The need for high-resolution imaging becomes particularly important in remote sensing image applications, such as ground-object identification. We introduce a Bayesian multiframe super-resolution algorithm that efficiently improves the imaging resolution of our micro–nano carbon satellite images. We begin by presenting a theoretical overview of the algorithm, and subsequently we conduct experiments in two phases. In the first phase, we compare the performance of our algorithm with those of other similar algorithms using a set of reference images. In the second practical application phase, we apply all the algorithms considered to panchromatic images obtained from our civilian micro–nano carbon satellite. Our results indicate that the proposed algorithm significantly outperforms the other algorithms, affording higher image resolution and greater image detail. We believe that our approach can contribute to the further development of satellite imaging systems.

This guest editorial introduces the Special Section on CubeSats and NanoSats for Remote Sensing.

The athermalized panchromatic imaging system (APIS) was the low-resolution refractive camera proposed by the Laboratorio de Instrumentación Espacial as a CubeSat payload. APIS flew on-board OPTOS CubeSat designed and developed by INTA using the methodology of European Cooperation for Space Standardization and space qualification tests. APIS had two main objectives: to analyze the performance degradation of commercial off-the-shelf (COTS) components due to space radiation and to verify in-flight functionality of the passive athermalization system. We summarize the design, manufacturing, and assembly integration and verification phases of the instrument, as well as the analysis of the radiation tests. Additional studies are included, such as thermal behavior, tolerances and sensitivity analysis, signal-to-noise ratio, and ghost images, as well as their implications during the design process. Three main goals were achieved during the mission lifetime: (1) the viability of a small refractive Earth observation camera on-board a CubeSat, (2) the validation for low Earth orbits of a passive athermalization system, and (3) the use of COTS elements, such as commercial glasses and detectors based on complementary metal–oxide–semiconductor technology, on a 2-year Earth observation mission.

The U.S. Naval Research Laboratory (NRL) has developed the Triple Tiny Ionospheric Photometer (Tri-TIP), an ultraviolet remote-sensing instrument based on the TIP. Tri-TIP measures emissions of atomic oxygen (Oi 135.6 nm) to determine plasma density in the nighttime ionosphere. The Tri-TIP design shrinks TIP to a 1U CubeSat form-factor and simplifies the mechanical design with a three-channel photometer system to isolate the target wavelength without a filter wheel. A heated strontium fluoride (SrF₂) filter eliminates incoming light at wavelengths shorter than 135.6 nm. The filtered light is divided between two matched photometers by a beam splitter with a magnesium fluoride coating over aluminum (AlMgF₂) deposited on 50% of the surface in a polka-dot pattern. The third photometer monitors dark count noise for later subtraction. One Tri-TIP configuration uses a beam splitter with a sapphire (Al₂O₃) substrate, which is opaque to wavelengths shorter than ∼140  nm, to later subtract contaminating emissions at wavelengths longer than 140 nm. A second Tri-TIP configuration uses a MgF₂ substrate beam splitter to simultaneously measure Oi 135.6 nm from two adjacent fields-of-view. The performance of both beam splitters has been tested at NRL, and the results are presented.

RainCube (Radar in a CubeSat) is a technology demonstration mission to enable Ka-band precipitation radar technologies on a low-cost, quick-turnaround platform. The 6U CubeSat, currently in orbit, features a radar payload built by the Jet Propulsion Laboratory and a spacecraft bus and operations provided by Tyvak Nano-Satellite Systems. Following the deployment of the half-meter parabolic antenna, the radar first observed rainfall over Mexico. The mission continues to operate and has met all requirements through repeated observations of precipitation in the atmosphere. RainCube is funded through the Science Mission Directorate’s Research Opportunities in Space and Earth Science 2015 In-Space Validation of Earth Science Technologies solicitation. We report on the first radar observations of precipitation.

