Search

Deep learning (DL) has been proven to be a powerful method in computer vision and is receiving increasing attention in remote sensing. It is important to analyse the research progress, hotspots, trends and methods in the field of deep learning in remote sensing. First, the main research countries (11), research institutions (20), researchers (20), and the most cited references (20) and hotspots (8) in this field were identified by analysing a total of 2,467 published papers with the bibliometric and visualized analysis (BVA) method. Then, based on the above analysis results, the research basis and the progress of hotspots in this field were summarized by reading a total of 181 relevant papers in detail with the traditional literature combing (TLC) method. The results indicate that deep learning is becoming an important tool for remote sensing and has been widely used in the vast majority of remote sensing tasks related to image processing. Among the following deep learning methods, the convolutional neural network (CNN) is undoubtedly the most widely used model. [ABSTRACT FROM AUTHOR]

Underwater biological, optical and physical measurements have been carried out at the Acqua Alta Oceanographic Tower (AAOT) in the northern Adriatic Sea on behalf of the European project 'Coastal region long-term measurements for colour remote sensing development and validation' (COLORS). This project is aimed at studying the variability of bio-optical properties and testing existing bio-optical algorithms for the major water parameters. In particular, this paper deals with the analysis of underwater profiles of optical and physical parameters, which are aimed at studying the seasonal development of stratification in the water column. The first part of this paper describes the analysis of long-term series of physical measurements, such as temperature, salinity and conductivity, to enhance the development of stratification. The influence of stratification on the optical properties of the water column is discussed in the second part of the paper, and illustration is made of the application of the previous results to the analysis of the underwater profiles of downwelling irradiance and upwelling radiance. [ABSTRACT FROM AUTHOR]

Due to background clutter in synthetic aperture radar (SAR) images, the detection of dense ship targets suffers from a low detection rate, high false alarm rate, and high missed detection rate. To address this issue, an FSM-DFF-YOLOv5+Confluence algorithm is proposed in this paper for the detection of near-shore ship targets in SAR images with complex backgrounds. First, based on the YOLOv5 target detection algorithm, two improvements are made in the feature extraction network: feature refinement and multi-feature fusion; in the feature extraction network, deformable convolutional neural networks are adopted to change the position of the target sampling points of the convolution to improve the feature extraction capability of the target and the detection rate of ship targets in SAR images with a complex background; in the multi-feature fusion network structure, cascading and parallel pyramids are used in the multi-feature fusion network to realize feature fusion at different levels; the visual perceptual field of feature extraction is expanded by using null convolution to enhance the adaptability of the network to detect near-shore multi-scale ship targets with complex backgrounds and reduce the false alarm rate of ship target detection in SAR images with complex environments. In this way, the DFF-YOLOv5 near-shore ship target detection algorithm is established. Meanwhile, to address the problem of missed detection in near-shore dense ship target detection, this paper adds rectangular convolution kernels to the convolution of the feature extraction network to better realize the feature extraction of dense ship targets in SAR images with complex backgrounds. Besides, the Confluence algorithm instead of non-maximum suppression is used in the prediction stage. Through experiments on the constructed complex background near-shore ship detection dataset, it is indicated that the average accuracy of the FSM-DFF-YOLOv5+Confluence detection algorithm reaches 88.96%, and the recall rate reaches 88.80%. [ABSTRACT FROM AUTHOR]

An introduction is presented in which the editor discusses articles in the issue on topics including peer-review report, strong empirical evidence and research journals.

Road vector data are an important part of geographic information databases and play a leading role in social and economic development. Among them, the accuracy and current situation of vector data are the key values of their application. In recent years, with the development of remote sensing technology, remote sensing images contain increasingly more road information, which provides a large number of available features for the establishment of vector correction models. Therefore, the vector-to-image vector correction method based on remote sensing images has become an important means to ensure the accuracy and current situation of vector data. However, in actual scenes, there are great differences in the design structure and design index between urban roads and rural roads, which is reflected in the serious spatial heterogeneity between them in the remote sensing images. Therefore, when the existing methods use remote sensing images for vector correction, the universality is limited by the change of regional scenes, so it is difficult to realize the simultaneous correction of urban road and rural road vectors, with poor applicability and low compatibility. To solve this problem, this paper proposes a vector data partition correction method supported by deep learning. First, the U2-net model and line segment sequence method are used to generate the feature set. Second, according to the characteristics of the poor quality of urban road extraction results, the regular geometric form of vector data and the poor structural information of road images, the methods of vector line decomposition, subvector line correction, subvector line centralization and vector line synthesis are proposed to correct the vector lines of urban roads. Finally, according to the characteristics of high-quality rural road extraction, the irregular geometric form of vector data and the strong information of road image structures, this paper proposes road edge extraction, a road centreline reasoning model and a vector line connection method to complete the correction of rural vector road data. The quantitative analysis of the experimental results shows that compared with other methods, the urban road vector correction method in this paper is much better than the comparison methods and has achieved better correction results in rural areas. This method has the advantages of compatibility and better accuracy for roads in different areas. [ABSTRACT FROM AUTHOR]

The imaging spectrometer MOS on IRS-P3 was launched in March 1996 as the first example of a new generation of ocean colour sensors. It consists of three different spectrometers in the visible/near-infrared spectral region with 18 channels. The IRS-P3 mission is focused on the remote sensing of case 2 water, particularly the derivation of different water constituents in coastal waters. Due to the more complex spectral behaviour of case 2 water, a new methodological approach was developed which works directly with satellite measured top-of-atmosphere radiance and accounts for the correlation of the different water constituents as well as for the spectral shape. This paper gives an overview of the mission, the scientific goals and the development and improvement of the retrieval algorithms. The potential of the algorithm is demonstrated and examples of selected European coasts are shown. Derived maps of water constituents are presented. [ABSTRACT FROM AUTHOR]

