV) and the rationality of route planning, a self-organizing optimal classification method based on principal component analysis (PCA) and the improved back-propagation (BP) neural network is proposed for candidate geomagnetic matching areas. Methods: This paper unifies the classification of candidate geo-magnetic matching areas into the framework of pattern recognition. Firstly, PCA is used to linearly trans⁃ form some geomagnetic characteristic parameters to obtain the independent characteristic parameters of principal components. Secondly, the initial weights and thresholds of the BP neural network are optimized by the genetic algorithm (GA) to improve the classification accuracy of the matching suitability of candidate geomagnetic matching areas. Finally, the correspondence between the geomagnetic characteristic parameters and matching performance is established based on PCA and the GA⁃BP neural network for the automatic recognition of geomagnetic matching areas. Results: Simulated experimental results show that the self-organizing optimization classification method has a higher classification accuracy and reliability in the se⁃ lection of the matching areas for geomagnetic navigation and the accuracy of integrated navigation systems is also improved. Conclusions: The proposed method can provide important support for AUV route planning, which is an effective guarantee for the high-precision and long-voyage autonomous navigation of AUVs [ABSTRACT FROM AUTHOR]

Objectives: Deep space antennas are widely used in deep space and radio astronomical observations, and their precise position coordinates are required for research needs. In many cases, this precise position is given by very long baseline interferometry (VLBI) based on an initial position with millimeter-scale accuracy. Global positioning system (GPS) alone is merely not sufficient to obtain the initial position of newly-built antennas. Therefore, a method using other objects as reference is needed to calibrate observatory positions, especially antenna arrays, and this paper presents such a method for deep space and radio astronomical observations. Methods: This method employs an azimuth-elevation telescope with an accurate position as a reference object to measure the center position of a target telescope. Other reference objects with satisfying position accuracy can also be used in this method. GPS receivers are adopted to establish a reference network with given scale and orientation by differential measurement. The next step is measuring the local positions of two telescopes (or objective and reference) in the reference network using total station instruments deployed in several different locations. Calculations are induced during the transfer of positions in the reference system to unify data from those measurements. The position of the target to the reference object in a local coordinate system is obtained. The final step is the calculation of the coordinates of the tar⁃ get telescope through the accurate position of the reference telescope. A series of processes are used to con⁃ vert coordinate positions from the local reference system to the international terrestrial reference frame 2014 (ITRF2014) reference system. Results: Experimental measurement was carried out to obtain the coordinates of the 40 m diameter radio telescope in national observatory of China (NAOC) Miyun observatory with the 50 m diameter radio telescope also there as a reference object. Total station measurement was con⁃ ducted at three locations around the reference telescope and four locations around the target telescope. Two reflecting prisms calibrated by GPS devices were used as scale and orientation reference. Total station targets were placed on azimuth and elevation structures of both telescopes. The target position data were obtained when the telescope rotated and tilted at multiple angles. More than 400 times of measurement were taken during a two ⁃ day field experiment. Round ⁃ fitting and sphere ⁃ fitting achieved by the trust-region method were used to process measured points, and the coordinates of the 40 m diameter radio telescope in NAOC Miyun observatory were given with a calibration error of 2.312 mm (root mean square error， RMSE) in the ITRF2014 system. Conclusions: This method provides center point location in mm or sub-mm-scale precision in complex terrain situations with simple tools and low time cost. The measurement of total station instruments towards reference objects or target objects could be the direct or centering measurement. This method can be applied to obtain positions of single or multiple antennas for observation and further calibration. Observatories can apply this method in complex terrain locations such as mountains, rocky fields, and even a meteor crater on an extra-earth planet. Subsequent observations show that this calibration has met the need. [ABSTRACT FROM AUTHOR]

Objectives: In the field of human-related geographic information systems (GIS), the spatiotemporal analysis of character information has received increasingly more attention. It is important in that it helps GIS users to generate various thematic maps and achieve the visualization of human geographic con⁃ tent. For adaptation to the development direction of GIS intellectualization, it is of great significance to combine GIS requirements with natural language processing (NLP) methods and build a character information model. Methods: Firstly, we expound the research status of character information models in GIS and NLP and put forward the concept of life-track, which is mainly composed of a series of character event mentions. Secondly, considering the feasibility of the existing information extraction methods, a conceptual character life-track information model is determined. This model focuses on event information to highlight character spatiotemporal elements and also includes character attribute and relationship information. Finally, a complete information extraction process is designed for the model with online character encyclopedia pages as the data source. This paper focuses on two sub-tasks in the process: One is to use time features and Open How Net semantic calculations to identify event mentions, and the other is to use linguistics features and the conditional random field (CRF) model to extract location information. Results: Experiment results show that the method of event mention identification has an accuracy of 91.8%. Although the average F1 value of location information extraction is only 78% under the condition of a limited labeling corpus, some valuable experimental conclusions have been obtained by analyzing the weight of the transmit matrix of the CRF model: (1) The location phrase and its adjacent words have obvious characteristic effects. (2) The dependency syntactic parsing and the relative position of the word in the sentence have little influence on the extraction. (3) The target of location information extraction is the place where the event occurred, but in a few cases, some location phrases are irrelevant to the location of the event. This is the main reason for the low accuracy. Conclusions: Combining GIS with NLP, intelligent GIS development will be promising. The character life-track information model provides an example of the large-scale use of ubiquitous internet information. Improving methods applied in the extraction process and applying those methods to more online text types are the focus of our team's subsequent research. [ABSTRACT FROM AUTHOR]

