Search

With the rise of IoT and edge computing, deploying neural networks (NNs) on low-power edge computing devices is drawing more and more attention. In NNs, convolutional layers take up the majority of the computing cycles, especially when NNs are implemented on ARM processors. Therefore, it is necessary to optimize the convolutional implementation on ARM Cortex-M MCUs. This paper proposes an efficient im2row-based fast convolution algorithm with two innovations. First, a novel im2row method for reusing the data of adjacent convolutional windows is presented. This method utilizes a reusable im2row buffer for data reuse, significantly reducing the amount of data copied during im2row and improving efficiency. Second, in algorithm implementation, a q7_t to q15_t data type extension technique that avoids data reordering is employed. This technique eliminates data reordering instructions, thus reducing the runtime of the algorithm. We evaluate our algorithm in separate convolutional layers and NNs. The results for convolutional layers show that, compared to baseline, the proposed algorithm speeds up the convolutional layer by an average of $1.42\times $ , and the maximum speedup is up to $2.9\times $ . Experiments on different NNs demonstrate that our algorithm can speed up the overall NN by up to $2.15\times $ .

Detailed information regarding the status and dynamics of farmlands is essential to protect food security, particularly for areas with a limited portion of farmlands like Fujian province. Since Fujian has a mountainous terrain and subtropical climate, it is challenging to acquire cloud-free remote sensing images to extract land information from remote sensing images. This study investigates the use of Google Earth Engine (GEE) to monitor farmland changes from remote sensing images, and Geographical Detectors to analyze the underlying driving factors in the Fujian province. We trained an online Deep Neural Network (DNN) model via Google Colaboratory, and then applied to it classify land-use types from Landsat images during 2000-2020 derived from the GEE platform. The obtained changes of land-use types were then inputted to Geographical Detectors, integrated with population, meteorological, and socio-economic data, to analyze their driving forces. The results showed that the proposed method produced land use maps at intervals of three years with higher overall accuracy (0.91±0.01). Based upon Geographical Detectors, we found that the farmland changes mainly caused by factors regarding elevation, population and slope, followed by factors of soil type, temperature and gross domestic product (GDP). We conclude that the GEE platform combined with deep learning models is of high potential to extract land cover and multi-temporal land use maps over large regions, and the Geographical Detectors are suitable for analyzing land changes using remote sensing derived products.

Tobacco is an important economic crop in the southern part of China, e.g., Fujian Province. Detailed spatial information of tobacco planting is essential for a good agriculture plan and sustainable management of tobacco. Optical remote sensing images acquired in the Fujian region are heavily affected by cloud coverage due to a subtropical climate. In this study, we investigate the use of time series C-band Sentinel-1 (S1) SAR data to extract tobacco planting areas. We use a Long Short-Term Memory (LSTM) model to quantify the relations between tobacco’s phenological information and the time series of features extracted from S1 SAR data. More specifically, the VH polarization channel was used to create the time series of feature datasets. Experiments were conducted on the S1 SAR dataset acquired during the growth cycle of tobacco from 2019 to 2020 in Nanping, Fujian, China. To evaluate the effectiveness of the proposed method, we compared the extraction results with that of the conventional machine learning method, i.e., Light Gradient Boosting Machine (Light GBM). Results show that the tobacco areas extracted by the proposed LSTM method have an overall accuracy of 82.9%, based on validation samples derived from very high resolution remote sensing images and a field survey conducted in 2020. The obtained extraction accuracy is higher than that of the Light GBM method, i.e., 78.6%. We conclude that the proposed LSTM method has a high potential for mapping tobacco planting in (sub)tropical regions using time series of S1 SAR data, and can be used as an alternative method for mapping the planting of other crop types from remote sensing images.