The present paper proposes an automated approach to estimate the aerosol reflectance at the Advanced Very High Resolution Radiometer (AVHRR) red channel. The aerosol dominant pixels were separated through two orthogonal transforms. The aerosol reflectance ratio at these pixels was estimated through regression. The results are validated with in situ measurements. The retrieved water-leaving reflectance matched the modelled values with a relative error below 45%. The smallest error values were at the stations with the closest sampling time to image acquisition. However, a weak correlation of 16% was found between water-leaving reflectance and aerosol signals. This suggested that these errors could be attributed to the spatial and temporal variability between the two sampling methods (ship measurement and pixel reflectance). [ABSTRACT FROM AUTHOR]

This paper presents a study of the characteristics of extra-tropical oceanic Rossby waves from datasets of Sea Surface Height (SSH), Sea Surface Temperature (SST) and ocean colour. The main focus is on the propagation speed of the waves and a comparison is made between the observational results and the speeds predicted by the classical theory and by the most recent extended theory of Rossby waves. There is also discussion, with an example, of the additional information that can be derived by a comparison of the wave signatures in the different datasets. [ABSTRACT FROM AUTHOR]

Light Detection and Ranging (LiDAR) intensity is associated with the target surface material, which could help the points cloud classification. However, the intensity is also associated with the laser beam incident angle and the transmitting distance, which obstructs its further application in points cloud classification. Motivated by this problem, this paper proposed a practical method for employing the LiDAR intensities in points cloud classification without distance and incident angle calibration, specifically, ratio values between different spectral channels from a newly invented Hyper-spectral LiDAR (HSL) were defined and calculated for generating robust spectral features. Since the HSL different channels had the same transmitting distance and incident angle, therefore, the ratio values were independent on the laser pulse transmitting distance and laser beam incident angle. An indoor experiment was conducted for fully assessing the proposed method. The HSL had eight different spectral channels with spectral wavelength covering from 650 nm to 1000 nm. In the experiments, papers with different colours were pasted on a flat glass; the HSL scanned them at four distinctive positions with 60 cm displacement. The spectral ratio values between different channels at each position were calculated using the obtained multiple spectral profiles from the HSL. The results showed that the points cloud scanned at different incident and distance could be classified though the spectral ratio values without complex distance and incident angle calibration. [ABSTRACT FROM AUTHOR]

The aim of this research is to automatically detect and visualize dynamic ocean colour phenomena such as algae blooms, fronts and eddies from a sequence of cloudy satellite images. The composite front map methodology has been extended to combine feature observations from multiple ocean colour and temperature products in a single map, to explore the interaction between physical and biological oceanic processes. Sample maps showing chlorophyll, sediment and sea surface temperature fronts are presented, derived from long sequences of cloud-affected Advanced Very High Resolution Radiometer (AVHRR) and Sea viewing Wide Field of view Sensor (SeaWiFS) data. The successful detection and animation of many ocean colour features suggests a valuable application to summarize the increased multi-spectral data provided by the National Aeronautics & Space Administration's (NASA) Moderate Resolution Imaging Spectroradiometer (MODIS) instrument. [ABSTRACT FROM AUTHOR]

The development and validation of ocean colour products require the availability of in situ atmospheric and marine data. Since 1995, the Coastal Atmosphere and Sea Time Series (CoASTS) Program ensures the collection of an in situ comprehensive set of measurements taken from an oceanographic tower in the northern Adriatic Sea. An example of a validation exercise is presented through the comparison of in situ and satellite-derived aerosol optical thickness at various wavelengths in the 443-865 nm spectral interval. [ABSTRACT FROM AUTHOR]

A long-term time series of Advanced Very High Resolution Radiometer (AVHRR) (1981-1999) data has been used to assess the main physical features in the Adriatic Sea. Individual images were processed to estimate Sea Surface Temperature (SST) values, to create long-term composite fields (weekly, monthly, seasonal scales), and to derive basic statistics for the Northern, Central and Southern regions, each split again into an Eastern and a Western section. At the basin scale, an apparent general warming trend can be recognized in the time series. The linear fit to the seasonal cycles suggests an increase of about 2°C in 20 years, essentially due to a steady rise of summer values. A general north-south gradient can be found during winter, the Northern sections being colder than the Southern ones. An east-west gradient appears during summer, the Western sections being warmer then their Eastern ones. The Northern Adriatic exhibits substantial fluctuations, possibly linked to the cycle of winter cooling and summer warming in the relatively shallow sub-basin. The North Western section shows larger fluctuations than the North Eastern one, with lower winter SST, probably due to the freshwater inflow from the Po River delta. The Southern Adriatic exhibits less variability, possibly influenced by the periodic water exchanges with the Ionian Sea. The South Eastern section shows somewhat larger fluctuations than the South Western one, with higher winter SST, probably due to the inflow of warmer waters from the south. The two Central sections reveal patterns similar to the ones of the whole basin. The observed temperature patterns appear to follow the classical Adriatic cyclonic circulation scheme. [ABSTRACT FROM AUTHOR]

A combination of careful harmonic analysis and high-resolution numerical modelling of tides in altimeter data from TOPEX/Poseidon and ERS-2 satellites allows detailed calculation of seasonal and interannual variability of sea level in coastal areas of the north-east Pacific Ocean. These calculations show large changes in sea level in marginal seas during the 1997-8 El Niño and La Niña events, as well as providing information on the origin and life cycle of large anticyclonic eddies, observed to be generated off the west coast of Canada. [ABSTRACT FROM AUTHOR]

A sensor installation and evaluation programme at the Institute of Ocean Sciences, Sidney, BC, Canada, shows that the weather buoys operating round Canada's coasts can be used as important new sources of surface information for satellite sensor calibration and validation. The standard instruments on the buoys provide real-time measurements to forecasters--of wind, waves, water and air temperature and air pressure. The present programme is evaluating the addition of sensors measuring chlorophyll fluorescence, Photosynthetically Available Radiation (PAR), salinity and water-leaving radiance. Time-series of these data are also of value for a variety of fisheries, ecosystem and climate studies. Data can be viewed on the web at http://www-sci.pac.dfo-mpo.gc.ca/ecobuoys. [ABSTRACT FROM AUTHOR]

