We consider the coherent detection problem for maneuvering targets with jerk motion. The range migration and Doppler frequency migration are addressed within the coherent interval. An efficient coherent integration method based on time reversal transform (TRT) and scaled nonuniform fast Fourier transform (SNuFFT) is proposed. First, the effects of velocity and jerk are eliminated by multiplying the signals before and after TRT operation. Then, the SNuFFT is newly defined to achieve coherent integration. In comparison with several existing methods, the proposed method has much lower computational complexity by employing fast Fourier transform and nonuniform fast Fourier transform, which guarantees the real-time processing. Simulations and real data processing results validate the effectiveness of the proposed method.

We present the scatterometer-derived snowmelt and freeze variability data for three river basins in the Indian Himalayas (the third pole)—knowledge about which is important for studying the climate and water resources of the area. We document the timing and duration of snowmelt/freeze across the Indian Himalayas. The analysis includes data from QuikSCAT (2000 to 2009), Oceansat Scatterometer (OSCAT) (2010 to 2013), Advanced Scatterometer (ASCAT) (2014 to 2016), and SCATSAT (2017). We utilize Ku-band HH-polarized data from QuikSCAT, OSCAT, and SCATSAT and C-band VV-polarized data from ASCAT and use the adaptive/variable threshold method based on winter mean and standard deviation. The analysis is based on the sensitivity of the backscatter coefficient to the presence of liquid water content in the snow caused due to melting. Past air temperature data available for seven Himalayan stations are used to validate the methodology. The melt/freeze detection is limited to areas where snow is available in that year. The majority of melt onset in the study area takes place in April. Melt onset for higher-elevation zones for the Indus, Ganges, and Brahmaputra basins varies from April 19 to May 20, April 17 to April 30, and April 7 to April 28, respectively. Major freeze onset for these basins is in October. The outputs generated from the analysis will be available on Bhuvan portal under the National Information System for Climate and Environment Studies program of Indian Space Research Organisation.

Advanced spaceborne remote sensing techniques are currently available for the generation of high-resolution digital elevation models (DEMs) aiding in the advancement of Earth’s topographic modeling. DEM assimilation using data-driven models is a technique that can effectively overcome the limitations of individual DEMs generated from different techniques and sources of remote sensing data. The assimilation of DEMs generated from optical stereo pairs (Cartosat-1) and InSAR pair (ALOS PALSAR-1) was done for improving the DEM quality. Further, by using data assimilation techniques, the openly accessible DEM products derived from remote sensing datasets were used for generation of improved high-quality DEMs. The feature level fusion and the Kalman optimal interpolation techniques were used for the assimilation of these DEMs. Three sites, namely Dehradun, Jaipur, and Kendrapara, were selected on basis of topographic conditions, i.e., rugged, moderate, and gentle. RMSEs calculated for Dehradun and Jaipur have shown significant improvement in assimilated DEMs. Assimilation of CartoDEM V3 R1 and ALOS PALSAR RTC HR DEMs at the Jaipur site has resulted in highly improved RMSE in elevation of 1.22 m using the Kalman optimum interpolation model. However, in case of the plain region of Kendrapara, the CartoDEM V3 R1 data are found to have the highest accuracy of 2.00 m, indicating importance of selective use of input DEMs and assimilation methods.

We develop a vegetation scattering model to eliminate the effect of vegetation and surface roughness on the radar signal. The canopy water content is an important variable associated with the scattering effect of vegetation, and it can be calculated using the leaf area index, which is retrieved from PROSAILH optical model based on Landsat-8 images. The scattering model introduced the direct scattering contribution of underlying ground into the water cloud model. The experimental correlation length was replaced by a fitting parameter from C-band RADSARSAT-2 radar data to calculate the scattering contribution of underlying ground. Results demonstrate that the vegetation scattering model has a good performance in soil moisture retrieval with R² of 0.805 and root-mean-square error of 0.039  m³  ·  m^−  3. The applicability and capability of the scattering model will provide the operational potential of C-band radar data for soil moisture retrieval in an agricultural region.

The rapid development of remote sensing technologies provides interesting possibilities for the further development of nationwide mapping procedures that are currently based mainly on passive aerial images. In particular, we assume that there is a large undiscovered potential in multitemporal airborne laser scanning (ALS) for topographic mapping. In this study, automated change detection from multitemporal multispectral ALS data was tested for the first time. The results showed that direct comparisons between height and intensity data from different dates reveal even small changes related to the development of a suburban area. A major challenge in future work is to link the changes with objects that are interesting in map production. In order to effectively utilize multisource remotely sensed data in mapping in the future, we also investigated the potential of satellite images and ground-based data to complement multispectral ALS. A method for continuous change monitoring from a time series of Sentinel-2 satellite images was developed and tested. Finally, a high-density point cloud was acquired with terrestrial mobile laser scanning and automatically classified into four classes. The results were compared with the ALS data, and the possible roles of the different data sources in a future map updating process were discussed.

The importance of lake water surface temperature has long been highlighted for ecological and hydrological studies as well as for water quality management. In the absence of regular field observations, satellite remote sensing has been recognized as a cost-effective way to monitor water surface temperature on large spatial and temporal scales. The thermal infrared sensors (TIRS) onboard of Landsat satellites (since 1984) are adequate tools for monitoring surface temperature of small to medium sized lakes with a biweekly frequency, as well as for performing retrospective analysis. Nonetheless, the satellite data have to deal with effects due to the atmosphere so that several approaches to correct for atmospheric contributions have been proposed. Among these are: (i) the radiative transfer equation (RTE); (ii) a single-channel algorithm that depends on water vapor content and emissivity (SC1); (iii) its improved version including air temperature (SC2); and (iv) a monowindow (MW) algorithm that requires emissivity, atmospheric transmissivity, and effective mean atmospheric temperature. We aim to evaluate these four approaches in a river dammed reservoir with a size of 12  km² using data gathered from the band 10 of the TIRS onboard of Landsat 8. Satellite-derived temperatures were then compared to in situ data acquired from thermistors at the time of Landsat 8 overpasses. All approaches showed a good performance, with the SC1 algorithm yielding the lowest root mean square error (0.73 K), followed by the SC2 method (0.89 K), the RTE (0.94 K), and then the MW algorithm (1.23 K). Based on the validation results, we then applied the SC1 algorithm to Landsat 4, 5, and 8 thermal data (1984 to 2018) to extend data series to past years. These data do not reveal any warming trend of the reservoir surface temperature. The results of this study also confirm how the 100-m spatial resolution of TIRS is valuable as an additional source of data to field-based monitoring.

The constant false alarm rate (CFAR) detector is a classical algorithm for ship detection with synthetic aperture radar (SAR). However, the algorithm is susceptible to the accuracy of sea clutter modeling and the desired probability of false alarm, thus reducing detection performance. Therefore, a goodness-of-fit test and a certain number of ship detection experiments and theoretical analysis of false alarms have been extensively practiced as prior knowledge. Compared to earlier SAR sensors, the newly launched Sentinel-1 has nearly uniform signal-to-noise ratio, and the distributed-target-ambiguity ratio may provide additional capabilities for ship detection. Owing to the complex interaction between SAR system and sea surface, the previous work may not be completely suitable for Sentinel-1. As its application is in the beginning, further research is needed. We evaluate the effectiveness of model fitting among five commonly used distributions, and the influences of incident angle, polarization, and sea state on modeling are analyzed. In addition, CFAR detectors constructed by these distributions carried out the ship detection experiments. Moreover, the false alarms that are inevitably caused during the ship detection are classified and statistically analyzed. These aforementioned works can provide an important reference for Sentinel-1 to implement large-scale, as well as long-term, ship detection activities and for further improvement.