The explosive growth of medical images brings about massive amounts of high-dimensional images. The medical image retrieval technique is effective in selecting some similar medical images for medical analysis. A vast majority of existing medical image retrieval methods aim to extract a single feature to represent a medical image. Although these methods have improved the performance, they ignore that different characteristics express different information and they are an indispensable part of content information. Moreover, conventional features consume much storage, which is not conducive to data storage. Therefore, to reduce the loss of information and save storage space, we propose a novel jointly subspace hashing (JSSH) method for medical image retrieval. We solve the high-dimensionality of medical by the simplest image segmentation technology, and then study different subspace projection matrices. Finally, we integrate these matrices into the hash learning model, and build an objective function and learn a series of discriminative binary codes. By conducting comprehensive experiments on three medical image benchmark datasets, we demonstrate the effectiveness of our proposed JSSH.

The synchronous rectification (SR) is an effective method to further improve the efficiency of the CLLC bidirectional resonant converters. However, the SR driving signals are hard to generate. This paper proposed a novel SR scheme based on resonant inductor voltage sensing. The inductor voltage is sensed by a voltage sensing winding integrated in the inductor, and is then integrated to obtain the resonant current. The resonant current direction is used to generate SR driving signals, and a widely applicable control strategy is proposed to cover all operation conditions. The proposed method does not need any loss, costly and bulky current sensors, while maintaining the benefits of current sensing based SR methods. Finally, the proposed SR scheme is verified on a 380V/270V CLLC bidirectional resonant converter.

It is a fast and effective method to evaluate the ecological environment of mining area using remote sensing technology. In this paper, Landsat series data from 1992 to 2017 were used to monitor the change of ecological environment of Zijin Mine, located in the west of Fujian province, China. After radiometric calibration, the normalized difference vegetation indexes (NDVI) of Landsat TM or OLI from 1992 to 2017 were calculated. Based on NDVI data, the change of vegetation coverage of environment around Zijin mountain gold-copper deposit was analyzed. The results indicated that: （1）The area of Zijin mountain gold-copper deposit was increasing from 1992-2017. During 1996-2001, the area of mine was increasing rapidly because of large-scale blasting for removing loading strata in 1997, and it became obviously larger from 2014-2017 due to the copper price increases. （2）the NDVI around Zijin mountain raised from 1992 to 2014, indicating a better vegetation recovery, but it reduced from 2014 to 2017, indicating a worse vegetation recovery. low NDVI mainly exist in the middle of Zijin deposit, which is relate to mining activities. These results can give some scientific references for assessment of Zijin gold-copper deposit ecological environment.

Tea is the most popular drink in China. The spatial distribution information of tea plantation is useful for local government management. Lantian Country, with an area of 99.77km2, located in the midwest of Anxi County, which is famous for Oolong Tea, was chosen as study area, and image from Chinese high resolution satellite Gaofen-2 acquired on Jan 22, 2015 was used to study the method of tea plantations extraction. In order to construct best features for classification, optimum index factor (OIF) were firstly calculated on different original spectral bands combinations and the one with max OIF was chosen. Secondly, spectral enhancement was carried on multi-spectral bands.Difference between two vegetation indexes, namely, normalized difference vegetation index and modified normalized difference vegetation index was calculated and named as DNDVI. In DNDVI image, the brightness difference between tea plantation and background was improved and shadowed area in either index image was reduced. Thirdly, gray level co-occurrence matrix (GLCM), Gabor filter, local binary patterns (LBP) extraction, and method combined LBP and Gabor was carried on pan image to construct texture features. Among eight common features based GLCM, contrast, dissimilarity, entropy, variance, tea plantation area was darker. In homogeneity and angular second moment, this phenomena is just the opposite. In mean and correlation, there was no obvious difference between target tea plantation and background. So the gray level co-occurrence texture (GLCT) subtract sum of second two features from sum of the first four feature was used as final GLCM feature, and window size for GLCM set to be 15 was preferred. Multi-scale and multidirectional Gabor texture with max frequency set to be 1HZ was derived. For LBP, the operator LBP16, 2 with rotation invariance was tested to be the best. Finally, five schemes combine these spectral and textural features as inputs of classifier were evaluated in term of classification accuracy. Six categories including tea plantation, forest, roads, water, build-up, bare soil, shadows were classified by support vector machine. The result showed that overall accuracy range from 75.55% to 89.11%, Kappa coefficient range from 0.613 to 0.843, for plantation, user accuracy range from 84.95% to 100%, producer accuracy range from 53.29% to 91.53%. Gaofen-2 show its capacity to map the tea plantation area accurately. Schemes utilized spectral and textural features together perform much better than that utilized spectral only. The scheme combination of band1, band 3, ban4, DNDVI, LBP_Gabor outperformed other Scheme, with the highest overall accuracy and Kappa coefficient. The textures feature of high resolution image helps to improve the accuracy, and the way to construct suitable texture feature and merge different texture feature deserved study more. The proposed method to extract tea plantation is applicable at administrative level of country.