Remote inspection can supplement instrumentation to holistically characterize vehicle state, but for space vehicles in flight, inspection is often either risky or costly. We propose using passive inspection CubeSats (PICs) to reliably and affordably image the launch vehicle with fine resolution, moderate coverage, and rapid response. The PICs mission, developed by Brigham Young University undergraduate students and manifested for launch in 2019, is designed to demonstrate this approach using two redundant CubeSats. Each CubeSat uses a spherical array of six cameras, 12 light-emitting diode flash units, and a field-programmable gate array to rapidly boot after separation and capture spherical images of its entire surroundings, including the launch vehicle and other released CubeSats. The PIC bus system includes several custom components, including a compact 3500-mAh battery system, 1 to 2 W solar power system, chassis, omnidirectional antenna system, radio, and a small integrated remove-before-flight-pin and system access port. Preliminary test results suggest the PICs are able to boot within 1.1 s of separation and inspect 1-mm features on each other and their launch vehicle, regardless of the spin developed at separation.

Satellite wind vector data are integral to atmospheric models and forecasts, but current measurement limitations make some synoptic, mesoscale activities difficult to observe. Using miniaturized electronics and advanced deployable mechanisms, new satellite wind scatterometers may be possible that increase spatial, temporal, and wind resolution and coverage. We propose a simple parametric model of the space of satellite wind scatterometer designs, their performances, and the transformation between design and performance, together called a solution space model. We explore two applications of this model: understanding how advances have expanded the design space and searching for alternative approaches to satellite wind scatterometry. Recent advances enable a greater capability-to-volume ratio, which enables constellations of small, low-cost scatterometers that co-operate in a variety of modes. We present two example concepts for constellations of co-operative satellite wind scatterometers. We estimate that a constellation of CubeSat scatterometers may affordably measure global ocean vector winds every three hours.

CubeSats have become very popular science platforms in the past decades, leading to a continuously increasing number of developers in the academic field. For science missions, customized payload electronics have to be developed, depending on measurement tasks and requirements. Especially for the deployment of complex remote sensing payloads, state-of-the-art performance is needed to provide operational control and specific data processing, e.g., for image sensors. Highly integrated system-on-module (SoM) architectures offer low resource requirements regarding power and mass, but moderate to high processing power capabilities. However, a requirement to use a standard SoM in a satellite is to quantify its radiation tolerance. The radiation environment has been modeled, estimating the hazards at module level and reducing the risks to an acceptable level by applying appropriate mitigation techniques. This approach results in a sensor electronics design that combines hardware and software redundancies to assure system availability and reliability for long-life science missions in low earth orbits. Integrated in a miniaturized limb sounding instrument for atmospheric remote sensing imaging, the payload electronics will be deployed on a technology demonstration satellite for in-orbit verification.

The CubeSat Infrared Atmospheric Sounder (CIRAS) is a proposed small satellite IR sounder developed in response to the challenges of high cost and possible data gaps for currently operational hyperspectral IR sounders, such as Atmospheric Infrared Sounder, Infrared Atmospheric Sounding Interferometer, and Crosstrack Infrared Sounder. A key objective of CIRAS is to demonstrate the technologies required for a low cost-to-orbit, operational IR sounder. CIRAS was designed by National Aeronautics and Space Administration’s Jet Propulsion Laboratory, with 625 channels measuring upwelling IR radiation of the Earth in the mid-wavelength IR spectrum region. Through observational system simulation experiments, the impact of assimilating CIRAS on global numerical weather prediction was assessed. CIRAS was simulated by applying the Community Radiative Transfer Model to profiles extracted from the Goddard Earth Observing System Model, Version 5 nature run for an afternoon polar orbiting sensor, and then assimilated by the National Centers for Environmental Prediction Global Data Assimilation System. Assimilating CIRAS improved global analysis and forecasts, when added to the currently operational observing configuration and to a data gap scenario.