The aim of this study is to investigate the performance of 15 automatic thresholding methods, namely Huang and Wang’s fuzzy thresholding method, intermode thresholding method, isodata thresholding method, Li and Tam’s thresholding method, maximum entropy thresholding method, mean thresholding method, minimum error thresholding method, minimum thresholding method, moment-preserving thresholding method, Otsu’s thresholding method, percentile (p-tile) thresholding method, Renyi’s entropy thresholding method, Shanbhag’s thresholding method, triangle thresholding method, and Yen’s thresholding method, for mapping water body using Sentinel-2 data based on normalized difference water index. Three different types of surface water bodies, such as a natural lake (Lake Burdur), a dam reservoir (Aslantaş Dam Reservoir), and a part of a river (Aras River), are chosen to reveal the potential of 15 thresholding methods. The reference water body maps of each test site were generated by manual digitization of high-resolution Google Earth images. The thresholding methods were assessed using the statistical measures, namely overall accuracy (OA), Kappa coefficient, producer’s accuracy, user’s accuracy, and misclassification error (ME). The accuracy analyses of 15 thresholding methods were carried out separately for each test site, and then the overall accuracies were calculated to determine the best method. The obtained OA results showed that minimum thresholding method was the best method among these 15 algorithms with 0.0008 ME, 99.92% OA, and 0.9758 Kappa coefficient. On the other hand, Shanbhag’s method provided the lowest overall accuracies as 0.3133 ME, 68.67% OA, and 0.3190 Kappa coefficient.

Investigating time series variations of total precipitable water (TPW) vapor is of great importance in meteorology and hydrology. The objective of this study is to use two nonparametric tests including Mann–Kendall and Sen’s slope estimator to investigate the annual and seasonal trend of TPW changes using two remote sensing datasets and a ground dataset in western Iran. The daily near-infrared product of TPW from Moderate Resolution Imaging Spectroradiometer (MODIS), MOD05, and a daily TPW extracted from a developed regional TPW MODIS algorithm were considered as remote sensing data. The remote sensing estimations were utilized in six climatic zones, including hyperarid, arid, semiarid, semihumid, humid, and hyperhumid. For the ground data, the daily TPW was prepared from five radiosonde stations. At a significance level of 0.05, in the annual trend, MOD05 results showed a significant increase in the hyperarid, semiarid, semihumid, and hyperhumid climates, and regional MODIS TPW algorithm results showed a significant increase in all six climates. In the seasonal trend of spring in the Ahvaz radiosonde station, as well as in the seasonal trend of summer of MOD05 results in the hyperarid, arid, semihumid, and humid climates, there was a significant increase. Regional MODIS TPW algorithm results demonstrated a significant increase trend in all six climates in the spring and summer. Generally, the results proved the TPW incremental trend in the study area.

Advancing the sustainable use and conservation of marine environments is urgent. Tons of debris including macro- and microplastics generated on land are entering the oceans, marine resources are decreasing, and many species are facing extinction. Though satellite remote sensing techniques are commonly used for global environmental monitoring, it is still difficult to detect small objects such as floating debris on the vast ocean surface, and the ecosystems deep in the oceans where light does not reach are unobservable. An autonomous monitoring system consisting of optimally controlled robots is required for acquiring spatiotemporally rich marine data. However, object detection in marine environments, which is a necessary function the robots should have for underwater and aerial monitoring, has not been extensively studied. Here, we argue that state-of-the-art deep-learning-based object detection works well for monitoring underwater ecosystems and marine debris. We found that by using the deep-learning object-detection algorithm YOLO v3, underwater sea life and debris floating on the ocean surface can be detected with mean average precision of 69.6% and 77.2%, respectively. We anticipate our results to be a starting point for developing tools for enabling safe and precise acquisition of marine data to elucidate and utilize this last frontier.

In recent years, land cover classification technology based on multispectral remote sensing images has been applied to many fields of environmental monitoring. The existing methods generally rely on the information of spectral bands. However, the spectral information of monotemporal multispectral images cannot be generalized to describe the spectral characteristics of the ground objects at different times. In addition, the time-series samples will slightly change over time, and it is difficult to maintain classification performance continuously. To address these problems, we propose a method of land cover classification for multispectral images based on the time-spectrum association feature and multikernel boosting incremental learning. Our method is conducted in two main stages. (1) We propose the time-spectrum association features to acquire the seasonal spectral characteristics different ground objects. (2) To design the classifiers, we propose multikernel boosting method and introduce a multikernel boosting classification learning, which uses continuous new samples to update the weights of the classifier by low-computational and small-scale incremental learning. We test the proposed method on a public multispectral dataset from Landsat-5. The experimental results show that the extracted time-spectrum association features can better characterize the differences of different ground objects, and the proposed classifier can reach more accurate classification with gradually increasing samples over time.

RapidEye multispectral imager is the first satellite that consists of red-edge (RE) band. This study aims to analyze the effect of incorporating RE band information on land use land cover (LULC) classes. It further investigates the impact of integrating the most common vegetation indices: normalized difference vegetation index (NDVI) and its RE adaption, i.e., NDVI-RE in the classification process and sensitiveness of RE band information on LULC classes. In addition, this study also examines the potential of an advance ensemble technique, i.e., extreme gradient boosting (XGBoost) in comparison to the two state-of-the-art machine learning algorithms, random forest (RF) and support vector machine (SVM). A systematic comparison is performed using machine learning classifiers in case of inclusion and exclusion of RE and inclusion of NDVI versus NDVI-RE. Results show that inclusion of RE band improves accuracy of classification by +2.89  %  , +3.49  %  , and +3.03  %   over without RE using XGBoost, RF, and SVM, respectively. The results obtained using all classifiers confirmed the effectiveness of NDVI-RE over NDVI for LULC classification. However, XGBoost outperformed RF and SVM by achieving highest overall accuracy of 92.41% when NDVI-RE is included as an input feature. For class-specific performance, incorporation of RE shows a significant rise in accuracy of all vegetation classes, whereas the inclusion of NDVI-RE reported a maximum increase in the accuracy of shrubland. Furthermore, results clearly indicated that XGBoost has great potential, and it may be beneficial for more complex classification problems.

Characterization of polarimetric backscattering phenomenon is investigated through the ground-based synthetic aperture radar (GB-SAR) data that are collected from a test site consisting of manmade targets distributed over a vegetated terrain at L-band frequencies. The characterization of either synthetic or natural targets has been made by applying two main polarization interpretation schemes, namely amplitude-based interpretation and target-based interpretation in extracting the features of those targets. Polarimetric backscattering signatures of objects are analyzed by assessing the Pauli, total power, entropy/mean-alpha (H  /    α¯  ) images of the terrain for the goal of identification and classification of scattering mechanisms. After applying the classification methodology presented, obtained polarimetric images have demonstrated that target features can be effectively discriminated from each other providing a successful characterization of natural and manmade objects based on GB-SAR measurements. Specifically, H  /    α¯   classification results are shown to be well capable of clearly identifying the distinct scattering mechanisms of the terrain. Full-polarimetric measurements of this particular scene confirmed the ability to retrieve the physical target features to a certain extent using the high-resolution GB-SAR imagery and the relevant polarimetric analyses.

Feasibility of precipitable water vapor (PWV) retrieval from the ground-based infrared measurements in clear sky conditions was performed based on the radiative transfer simulation and retrieval experiments. The effects of aerosol, view zenith angle, and instrument spectral response function (SRF) on clear-sky infrared brightness temperature (T_sky) were analyzed. The results showed that atmospheric aerosol and SRF have obvious influence on T_sky measurements and PWV retrievals. The relationship between T_sky and PWV under low aerosol loading conditions is better than that of high aerosol loading conditions. Aerosol information is necessary for the inversion of high-precision PWV using a single-angle T_sky measurement. T_sky at the infrared atmospheric window (i.e., 10 to 12  μm) has a better exponential relationship with PWV than that covers 6.3  μm water vapor and 9.6  μm ozone absorption bands. Furthermore, a neural network (NN)-based PWV retrieval algorithm was proposed using dual-angle T_sky measurements and near-surface air temperature (T_air). The results showed that the introduction of multiangle T_sky can effectively reduce the influence of aerosol on PWV retrieval and improve the PWV retrieval accuracy. The determination coefficient, root-mean-square error, and bias of the NN model using dual-angle T_sky (i.e., 0 deg and 30 deg) and T_air were 0.989, 0.191, and 0.002 cm, respectively.