Tarim River Basin is located at the core area of The Belt and The Road at present and it once has been an essential transportation hub of the ancient Silk Road. However, Tarim River Basin belongs to arid and semi-arid region where water resource and vegetation cover are scarce, leading to soil erosion, desertification, and environmental degradation. So it is meaningful to study the variation of fractional vegetation cover (FVC) in Tarim River Basin. MODIS13Q1 data of Tarim River Basin from 2000 to 2018 were collected, and then, the pixel dichotomy model and Sen Mann-Kendall were applied in this research.The results were as follows: (1) The area ratio of different FVC types fluctuated with time, but the overall trend showed a slow growth trend. (2) In the trend analysis results, Significant improvement was concentrated in the low hills and oasis areas where the water resources in the basin are sufficient. In contrast, Severe degradation presented a feature of plaque-like sporadic distribution. (3)There are obvious seasonal differences FVC change. More specifically, FVC in spring and winter was considerably low, and yet FVC in summer and autumn was improved dramatically. Severe degradation and Significant improvement mainly occur in summer and autumn, and remain basically unchanged in spring and winter.

Land degradation is an important factor threatening human survival, social stability and sustainable development.This study aims to spatially analyse the land degradation sensitivity in Korla,Xinjiang in 2018, and provide a basis for identifying areas of land degradation.Based on the MEDALUS framework and the characteristics of the study area,four main factors of soil, climate, vegetation and management were considered, and several indicators affecting the quality of each factor were determined.The sensitivity of the areas to Land degradation risk was evaluated by LDSI(Land Degradation Sensitivity Index).As a whole, 84.4% of the districts in Korla have the possibility of land degradation. The ecological environment of Korla is relatively fragile.

Due to the high variability in tumor morphology and the low signal-to-noise ratio inherent to mammography, manual classification of mammogram yields a significant number of patients being called back, and subsequent large number of biopsies performed to reduce the risk of missing cancer. The convolutional neural network (CNN) is a popular deep-learning construct used in image classification. This technique has achieved significant advancements in large-set image-classification challenges in recent years. In this study, we had obtained over 3000 high-quality original mammograms with approval from an institutional review board at the University of Kentucky. Different classifiers based on CNNs were built, and each classifier was evaluated based on its performance relative to truth values generated by histology results from biopsy and two-year negative mammogram follow-up confirmed by expert radiologists. Our results showed that CNN model we had built and optimized via data augmentation and transfer learning have a great potential for automatic breast cancer detection using mammograms.

Interruption property of vacuum interrupter with four different contact materials were researched by experimental method. The diameter of Wan-type contact and Spiral-type CuCr25 contact was 55mm. The diameter of Spiral-type CuCr35 and CuCr30 was 58mm. The contact thickness was 10mm before breaking test and the contact thickness was about 9mm after breaking test. The breaking current was 31.5kA and the rated voltage was 12kV. The interruption times of Wan-type contact with arc melting CuCr35 material was 56 times. The interruption times of Spiral-type contact with arc melting CuCr35 material was 22 times. The interruption times of Spiral-type contact with powder metallurgy CuCr25 material was 100 times. The interruption times of Spiral-type contact with casting CuCr30 material was 36 times.