Object: The influence of human factors on wayfinding performance is essential to improve the user experience of intelligent navigation service. Method: Based on four cognitive abilities, i.e. mental rotation, abstract reasoning, spatial short-term memory and spatial perception, this paper conducts several spatial cognitive experiments to analysis the correlation between spatial cognitive abilities and wayfinding performance. Results: The results indicate that: (1) These four spatial cognitive abilities are weakly correlated with each other, and can be separated partly with reference to the scale of recognition space. Specifically, cognitive group with mental rotation ability and abstract reasoning ability, and cognitive group with short ⁃ term memory ability and spatial perception ability, have stronger correlation respectively concerning the same space scale. While those groups where cognitive abilities cross different scale space have week correlation. (2) Mental rotation ability and abstract reasoning ability have higher significance and goodness of regression fitting on wayfinding performance. Considering the combined effect of multiple cognitive factors, in virtual scene, the mental rotation ability shows the highest influence while the spatial perception ability has the lowest influence; in real scene, the spatial perception ability shows the highest influence and the short-term memory ability has the lowest influence. (3) The wayfinding performance has significant scene effect with respect to the real and virtual scenes, i.e. the determinant coefficients of cognitive factors for way⁃ finding performance fitting varies with the scene type. Conclusions: These results provide an empirical reference for the effectiveness of pedestrian navigation research based on virtual scene. It can also serve as a basis for the design of navigation service considering human factors. [ABSTRACT FROM AUTHOR]

Objectives: Aiming at the problems that current address matching relies heavily on word segmentation dictionary and cannot effectively recognize address elements in addresses and their types, a Chinese address parsing method based on deep learning is proposed. Methods: The model combining robustly optimized bidirectional encoder representations from transformers (BERT） transformers pretraining approach (RoBERTa), bidirectional long short-term memory (BiLSTM), and conditional random field (CRF) is used to parse Chinese addresses. Firstly, the RoBERT a model is used to obtain the word vector representation of the address. Secondly, BiLSTM is used to learn the address model features and contextual in⁃ formation. Finally, CRF is used to construct the constraint relations between the labels. Results: Through the comparison and evaluation of different word vector representations of addresses and different sequence labeling models, the proposed method in this study achieves the maximum value of 0.940 on the generalization ability test dataset, and the precision, recall rate, and F1⁃score of the corresponding test dataset reach 0.962, 0.974, and 0.968.Conclusions：The method proposed in this paper does not need to extract the ad⁃ dress model features, nor does it rely on word segmentation dictionary for address segmentation. Address elements are recognized by learning address context information and address model features. The model generalization ability test dataset used in this study can effectively test whether the model is overfitted. There is little difference between bidirectional gated recurrent unit (BiGRU) and BiLSTM for Chinese ad⁃ dress resolution, and the sparse attention mechanism helps to improve the accuracy of address resolution. [ABSTRACT FROM AUTHOR]

Objectives: The rapid development of cartographical and communication technologies makes the public free to create, publish, edit and share their map image resources and products with various platforms and tools. These map images are remarkably ubiquitous in terms of map content, mapping standards, and other aspects, which poses a big challenge to establishing massive well-annotated map image data. Thus, although the state⁃of⁃the⁃art deep learning methods have made a breakthrough in recognizing the standardized maps, they are still intrinsically unqualified to effectively address map image recognition and under⁃ standing due to the inadequate well-labeled map images. Methods: This paper summarizes the progress and challenges regarding map image recognition and discusses the theoretical configurations and potential geospatial artificial intelligence (GeoAI) techniques for efficient map image recognition and understanding. We propose the map features for deep learning models to represent map image content. Then, we explore the promising methodologies for map image content recognition and the possible semantic analysis methods for map image understanding. Subsequently, we prospect several implementations regarding map image recognition and understanding and their future potentials. Results: In our opinion, further investigation on theories and methods for map image representation is essential. Moreover, fully utilizing the values of map images depends upon recognizing the explicit content (map image perception) and mining the hidden semantics (map cognition). Conclusions: We hope our exploration can contribute to the cartographical community offering a GeoAI and data representation integrated perspective on map image utilization. [ABSTRACT FROM AUTHOR]