Sugarcane lodging monitoring in large areas is a fundamental requirement for damage reduction, yield prediction, and loss evaluation. The polarimetric features extracted from polarimetric synthetic radar aperture have proved to be sensitive to wheat lodging, but for sugarcane lodging monitoring, there are three influences that make it more complicated: heterogeneity of sugarcane scattering, variations of sugarcane growing conditions, and sugarcane growth. A process was proposed to identify sugarcane lodging. First, object-based image segmentation was applied to extract the homogeneous sugarcane fields. Complex Wishart classifier was used to identify sugarcane fields from other land cover types, and 18 polarimetric features were extracted from six consecutive RADARSAT-2 datasets. Then, monthly variations of the polarimetric features were used to overcome the influence of variations of sugarcane growing conditions. We found that it was impossible to identify the lodged sugarcane based on the data from a single date. Finally, the influence of sugarcane growth was reduced by finding the polarimetric features, for which the monthly variations before and after lodging happened were opposite. This study indicated a great potential capability of C-band RADARSAT-2 data for monitoring sugarcane lodging.

Earthquake disasters and postdisaster reconstruction have profound impacts on human society. We use time series nighttime light images collected by the defense meteorological satellite program-operational linescan system sensors as a proxy data source for human activities (HAs). After calibration, a time series analysis method is used to study the distribution and intensity of the changes in HAs after an earthquake. We consider the Haiti earthquake an example to analyze the HA sequence patterns and the aggregation patterns of the HA centers. The results show the following: (1) postearthquake recovery and reconstruction efforts led to increases in the long-term HAs, but the level of increased HA was usually low. (2) The level of postearthquake HAs in the most affected areas (seismic intensity zone VIII+) increased, whereas the postearthquake HAs in severely affected areas (seismic intensity zones VI and VII) were more fragmented. (3) The recovery of HAs in seismic intensity zone VIII+ (mainly Port-au-Prince) required 2 years, but the actual time may be far longer.

A segment of the field of precision agriculture is being developed to accurately and quickly map the location of herbicide-resistant and herbicide-susceptible weeds using advanced optics and computer algorithms. In our previous paper, we classified herbicide-susceptible and herbicide-resistant kochia [Bassia scoparia (L.) Schrad.] using ground-based hyperspectral imaging and a support vector machine learning algorithm, achieving classification accuracies of up to 80%. In our current work, we imaged kochia along with marestail (also called horseweed) [Conyza canadensis (L.) Cronquist] and common lambsquarters (Chenopodium album L.) and the crops barley, corn, dry pea, garbanzo, lentils, pinto bean, safflower, and sugar beet, all of which were grown at the Southern Agricultural Research Center in Huntley, Montana. These plants were imaged using both ground-based and drone-based hyperspectral imagers and were classified using a neural network machine learning algorithm. Depending on what plants were imaged, the age of the plants, and lighting conditions, the classification accuracies ranged from 77% to 99% for spectra acquired on our ground-based imaging platform and from 25% to 79% on our drone-based platform. These accuracies were generally highest when imaging younger plants.

We utilize a tomography-based persistent scatterer interferometric synthetic aperture radar (Tomo-PSInSAR) approach for monitoring the deformation performance and analysis of the thermal expansion of 33 high-rise buildings in Shanghai Lujiazui. For the purpose of this study, we use 31 stripmap acquisitions from TerraSAR-X missions spanning from December 2009 to February 2013. Considering that thermal expansion, concrete creep, and shrinkage are three long-term movements that occur in high-rise buildings with concrete structures, we use an extended 4-D SAR phase model, and three parameters (height, deformation velocity, and thermal amplitude) are estimated simultaneously. Additionally, we apply a two-tier network strategy to detect single and double PSs with no need for preliminary removal of the atmospheric phase screen in the study area. The thermal amplitude map demonstrates that the derived thermal amplitude of the buildings is highly related to the building height due to the upward accumulative effect of thermal expansion. The deformation velocity map reveals that the Shanghai World Financial Center is subject to a great deal of deformation during the newly built period due to concrete creep and shrinkage effects, which are height-dependent downward movements. In addition, the deformation rates caused by concrete creep and shrinkage are largest at the beginning and gradually decrease, finally achieving a steady state as time goes to infinity. In contrast, the deformation velocity map shows that the Jin Mao Tower is almost stable during this period. We observe a negative correlation between deformation characteristics and the time of completion of high-rise buildings, inferring that the newly built concrete buildings are more sensitive to concrete creep and shrinkage effects than the older concrete buildings. Our study emphasizes the potential of the Tomo-PSInSAR solution for monitoring the deformation performance and analysis of the thermal expansion of high-rise buildings.

Based on the optimized random forest (ORF) model proposed in our research, data of the Landsat-OLI, Landsat-TM images, and digital elevation model were used by us to obtain the land cover maps during several periods in the Napahai Basin. Model performance testing results show that producer accuracy and kappa coefficient of the ORF approach are 0.916 and 0.903, which are higher than the traditional maximum likelihood method. Furthermore, we analyzed the characteristic of land cover dynamic change and seasonal variation of wetlands landscape in 2002 and 2017, and revealed their driving mechanisms and ecological response process. The conclusions are as follows: (1) Lake area expanded in the past 15 years because of increased glacial runoff generated from global warming. Construction land and farmland replaced wetland areas in the middle and downstream regions. Moreover, due to the limitation of hydrothermal conditions in the canyon, few forests in the northeastern basin transferred to shrubs. (2) Dam, which was constructed in northern Napahai lake, controls water storage and made the lake area stable after 2013, so, area change percentage in 2017 from wet to dry season is lower than 2002. The wetland region in the central and downstream basin was synthetically affected by climate change and drainage projects, so its environment condition becomes dryer, and vegetation communities’ succession shows a reversed process. (3) Ecological problems in the Napahai Basin were mainly reflected at aspects of agricultural nonpoint source pollution, overgrazing, and urbanization. These factors affected lake, soil, and groundwater in the Napahai Basin and destroyed the ecological health of wetland landscapes. Therefore, in order to realize sustainable development environmental resources, local government should propose the effective policies to restore the degraded ecosystem and protect natural wetland.

Southeast Asia (SEA) has the largest mangrove forest area in the world, which plays an important role in the global carbon cycle and is helping to mitigate climate change. In order to manage the mangrove forests in SEA, their total biomass needs to be determined. However, development of a biomass dataset based on field survey is time consuming. An aboveground biomass (AGB) dataset of mangrove forests was developed for SEA based on ALOS PALSAR 25-m mosaic. Specifically, ALOS-PALSAR 25-m images were first retrieved for SEA from the Kyoto and Carbon Initiative projects and then converted from a digital number to a normalized radar cross-section format in decibels. Samples of mangrove forests in SEA were collected as regions of interest from ALOS PALSAR data based on visual interpretation using Landsat data and Google Earth imagery. A rule-based classification method based on mangrove backscattering characteristics was then used to classify mangroves and nonmangroves in the region. Subsequently, an empirical model was adopted to estimate the AGB of the mangrove forests and an AGB dataset was developed. The results indicate that the spatial distribution of mangrove forests over SEA is 5.1 million hectares, and the estimated average AGB is 140.5  ±  136.1  Mg  /  ha.