In this study, we applied a process-based model, boreal ecosystem productivity simulator (BEPS) to simulate net primary productivity (NPP) over Fujian Province in 2015, and the responses of NPP to various environmental factors (e.g. temperature, elevation, slope, aspect) was analyzed. The results showed that: (1) the total NPP and mean NPP of Fujian province in 2015 are 43.13 × 106 tC and 540.75gC m−2yr−1, and the rank of annual mean NPP of different forest types was as follows: Evergreen Broadleaf Forest(EBF)>Evergreen Needleleaf Forest (ENF)>Needle-Broadleaf Mixed Forest(NBMF)>Deciduous Needleleaf Forest (DNF)>Deciduous Broadleaf Forest(DBF), with the values of: 625.6gC m−2yr−1, 525.3gC m−2yr−1, 436.6gC m−2yr−1, 412.4gC m−2yr−1, 399.9gC m−2yr−1, respectively; (2) monthly mean NPP show “M” distribution temporally due to the high temperature in summer;(3)a significant correlation existed between NPP and terrain factors.

To effectively carry out soil and water conservation work and monitor soil erosion in Fujian Province, southeast of China, Universal Soil Loss Equation (USLE) model has been adopted. Landsat8 OLI images, combined with topographical data, soil data and rainfall data were used. Then the soil erosion modulus of Fujian Province was calculated quantitatively based on USLE and classified into different levels of soil erosion intensity. The spatial distribution characteristics of soil erosion in Fujian Province were analyzed. The results show that: The area of soil erosion in Fujian Province in 2015 is 10562.973 km2, accounting for 8.51% of the area of the whole province. The soil erosion intensity concentrated in light mainly, accounting for 82.64% of total erosion area. The dominant serious and above serious soil erosion areas are located in the southeastern of Fujian Province, such as Quanzhou and Zhangzhou.

In this paper, we propose a modified Boolean Particle Swarm Optimization velocity mutual (GEBPSO-vm) algorithm to design thinned planar array, and then optimize the side lobe level of thinned planar array. This new algorithm introduces the concept of feedback factor and G inertia weight factor. The feedback factor will generate an additional speed factor, which makes the evolution direction of the population have adaptive regulation function. The G inertia weight factor allows each particle swarm has a unique local adaptive weight. It is shown that the side lobe level, the sparsity and convergent speed of the planar array are improved by GEBPSO-vm. The simulation results of side lobe level, the sparsity and convergent speed are presented.

Taking the Zhuxi watershed in Changting County of China as the study area, soil erosion is estimated by using SWAT model. The runoff sediment data from 2013–2015 years of Hetian hydrologic station are used for the model's calibrations and correction. The experimental results showed that the simulated results from the SWAT model with the correction parameters could reflect the actual situation of the study area objectively. The estimation results show that the main soil erosion types in the watershed are mainly micro and light erosion, the area of moderate and serious erosion is small and these areas are mainly distributed in the central region. Among all kinds of land use types, the soil erosion modulus of bare land is maximal, while the absolute value of sediment yield in forest land is largest for its largest area.

During recent years, with the rapid development of wireless sensor technologies, Wireless Sensor Network (WSN) systems are widely used in various applications, such as environmental monitoring, smart home and so on. A WSN system is a self-configuring network of sensor nodes connected by wireless links. The union of sensor nodes forms an arbitrary topology. Routing in a WSN system is a critical task. Therefore, the routing protocol has much effect on the throughout performance and energy consumption of the whole network. Most available routing protocols employ distributed structure that is complex and has high energy consumption. Compared with these distributed protocols, a novel centralized routing protocol with a central node is proposed. And all kinds of routing management of the whole WSN systems are centralized in the central node. The centralized management strategy reduces routing overhead and energy consumption of the whole network. Finally, a WSN system with Raspberry Pi and OpenWrt is designed to evaluate the new method. Experiment results show that it is very efficient.