Objectives: With the continuous popularity of social network products and applications, there are more and more ways to obtain location check-in data. The judgement of friendship based on check-in data is one of the most popular research directions of location based social networks (LBSN). However, social network data is often difficult to obtain in advance. In order to solve the problem of friend relationship judgement based on location check-in data only, this paper proposed a friend relationship judgment method based on an improved gravity model. Methods: Firstly, the information gain was used to calculate the influence of different feature parameters on friendship judgment, and two feature parameters, residence distance and spatial temporal co ⁃occurrence zone, were selected as consequences. Secondly, the gravity model was improved according to the selected feature parameters, and the model's value range was mapped from 0 to 1 by using Sigmoid function, to facilitate the judgement of friend relationship and calibration of model parameters. Finally, the parameters of the model were calibrated by logistic regression, and the prediction of friend relationship was realized on the Gowalla and Brightkite datasets. Results: Multiple experiments were conducted. Part of check-in data were selected to test the validity of feature parameters. The AUC (area under the curve) value of the spatial temporal co-occurrence zone parameter on the Gowalla dataset was 0.710, and the AUC value in the Brightkite dataset was 0.760. The residence distance parameter on the Gowalla dataset got an AUC value of 0.634, and 0.647 on the Brightkite dataset. By setting different degrees of balance, the improved gravity model was used to conduct friend relationship prediction experiments. As the data imbalance increased, the accuracy increased, while the recall rate and Fvalue decreased. However, the model still had a high accuracy and recall rate even with an imbalance of 1∶150. The accuracy rate was 0.846 1 and the recall rate was 0.788 5 when the Gowalla dataset as the training dataset while the test was performed on the Brightkite dataset. The accuracy rate was 0.859 1, and the recall rate was 0.853 7 when the Brightkite dataset as the training dataset while the test was performed on the Gowalla dataset. The contrast experiment with the friend relationship probability model shows that the recall rate of the contrast model dropped from 0.75 to almost 0 with the increase of the threshold. The experimental results above show that the selected feature parameters and the proposed model have a good predictive effect on the judgment of friend relationship. Conclusions: The proposed method can realize friend relationship judgment under the condition of only relying on location check-in data. The model has better stability between datasets from different sources, and the overall performance of this method is significantly higher than that of the comparison method. [ABSTRACT FROM AUTHOR]

Objectives: CAE (convolutional autoencoder) combines the autoencoder neural network structure and the convolutional network structure. This paper processes the gravity data based on the deep learning, and reconstructs the gravity contour image from the noisy gravity data with the CAE. The autoencoder structure contains the equal dimensional input and output data, which goes for the gravity data processing. The convolutional network structure is widely used in image recognition recent years, which can learn and recognize the specific objects in an image. Methods: To create the training set, we generate 1000 hexahedrons with the random triaxial length parameters and the random density to simulate the natural gravity source bodies. The 2D gravity data set for these bodies is fast computed with the gravity forward formula and the noisy gravity data set is generated by adding Gaussian noise to the computed data. Meanwhile the accuracy of the length parameters and density is limited in order to improve the representativeness of the training data set. Hence the 2D gravity data and the noisy data will be output and input data of the CAE. We design 5 layers CAE. The input and output layers both are 26×26. 32 and 64 feature maps are generated with 3×3 convolution kernels in the inner layers. The training is executed using RMSProp(root mean square propagation) optimizer. Results: To test the generalization of the trained CAE, the testing set with 500 samples is generated in the same way with the training set. The relative error histogram on the testing set shows that the reconstructing error is less than 5% and most of the error is around 0. To test the recognition of the gravity features in the 2D gravity, we test the CAE with simulated gravity data which contain 2 and more gravity anomalies. The results show that the CAE can recognize all gravity anomalies in an noisy gravity image and reconstruct them and the output image shows fairly fine reconstruction from the gravity with 10% noise. The measured gravity data of Kauring testing ground are used to test the generalization performance of the CAE. The reconstructed image is compared with the traditional FFT(fast Fourier transform) filtering and db wavelet filtering results. To increase the reconstruction difficulty, the Kauring data is added with 10% Gaussian noise. The results show that the trained CAE performs significantly less error than the FFT filter and the wavelet filter. Conclusions: In terms of reconstructing performance, the CAE can avoid over processing that may eliminate the real gravity anomaly, which is often seen in filtering especially for FFT filtering. The CAE reconstructs all of the 3 gravity anomalies of Kauring while the FFT filtering reconstructs 2. And the over-smoothing artifact is less than filtering methods. The CAE can process 2D gravity data and show lower error and generate less artifacts than filtering methods in our tests with testing set, simulated gravity data and measuring data. Beyond that, this article reveals that the deep learning can learn gravity image. With the convolution structure in a neural network, the network can learn single gravity anomaly samples and extensively recognize all gravity anomalies in a gravity image, which signifies that the gravity can integrate with the artificial intelligence technology. [ABSTRACT FROM AUTHOR]