A comparison of Synthetic Aperture Radar (SAR) images acquired by the ERS-2 and RADARSAT-1 satellites with Landsat Thematic Mapper (TM) images acquired close in time is presented. Similarities and differences on the kind of ocean information they provide are discussed. The images were acquired over the Brazil-Malvinas (Falkland) Confluence under different solar elevation and azimuth angles. Ocean features observed by both types of sensors are discussed. These observations show that (1) at high solar angles, sun glint prevails over upwelled water-leaving radiance resulting in optical images that tend to provide dynamical information similar to that obtained from SAR imagery and (2) at low solar zenith and azimuth angles, upwelled radiation tends to prevail over sun glint, providing typical ocean colour and water quality information that can still complement and aid in the interpretation of SAR observations. [ABSTRACT FROM AUTHOR]

In summer 1999, an extensive survey of the tropical Atlantic ocean was operated during the EQUALANT99 cruise. Several experiments were carried out on board including, for the first time in that area, eXpendable Conductivity-Temperature-Depth (XCTD) launches. The aim of this experiment was first to evaluate XCTD temperature and salinity profiles, and secondly to check how XCTD could be used together with altimetric satellite missions in order to capture a good picture of the tropical Atlantic surface circulation. [ABSTRACT FROM AUTHOR]

The technique of vicarious calibration is used in connection with an atmospheric correction to improve the Sea viewing Wide Field of view Sensor (SeaWiFS) normalized water-leaving radiance by the first determination of mean vicarious calibration coefficients from in situ measurements in the Baltic Sea. A necessary adjustment of the SeaWiFS pre-flight calibration slope was found to be +3.5%, +0.3%, -1.7%, -0.4%, +0.8% and -1.3% for the first six SeaWiFS channels. The derived mean vicarious calibration coefficients are higher than the coefficients in the standard SeaWiFS Data Analysis System (SeaDAS) software but with similar shape and good agreement with other research results. The coefficients were used to obtain better normalized water-leaving radiance from SeaWiFS measurements in the Baltic Sea. The deviations of calculated to measured radiances in the open Baltic Sea are between 3% and 47% in the channels 412 to 670 nm, with the trend of higher deviations in the blue channels. The objective of radiance determination in all SeaWiFS channels with a maximum uncertainty of 5% in clear water regions is probably not reachable in the turbid water of the Baltic Sea. [ABSTRACT FROM AUTHOR]

Similar seasonal courses of pigment concentrations were found in all the regions of the Adriatic, from the Coastal Zone Color Scanner (CZCS) data. Maximum pigment concentrations were observed in the late autumn and in early winter. Individual chlorophyll measurements, from particular cruises, correlated well with the pigments retrieved from the satellite data. However, the monthly means, from longer observation periods, did not correspond well and it was found that the pigment values retrieved from CZCS were considerably higher than in situ measured chlorophyll concentrations for these months in the whole of the Adriatic. This is a result of hydro-geo-meteorological factors, namely mixing of the sediments, especially in the shallow northern Adriatic after strong winds, and the influence of coloured substances brought by the rivers. [ABSTRACT FROM AUTHOR]

The bio-optical relationships between inherent and apparent optical properties, and between optical properties and phytoplankton pigment concentration ( C ) averaged in a layer (Δ Z ), were derived from analysis of data collected during the period 1996-1998 in the Gulf of Aqaba (Eilat). Parametrization of these relationships was based on radiative transfer theory, Gershun's equation, minimization of model errors by least-square fitting, and on known optical models relating underwater remote sensed reflectance ( R rsw ) with the ratio of backscattering ( b b ) to vertical attenuation coefficient ( K d ) [or to absorption coefficient ( a )]. These relationships explain a frequently used form of remote sensing algorithms for C estimation using ratio of water-leaving radiances measured at two or more wavelengths ( λ ). In this study, the possibility of using for this purpose a single wavelength in the blue range ( λ =443 nm) within the framework of in situ and remote sensing algorithms for Case 1 waters was assessed. [ABSTRACT FROM AUTHOR]

Internal wave distribution maps have been compiled from more than one hundred ERS-1/2, RADARSAT and Space Shuttle SAR images in the South China Sea (SCS) from 1993 to 2000. Based on these distribution maps, most of the internal waves in the north-east part of the SCS were propagating westward. The wave crest can be as long as 200 km with an amplitude of 100 m, due to strong current from the Kuroshio branching out into the SCS. In a recent SCS internal wave study, moorings were deployed in April 1999 and 2000. Simultaneous SAR coverage from ERS-2 and RADARSAT were also collected. The ERS-2 high-resolution SAR images collected from the Taiwan ground station were processed in near real-time to coordinate the field test. SAR data are then used to compare with mooring data. SAR data and Acoustic Doppler Current Profiler (ADCP) data show internal solitons induced by the semi-diurnal tides with a wave speed of 2.4 m s -1 . [ABSTRACT FROM AUTHOR]

Spectral in-water measurements of the downward irradiance, E d and the upward irradiance, E u are sufficient in order to calculate the following spectral quantities: the irradiance ratio R , the two scalar irradiances E o and E od , the two average cosines μ and μ d , the light absorption coefficient a , the backscattering coefficient b b and the vertical attenuation coefficients for the above-mentioned downward and upward irradiances. The algorithms are valid for horizontally stratified and optically deep sea waters. [ABSTRACT FROM AUTHOR]

A new generation of airborne salinity imager, the Salinity Temperature and Roughness Remote Scanner (STARRS), has been designed and is currently under construction with delivery expected shortly. With both significantly lower noise levels and enhanced capability to measure secondary parameters, including sea surface temperature and roughness, STARRS will serve as both an operational survey tool and as a test-bed for refinement of retrieval algorithms. [ABSTRACT FROM AUTHOR]

The hydrographic conditions in the south-east Atlantic were investigated during the cruise of R/V Poseidon in April 1999. The area of investigation covers the two major eastern boundary currents of the Southern Atlantic and its convergence in the Angola-Benguela Frontal Zone (ABFZ). Hydrographic measurements have been carried out in combination with current measurements and are supplemented with remote sensing data of sea surface temperature (SST) and wind stress. The Angola Current transported in its surface part less saline water from the great rivers in the north towards the ABFZ. At the same time, south of the front, strong coastal upwelling in the Benguela was observed. At the beginning of April 1999 the ABFZ was found at an unusual southern position. The rapid dynamic response of the ABFZ to wind forcing is discussed on the basis of satellite SST images combined with hydrographic measurements. [ABSTRACT FROM AUTHOR]