The current remote sensing systems designed to measure soil moisture have a relatively coarse spatial resolution ranging from 25 to 50 km. The Advanced Microwave Scanning Radiometer 2 (AMSR2) is a passive sensor that measures soil moisture through C-band (6.9 and 7.3 GHz) observation of brightness temperature (BT). AMSR2 uses land parameter retrieval model to retrieve surface soil moisture and vegetation optical depth. This model partitions the microwave observation into its respective soil and vegetation emission components. AMSR2 loses sensitivity to soil moisture as vegetation density increases, during the growing season. Field observations show that AMSR2 tends to overestimate the soil moisture when the vegetation intensity increases and covers the soil. We address two existing issues in the use of soil moisture products of the AMSR2: (1) spatial resolution of the soil moisture product and (2) the impact of vegetation cover on the radiative transfer. We used a vegetative index to estimate when soil moisture retrieval is not sufficiently accurate and how the optical data can be used to improve soil moisture estimation. The land surface temperature and vegetation index products of the Visible Infrared Imaging Radiometer Suite are used to downscale the AMSR2 soil moisture products to 1 km. A series of soil moisture data collected in the field were used to analyze the accuracy of the downscaled soil moisture values and the results indicated that introduction of the BT in the downscaling model improves the accuracy of the soil moisture products over the vegetated areas. The mean absolute error (MAE) of the downscaled soil moisture values is ∼5.6  %  —an improvement to 8.0% of original AMSR2 soil moisture products. In the presence of vegetation, however, the MAE is still greater than that of a similar analysis of AMSR2 product over semiarid areas.

High-resolution red-green-blue (RGB) images from remote sensors, such as those carried on aircrafts, UAVs, satellites, and the International Space Station (ISS), are cost-effective data sources for real-time emergency response applications. We describe an assessment undertaken on spectral behaviors to evaluate the effectiveness of two high-resolution RGB image datasets for mapping and monitoring of floodwater extent in dense urban areas. The assessment was as part of a case study of the Calgary 2013 flood event. The input imagery included very high-resolution aerial photos and imagery acquired with the SERVIR Environmental Research and Visualization System (ISERV) carried on the ISS. The results demonstrate the complementary nature of these two RGB image sets in providing effective urban floodwater mapping for real-time response. The aerial photos with higher spatial resolution and less atmospheric effect can provide the details about the floodwater distribution; the images from ISERV-ISS can provide the temporal variation of floodwater distribution.

This study concerns remote detection of alcohol vapors in cabins of moving vehicles. By means of a simplified description of propagation of two combined laser beams differently absorbed in alcohol vapors and in vehicle windowpanes, difficulties in detecting alcohol vapors in vehicle cabins using classical differential absorption method are demonstrated. Unreported in the literature, this method of remote detection of alcohol vapors in the air exhaled by a driver or a passenger in moving vehicles is proposed. The method uses two combined laser beams that illuminate different zones in the car cabin through the same windowpane. Based on the correlation between the spread of the transmission of these zones and the alcohol concentration in the air exhaled by a drunken person, it is possible to determine if the alcohol exists in the exhaled air. The effectiveness of the presented method is confirmed experimentally.

Change detection analysis for monitoring and modeling riverine systems requires detailed spatial and temporal surveying of river morphology dynamics. Accurate digital surface models are an essential part of the studies of morphodynamic modeling of rivers and river channel evolution. An accurate high-resolution surface model of the river channel and floodplain enables a more comprehensive view of the riverbed evolution and allows monitoring of morphological changes of the entire river channel more precisely compared with traditional methods. In current practice, unmanned aerial systems (UAS) are widely used for various applications in photogrammetry, especially for three-dimensional surface modeling and large-scale mapping of riverine environments. UASs offer many advantages in terms of cost and image resolution and approaches for mapping across varying spatial extents when compared with terrestrial photogrammetry and satellite remote sensing systems. UAS-based aerial mapping was implemented in order to produce very high resolution surface models of a flood-sensitive region based on structure from motion technique. Ultimately, multitemporal topographic data were produced, and morphodynamic processes in the lower course of the Büyük Menderes River in Turkey were modeled by digital shoreline analysis system and digital elevation model of difference methods. These methods were employed to examine the changes in the shoreline and to analyze the size of geomorphological changes and spatial patterns. The results showed that the change in the shoreline of the meanders varied from 3 to 27 m, and the water levels varied between ∼0.3 and 3 m. Although there was both sediment erosion and deposition along the shoreline, the predominant process was identified as deposition in the shoreline. A significant correlation was found between the deposited sediments and the sinuosity index values (r  =  0.88) according to the changes in water level over the months. In addition, major changes on the deposition rate were found to occur mostly after the summer season. This research showed that UASs could provide a suitable measurement model for determining areal and volumetric eroded/deposited sediment quantities along the meandering fields.

Herbicide injury on weeds is quantified through responses such as mortality, biomass, and visual scores, which are time consuming and/or subjective, making it hard to evaluate injury in experiments and making the site-specific management impracticable in major crops. The objective of our work was to build prediction models of glyphosate injury on weed species of economic importance in the Brazilian Cerrado (savana) through multispectral digital images. Two weed species with different levels of susceptibility to glyphosate, Eleusine indica (L.) Pers. and Brachiaria decumbens L., were cultivated in pots and subjected to increasing doses of glyphosate, simulating a gradient of injury. Visual scores were given by three trained evaluators during three weeks after the application. Images of the pots were acquired with a digital multispectral camera in the following bands of the electromagnetic spectrum: blue (450 nm), green (560 nm), red (660 nm), and near-infrared (850 nm), with a digital camera at 0.9-m height from the ground. Images were first segmented to separate plant from soil using the hue–saturation–value color space. Vegetation indices were calculated and used as explanatory variables for phytotoxicity in regression models. A field experiment was conducted in order to validate predictions. The injury pattern of glyphosate on E. inidica and B. decumbens was best modeled by an exponential function of the median of hue for plant pixels, which could be reproduced at field scale with prediction errors below 8%. Differences in sensitivity to glyphosate between the two species were detected with the kernel density of hue for plant pixels, with absolute correlation with phytotoxicity above 0.8.

We aim to investigate a potential impact of boosting saturation of aerial imagery on the performance of unsupervised human detection algorithms. The study is empirical since it is based on processing photographs taken during a full year experiment in the Izerskie Mountains (southwestern Poland) by a consumer-grade Canon S110 camera mounted onboard eBee, a fixed-wing micro-unmanned aerial vehicle (UAV). In the preliminary analysis, we used a few basic color adjustments (sharpening, hue–saturation–luminance, contrast, saturation, and vibrance) to process UAV-taken photographs prior to the automated human detection with the nested k-means algorithm. We found that saturation boost is an image preprocessing method that may potentially improve the performance of human detection. In the actual analysis, we investigate only the saturation effect by employing four saturation modification schemes (two versions of enhancements of unusual colors, additive boost of saturation, and multiplicative boost of saturation) and three human detection algorithms [three-dimensional (3-D) nested k-means on RGB, two-dimensional nested k-means on hue–saturation–value, morphological operations of erosion, and dilation with thresholding]. All the studied saturation boost techniques increase detection rates of the 3-D nested k-means on RGB, with particularly meaningful improvement for images acquired in the spring. Morphological operations of erosion and dilation are not found to be skillful in detecting persons. However, their performance is improved after the initial preprocessing by the original or modified enhancement of unusual colors.

The core of deep convective clouds (DCC) is one of the most consistent Earth targets. When viewed from space, the measurements have minimal impact from water vapor and aerosols and have been used as invariant scenes for calibration assessment for reflectance solar bands. DCC is also one of the coldest targets and can be used for assessing the calibration and product stability for infrared thermal emissive bands (TEB). The stability from the years 2003 to 2019 is analyzed for both Terra and Aqua moderate resolution imaging spectroradiometer (MODIS) long-wave infrared bands. Most bands exhibit very stable long-term trending. Terra band 30 shows the largest rate of change of 0.19  K  /  year. The Terra–Aqua difference is also analyzed. In general, for both Terra and Aqua, the long-wave infrared bands show good and stable detector uniformity. The impact of response change due to anomalies and events has been analyzed. Since the measurement bias over cold DCC is dominated by the offset term in the calibration equation, an assessment model is developed to estimate the offset bias between the mirror sides. The impact model is also developed to estimate the impact on higher brightness temperature measurements. The offset bias from the assessment modeling in this analysis is used as the input for impact modeling. The measurements over Dome-C, Ocean, and desert scenes are used to verify the impact model. These assessments are very helpful for MODIS TEB calibration and look-up table update procedure improvements.