Objectives: To overcome the problem that the current bathymetric filtering methods require manual intervention and are difficult to implement technically, a bidirectional cloth simulation filtering (BCSF) algorithm is proposed and implemented in this paper. Methods: Firstly, the transfer iterative trend surface is established to eliminate the negative anomalies and guarantee the continuous expression of the seafloor topography. Then, the filtering surface is established to solve the over-filtering problem of convex and concave seafloor topographies based on the proposed BCSF correction model. Finally, to further improve the effectiveness of the filtering, adaptive distance threshold is optimized and estimated. To evaluate the performance of the proposed algorithm, the BCSF algorithm is applied to shallow water multibeam bathymetry data. Results: The experimental results show that the BCSF algorithm can avoid the over-filtering. The elimination rate of the proposed BCSF algorithm is better than that of the CSF ( cloth simulation filtering) algorithm, which decreases from 12.87% to 0.76% for the whole study area and from 15.29% to 1.09% for local study area, respectively. Conclusions: Compared with the CUBE ( combined uncertainty bathymetry estimation) algorithm, the BCSF algorithm is more easily to implement and can retain more terrain de tails. Consequently, the BCSF algorithm has strong robustness and application prospects for multibeam bathymetry data. [ABSTRACT FROM AUTHOR]

Objectives: When using variance component estimation to determine the relative weight ratio, the least square solution is generally used as the initial value of the iteration. In geodetic joint inversion, the least squares method will cause ill ⁃ posed problems, so the regularized solution is used instead of the least square solution. Regularization introduces bias to reduce variance, but when using variance component estimation to determine the relative weight ratio, the influence of bias is not considered, and the introduction of bias will cause inaccurate variance component estimation. Methods: This paper adopts the bias - corrected variance component estimation method to eliminate the influence of bias introduced by regularization. The residual - based bias - corrected variance component estimation and the variance component estimation method are used for simulation experiments, the Visso earthquake and Norcia earthquake are used for verification. Results and Conclusions: Simulation experimental result shows that the variance component estimation method after bias correction can better reverse the slip distribution. The bias-corrected variance component estimation method takes into account the bias introduced by the iterative initial value, and the theory is more rigorous. Two real earthquakes results show that the bias correction is reasonable and advantageous. [ABSTRACT FROM AUTHOR]

Objectives: The lunar libration parameters are important parameters that affect the positioning accuracy of the Chang'E ⁃3 lander. Studying the lunar balance dynamic parameter estimation methods is of great significance to the development of lunar exploration projects in China. At present, the lunar libration parameters is calculated by the ephemeris method, but this method cannot guarantee the positioning accuracy of the lander. Methods: We first establish the expressions of the very long baseline interferometry (VLBI) time delay and delay rate with the changes of the lander's position, velocity, then based on VLBI technique, the lunar physical libration parameters are calculated by using weighted least square adjustment method. Results: Compared with the interpolation results of the DE421 ephemeris, the improvements of the Euler angles Ω, i and μ are compensated to -0. 692 4″, 0. 009 6″ and -0. 009 7″ respectively. Based on the lunar physical libration parameters,the position and velocity of the lander were solved, the results show that the positioning accuracy of the lander can be improved in X,Y,Z directions by 24. 204 m, 0. 405 m, 1. 996 m respectively. The estimation accuracy of the velocity error is also improved by 0. 010 6 m/s, 0. 013 5 m/sand 0. 007 2 m/s in X,Y,Z directions. Conclusions: This paper provides strong reference for the related research on the calculation of the physical libration parameters in the future. [ABSTRACT FROM AUTHOR]

Objectives: The Poincaré force related to the precession of Earth's axis is usually omitted in the solution of the dynamo equations as its effects are minimal. But in fact, the period of Earth's axial precession is still a term that is worth considering compared to the period of Earth's magnetic field reversal. Methods: This paper compares and analyzes different dynamo models with Ekman number and Rayleigh number by introducing Earth's axis precession velocity with a period of 25 960 years. Results: It is found that Earth's axial precession stabilizes the kinetic energy and magnetic energy fluctuation of spherical shell magnetic fluid in a smaller range than the benchmark dynamo, and increases the toroidal kinetic energy of spherical shell magnetic fluid by more than 10% - 20%, which leads to significantly accelerated westward drift rate of the magnetic field. Based on the magnetic field strength and magnetic Reynolds number at the core ⁃mantle boundary of those dynamos, it is found that the introduction of the precession item of Earth's axis makes magnetic energy more tend to kinetic energy conversion. And by comparing the core ⁃mantle boundary dipolarity, it is found that the introduction of the precession item will reduce the dipolarity. But for the dynamos in this article, the influence is not strong enough to make it become less Earth ⁃ like. Conclusions: From the comparison and analysis of the dynamo models, the introduction of Earth's axis pre⁃ cession would possibly change the dynamo model with a small Ekman number to a multipole dominant dynamo model. It can be inferred that the term of Earth's axis precession should be considered in more precise research of the numerical dynamo simulation for more. [ABSTRACT FROM AUTHOR]