The effect of ceramic metallization process on the performance of vacuum interrupter was researched by X-ray thickness test, metallographic analysis, tensile strength test and air tightness test. Electroplating Ni and coating Ni metallization processes were used in this research. The thickness of electroplating Ni is 3.5μm-6μm, and that of coating Ni is 4μm-6μm. The Ni layers of electroplating Ni and coating Ni are both continuous, but electroplating Ni is not uniform. A little Ni layer can be seen from electroplating Ni ceramic after vacuum brazing. The tensile strength difference of standard tensile pieces between them is little. The air tightness difference of vacuum interrupter between them is not obvious.

As the feature size of the semiconductor process is scaling down to 10nm and below, it is possible to assemble systems with high performance processors that can theoretically provide computational power of up to tens of PLOPS. However, the power consumption of these systems is also rocketing up to tens of millions watts, and the actual performance is only around 60% of the theoretical performance. Today, power efficiency and sustained performance have become the main foci of processor designers. Traditional computing architecture such as superscalar and GPGPU are proven to be power inefficient, and there is a big gap between the actual and peak performance. In this paper, we present the MaPU architecture, a novel architecture which is suitable for data-intensive computing with great power efficiency and sustained computation throughput. To achieve this goal, MaPU attempts to optimize the application from a system perspective, including the hardware, algorithm and corresponding program model. It uses an innovative multi-granularity parallel memory system with intrinsic shuffle ability, cascading pipelines with wide SIMD data paths and a state-machine-based program model. When executing typical signal processing algorithms, a single MaPU core implemented with a 40nm process exhibits a sustained performance of 134 GLOPS while consuming only 2.8 W in power, which increases the actual power efficiency by an order of magnitude comparable with the traditional CPU and GPGPU.

As an important assisted feature for spectral one, texture plays an important role in image analysis and automatic recognition, especially in high spatial resolution remotely sensed images. Meanwhile, wavelet is an effective method of extracting multi-scale features of ground objects in images. Then in this research, the extraction approach of wavelet-domain fractal texture (WDFT) was proposed, and it was implemented to improve the image classification of QuickBird of Fuzhou City. WDFTs of QuickBird image were computed on different window sizes and decomposition layers, and three texture images were selected from the viewpoint of image classification and thematic extraction of buildings, the different box-counting (DBC) cap features of CA1 (the coarse image of the first decomposed layer of QuickBird image) on 64 ∗ 64 and 16 ∗ 16 windows and the multi-fractal feature of CA1 on 16 × 16 window. The experiment results implied that: because of the addition of WDFT information, the aquafarm, major roads and bare land confused with buildings were distinguished well; the supervised classification based on spectral feature was modified, and its total classification accuracy and Kappa coefficient became better (from 76.17% to 81.25%, and from 0.7006 to 0.7587, respectively), and also made the extraction accuracy (user one and mapping one) of buildings better (from 80.70% to 82.54%, and from 65.71% to 74.29%, respectively). It proves that the WDFT was effective. Addressed on the disadvantages in the research, the WDFT extraction on rectangular window and adaptive sizes will be studied and more decomposition layers information will be integrated in the next work.

Jiulong River has been polluted to varying degrees. The water environmental quality will decrease sharply when the quantity of pollutant and waste exceeds the consumption capacity. Thus, It is of great significance to evaluate the water environment consumption capacity. We put forward a consumption capacity assessment model and a new calculation unit so-called landscape watershed unit. The water environment pollution consumption capacity is calculated based on the landscape watershed unit and the administrative unit respectively. The results tend to polarize according to landscape watershed unit. The results reflect the water environment consumption potential well, which pointed out that the environment of most of the area in the Jiulong river is in a certain acceptance potential of Chemical oxygen demand (COD) and ammonia nitrogen, while in some important local urban areas, potential consumption capacity is close to saturated condition.

We consider the self-calibration problem (estimation of location and orientation of multiple camera sensors), in visual sensor networks equipped with RGB-D cameras. We propose two algorithms based on feature matching and relative pose estimation. First one uses Floyd-Warshall algorithm, and can accurately estimate the camera locations. On the other hand, the second algorithm is more scalable than the first one, and can be used in large networks if high accuracy is not an issue. Numerical results demonstrate that both algorithms, depending on the network topology, can be used for self-calibration in RGB-D equipped visual sensor networks.