The Space Information Laboratory (SIL) of the Tropical Center for Earth and Space Studies of the University of Puerto Rico at Mayagüez (UPRM) has been collecting and processing satellite data since December of 1996. Satellite imagery from the Advanced Very High Resolution Radiometer (AVHRR) and the Sea viewing Wide Field of view Sensor (SeaWiFS) provides us with a new understanding of phytoplankton dynamics in the Caribbean region. SeaWiFS shows the intrusion of waters into the eastern Caribbean Sea from the Orinoco River during fall and from the Amazon River during spring-summer. Strong coastal upwelling in Venezuela produced by the trade winds during winter-spring is detected with the AVHRR. The satellite data suggest that these seasonal events may play an important role in phytoplankton fertilization of the eastern Caribbean Sea. SeaWiFS and hydrological data are also combined to evaluate the impact of hurricanes on phytoplankton distribution. The development of models for estimation of ocean primary productivity using SeaWiFS and AVHRR data is now in progress. [ABSTRACT FROM AUTHOR]

Satellite and in situ observations are used to test model dynamics for the California Current System (CCS). The model and data are combined to reconstruct the mesoscale ocean structure during a given three-week period. The resulting physical flow field is used to drive a 3D ecosystem model to interpret SeaWiFS and in situ chlorophyll- a (chl- a ) variations. With this approach a more complete and consistent picture of the physical and ecosystem processes of the CCS is obtained, providing the basis for addressing fundamental questions about dynamics and predictability of the coastal ocean. [ABSTRACT FROM AUTHOR]

The method used to identify sea surface thermal features for estimating surface currents is presented. Results given have been obtained from investigations made in the Central Adriatic Sea off the coast of Monte Conero in the Marche Region, Italy, using Advanced Very High Resolution Radiometer (AVHRR) observations of NOAA-14, made simultaneously with the HF coastal radar system installed there. [ABSTRACT FROM AUTHOR]

The first two years of SeaWiFS (Sea viewing Wide Field of view Sensor) data (1997-1999) are used to document the variability of large-scale surface chlorophyll patterns within the coastal region along the full latitudinal extent of each of the four major global eastern boundary currents; the California, Humboldt, Benguela and Canary Currents. Seasonal chlorophyll patterns are compared to coincident seasonal cycles of Ekman transport calculated from satellite scatterometer data. In all four regions, maximum chlorophyll concentrations are generally temporally and latitudinally coincident with the seasonal maximum in upwelling (offshore Ekman transport) over most of their latitudinal range, but exceptions are documented. Interannual differences are evident in each region, most notably in the two Pacific regions where the 1997-1998 chlorophyll seasonality was affected by El Niño conditions. Significant differences between previously published chlorophyll seasonality deduced from the relatively sparse coverage of the Coastal Zone Color Scanner (CZCS) and the more complete coverage of SeaWiFS in both Southern Hemisphere regions are evident. [ABSTRACT FROM AUTHOR]

The aim of the research reported here is to evaluate Synthetic Aperture Radar (SAR) capability to estimate the wind vector and associated directional wave spectrum. Two ERS-2 SAR images of the Mediterranean Sea, one over the Sicily Channel and one over the Ligurian Sea, were selected as case studies. Wind speed was estimated using SAR calibrated backscatter response, in conjunction with empirically derived ERS scatterometer models such as CMOD4 and CMOD-IFREMER. The predictions of these models were then compared with the actual sea surface wave spectra either provided by in situ measurements or resulting from the inversion of the SAR image spectrum. SAR-detected effects of both wind and wave features, induced either by atmospheric boundary layer instability or by land shadowing, were also used as reliable indicators of wind direction. [ABSTRACT FROM AUTHOR]

The salinity of the open ocean is important for understanding ocean dynamics and for modelling energy exchange with the atmosphere. But existing data are sparse and much of the ocean is unsampled. Sea surface salinity can be measured remotely with passive microwave sensors operating near 1.4 GHz (L-band). Salinity differences have been observed from space and aircraft instruments have demonstrated that salinity can be measured with an accuracy of better than 1 psu. Sensor technology has improved sufficiently to seriously propose a satellite system to map salinity over the open oceans. [ABSTRACT FROM AUTHOR]

The Tropical Rainfall Mapping Mission Microwave Imager (TMI) instrument Sea Surface Temperature (SST) product (v1.0) is compared with in situ observations obtained in the Atlantic Ocean. The TMI SST has a mean warm bias of 0.25 K±0.7 K when compared to in situ SST at a depth of 7 m. When TMI SST are compared to in situ skin SST measurements, the bias is 0.6 K±0.5 K. A limited global comparison between TMI SST and co-incident ERS-2 Along-Track Scanning Radiometer (ATSR/2) skin SST demonstrates a bias of 0.6 K±0.6 K consistent with the result obtained using in situ observations. These results are consistent with the predicted accuracy of the TMI SST data products. Based on these results, a simple method to merge the TMI and ATSR data is proposed. [ABSTRACT FROM AUTHOR]

Synoptic coverage of the temporal and spatial variability of aerosol distribution patterns can only be achieved with satellites. Results from the first ocean colour sensor, the Coastal Zone Color Scanner (CZCS), indicate an annual cycle of the major mineral aerosol plumes that is consistent with the published literature. Seasonality and interannual aerosol variability observed with the CZCS agrees well with that found by ground data measurements and other satellite platforms used to monitor aerosols. The successor to the CZCS--the Sea viewing Wide Field of view Sensor (SeaWiFS)--provides estimates of aerosol load and particle size, both on a global scale. Seasonal maps of both of these aerosol optical properties are in accord with well-known distribution patterns and also with independent satellite estimates. These results indicate that ocean colour sensors are capable of monitoring the variability of global aerosol loads and, more recently, with the retrieval of aerosol particle size, they can be used to characterize different aerosol events. [ABSTRACT FROM AUTHOR]