Objectives: With the development of remote sensing technology, the spatial resolution of remote sensing image keeps improving, which brings both opportunities and challenges for the traditional remote sensing image classification and change detection. In order to improve the reliability of change detection of high resolution remote sensing image, this paper proposes a method of change detection of remote sensing image based on fuzzy comprehensive evaluation. Methods: Firstly, images of two‑phases were overlapped to a new image which can be segmented at multi‑scale. Secondly, a fuzzy comprehensive evaluation model was established for objects in certain scale to calculate membership of pixels in each object, which took the spectral and texture characteristics of two‑phase remote sensing image objects into comprehensive consideration. Finally, the entropy method was used to fuse the fuzzy evaluation membership degree of each pixel in different scales based on information entropy. Results: Taking two groups of high resolution images with different phases as examples, we realized the effective fusion of multi‑scale change detection based on fuzzy logic, which made full use of multi‑level pixel features and consequently improved the overall effect of single scale object‑oriented change detection. Conclusions: The proposed method provides a new idea for the exploration of multi‑scale change detection. [ABSTRACT FROM AUTHOR]

Objectives: To determine the geologic age of the lunar surface is the foundation of the research on the formation and evolution of the Moon, and also the foundation of the inversion process of lunar geological events. Lack of lunar rock and soil samples limits the range of geological units that can be used for isotopic dating. Therefore, the dating using crater size‑frequency distribution (CSFD) is employed to obtain the geologic ages of broader regions on the lunar surface. However, the presence of secondary craters will lead to a deviation in geologic age which is obtained by CSFD method. Thus, secondary craters should be eliminated to get a more accurate geologic age. Methods: This paper presents a method to identify secondary craters based on the distribution of iron element on lunar surface. First, the method assumes that the iron content in secondary craters is close to that in lunar regolith. Then, band ratio method is utilized to acquire the iron content. Finally, secondary craters are distinguished from primary ones in terms of the difference of iron content between craters and lunar regolith. Results: The effectiveness and robustness of the proposed method are tested by using the multispectral data obtained by multi‑band imager of the Japanese Selene Mission. The experimental results show that the geologic age deviation compared with the known isotopic dated ones is less than 0.1 Ga, which shows good consistency. Conclusions: Compared with other secondary craters identification methods, the proposed method is proven to be more effective and robust. [ABSTRACT FROM AUTHOR]

Objectives: Geometric misalignment is one of the key problems for orthophoto mosaicking. To solve this problem, this paper proposes a regional seamline detection algorithm based on dense optical flow. Methods: Firstly, the cost image is constructed with dense optical flow, gradient information and gray information. And the cost image is regarded as a weighted undirected graph. Secondly, the principle of maximum flow and minimum cut is used to search for seamline based on graph‑cut model. Results: The experiment results show that the seamline cost of the sparse building area greater than 100 pixels obtained by the proposed method only accounts for 0.7% of the path length. And the search efficiency of the proposed method is increased by 17% compared with OrthoVista software. Conclusions: The proposed method can automatically avoid passing the building area and the areas with large projection in digital orthophoto map, greatly reduce the probability of geometric dislocation phenomenon of image mosaicking, and realize the intelligence of orthophoto seamline searching. [ABSTRACT FROM AUTHOR]

Objectives: On 14th and 15th January 2022, the Hunga Tonga‑Hunga Ha̓apai (HTHH) underwater volcano in Tonga in the South Pacific Ocean violently erupted and caused a tsunami, which has attracted widespread international attention. In this paper, we attempted to rapidly investigate this oncein‑a‑millennium massive event using integrated remote sensing techniques. Methods: We comprehensively used multi‑temporal satellite optical images, radar images, global navigation satellite system (GNSS) observations and other datasets to quickly interpret and analyze the volcanic eruption process and its impacts. The damage assessment in parts of Tonga was also performed. Based on the integrated remote sensing techniques, we presented a technical framework for the rapid interpretation and analysis of volcanic eruptions, which comprises of six components, namely multi‑source data acquisition, geomorphology monitoring, deformation monitoring, environmental response detection, damage assessment and recovery decision. Results: Our results show that the HTHH underwater volcano in Tonga had exhibited obvious surface displacements in the satellite radar line of sight direction with a maximum accumulated displacement of 6.0 cm since June 2020, and erupted since around 22nd December 2021. During the volcanic eruption, the local magnetic field and ionosphere exhibited obvious abnormal signals, and the continuous GNSS data in Tonga suggested clear surface displacements with an uplift up to 50.2 cm. Most areas of Tonga were blanketed in ash, but it appeared that, in northern part of Tonga̓s capital, little damage was observed in the coastlines and the main buildings and roads remained intact. Conclusions: This study shows how to use integrated remote sensing techniques to quickly respond to underwater volcanic eruptions, including the topography evolution of the HTHH volcano and the damage assessment of its eruption. It is believed that the remote sensing integrated technical framework shall not only contribute to the speedy recovery and reconstruction of the society after the HTHH volcanic eruption, but also to the mitigation of future geohazard events. [ABSTRACT FROM AUTHOR]