STRaND and PhoneSat programs have attracted the attention of the aerospace community by applying, in CubeSats, commercial off-the-shelf smartphones based on Google’s Android. In Android, the development commonly applies Java hence this language is a natural candidate for the attitude and orbit control subsystem (AOCS). Moreover, such AOCS can be designed with success by linear control theory; however, the linearized models are not able to represent all the effects of the nonlinear terms present in the dynamics. Therefore, nonlinear control techniques can yield better performance. An example is the Nano-Satellite Constellation for Environmental Data Collection, used as the reference in this work, a set of remote sensing CubeSats from the Brazilian National Institute for Space Research, in which the AOCS must stabilize the satellite in three-axes. We present the investigation of a state-dependent Riccati equation (SDRE) controller, a nonlinear controller, based on attitude errors given by quaternions. The investigation uses Java, accordingly, it can run on an Android operating system in a CubeSat, and it has low cost. Two controllers (linear and SDRE) were evaluated using a Monte Carlo perturbation model. The initial results show that the SDRE controller provides better performance.

The purpose of this work is to develop a low-cost space telescope for Earth remote sensing that can be housed in a 12U CubeSat with a ground sampling distance (GSD) of 60 m at long wavelength infrared. It will use the three-mirror anastigmatic concept and commercial off-the-shelf components whenever possible. The main mission of the instrument is oil spill detection, especially in the ocean environment. Apart from the obvious functionality of this main mission, oil spill detection can be very useful for detecting aircraft that have crashed in the ocean. As secondary missions, this instrument can be used to map deforestation, find lost ships in the ocean, and help fight drug smuggling by finding camouflaged man-made objects. In order to fulfill these missions, the instrument will extract polarization information from the target. The optical design indicates that the optical performance of the instrument has a modulation transfer function >53  %   at Nyquist frequency and a native GSD of 60 m and that the instrument is diffraction limited (wavefront error of 0.036 wavelengths and Strehl ratio of 95.77%). The mechanical design indicates that the maximum deflection caused by gravity induces a minimal error (smaller than one over one hundred and forty) when compared to the mean wavelength of operation. The radiometric analyses indicate that the instrument has an NEDT of 56.30 mK.

The Micro-Sized Microwave Atmospheric Satellite (MicroMAS-2) is a 12-channel passive microwave radiometer on a CubeSat developed by Massachusetts Institute of Technology Lincoln Laboratory. MicroMAS-2 observations (1) enhance the current observing system capabilities and (2) mitigate the potential loss of some existing observations. Observing System Simulation Experiments (OSSEs) quantitatively estimate the expected impact of adding the MicroMAS-2 to the current global operational analysis and forecast systems or to a data gap scenario. First, MicroMAS-2 data on one low Earth-orbiting platform were simulated by the Community Radiative Transfer Model from the NASA GEOS-5 Nature Run, which serves as the truth in the OSSE. The assimilation of the MicroMAS-2 observations into the research version of the January 2015 NOAA Global Data Assimilation System in addition to the current operational observation system and in addition to a data gap scenario shows improvement on both analysis and forecast performance skills.

Hyperspectral infrared sounding in a CubeSat will provide a new dimension to the current suite of IR sounders by allowing measurements at multiple times of day and enabling formation flying of IR sounders for new data products such as atmospheric motion vector winds. We focus on technology development during the CubeSat Infrared Atmospheric Sounder (CIRAS) project sponsored by the NASA Earth Science Technology Office (ESTO) and coincident studies by the National Oceanic and Atmospheric Administration (NOAA) Office of Projects, Planning, and Analysis (OPPA). The CIRAS approach incorporates key instrument technologies, developed at the Microdevices Lab (MDL) at Jet Propulsion Laboratory (JPL), including a two-dimensional array of High Operating Temperature Barrier Infrared Detector (HOT-BIRD) material, selected for its high uniformity, low cost, low noise, and higher operating temperatures than traditional materials. The second key technology is a mid-wavelength infrared grating spectrometer designed by Ball Aerospace with a JPL MDL slit and immersion grating to provide hyperspectral infrared imaging in a CubeSat volume. The third key technology is a blackbody calibration target fabricated with MDL’s black silicon to have very high emissivity in a flat plate construction. JPL has completed design and breadboard of the mechanical, electronic, and thermal subsystems for the CIRAS payload including a HOT-BIRD focal plane assembly, with filters in a dewar and a breadboard of the electronics and scan mirror assembly. Blue Canyon Technologies, developer of the CIRAS 6U CubeSat, completed the Final Design Review for the spacecraft. NOAA is sponsoring the continued development of the CIRAS Proto-Flight Model (PFM) instrument at JPL using many of the existing subsystems. Completion of the PFM is expected in mid 2021, with launch no earlier than 2022.