The 1997-1998 ENSO (El Niño-Southern Oscillation) was not only the largest event of the century but also the most comprehensively observed. Satellite data were employed for ocean colour, sea level, winds, sea surface temperature (SST), and outgoing longwave radiation (OLR) were used to describe the response of the surface marine ecosystem associated with the ENSO event. Some of the large-scale anomalies in ocean colour include elevated biological activity to the north of the Equator in the Pacific coincident with lower sea levels associated with the classic ENSO-horseshoe pattern ecosystem response to the anomalous upwelling in the eastern Indian Ocean caused by the 1997-1998 dipole event, and the dramatic eastward propagating feature in the Equatorial Pacific in response to the La Niña dynamics. Ocean general circulation model (OGCM) experiments show that capturing the high-frequency wind changes is crucial for simulating the La Niña and the coupled biological-physical model (OBGCM) runs clearly show that higher frequency winds are also important for capturing the mean upwelling and nutrient supply into the euphotic zone. Thus, the QuickSCAT winds are expected to play a major role in ecosystem modelling in the future. This study shows the utility of satellite data for understanding not only ocean circulation but also the coupled ecosystem variability. Morcover, it is also shown that spatio-temporal resolution of the satellite winds will directly affect the accuracy of oceanic and ecosystem simulations. [ABSTRACT FROM AUTHOR]

The Clean Seas project focused on the role that existing Earth observing satellites might play in monitoring marine pollution. Results are presented here from August 1997, for the North Sea test site, using sea surface temperature (SST), colour and synthetic aperture radar (SAR) images in conjunction with a hydrodynamic model. There was good correlation between data sources, e.g. between SST and ERS-2 SAR images. Both datasets showed the development of fine plume structures close to the Rhine outflow, apparently associated with the outflow, and possibly caused by tidal pulsing of the Rhine Plume. The model reproduced general temperature and sediment distributions well, but fine structures were not reproduced. Model sediment distribution patterns were verified using 'chlorophyll concentration' data from colour sensors, representative of sediment concentration in turbid water. In conjunction with the visible channels of the Advanced Very High Resolution Radiometer and Along-Track Scanning Radiometer, they give an uncalibrated measure of the sediment load. The model gives a more complete picture of the temporal dispersion of the Rhine Plume over time than is evident from the remotely sensed data alone. [ABSTRACT FROM AUTHOR]

Various oscillatory modes of sea surface temperatures (SSTs) observed over a period of 8.8 years with the NASA Nimbus 7 Scanning Multichannel Microwave Radiometer (SMMR) and for 13 years with the NOAA Advanced Very High Resolution Radiometer (AVHRR), the latter sensing in the thermal infrared band, are described for the Pacific Ocean. The various modes are isolated by a combination of techniques designed also to accommodate non-stationary phenomena. After detrending and removing the seasonal cycle from each grid map element of the data, singular value decomposition (SVD) is used to separate the data into spatial and temporal parts to facilitate the modal analysis. Empirical Mode Decomposition is then used to separate the temporal parts of the data into approximately seven intrinsic modal functions (IMFs) for the temporal parts of the first five principal components (PCs) resulting from the SVD. A filtered time sequence of SST grids is then obtained by selecting IMFs with periods longer than 1.5 years and then reconstructing the SST grid maps from the filtered PCs. The time sequence of SMMR SSTs in the Pacific Ocean shows El Niño Southern Oscillation (ENSO) oscillations not only along the Equator, but also in both the North and South Pacific, with, in fact, even larger amplitudes than along the Equator. A similar analysis was applied to the SST record from the AVHRR instrument. During the period of overlap with the SMMR record, similarities occur in the equatorial region, but the records are by no means identical. The AVHRR SSTs do not show any strong oscillations in the South Pacific. [ABSTRACT FROM AUTHOR]

SeaWiFS (Sea viewing Wide Field of view Sensor) imagery is used to monitor the birth, development and termination of phytoplankton blooms in the north-east Atlantic. Extensive patches of water giving high reflectance of visible light were observed from the beginning of February to the end of July 1998 in the Celtic and Armorican Shelf regions between 45° and 60° N and interpreted as phytoplankton. Eighteen relatively cloud-free SeaWiFS images have been analysed in order to detect the spatial, temporal and spectral development of the blooms within the two study areas. Consistency of the spectral signature and the very high broadband reflectance leads to the conclusion that these are blooms of the widespread coccolithophore Emiliana huxley , although no in situ data during the study period are available to validate this hypothesis. The utility of SeaWiFS data for studying the life history of phytoplankton blooms is confirmed. [ABSTRACT FROM AUTHOR]

Variability of chlorophyll concentration in the ocean is one of the most important components of the primary production process on the planet. To preserve the biosphere and to make appropriate use of it, it is imperative to have a deep insight into the long-term dynamics of the primary production on the planet. To investigate the dynamics of chlorophyll concentration based on satellite data, the Institute of Biophysics (Russian Academy of Sciences, Siberian Branch) has developed a geoinformation system. It was used to detect the areas in the global ocean that are quasistationary in relation to seasonal dynamics of chlorophyll concentration. Areas such as these, found in the Indian Ocean, are described in this work. These areas form the basis for the analysis of long-term dynamics of chlorophyll concentration. In these quasistationary zones systematic monitoring of phytopigment concentration is conducted by space-borne and marine craft. The work presents long-term satellite-based data on the space distribution of anomalous deviations of chlorophyll concentration in the ocean. An anomaly criterion is proposed and maps of Pacific Ocean areas with a high probability of anomalies are presented. [ABSTRACT FROM AUTHOR]

The first images from MODIS (Moderate Resolution Imaging Spectroradiometer) are now being used to evaluate information on the in vivo fluorescence peak near 685 nm from chlorophyll- a , stimulated by sunlight. The Fluorescence Line Imager (FLI) airborne imaging spectrometer was used in the 1980s to demonstrate the mapping of this signal from an aircraft, showing that it gave significant rejection of confusing signals from atmospheric radiance. For imaging from space, a major limitation is sensor sensitivity, which tends to restrict imaging to relatively high concentrations under good solar illumination. Noise-equivalent chlorophyll concentrations for MODIS and MERIS (Medium Resolution Imaging Spectrometer) have been estimated as 0.07 and 0.1 mg m -3 , respectively, under zenith Sun, clear sky conditions. Although MERIS has slightly poorer sensitivity, it has the advantage of flexible band placing and presence of a band at 709 nm in the baseline band-set that allows better definition of the continuum spectrum above which fluorescence is measured. This band should also allow detection of bright plankton blooms (red-tide events) through the peak radiance near 709 nm caused by a combination of in-water scattering and absorption. [ABSTRACT FROM AUTHOR]