Objectives: The temporal‑spatial variation characteristic of the glacier surface velocity is an important indicator of dynamic change of glacier and can provide important information for glacier mass balance. Surface temperature, glacier bed slope and glacier altitude are all important factors that affect glacier surface velocity, and their respective effects need to be analyzed. Methods: In this paper, the offset ‑ tracking method was used to measure the surface velocity of the Gangnalou Glacier, based on high‑resolution Sentinel‑1A synthetic aperture radar (SAR) images ranging from January 2016 to December 2016. We divided the surface velocity into freezing period and ablation period based on whether surface temperature was greater than 0 °C or not. For further analysis, we extracted the central line glacier velocity, and then jointly used surface temperature and digital elevation model (DEM) to analyze the influences of temperature and glacier bed slope. Results: The results showed that surface velocity of freezing period was more stable and central line glacier velocity was faster than that of boundary which tended to 0 m/a. SAR measurements were affected by glacier ablation, accumulation and motion, and showed that the maximum surface velocity was 65.43 m/a during freezing period and 46.28 m/a during ablation period. This surface velocity field showed a partitioning phenomenon in which the velocities of the middle and lower parts of glacier tongue were significantly different from that of the upper part during ablation period. Conclusions: The surface velocity of the Gangnalou Glacier was affected by many complex factors, of which glacier ablation and glacier bed slope are main factors. The partitioning phenomenon in the ablation period could be better explained by increasing surface temperature and changing glacier altitude, which directly affected glacier ablation and accumulation. Additionally, by analyzing the effect of the glacier bed slope, we found that the Gangnalou Glacier followed the compressive and extending flow mode in the freezing period. [ABSTRACT FROM AUTHOR]

Objectives: For multiplicative error model, the parameters estimated by weighted least squares (WLS) are nonlinear functions of observations, and the weights of observations are nonlinear functions of the estimated parameters. The existing parameter estimation methods of multiplicative error model can theoretically achieve second‑order unbiased, but the precision of uncertainty can only achieve first‑order unbiased. Therefore, a new method is used to improve the accuracy of uncertainty. Methods: The effect of nonlinear iterative processes on WLS parameters is considered, and the relationship between the estimated parameters and the observations in iterative WLS process is regarded as a nonlinear nested function. The derivative‑free Sterling interpolation method with symmetric sampling is used to calculate the expectations of estimated parameters and the standard deviation. Results: From the analysis of two synthetic experiments, the following results are obtained: (1) Considering the impact of the randomness of each iteration on the parameter estimation, the proposed Sterling interpolation method can get better estimated parameters than the existing methods. (2) When the nonlinearity of model is high, the effectiveness of the Sterling interpolation method is more significant. (3) The uncertainty estimation method in this paper can achieve secondorder precision. Conclusions: The feasibility and effectiveness of the Sterling interpolation method for parameter estimation and precision estimation of multiplicative error model are verified. [ABSTRACT FROM AUTHOR]