Image fusion plays an important role for integrated usage of remotely sensed data from multi-sensors. Spectral response functions of different satellite sensors are used to normalize and to eliminate the difference between sensors. Then the ESTARFM model are applied to yield the time series high frequent images with fine spatial resolution. In the experiments, one cloud-free IRS P6 LISS image and two Landsat TM images are used as fine-resolution data, and 11 scenes of cloud-free MODIS L1B images with a spatial resolution of 250m are used as time series data. Fused results are evaluated both qualitatively and quantitatively. The results show that, the fusion algorithm based on spectral response function and ESTRAFM can be well used for merging MODIS data and fine-resolution data from multisensors. The predicted image integrated the spatial advantages of the high spatial resolution images and temporal advantages of the high temporal resolution images.

The building heights and the heights' distribution of a city are important aspects of urban remote sensing researches. Building heights are estimation from the shadow on Chinese No.3 resource satellite image (ZY3) in Pudong area of Shanghai, China. The morphological building/shadow indexes (MBI and MSI) derived from remotely sensed data and object-based information extraction methods are used to classify the building and its shadow. Based on the formula derived from the relationship among the sun, the satellite and the building, building heights are calculated. The most of the calculated building heights are reliable. Absolute errors below 3m are about 67.5%, the average relative error is 3.7%, the errors below 5% are about 73.7%. The proposed method could well achieve the building extraction and building heights estimation in high-resolution images.

This work presents a new method to efficiently communicate the correlated visual and depth information captured by multiple mobile RGB-D camera sensors. Our proposed scheme adopting the special properties of the depth map can extract the uncorrelated information between two images without conducting computationally heavy motion estimation. Moreover, a novel entropy coding scheme is proposed to further compress the depth data.

In mobile visual sensor networks, relative pose (location and orientation) estimation is a prerequisite to accomplish a wide range of collaborative tasks. In this paper, we present a distributed, peer-to-peer algorithm for relative pose estimation in a network of mobile robots equipped with RGB-D cameras acting as a visual sensor network. Our algorithm uses the depth information to estimate the relative pose of a robot when camera sensors mounted on different robots observe a common scene from different angles of view. To create the algorithm, we first developed a framework based on the beam-based sensor model to eliminate the adverse effects of the situations where two views of a scene each are partially seen by the sensors. Then, in order to cancel the bias introduced by the beam-based sensor model, we developed a scheme that allows the algorithm to symmetrize across the two views. We conducted simulations and also implemented the algorithm on our mobile visual sensor network testbed. Both the simulations and experimental results indicate that the proposed algorithm is fast enough for real-time operation and able to maintain a high estimation accuracy. To our knowledge, it is the first distributed relative pose estimation algorithm that uses the depth information captured by multiple RGB-D cameras.

We present a survey of recent research works on multiview image compression and transmission techniques developed for Wireless Multimedia Sensor Networks (WMSNs). We classify them into two categories with respect to the coding methods adopted: (i) in-network processing with joint coding schemes, and (ii) distributed source coding schemes. The survey also includes a comprehensive evaluation of the limitations of each approach. Based on the review results, we discuss future research directions, and identify the ways of more efficient transmission of the spatially correlated and redundant data in WMSNs.

The building envelope heat transfer coefficient is an important measurement of building energy efficiency. The detection of the heat transfer coefficient is always impacted by surrounding environment and noises. Meanwhile it is impractical to accumulate rich enough data as input to calculate the heat transfer coefficient. The Particle Filter (PF) and Extended Kalman Filter(EKF) are employed in this paper in predicting the heat transfer coefficient based on the temperature control box-heat flow model. With the comparison of the two predicted values with the real measured one, the Particle Filter shows high efficiency and better accurate than Extended Kalman Filter. The simulation results show that the accuracy of PF is high. The budget result of PF is more close to the real values. Then the estimated calculation according to Particle Filter is used to calculate wall body heat transfer coefficient.