Frost in the uplands of northeast India frequently damages maize, leading to poor yields or even complete crop failure. We studied the feasibility of using reflectance spectra from a handheld radiometer and a camera on board a drone at 80 m above the crop to assess physiological changes, differential responses of crop to nutrient managements (N, P, and NPK) and also predicted the likely losses of grain yield caused by one night’s frost damage in 90-day-old maize crop in Meghalaya, northeast India. A portable single-beam field spectroradiometer (model SVC-HR-1024) was used to measure radiation in the range of 350 to 2500 nm with varying bandwidth at 1 m above the crop. All measurements (both before and after frost) were made strictly between 11:00 and 13:00 h, when the sun was almost overhead. Similarly, a Parrot Sequoia multispectral camera sensing in four wavebands was mounted underneath a drone (DJI Matrix 600 model). We compared the measured spectra with those recorded on the same crop before the frost and also measured several biochemical constituents in the leaves on the two occasions for comparison. The frost damage increased the reflectance in the photosynthetically active visible and infrared regions with a strong peak at 2100 nm while it caused a sharp decline in the near-infrared (between 720 and 1350 nm) and shift in the red edge. It caused a decrease in the normalized difference vegetation index as measured by the radiometer from an average of 0.36 to 0.13 and a decrease from 0.5 to 0.31 as measured from the drone. The yield in frosted year was on average 870  kg ha^−  1 less than in the previous frost-free year on the same plots. Frost injury reduced leaf pigments (chlorophyll and carotenoids), nitrogen, and grain yield by a greater magnitude in P-stressed and NPK-deficient maize crops over N-stressed and stress-free (with NPK) plants. Spectral reflectance and vegetation indices also varied significantly. We could also assess quantitatively the crop’s loss of photosynthetic potential from the hyperspectral reflectance measurements and from the low-flying drone. The latter technology should enable farmers to monitor and assess frost damage immediately and to predict the likely losses of yield.

Due to the effects of rugged terrain and vegetation cover, improving the locating and monitoring accuracy of time series interferometric synthetic aperture radar (TS-InSAR) in detecting the development of landslides in mountainous and valley areas has become a problem that needs to be solved. Our study uses the ground local incident angle (GLIA) to locate noneffective monitoring areas that are affected by steep terrain as well as to improve the locating accuracy of TS-InSAR technology in the effective coverage area. At the same time, the improved small baseline subset InSAR (SBAS-InSAR) method is used to improve the density of interferometer targets as well as to solve the problem of insufficient permanent scatterers due to the effects of vegetation coverage. A geometric model of GLIA was established. Ascending and descending synthetic aperture radar (SAR) data from the Advanced Land Observation Satellite (ALOS) were used for calculating the value of GLIA. The relationships of the GLIA and the interference characteristics of InSAR data were analyzed, and the locations of noneffective areas were determined. Finally, 19 ascending pass SAR data scenes acquired by the ALOS-1 satellite were processed through using the improved SBAS-InSAR method, and the detailed time series moving displacement information from 2007 to 2011 was mapped in Wudongde hydropower reservoir. Furthermore, the potentially moving landslide areas and landslide hazard areas were located in the effective interference areas. The active Jinpingzi landslide was investigated in detail, and the single-point monitoring results of SBAS-InSAR were compared with the electronic total station (ETS) measurement results. The conclusion shows that the monitoring results of SBAS-InSAR were in agreement with the ground field survey and the single-point ETS monitoring, and the applicability of the model and algorithm proposed was proved. Our study provides a reference for the early identification and high-precision monitoring of landslides in mountainous and valley areas.

In view of current land cover classification methods using the full-waveform characteristic parameters from spaceborne lasers, land cover categories are misclassified because the extracted features cannot accurately distinguish various types of features. We propose a land cover classification method that considers the contribution of waveform feature parameters and the pooling scale. First, the waveform after preprocessing and Gaussian decomposition is characteristic extracted and roughly classified, and the importance of each waveform feature parameter is evaluated and sorted. The best feature parameter set is determined by backward selection of the waveform feature parameters. Moreover, the original waveforms are pooled at different scales, and the correlations between ground objects after pooling are compared with the correlations between the original waveforms to ensure the stability of the correlation between various waveforms of ground objects and the peak information of waveforms while reducing the amount of data. Finally, random forest (RF) and support vector machine classifiers are selected to complete the land cover classification experiment based on the optimal waveform parameter set and the optimal pooling scale. We use ICESat (ice, cloud, and land elevation satellite) and Geoscience Laser Altimeter System data from the Beijing–Tianjin–Hebeii region for verification. The results show that, based on the RF classifier, the overall classification accuracy was 86.95%, and the use of spaceborne laser waveform characteristic parameters compared with the classification method greatly improved the accuracy, fully verifying the feasibility of the method in classifying land cover types. These findings can provide a technical basis for the application and promotion of GaoFen 7(GF-7) and other follow-up, satellite-based laser altimetry data in the fine-grained classification of land cover types.

Land subsidence resulting from groundwater extraction is a widely recurring phenomenon worldwide. To assess land subsidence, traditional methods such as numerical and finite element methods have limitations due to the complex interactions between the different constructor factors of aquifer in each area. We produced a groundwater-induced subsidence map by applying the geological and hydrogeological information of the aquifer system using an artificial neural network (ANN) combined with interferometric synthetic aperture radar (InSAR) and geospatial information system. The main problem with neural networks is providing the ground-truth dataset for training step. Therefore, the subsidence rate used as the network output was estimated using the InSAR time series analysis method. This study indicates the ANN approach is capable of knowing the mechanism of the land subsidence and can be used as a complementary of InSAR method to estimate the land subsidence with effective parameters and accessible data such as groundwater-level data especially in those areas in which measuring the subsidence was not feasible using InSAR. However, the results indicated that average groundwater depth and groundwater level decline were the most effective factors influencing subsidence in the study area using sensitivity analysis.

Due to their low computational complexity, greedy pursuit algorithms are widely used in sparse signal reconstruction. An improved greedy iterative algorithm, called the expanded subspace pursuit, is proposed. By incorporating a simple backtracking technique, the proposed algorithm removes the mismatching atoms to refine the estimated support set effectively. Furthermore, the proposed algorithm can achieve blind sparse reconstruction even without the prior of the sparsity degree. Compared with other greedy algorithms, the proposed algorithm exhibits superior reconstruction accuracy and lower computational complexity. Finally, numerical results are presented to demonstrate the validity of the proposed algorithm.

A method of super-resolution reconstruction of remote sensing images based on convolutional neural network is proposed to address the problems of low-resolution and poor visual quality of remote sensing images. In this method, a sample database with high-resolution (HR) and low-resolution (LR) remote sensing images was constructed. A convolutional neural network was then obtained by determining the intrinsic relationship between HR and LR remote sensing images in the sample database. Multiple pairs of HR and LR images were selected from the sample database and sent into a six-layer convolutional neural network. The experimental results showed that compared with other learning-based methods, such as the fast super-resolution convolutional neural network (FSRCNN), the image quality obtained by our method is improved both subjectively and objectively. Moreover, the training time was ∼17  %   less than in the FSRCNN method.

Compared with speckle noise, the targets in synthetic aperture radar (SAR) images have strong directionality. Since the target and noise are different in the directional sub-bands on the same scale of nonsubsampled contourlet transform (NSCT), there are obvious differences in characteristics of the NSCT coefficients. Taking advantage of the differences, a denoising method for SAR image based on intrascale correlation of NSCT is proposed. The coefficients in different directional sub-bands in NSCT field are analyzed and the distribution law of the difference between maximum and minimum coefficients on the same scale is presented. Then, a threshold-determining strategy is proposed for identifying noise from targets. Finally, the proposed method is compared with some state-of-the-art denoising methods. It is observed from the results that our method presents the best performance in balance of noises and edge-preserving.