Objectives: There are obvious seasonal variations in the GPS height time series, which affect the improvement of precision and can be corrected by both mathematical modelling and geophysical mechanisms. Compared to least square fitting, singular spectrum analysis (SSA) can extract random seasonal signals effectively through signal reconstruction, which is unaffected by the assumed sinusoidal waves. According to the elastic loading theory, the gravity recovery and climate experiment (GRACE) can be used to calculate the vertical surface displacement caused by changes in terrestrial water storage. Methods: This paper mainly studies the feasibility of correcting the seasonal variations in GPS heights using SSA and GRACE inversion results. The height time series of 27 GPS stations in Australia with a time span of from 5 to 10 years were chosen and combined with GRACE simultaneous inversions. Results: Because the spatial resolutions of GRACE are coarse and the loading displacement is much more sensitive to near‑field mass changes than far‑field ones, the amplitudes of GRACE ‑inferred hydrological loading deformations are significantly smaller than GPS. The weighted root mean square (WRMS) are reduced at 22 stations after GRACE ‑inferred displacement corrections, and the correlation coefficients between deformations estimated by GPS and GRACE range from 0.12 to 0.78 with a mean value of 0.43, indicating that GPS and GRACE results have good consistency and correlation. SSA is used to extract the annual signals of vertical displacements derived from GPS and GRACE, and contribution rates of singular spectral variance of annual signals are 21.60% and 34.48%, respectively, expressing that annual signals are the main components of GRACE inferred results. Geographical climatic conditions have a significant impact on the consistency of annual signals derived from GPS and GRACE. Compared with the arid areas in central and western Australia, the amplitude and phase of annual signals derived from GPS and GRACE are more consistent in the northern region with seasonal rainfall. Furthermore, cross wavelet transform (XWT) finds that the vertical displacement series derived from GPS and GRACE of each station have a significant resonance period of one year. The circular average phase angles of GPS/GRACE at the period closet to 1 cycle per year (cpy) outside the cone of influence range from − 74.03° to 67.23° . The mean XWT ‑ based semblances range from 0.28 to 0.99 with an average value of 0.79, showing that there is a significant positive correlation between the annual variations derived from GPS and GRACE. Conclusions: Overall, GRACE‑inferred deformations can explain the annual variations of GPS‑derived displacements, particularly in areas with high hydrological loading. It is possible to correct the annual signals of GPS heights by GRACE inversions, but the effect is not as good as the SSA‑filtered annual signals. [ABSTRACT FROM AUTHOR]

Objectives: Nowadays, a vast number of sensors and observation platforms bring a significant challenge for city perception. In the study, we focus on developing of a smart city tempo‑spatial perception system. The concept of smart city awareness base station (SCABS) is proposed in this study, and the SCABS architecture considers the above problems. Methods: Firstly, we systematically reviewed the requirements and the complexity of smart city perception and summarized the development process and network architecture of smart city perception technology systems. On this basis, the challenges of ubiquitous access, trusted perception and intelligent management in future smart city perception are addressed. Then, this paper sorted out the demand for future city perception services and constructed the architecture of SCABS. Finally, we standardized the technical indicators and service capabilities of SCABS. Results: We constructed the prototype of SCABS. According to the qualitative analysis and experiments based on real environment, we found that SCABS with unified access, integrated management, and edge intelligence can meet the access and management ability of multi‑platform and high‑density sensing resources. Conclusions: SCABS would act an essential role in the smart city perception in the future. As a cyber‑physical infrastructure, it expects to cover various scenarios in the city with multiple heterogeneous perception access and hybrid access of multi‑protocols, and to automatically support the multi‑platform coordinated control. The SCABS meets global, accurate, heterogeneous, and multiple sensory perception. The SCABS is expected to be one of the critical infrastructures of the future smart city. [ABSTRACT FROM AUTHOR]

Objectives: The global navigation satellite system (GNSS)radio occultation (RO) signals can be reflected on the earth surface during the propagation process. It is of great value to explore atmospheric information contained in the reflected occultation signal for improving the accuracy of weather forecast and climate monitoring. Methods: Using the constellation observing system for meteorology, ionosphere and climate (COSMIC) occultation data collected from January 2011 to December 2015 in China, this paper extracted the occultation profile containing the reflected signals, which were strong at oceans and weak at inland and also showed seasonal changes. It was found that the inland snowy areas and the ocean surface were easier to reflect the occultation signals. Further, the offshore areas of China were set into 5°×5° grids and the correlation between the percentage of occultation reflected signals and the reanalysis data from European Centre for Medium‑Range Weather Forecasts in each grid was analyzed. Results: Around the offshore areas to the north of 10°N, the water vapor and temperature are negatively correlated with the percentage of occultation profile reflected signal, with a correlation coefficient from − 0.28 to − 0.58. However, around the offshore areas to the south of 10°N, affected by the monsoon at the South China Sea, the percentage of occultation profile reflection signal is positively correlated with the reanalyzed water vapor data, with a correlation coefficient from 0.32 to 0.34, but not significantly correlated with the reanalyzed temperature data. Conclusions: The RO reflected signals associate with atmospheric conditions, and it can be applied to the direct signals as supplementary data. [ABSTRACT FROM AUTHOR]