ERS-2 SAR images acquired over the Sulu Sea reveal that there are at least three areas where internal solitary waves are generated: (1) at the sill between Dok Kan Island and Pearl Bank; (2) at the sill between Pearl Bank and Talantam Shoal; and (3) at the sill between Talantam Shoal and Sentry Bank. It is observed that the internal solitary waves generated at different source regions merge into a single solitary wave system. When the solitary waves propagate into shallow water, the distance between the solitary waves in a wave packet decreases. Furthermore, when the water depth decreases in the direction of the soliton's crest line, the crest line is bent towards the shallow water region. These observational facts are explained by a wave refraction model which is based on the Korteweg-de Vries equation which is also valid for large amplitude internal solitary waves provided the pycnocline is sufficiently broad. [ABSTRACT FROM AUTHOR]

For the year 1999 all Sea viewing Wide Field of view Sensor (SeaWiFS) scenes of the Danish waters from the North Sea to the Baltic Sea were browsed, and a total of 47 SeaWiFS scenes with reasonably low cloud cover and, therefore, potential in situ match-ups were found and processed. The in situ data used as match-ups were collected on routine monitoring cruises by Danish and Swedish environmental authorities. A few stations in the North Sea, Skagerak and the western Baltic Sea were sampled, while most stations were located in Kattegat and the inner Danish waters. A turbid water SeaWiFS atmospheric correction algorithm was applied, since the standard SeaWiFS algorithm for chlorophyll- a (CHL) has been shown to be fairly inaccurate in turbid coastal waters. This is due to both inaccurate atmospheric and to relatively high and variable abundance of yellow substance. The application of the turbid atmospheric correction substantially improved the SeaWiFS CHL estimates. Regressions between SeaWiFS estimates using the OC2 and OC4 algorithms used in the SeaDAS software (versions 3.3 and 4.0, respectively) and in situ CHL values were made as well, and regression with a number of other possible reflectance ratios with SeaWiFS channels. The best correlation was found to be R 2 =0.54 using a double-ratio algorithm using both R510/R555 and R443/R670, while the OC4v4 had the second best correlation of R 2 =0.39. Among other single ratios, the R510/R555 had the highest correlation with CHL, which was expected since this is also the ratio that OC4v4 most often switches to in the waters investigated here. The range of CHL concentrations in this study was rather limited (all but three points from 0.5-3 mg m -3 ) so there is a need for inclusion of more data to expand the concentration range. This should be possible using also data from 2000, 2001 and onwards and, hereafter, a more 'stable' empirical algorithm can be derived for the Danish waters. [ABSTRACT FROM AUTHOR]

The National Aeronautics & Space Administration (NASA) Sea viewing Wide Field of view Sensor (SeaWiFS) began operational measurement of ocean colour in September 1997. Upgrades to the SeaWiFS data processing system (SeaDAS) have occurred frequently and the effects of these revisions on the remotely sensed estimates of chlorophyll- a concentration (chl- a ) have been significant. Measurements of chl- a from research work in the Bay of Biscay and Gulf of Cadiz during 1998-1999 are used to validate the SeaWiFS chl- a product generated using the current version of SeaDAS (version 4.1). The validation data cover coastal and offshore waters, including those dominated by inorganic suspended sediment, and an intense dinoflagellate bloom where shipboard chl- a measurements exceeded 50 mg m -3 . The standard SeaWiFS chlorophyll algorithm (OC4v4) generally performed well, but significantly over-estimated chl- a where inorganic suspended sediment was present. The algorithm is only applicable to chl- a values up to 64 mg m -3 , which was less than chl- a at the centre of the bloom. A novel algorithm for chl- a , which first estimates the inherent optical properties of the water, was applied to the SeaWiFS measurements but failed on over 90% of the pixels, perhaps because SeaWiFS is under-estimating water reflectance at the extreme blue end of the visible spectrum. [ABSTRACT FROM AUTHOR]

Based on satellite imagery in combination with ground data measurements and numerical modelling, the distribution patterns of water masses along the German Baltic Sea coast of the state of Mecklenburg-Vorpommern (M-V) were investigated. The dynamic processes forming these patterns are driven by the local wind. Distinct wind directions induce typical features in several regions of the M-V coast. The processes considered are the distribution of Oder river water in the Pomeranian Bight, the upwelling off Hiddensee and off the coast of the Mecklenburg Bight including the Warnow river outflow, and the special dynamic regime in Lübeck Bay. In the Pomeranian Bight, a strong relation was found between the wind and transport direction of river water which was also applied to interpret the distribution of inorganic and organic pollutants. During easterly winds, different upwelling cells arise with the highest intensity in the Hiddensee area. The cells merge by increasing wind velocity. Lübeck Bay is mainly excluded from water exchange during the dominant wind directions. The first results of this systematic remote sensing study along the entire coast of M-V are summarized and will be continued to support the coastal monitoring programme of regional authorities. [ABSTRACT FROM AUTHOR]

A large database of in situ bio-optical measurements were collected at the LEO-15 (Long-term Ecosystem Observatory) off the southern coast of New Jersey, USA. The data were used to quantify the impact of coastal upwelling on near-shore bulk apparent (AOP) and inherent (IOP) optical properties. There was good qualitative agreement between the AOPs and IOPs in space and time. The measured IOPs were used as inputs to the Hydrolight radiative transfer model (RTE). Estimated spectral AOPs from the RTE were strongly correlated (generally R 2 >0.80) to measured AOPs. If optical closure between in-water measurements was achieved then the RTE was used to construct the spectral remote sensing reflectance. The modelled remote sensing reflectances were compared to satellite-derived reflectance estimates from four different algorithms. Quantitative agreement between the satellite-measured and in-water modelled remote sensing reflectance was good but results were variable between the different models. The strength of the correlation and spectral consistency was variable with space and time. Correlations were strongest in clear offshore waters and lowest in the near-shore turbid waters. In the near-shore waters, the correlation was strongest for blue wavelengths (400-555 nm) but lower for the red wavelengths of light. [ABSTRACT FROM AUTHOR]