Till now, there were several methods of fractal texture extraction developed for high-resolution remotely sensed images: triangular prism, fractal Brownian motion, DBC (different box-counting) and multi-fractal. However, one proper method should be selected in application. Upon that, in the paper, taking QuickBird image as example, based on a rather entire classification system founded for the image, fractal dimensions and correlative parameter were utilized to analyze features of those methods; mean, variance and computation time were utilized to compare difference of them. Then several conclusions were drawn through the experiment: fractal texture could be an effective feature to distinguish ground objects; among those methods, triangular prism couldn't do well on ground objects' distinguishing, FBM wasn't accurate (without no-scale part), but DBC and multi-fractal represented better on distinguishing ground objects and computing efficiency.

This paper presents an efficient link quality prediction method for OFDM systems. It is based on effective SINR mapping (ESM) approach. Being different from conventional ESM methods, the novel scheme employs constrained linear averaging to multiple instantaneous channel indicators. Its computation complexity is lower than conventional ones. Because there is no two-step mapping processing with invertible mapping functions. Its memory requirement is also much less because no additional off-line ESM look-up table is required. Finally, its performances in terms of BLER and throughput are evaluated in link adaption. Simulation results show that it is efficient.

The ability to record single unit neural activity from the brain of awake and naturally behaving animals will open new frontiers in neuroscience research. We present the prototype design of a battery powered, high bandwidth and miniature neural telemetry system operating in the 5.8 GHz ISM band. It was designed for spike recordings from freely moving marmosets (Callithrix jacchus) while they engaged in natural vocal behavior. The wireless system was tested on chronically implanted adult marmosets with multi-electrode arrays in the frontal cortex. With a baseband bandwidth of 3 MHz this system allows simultaneous transmission of 15 channels of neural data to a receiver 10 meters away. It has additional bandwidth for behavioral data such as multi-channel acoustic information. The transmitter measures 2.5×2×1.3 cm and weighs 10g. Our system provides researchers with design flexibility compared to expensive commercial alternatives and proved comparable to tethered systems during in vivo tests.

To make full use of the advantages of the high spatial resolution of Landsat, along with the high temporal resolution of MODIS, we applied the spatial and temporal adaptive reflectance fusion model (STARFM) by means of comparing the difference of the spectral response functions, as well as normalizing the similar bands between these two types of sensors, to construct a time series images with high spatial resolution in this paper. In this research, the experiments were carried out on the TM and MODIS images in the Yongtai area of Fuzhou city. The results show that: 1) in terms of spatially manner, the predicated data well reflected the spatial details difference of the high resolution remote sensing imagery by analyzing the difference in histogram between the real data and the predicated data; 2) in terms of timely manner, the predicated data retained the time change trend of high temporal remote sensing imagery by comparing the values of NDVI between the MODIS image and the predicated image.

This paper presents a genetic algorithm (GA) approach for parameters optimization of support vector machine, which is used for the object-oriented classification of high spatial resolution images over urban area. The proposed method is a three-step routine involves the integration of 1) image segmentation, 2) GA-based parameter optimization of Support vector machine (SVM), and 3) objected-based classification. Experiments conducted on multi-spectral Quick-Bird image fused with panchromatic image in Fuzhou city. In addition, a traditional parameter searching method, Grid algorithm, was investigated to evaluate the effectiveness of the proposed approach. The results show that our proposed GA-based approach significantly outperforms the Grid algorithm both in terms of classification accuracy and time efficiency.

As the third profits source, logistics plays a more and more important role in economic development. Service innovation is a strong driving force for the development of logistics enterprise, and the research on characteristics of service innovation will provide a theoretical basis to implement innovation activities in logistics enterprise. This paper starting with the characteristics of general service innovation and combing the characteristics of logistics services, analyses the characteristics of service innovation in logistics enterprise systemically. Then it summarizes four characteristics of service innovation in logistics enterprise: complicated process, high cost, long cycle and systematic. This study enriches the basic theory research of the service innovation, which is useful to carry out the service innovation effectively and scientifically in logistics enterprise.