Objectives：Due to the sharp decline of global ecosystem services (ESs), there is an urgent need for correct environmental governance policies. The two problems in the current ecological environment governance in China need to be solved: (1) The lack of overall planning based on ecological function areas, (2) The shortage of in⁃depth synergy and tradeoff understanding among ESs. This study takes ecological function areas as the research units to measure the spatio temporal changes of four ESs (grain production (GP), carbon sequestration (CS), outdoor recreation (OR), and biodiversity conservation (BC)). It also explores the formation mechanism of the synergy and tradeoff among ESs and its spatially non⁃stationary response of the synergy and tradeoff to the five influencing factors of temperature, precipitation, sunshine, altitude, and urbanization. Methods: Various models and multisource data are integrated to estimate the yield distribution of ESs in Fujian Province. A difference comparison method is proposed to determine the spatial distribution of the synergy and tradeoff among ESs. The Moran􀆳s I applied is to determine the spatial autocorrelation of the synergy and tradeoff among ESs. The influencing factors are selected by principal component analysis, and the geographical weighted logistic regression (GWLR) determined the spatial response of the synergy and tradeoff to the influencing factors. Results: The main results are as follows: (1) The four ESs selected in the study (GP, CS, OR, And BC) have spatial heterogeneity. Among them, the yield distribution of GP is significantly related to the administrative region, and the highest values of BC, OR and CS appear in the ecological functional areas of biodiversity and water conservation. (2) All the Moran􀆳s I values applied are greater than 0, which means the synergy and tradeoff among ESs have significant spatial autocorrelation and are spatially clustered. (3) In space, there are synergy and tradeoff, instead of a single relationship between two ESs. (4) When the yield spatial distribution of two services is similar, such as BC and OR, the spatial distribution of synergies and tradeoffs of other services is also very similar. (5) The synergy and tradeoff among ESs showed significant responses to climate, topography, and urbanization. (6) The responses of the synergy and tradeoff among ESs to the influencing factors are spatially heterogeneous, and their positivity, negativity, and intensity varied with space. Conclusions: The diagnosis results of the model show that GWLR is superior to the global regression model in explaining the correlation of synergy and tradeoff among ESs with the influencing factors. By summarizing the law of the occurrence of synergy and tradeoff among ESs, this paper refines the existing formation mechanism of the synergy and tradeoff among ESs: In a certain area, the area proportion of supporting type of land and non⁃supporting type of land, and the degree of their support effect for each service will divide the two services into superior service and inferior service within the region. Due to the changes in land use, there are a dynamic gap between the superior service and the inferior service, and the increase in the gap may lead to tradeoff. [ABSTRACT FROM AUTHOR]

Objectives: The random volume over ground (RVoG) model is widely used in forest height inversion with polarimetric interferometric synthetic aperture radar (PolInSAR). The model assumes that the forest is a random uniform homogeneous body and the extinction coefficient in the model is a constant without considering the effects of forest vertical heterogeneity and terrain slope. This paper proposes a promising multi⁃baseline (MB) algorithm for forest height inversion based on a slope⁃RVoG (S⁃RVoG) model with linearly varying extinction. Methods: The effects of terrain slope and forest vertical heterogeneity on forest height inversion with the RVoG model are considered in the proposed algorithm. Firstly, the terrain slope is introduced to rectify the effective vertical wavenumber, and the S⁃RVoG model is derived on the basis of the basic RVoG model. Secondly, the linearly varying extinction coefficient, which is assumed to vary linearly with the forest height, is introduced into the S⁃RVoG model, and it can be solved by the Gaussian error function. Finally, MB PolInSAR datasets are used to solve the parameters of the S⁃RVoG model with linearly varying extinction, and the forest height can be obtained by the MB three⁃stage algorithm with coherence separation product criterion. The P⁃band F⁃SAR airborne PolInSAR datasets obtained by the 2016 AfriSAR campaign of the European Space Agency are selected for experimental verification. Results: The results of four MB algorithms, namely MB RVoG, MB S⁃RVoG, MB RVoG with linearly varying extinction, and MB S⁃RVoG with linearly varying extinction, are compared. The root mean square error (RMSE) and the relative error are used to evaluate the accuracy of the obtained forest height. (1) The forest height calculated by the MB RVoG algorithm is a significant overestimation, with RMSE of 6.57 m and relative error of 16.8%. (2) The RMSE of the MB S⁃RVoG algorithm is 5.97 m, and the relative error is 15.1%. The accuracy is improved by about 10% with the addition of terrain slope correction. (3) The MB RVoG algorithm with linearly varying extinction has RMSE of 4.71 m and relative error of 11.0%. Compared with the conventional MB RVoG algorithm, it improves the accuracy by about 28.3%. (4) The RMSE of the MB S⁃RVOG algorithm with linearly varying extinction is 4.27 m, and the relative error is 9.9%. Compared with the results of the MB RVoG algorithm and the MB S⁃RVOG algorithm, the accuracy is improved by about 35% and 28.4%, respectively. Conclusions: The RVoG model is widely used in PolInSAR forest height inversion. The MB S⁃RVOG algorithm with linearly varying extinction considers the effects of terrain slope and forest vertical heterogeneity simultaneously and introduces linearly varying extinction and terrain slope to correct the model, which makes up for the deficiency of the traditional RVoG model. The results show that the S⁃RVOG model with linearly varying extinction performs better in tropical forests with high forest density and great forest height. [ABSTRACT FROM AUTHOR]