Biological soil crust is an important feature of desert ecosystem composition and surface landscape. Determining the role of biological soil crust in the energy flow and logistics cycle in the desert ecosystem is one of the frontier areas of ecological restoration in arid and semi-arid areas. Obtaining biological soil crust information from drone images is an efficient, fast, and low-cost method. However, due to the scattered and uneven growth of biological soil crusts and the complexity of the field environment, it is difficult to accurately extract biological soil crusts. In view of this, this study used the improved UNet++ model to extract biological soil crusts based on UAV image data. Firstly, the optimal Epoch, Backbone, and Loss function are selected and trained based on the network structure of UNet++ model. Then, the improved UNet++ model proposed in this paper, which takes ResNeXt as the Backbone and Soft Cross-Entropy Loss+Dice Loss as the Loss Function, is obtained. Finally, the test results of UNet++, U-Net, LinkNet, FPN, and PSPNet are compared with the improved UNet++ model in this paper. The results showed that the improved UNet++ model had the best segmentation effect, and the precision, recall, F1-Score, and IoU were 0.9788, 0.9501, 0.9495, and 0.9309, respectively. UAV image biological soil crust recognition based on the improved UNet++ model in this paper can obtain high-precision extraction results, provide good data support for studying the development of biological soil crusts in arid areas, and provide a new method for precise segmentation of different features in arid areas. In particular, it is of great significance to evaluate the governance effect of ecological restoration projects. [ABSTRACT FROM AUTHOR]

Remote sensing images have been widely used in various fields due to the great success of deep learning. However, as the basis of other tasks, remote sensing image object detection is still challenging. In this paper, we propose a Spatial-Coordinate Attention (SCA) and Multi-Path Residual Block (MPRB) guided oriented object detection algorithm based on an extended ResNet-18 backbone network. For specially, we first modify the coordinate attention module and spatial attention module through a multi-branch manner, and then combine them to generate the new Spatial-Coordinate Attention module. In this module, the channel information, location information and spatial information are integrated to enhance the representative capability of the object. After this, in order to further improve the feature representative power of a shallow CNN, such as ResNet-18. We increase the residual connection in the Residual Block and designed a Multi-Path Residual Block to make it more powerful. Moreover, we also introduce a new skip-connection architecture into the upsampling process to ensure that the predicted feature map can integrate high-level and low-level semantic information. We conduct experiments on the optical remote sensing image dataset UCAS-AOD, HRSC2016 and the SAR aerial dataset SSDD. Experiments show the effectiveness and feasibility of the algorithm in this paper. [ABSTRACT FROM AUTHOR]

Cloud detection is a key pre-processing step for optical satellite remote sensing images. Convolutional neural networks (CNNs) are the most widely used and powerful cloud detection techniques in recent years, featuring the fusion of different levels of feature maps. However, existing methods use feature cascading, element addition, or element multiplication to fuse directly, ignoring the similarities and differences between features and their contribution to the detected cloud regions, resulting in a large amount of noise being introduced into the fusion process. To solve this problem, this paper proposes a feature-aware aggregation network (FAANet), which can make full use of valuable information for cloud detection through feature enhancement and feature aggregation. The feature enhancement module (FEM) consists of a high-level feature enhancement module (HFEM) and a low-level feature enhancement module (LFEM). HFEM is used to enhance the advanced semantic features to supplement the cloud features lost during the recovery of high-resolution feature maps; LFEM is used to enhance the image spatial detail features to further refine the predicted cloud maps. In addition, a feature aggregation module (FAM) is designed to selectively aggregate coded block features, decoded block features, and enhanced semantic features through a new aggregation mechanism to compliment the features diluted by up-sampling and improve the integrity of the generated cloud maps. Experimental results on three open-source cloud detection datasets and self-built Landsat-8 cloud detection dataset with higher spatial resolution show that the proposed method achieves competitive performance under different evaluation metrics. The advantages of the method proposed in this paper for saliency image segmentation are verified and provide a reference for other computer vision tasks. Code is available at . [ABSTRACT FROM AUTHOR]

Synthetic aperture radar (SAR) has all-weather and all-day observation capabilities and a certain ability to penetrate the surface; therefore, it has unique advantages in many aspects, which other remote sensing methods cannot achieve. However, owing to imaging principles, the interpretation of SAR images is complicated. Therefore, converting SAR images into optical remote sensing images is one method of SAR image interpretation. Therefore, this paper proposes an improved conditional generative adversarial network (cGAN) that includes a generator based on the encoder-decoder structure of the swin transformer feature extraction module. The generator uses a spatial pyramid structure and multi-scale depth feature extraction structure to extract the SAR image features at different scales. A multi-scale discriminator was used to discriminate images on different scales. The introduction of feature matching loss makes the discriminator's feedback to the generator richer and uses perceptual loss based on VGGNet to evaluate the image quality in depth. A series of experiments verified the effectiveness of the method in this study and proved the potential of the transformer structure in the field of SAR image translation. The method presented in this paper has obvious advantages and can translate SAR images into optical remote sensing images. [ABSTRACT FROM AUTHOR]