Hyperspectral remote sensing images (HRSIs) face the problem of extracting efficient discriminant features when foreign objects with similar spectra and outlier classes exist. For solving the problem, first, we derive a class pair (CP) form of the classical Fisher criterion, namely CP Fisher criterion, where the between-class scatter matrix and within-class scatter matrix are both decomposed into the form of CPs. And then, a CP-weighted criterion is proposed to weight the between-class scatter matrix and within-class scatter matrix of each CP according to its separability so that the separabilities of CPs are balanced in the CP-weighted feature subspace. The Bayesian classifier and k-nearest neighbors classifier are used to evaluate the feature extraction performance. Experimental results on three real HRSIs show that the presented CP-weighted subspace method improves the classification accuracies of the CPs with small separabilities so that it is superior to the linear discriminant analysis subspace method and the weighted pairwise Fisher subspace method.

Hyperspectral (HS) data are enriched with highly resourceful abundant spectral bands. However, analyzing and interpreting these ample amounts of data is a challenging task. Optimal spectral bands should be chosen to address the issue of redundancy and to capitalize on the absolute advantages of HS data. Partial informational correlation (PIC)-based band selection approach is proposed for feature selection-based classification of HS images. PIC measure appears to be more skillful compared to mutual information for estimation of nonparametric conditional dependency. In this proposed approach, HS narrow bands are selected in an innovative way utilizing the PIC. This approach is more efficient in terms of computational time and in generalizing the applicability of selected spectral bands. Further, these optimal spectral bands are used in the support vector machine (SVM) and random forest classifier for performance evaluation. The optimum performance is accomplished with SVM classifier, and the achieved average overall accuracies are 82.89%, 91.4%, and 91.29% for the Indian Pines, Pavia University, and Botswana datasets, respectively. The proposed band selection approach is compared with different state-of-the-art techniques. This methodology improves the classification performances compared to the existing techniques, and the advancement in performances is proven to be statistically significant.

The bilinear mixing model is a more realistic, generalized model that can represent a wide range of real-world hyperspectral images with tolerable accuracy. The use of a spectral library makes the problem more tractable. However, the high mutual coherence of the spectral library creates computational as well as performance issues in library-aided bilinear unmixing. Besides, the high mutual coherence of the spectral library reduces the accuracy of these unmixing methods, and the high cardinality of the spectral library increases the computational complexity. We propose a computationally efficient, two-phase library pruning approach for unmixing hyperspectral image, which also withstands a highly coherent spectral library. In this work, we first segregate the data into pixels generated due to linear and bilinear interaction using the subspace clustering method and subsequent rank estimation strategy. We subsequently reduce the mutual coherence of the spectral library and prune the linear interactions. In the next stage, we create a library corresponding to the bilinear components assuming that only the secondary reflections of the pruned library elements may be prevalent in these pixels. We perform pruning using a novel, low-rank based, sequential approach. Finally, we compute the abundance of the matrix by exploiting sparseness of the abundance matrix and include its low-rankness, and spatial structural similarity as regularization. We validate the overall advantages of our proposed framework on several real and synthetic data experiments.

As one of the most challenging tasks of optical remote sensing object detection, airport aircraft detection has attracted more attention from researchers. However, various scales and orientation of aircraft, and the complexity of airport regions leads to low detection performance. To address these problems, this article presents an effective aircraft detection method called the local context (LC) deformable parts model. First, aircraft candidate region fast extraction is developed to improve the detection efficiency. Then, the LC weight map is constructed by weighting the key points with the aircraft local characteristics. Afterward, gradient features are combined with the weight value to establish the LC histogram of oriented gradient feature pyramid. Finally, based on the established feature pyramid, aircraft candidate region is identified efficiently by orientation prediction of the suspected target. A series of comparison experiments on two datasets demonstrate that the proposed method has higher detection accuracy and efficiency than the existing detection methods.

Soil texture is a primary parameter affecting the soil–water retention characteristics, plant nutrient storage, and carbon sequestration capability of a soil. A rapid, nondestructive, and economical detection method needs to be developed. A near-infrared hyperspectral imaging (HSI) system with a wavelength range of 874 to 1734 nm was used to measure the soil samples from Songyang County, Zhejiang Province, China. Totally, 198 samples were classified into four different texture groups: clay, clay loam, loam, and sand, based on the International Soil Science Society standard. The optimal wavelengths were selected by the method of competitive adaptive reweighted sampling. The texture features (contrast, correlation, energy, and homogeneity) of the images at the optimal wavelengths were extracted by the gray-level co-occurrence matrix. The results showed that both the spectral and image texture information could classify soil texture to some extent. The model established by the combination of optimal wavelengths and texture features achieved the highest correct classification rate of 87.9%. It was concluded that HSI was an efficient technique to classify the soil texture into four groups of clay, clay loam, loam, and sand.

The decomposition of mixed pixels in Moderate Resolution Imaging Spectroradiometer (MODIS) images is essential for the application of MODIS data in many fields. Many existing methods for unmixing mixed pixels use principal component analysis to reduce the dimensionality of the image data and require the extraction of endmember spectra. We propose the pixel spectral unmixing index (PSUI) method for unmixing mixed pixels in MODIS images. In this method, a set of third-order Bernstein basis functions is applied to reduce the dimensionality of the image data and characterize the spectral curves of the mixed pixels in a MODIS image, and then the derived PSUIs (i.e., the coefficients of the basis functions) are calibrated by means of the abundance values of the ground features from the Landsat Enhanced Thematic Mapper Plus (ETM+)/Operational Land Imager (OLI) classification images corresponding to the date and region of the MODIS image. The proposed method was tested on MODIS and ETM+/OLI images, and it obtained satisfying unmixing results. We compared the PSUI method with conventional methods, including the pixel purity index, the N-finder algorithm, the sequential maximum angle convex cone, and vertex component analysis and found that the PSUI method outperformed the other four methods.

Semantic segmentation is one of the most popular and challenging applications of deep learning. It refers to the process of dividing a digital image into semantically homogeneous areas with similar properties. We employ the use of deep learning techniques to perform semantic segmentation on high-resolution satellite images representing urban scenes to identify roads, vegetation, and buildings. A SegNet-based neural network with an encoder–decoder architecture is employed. Despite the small size of the dataset, the results are promising. We show that the network is able to accurately distinguish between these groups for different test images, when using a network with four convolutional layers.

Bag of visual words (BoVW) model has been widely used in content-based image retrieval for its good performance. The k-means clustering algorithm plays an important role in BoVW, but there are still some problems that need to be solved. Many methods are proposed to improve the clustering performance, such as fast search and find of density peaks (FSFDP), which has been successfully applied in image classification or image retrieval for its high efficiency but suffers from some limitations when it is applied to remote sensing image retrieval. A remote sensing image retrieval method based on an improved FSFDP-BoVW model is proposed. A feature selection method combining a chi-square test and a term frequency-inverse document frequency is proposed to select the features that can best represent image content and reduce the redundancy of remote sensing image features; a natural nearest neighbor algorithm is utilized to redefine the density of data points to reduce the interference of human subjectivity in FSFDP. Experimental results show that the proposed method can effectively reduce the time spent in constructing a codebook and improve the retrieval accuracy.

Change detection is a challenging task that has received much attention in the remote sensing field. Whereas numerous remote sensing change detection methods have been developed, the efficiency of these approaches is insufficient to meet the real-world applications’ requirements. Recently, deep learning methods have been largely used for remote sensing imagery change detection, most of these approaches are limited by their training dataset. However, adapting a pretrained convolutional neural network (CNN) on an image classification task to change detection is extremely challenging. An automatic land cover/use change detection approach based on fast and accurate frameworks for optical high-resolution remote sensing imagery is proposed. The fast framework is designed for applications that require immediate results with less complexity. The accurate framework is designed for applications that require high levels of precision, it decomposes large images into small processing blocks and forwards them into CNN. The proposed frameworks can learn transferable features from one task to another and escape the use of the expensive and inaccurate handcrafted features and the requirements of the big training dataset. A number of experiments were carried out to validate the proposed approach on three real bitemporal images. The experimental results illustrate the superiority of the proposed approach over other state-of-the-art methods.