The use of ocean color satellite image data to investigate oceanic processes has become essential for oceanographic research and monitoring. In order to exploit information from multiple sensors, a new data merger, i.e. adaptive weighted averaging was employed to merge MODIS Aqua and MODIS Terra chla product in the East China Sea and its performance was assessed. Comparison between chla products pairs reveal consistency below 10mg/m3, but discrepancy was obvious when turned to high values. The method present here will minimize the difference between the merged pixels and their adjoin pixels having single source. This will make the merged output more continuous.

Vegetation monitoring with spectral vegetation index (VI) has made great advances over the last several decades, and numerous vegetation indexes have been developed as well. However, the monitoring accuracy with conventional VI in rugged mountains is disturbed some extent by topographic effects. This paper presents one novel topography — adjusted vegetation index (TAVI) to obtain correct vegetation information in rugged area, which is only based on near infrared and red wavebands data of optical remote sensing image, without the digital elevation model (DEM) data. The validation with Landsat TM image shows that the slope of linear regression equation (k) of TAVI and the solar incidence cosine and the correlation coefficient (r) between them are similar to those from NDVI ac_tc (normalized difference vegetation index after atmospheric correction and topographic correction). TAVI achieves good results in rugged mountain vegetation monitoring, especially in resistance to topographic effect in rugged terrain and works well without the support of DEM data, which greatly extends its potentially applicable scope.

Satellite remote sensing plays a significant role in algal bloom monitoring and detection over these decades. Different satellite data from SeaWiFS, MODIS/Aqua, MODIS/Terra and MERIS were processed for East China Sea in May 2008. Derived normalized water-leaving radiance (nLw), chlorophyll concentration (Chl) and fluorescence line height (FLH) were compared and analysed for these sensors. The comparison results show that nLw data from MODIS/Terra with SeaWiFS and MODIS/Terra with MERIS have more reliable correlation than those from MODIS/Terra with MODIS/Aqua. The Chl comparsions reveal the good capability of MODIS/Terra and MERIS better than others. Therefore, it would be feasible to continue the time series using multi-sensor data and improve the data merging in algal bloom detection.

Numerous vegetation indexes have been developed for assessing vegetation characteristics and are still of increasing importance in environmental application of remote sensing. However, few of them are targeted to reduce topographic effects in rugged area. A novel topography-adjusted vegetation index (TAVI) is presented in this paper by combination of near-infrared and red wavebands data with topography adjusting coefficient without digital elevation model (DEM) data. The validity of TAVI is illustrated with a case study in Changting county, a mountainous area in southeast of China. The experimental results show that the proposed algorithm can effectively eliminate topographic effects caused by rugged terrain. The slope of linear regression equation of TAVI - the solar incidence cosine, and the correlation coefficient between them are only around 0.01 respectively. It also indicates that this topography-adjusted vegetation index could be applied in other mountainous areas.

We describe a flexible wireless neural recording system, which is comprised of a 15-channel analog FM transmitter, digital receiver and custom user interface software for data acquisition. The analog front-end is constructed from commercial off the shelf (COTS) components and weighs 6.3g (including batteries) and is capable of transmitting over 24 hours up to a range over 3m with a 25microV(rms) in-vivo noise floor. The Software Defined Radio (SDR) and the acquisition software provide a data acquisition platform with real time data display and can be customized based on the specifications of various experiments. The described system was characterized with in-vitro and in-vivo experiments and the results are presented.

An all-fiber optical pulse cleaner operating at 1053nm is demonstrated based on nonlinear polarization rotation. The pulse contrast can be enhanced by more than four orders of magnitude.

Research was made on the design scheme and basic algorithm of a kind of voltage-feedback type digital stabilized voltage switching mode power supply combining internal resources of single-chip microcomputer with control technique. Take 110V switching mode power supply by electric locomotive as an example, it adopted advanced sensor in sampling data, applied PID control algorithm to adjust voltage and used overcurrent, overheating and short-circuit protection. Design process of control algorithm was given. Eventually the result according to design requirements was achieved by modeling and simulating on the switching mode power supply, meantime the deficiency existed was indicated.