Seagrass beds play a vital ecological role by regulating climate, maintaining biodiversity, and sequestering carbon. Achieving accurate classification of seagrass beds across both time and space is of significant importance for efficient monitoring. This paper addresses the challenges of seagrass bed classification in remote sensing imagery and the difficulty of achieving temporal and spatial transferability of classification model. Leveraging a long-term Landsat satellite image time series spanning from 1976 to 2022 in the Yellow River Delta, we enhance the DeepLab V3+ network and propose a cross-temporal-spatial domain seagrass bed transfer learning classification approach. We utilize selectively chosen images from a few years in partial regions for training, introduce freeze-training and-unfreeze-training during the model training phase, and then transfer the model to achieve cross-temporal and cross-temporal-spatial domain classification on images from other years and regions. Additionally, we discuss the factors influencing the classification results. The research results indicate that: (1)The method we proposed has shown good results, achieving classification accuracy of 83.17% and 80.99% for seagrass beds in both cross-temporal and cross-temporal-spatial domains. (2) The classification results indicate that this method achieves an average classification accuracy higher by 16.18% compared to networks such as UNet, PSPNet, HRNet, UNet++, SegFormer, etc. under cross-temporal conditions and higher by 17.56% under cross-spatiotemporal conditions for seagrass bed classification. (3) Based on the classification results, it is observed that from 1973 to 1986, the overall area of seagrass beds in the Yellow River Delta had an initial growth trend, with the largest seagrass bed area recorded in 1986, reaching 4281.289 hectares. Subsequently, the seagrass beds showed an overall decreasing trend, and by 2019 they had almost completely disappeared. The cross-temporal-spatial domain seagrass bed transfer learning classification method proposed in this paper can provide valuable support for the monitoring, restoration, management, and utilization of seagrass ecosystems. [ABSTRACT FROM AUTHOR]

Classification of hyperspectral image (HSI) is an essential step in various remote sensing tasks. Currently, transformer-based models have made significant contributions to HSI classification due to their remarkable ability to model the relation among long-range sequence data. However, HSI data exhibit strong nonlinear coupling and strong correlation, leading to a decrease in the classification performance of transformer-based models. The SpectralFormer network has achieved excellent performance in HSI classification, but its cross-layer adaptive fusion mechanism cannot effectively highlight critical information, affecting the representation ability of spectral-spatial features. To address the above issues, this paper proposes an HSI classification method based on local feature decoupling and SpectralFormer with a hybrid attention module (HAMSF) network. Specifically, this paper first uses the histogram of oriented gradient algorithm to preprocess HSI coupled data and achieve preliminary nonlinear decoupling of HSI data. Secondly, the hybrid attention module is embedded in a cross-layer adaptive fusion mechanism, generating refined neighbouring information before each information fusion to improve the spectral-spatial feature representation ability. Finally, the processed data are input into the HAMSF network for the final decoupling and classification of HSI. The experiment is carried out on three benchmark hyperspectral datasets, and the experimental results confirm that the proposed method has better classification performance compared to several transformer-based and classical HSI classification methods. [ABSTRACT FROM AUTHOR]

Initialization plays an important role in the accuracy of endmember extraction algorithms (EEAs) in linear hyperspectral unmixing (LHU). Random initialization can lead to varying endmembers generated by EEAs. To address this challenge, an initialization strategy has been introduced, encompassing vertex component analysis (VCA), automatic target generation process (ATGP), among others. These techniques significantly contribute to enhancing the accuracy of EEAs. However, complex initialization is sometimes less preferable, prompting the unexplored question of whether there exists an EEA robust to initialization. This paper focuses on analyzing this issue within the context of minimum simplex volume-based (MV) methods, which have received considerable attention in the past two decades due to their robustness against the absence of pure pixels. MV methods typically formulate LHU as an optimization problem, most of which includes a non-convex volume term. Additionally, many MV methods use VCA as an initialization strategy. Firstly, this paper demonstrates that the variable splitting augmented Lagrangian approach (SISAL), as a representative non-convex MV method, heavily depends on initialization. To our knowledge, the impact of initialization for MV methods has not been thoroughly analyzed before. Furthermore, this paper proposes an initially robust MV method by introducing a new convex MV term. Numerical experiments conducted on simulated and real datasets demonstrate its outstanding performance in accuracy and robustness to initialization. Throughout the experiments the proposed method proves to be the most stable, which is crucial in real scene where the ground truth is unknown beforehand. [ABSTRACT FROM AUTHOR]

Classification with a few samples of training set has been a long-standing issue in the field of polarimetric synthetic aperture radar (PolSAR) image analysis and processing. In fact, one of the most important challenges of the PolSAR image classification task is the number of labelled samples. In essence, training network by utilizing just few numbers of training samples cannot be led to the reliable result since the neural network is so sensitive to the number of training samples. In addition, in the real PolSAR image classification task, the huge number of training samples is not accessible. On the other hand, classification performance using just small number of labelled samples is not an accurate result. So, in this paper, aiming at the small number of training samples of the PolSAR image classification task, a novel ensemble self-supervised feature-learning (ESSFL) model is designed. The designed ESSFL can automatically extract PolSAR features conducive to PolSAR image classification with a small number of training samples. In addition, it can significantly decrease the dependence of neural network algorithms on large labelled samples of training set. First, to utilize the spatial–polarimetric features of PolSAR data perfectly, the EfficientNet-B0 is presented and utilized as the main section of the deep learning (DL) model to extract DL features of PolSAR data. Then, using an optimization function that constrains the cross-correlation matrix of various distortions of each sample to the identity matrix, the designed deep learning model can obtain the effective features of homogeneous samples gathering together and heterogeneous samples separating from each other in a self-supervised manner. Moreover, two ensemble learning (EL) models, feature-level and view-level ensemble, are proposed to increase the feature extraction capability and classification result by jointly using spatial features at different scales and polarimetric information at different bands. Finally, the stack of the obtained features and the main polarimetric information of PolSAR data can be input into classifier for the classification of PolSAR data. In this paper we use two different classifiers, the first one is the Random Forest (RF) classifier and the second one is the Support Vector Machine (SVM) classifier. The advantages of SVM include that it can be used to avoid the difficulties of using linear functions in the high-dimensional feature space, and the optimization problem is transformed into dual convex quadratic programs that are so usable for an actual image classification task. In addition, among all the available classification methods, random forests provide one of the highest accuracies. The random forest technique can also handle big data with numerous variables running into thousands. In fact, it can automatically balance data sets when a class is more infrequent than other classes in the data that is so important for a real PolSAR image classification task. Experimental results on three well-known PolSAR data sets illustrate that the designed ESSFL can extract more discriminant features using the designed deep learning model. In the end, the experiments prove that the designed ESSFL has a significant classification performance compared with different deep learning models in the case of small number of training samples, and also it can achieve a better result in the case of large number of training samples by comparison with the most of the deep learning models. [ABSTRACT FROM AUTHOR]