We present a modified local tangent space alignment (LTSA) algorithm, called supervised kernel discriminant local tangent space alignment (SKDLTSA), for radar target recognition based on high-resolution range profile (HRRP). SKDLTSA aims to extract intraclass geometric structure embedded in local neighborhoods, as well as to maximize interclass separability characterized by overall distances among different classes. It is formulated with kernel technique to extract nonlinear features, which helps to obtain better performance than its linear counterparts. Extensive experiments on measured HRRP data from three flying airplanes demonstrate that the proposed method can significantly improve the recognition performance. Further results also indicate its robustness to target attitude variations and noise corruption.

A multiresolution and velocity vector average processing algorithm is proposed to retrieve the wind field. When the radar echo area is divided into many square grids with sides of 5 km, the radar radial velocities on the edges and in the center of these grids are obtained through velocity vector averaging in the grid region. In order to solve the problem of a nonuniform wind field caused by different distances from echo to radar, the wind field is retrieved with a different angular resolution, whereas the Cartesian grid resolution remains unchanged. Through comparison and research, the correctness of this retrieval method is proven. At the same time, superposition application between geographic information and radar wind field is completed. The application based on this method can provide support for gale disaster early warning.

Semisupervised graph learning has a broad prospect in remote sensing (RS) image change detection. However, an improper graph model may result in a contradiction between the detection accuracy and computational efficiency. In order to effectively extract the structural information of changes and heavily reduce the computational burden, we propose a hybrid graphical model (HGM) for bitemporal RS image change detection. The HGM utilizes the hybrid superpixels (HSPs) as its vertices, and a hybrid graph kernel (HGK) function is proposed for measuring the similarities between the vertices. The HSPs are composed of the background superpixels and foreground isolated pixels of a subtraction image. The HGM effectively exploits the image structures, and the small graph size dramatically reduces the computational complexity. Moreover, the piecewise HGK function well detects the structures of the changed areas and heavily resists the background disturbances. A semisupervised label propagation algorithm is implemented with the HGK matrix for obtaining the final change detection results. Experimental results on real RS images demonstrate the effectiveness and efficiency of the proposed method and prove that it is a good candidate for RS image change detection.

Target detection is a crucial problem for the analysis of hyperspectral remote sensing image. Sparse representation-based methods have shown considerable potential in hyperspectral target detection. In a hyperspectral image (HSI), the sparse representation with respect to a certain pixel means the pixel can be sparsely represented by the linear combination of data vectors from the data dictionary. When applied to target detection, a training dictionary needs to be constructed, consisting of both target and background samples in the same feature space. Then the test pixels can be sparsely represented through the decomposition over the constructed dictionary. Sparse representation is considered to preserve the main information of most pixels in HSI target detection. However, some background pixels may possess spectra similar to the target pixels’ spectra, which may cause false detection. Therefore, more constraints are needed to smooth these pixels. We propose a local similarity constraint-based sparse algorithm to deal with this problem. Based on the assumption that pixels have both spectral and spatial similarity should have similar sparse representation, a local similarity constraint term is incorporated into the sparsity model. Then, an iterative sparse recovery algorithm is provided to obtain the recovered sparse vectors composed of sparse coefficients corresponding to both the target subdictionary and the background subdictionary. With the obtained sparse vectors, the residuals between the original test samples and the estimates recovered from the dictionary (including the target subdictionary and the background subdictionary) can be calculated and used to decide which class the test pixels belong to. The proposed algorithm has been applied on real HSIs to test the performance of detecting targets of interest. Experimental results demonstrate that the proposed model achieves better target detection performance than the state-of-the-art methods.

We present a Bayesian sparse representation (BSR) model for synthetic aperture radar (SAR) image classification. It involves the design of the sub-categorization scheme for the samples, the formulation of the BSR model, and the implementation of the classification decision. First, the training samples of each target category are sub-categorized into multiple subclasses. The discriminative features are then extracted from each sample within one subclass. These same features from all samples are gathered to construct a sub-dictionary. After collecting all sub-dictionaries from all features, the sparse reconstructions are performed for all features. A BSR framework is employed for such purposes. Finally, a fusion strategy is applied to the residuals to predict the class label of the test sample. By sub-categorizing the samples into multi-clusters, the proposed model decreases the intra-class variations between the samples and thus improve the representation ability of the features to different targets. The test results using real field data demonstrate that the proposed method has superiority to some state-of-the-art methods.

After fossil fuel burning, deforestation and forest degradation are the second largest contributors to greenhouse gas emissions to the atmosphere. In order to claim the carbon credit under the reducing from deforestation and forest degradation (REDD+) scheme, a United Nation’s Framework Convention on Climate Change initiative for climate change mitigation, developing countries are required to prepare national reference emission levels for forests on the basis of historic data and national circumstances. Part of developing reference emission levels includes quantifying location, pattern, and rate of historic forest degradation, which are also called in a word the activity data for forest degradation. Applying Monte-Carlo spectral unmixing technique to Landsat images in the CLASlite^® algorithm followed by a knowledge-based classification approach, this research quantified the activity data for forest degradation at Raghunandan Hills Reserve (6143 ha) in Bangladesh. Moderate spatial resolution Landsat images were able to detect the activity data for degradation in a spatially explicit manner with high accuracy (>90  %  ). The research approach and findings can serve as valuable information for any future national level initiative for developing activity data for REDD+ projects.

Our research group built a super-low altitude detection platform with power delta wings and mounted it with a HySpex hyperspectral sensor. This platform has great potential for extracting alteration information as it can provide high spectral and spatial resolution images. The aim of this study is to verify the reliability and accuracy of alteration information extraction using this platform. The seven-scene hyperspectral images of the Karatagh region in Hami of Xinjiang obtained by the platform in September 2017 were used to identify iron-bearing minerals. First, these images were preprocessed by splicing, atmospheric correction, and geometric correction. Second, the images were compared with the United States Geological Survey (USGS) standard spectra of similarities between the end-member spectra of goethite-, limonite-, and jarosite-altered minerals. Third, the three altered minerals were mapped in the test area based on the end-member spectrum. The results show that the correlations between the end-member spectra of three altered minerals, including goethite, limonite, and jarosite, obtained from the HySpex image and corresponding spectra in USGS reach 0.92873, 0.95098, and 0.8875, respectively. The end-member spectral reflectance values were higher than those in the spectral library. The original spectral data after the continuum-removal and the local quantitative display show that the spectrum at the feature absorption location was similar to the reference spectrum. On this basis, the distribution maps of goethite, limonite, and jarosite iron-stained altered minerals in the Hami Gobi area were produced and field verified. The data show that the information of identified minerals from the HySpex hyperspectral data were in accordance with the actual situation. Both methods can recognize jarosite well, the spectral angular mapping method is better than the back propagation (BP) neural network for goethite recognition, and the BP neural network of limonite recognition is better than the spectral angular mapping method. The above results prove that the alteration information can be effectively identified using the HySpex hyperspectral data integrated on a super-low altitude detection platform, which is of great significance to the rapid delineation of minerogenic prospects area.

Optical detection systems have the potential to get around some limitations of acoustic detection systems, especially with increased fleet and port security in noisy littoral waters. Identification of environmental effects especially tropical storms on underwater optical detection is a key to the success. A typhoon-influenced area is chosen in the western North Pacific Ocean with high ocean transparency and low seasonal optical variability. Underwater inherent optical properties (IOPs) such as the beam attenuation coefficient for 470 nm light are measured in the selected region from the U.S. Naval Oceanographic Office sea gliders deployed after super typhoon Guchol’s (June 7 to 20, 2012) passage from June 25 to 30, 2012, and with no typhoon activity from January 9 to February 28, 2014. The observed two sets (with and without typhoon) of IOPs are taken as the input into the Navy’s electro-optical detection simulator. The simulation shows low detection after the super typhoon Guchol-2012’s passage and high detection without typhoon passage.

High-quality land use and land cover data are important for monitoring and analyzing environmental changes in the background of global warming. This study assessed the spatial and areal inconsistencies in the four most recent multiresources land cover products in a complex manner using the common classification systems of International Geosphere-Biosphere Program (IGBP)-17, IGBP-9, IPCC-5, and TC (vegetation, wetlands, and others only). Based on inconsistencies and multitemporal land cover datasets, a synthesis of the study was triggered out on land use and land cover dynamics during 2001 to 2017 in the southeastern region of Bangladesh. The overall areal and spatial inconsistencies decreased from high to low levels of aggregation (IGBP-17 to TC), indicating that the inconsistencies are not only influenced by the level of thematic detail and landscape complexity but also related to the conversion uncertainties. Overall, areal inconsistency in the comparison of the FROM-GLC and GlobeLand30 datasets was the smallest among the six pairs, whereas the pair of MODISLC and LULC was observed the highest inconsistencies. Based on an overall lower inconsistencies classification system (IGBP-9), the synthetic land use cover changes in the study area were assessed. During the period of study, the areal distribution of forest cover, built-up areas, and water increased annually by 0.4%, 1.32%, and 0.3% respectively, whereas the croplands and wetlands decreased by 0.5% and 0.3%, respectively. The dynamic changes of croplands, forest, and artificial surface were identified the prime cyclic land cover change. This research is helpful in providing training areas for the producer of land cover products.

Land cover classification is a significant task in remote sensing that aims at land cover monitoring and adjustment. Instance segmentation in deep learning has been widely used in land cover classification. However, this method requires high quality, labor-intensive pixel-level annotations. A bisupervised pipeline is proposed especially for datasets with ambiguous samples, which reduces the requirement of labeling accuracy. In order to supervise learning with ambiguous annotation samples, the pipeline has two losses to feedback, namely main loss and auxiliary loss. The main loss is still responsible for the expected output. According to the difference degree between ambiguous samples, categories are constructed artificially. The auxiliary loss generates feedback through categories, sharing decoder layers of main output during the training process. In the prediction process, auxiliary loss is used to update the main loss results. The neural network adopts the structure of U-Net with the pyramid pooling module, in which the multiscale feature is used in the feature extraction process. We also compare five different backbones and choose Inception-V3 as the backbone to improve the feature extraction capabilities of network encoders. The use of transposed convolution instead of traditional upsampling in decoders can improve the segmentation details. Our bisupervise network obtains Jaccard index of 83.936% and F1-score of 91.17% on Gaofen-2 imagery dataset. The results demonstrate that the proposed method can achieve better classification performance in datasets with ambiguous annotations.

Investigating groundwater storage (GWS) could greatly better our understanding of groundwater dynamics and the factors that influence them. To generate changes in ΔGWS at a finer scale, we developed a statistical downscaling model and applied it to the Gravity Recovery and Climate Experiment (GRACE). The model is based on the relevance vector machine (RVM), which needs the few parameters (kernel function and kernel width) for regression. In addition, considering the anthropogenic influence on the GWS, the interferometry of synthetic aperture radar (InSAR) time series, which could inverse the land subsidence resulting from groundwater extraction, was introduced into the model. The model was evaluated and compared with one constructed using support vector machine. We obtained a 0.1 deg  ×  0.1 deg ΔGWS from the model. It is shown that the study area suffered from a sustained groundwater reduction with an expansion of space during 2007 to 2010. Furthermore, RVM is suggested to construct the GRACE downscaling model instantly, and InSAR can be considered an important indicator.

Land surface temperature (LST) data are available for several years at a high acquisition rate from sensors such as a moderate resolution imaging spectroradiometer (MODIS). Advances in remote sensing technology have exploited this large volume of LST image time series (LITS) data for a clearer understanding of land surface processes. However, remotely sensed LST observations from spaceborne platforms are likely to be influenced by severe weather conditions (e.g., clouds, haze, heavy aerosols, heavy rainfall and flooding, hailstorms, and snowfall). Presence of bad data images in LITS can reduce the credibility of results from LITS data analysis. An algorithm to detect bad data images in a long-term MODIS LITS is proposed. Bad data in MODIS LITS were filtered at various stages. Initially, poor quality pixels were removed from LST maps based on quality control information provided in corresponding quality assurance maps. In further stages, thin cloud-contaminated LST pixels were removed that were left undetected by MODIS official cloud screening procedure. Finally, to remove bad data LST maps, median-based statistical outlier detection method was employed, and outliers were detected in LST pixel time series (LPTS) vectors generated at consistent normalized difference vegetation index (NDVI) pixel (CNP) locations. CNPs were selected by employing Mann–Kendall test statistics on annually aggregated MODIS NDVI time series vectors. Decisions were made to identify bad LST maps based on number of outlier LST values appeared in LPTS vectors at CNP locations for each LST map. The proposed algorithm was applied over parts of Gondwana coalfields situated in Jharkhand and West Bengal states of India.

Currently, due to the limited amount of data and the difficulty of designing a network, there are few papers on constructing a new convolutional neural network for scene classification using the publicly available datasets of high-resolution remote sensing images. Considering the existing problems, the current scene classification methods of high-resolution remote sensing images are summarized, and the IMFNet model is constructed to classify scenes of high-resolution remote sensing images in this paper. The IMFNet is an end-to-end network, which can learn features from data automatically. The main characteristic of the IMFNet network structure is that the Inception module is used to extract the details of remote sensing images and the multifeature fusion strategy is proposed to ensure the integrity of information. In addition, optimization methods are adopted to improve the classification accuracy. In order to verify the effectiveness of the method proposed in this paper, the two benchmark datasets—the UC Merced dataset and the SIRI-WHU dataset were adopted for experiments. The classification accuracy of the two datasets reaches 92.14% and 90.43%, respectively. Experimental results show that the method proposed has certain advantages over the classification methods based on low-level and middle-level visual features and even some classification methods based on high-level visual features.

In recent years, remote sensing of phenology has become a useful tool to reveal the response and feedback of vegetation dynamics to global climate change due to its multitemporal phase, wide-coverage, continuous space coverage, and long-time series. Based on the Moderate Resolution Imaging Spectroradiometer data product MOD09A1, enhanced vegetation index (EVI) was calculated, and the time series EVI of the growing season in the Qinghai–Tibet Plateau (QTP) from 2001 to 2016 was reconstructed in light of the double logistic function filter. A dynamic threshold method was used to extract the starting date of the growing season (SOG) of vegetation. The characteristics of temporal–spatial variation of SOG of vegetation in QTP were analyzed, and the irregularity of the vegetation SOG in response to climate change was explored. The results show that: (1) the vegetation SOG in QTP varied between day of year (DOY) 110 and DOY 170 between 2001 and 2016. Affected by the spatial difference of hydrothermal conditions, the vegetation SOG is gradually delayed from the southeast to the northwest in the plateau, the latest of which in the southwest of Tibet. The SOG of vegetation in the east of QTP is in advance while it is delayed in the west. (2) the SOG of vegetation has the characteristics of vertical differentiation as well, especially in the range of 3600 to 5000 m, and the delaying rate of the SOG with increasing elevation is 10.7  day  /  km. (3) The time lag analysis between climatic factors and the vegetation SOG in QTP indicates that the SOG has a time lag of 0 to 15 days behind the air temperature and 15 to 30 days behind precipitation. The SOG of vegetation in southern Tibet and southern Qinghai is more sensitive to air temperature changes than precipitation, whereas the vegetation SOG in southwestern Tibet and southwestern Qinghai is more sensitive to precipitation changes than air temperature changes.

Rain affects radar backscatter to some extent, and the effect is complicated. Our article models the radar backscatter at low incidence angles under rain conditions by a semiempirical method. Starting from no rain conditions, we choose two optics models as backscatter models at low incidence angles. From the rain effect analysis, an empirical function is deduced by curve fitting. Then we add the empirical function with the backscatter models at low incidence angles under no rain conditions, and two backscatter models at low incidence angles under rain conditions are obtained. After comparing with measured backscatter, we select a quasispecular model combined with the empirical function as a new backscatter model. Further, a correction item is introduced to improve the model accuracy. Results demonstrate that the model can predict backscatter as rain rate ranging from −15 to 5 mm/h dB and wind speed ranging from 6 to 16 m/s with a root-mean-square error of 0.97 dB and a correlation coefficient of 0.81.