Hyperspectral image classification is an important task for land cover interpretation that gives input to tremendous remote sensing applications. To deal with difficult classification problems, sparse representation has shown its significant potential. However, its execution is computationally complex and its performance limits its application in time-critical scenarii. We aim to improve the performance of an existent classification algorithm based on the sparse representation of extended multiattribute profiles as spectral–spatial features. We propose a parallel and distributed implementation of this algorithm on the computing engine SPARK using the MapReduce model to ensure the scalability of the classifier over available computing resources. Experimental results obtained on real and synthetic hyperspectral datasets show that the proposed approach reveals remarkable acceleration factors while retaining the same classification accuracy with regard to the sequential version.

Secondary succession is a process that is often observed taking place in former agricultural ecosystems. Its characteristics are especially important in protected areas, for the purposes of monitoring and protective measures. Effective mapping of succession is facilitated by the development of automated methodologies based on remote sensing data, which are capable of complementing traditional field research. The objective of this work is to determine whether the classification of high-resolution hyperspectral and light detection and ranging (LiDAR) data with the use of the random forest algorithm enables us to produce an accurate succession species map. First, feature extraction techniques are applied to 1-m hyperspectral images and a ∼7  point  /  m² dense point cloud. Minimum noise fraction layers and vegetation indices are calculated from the hyperspectral data and geometry related indices from the LiDAR data. Finally, the recursive feature elimination algorithm is applied to the combined dataset and the reference polygons to select the optimal set of features for subsequent classification. The results indicate that the proposed methodology has the potential to be used operationally. The final classification product is characterized by a relatively high Cohen’s kappa value of 0.68, with single species classified with various accuracies, expressed by F1 scores ranging from 0.45 to 0.87.

Monitoring of crops during their growth period is essential to increase the profitability of production and reduce costs. Unmanned aerial vehicles (UAVs) have been introduced as an appropriate tool in precision farming, due to their affordability and high-spatial resolution data. The potential of plant height and vegetation indices derived from UAV-based RGB imagery at 12-leaf growth stage (V12) is investigated for prediction of corn forage yield. Using the structure from motion techniques, the crop height model (CHM) was created by subtracting crop surface model from digital elevation model. Next, by evaluating the CHM using ground measurements of plant height, it was shown that UAV imagery provides high capability in modeling the plant height (R²  =  0.85). Pearson’s correlation coefficient (R) between variables showed that the plant height had the highest correlation with forage yield (R  =  0.84). Furthermore, among the vegetation indices, visible atmospherically resistant index with R  =  0.77, normalized difference index with R  =  0.75, and excess red with R  =  0.74 were better than other indices. Finally, the partial least squares regression (PLSR) model was constructed using plant height and vegetation indices to predict forage yield. The beta regression coefficients of the PLSR model were used to identify the most relevant predictors, and the final PLSR model was rebuilt with relevant predictors. The PLSR analysis confirmed that the plant height was the most important predictor, although integrating the vegetation indices in the PLSR model has slightly improved the prediction model (R²  =  0.74, RMSE  =  0.316  kg  /  m², and relative RMSE  =  12.39  %  ), compared to the model that used only plant height (R²  =  0.702, RMSE  =  0.339  kg  /  m², and relative RMSE  =  13.29  %  ). Overall, these results indicate that the UAV-based images are efficient and cost-effective data to predict corn forage yield as well as plant height modeling.

Latent heat flux is the main indicator of regional water-heat balance and plays an important role in drought monitoring and water resource management. Here, we attempt to estimate latent heat flux using a two-source energy balance model (TSEB). The decomposition algorithm of soil surface temperature and vegetation canopy temperature is discussed, and it is a key factor for calculating the latent heat flux in the TSEB model. Temperature decomposition was conducted using two methods: one is based on a simple linear relationship between the canopy temperature and directional radiation temperature and the other is based on soil latent heat flux expressed by the Priestley–Taylor formula. Then, the soil temperature was estimated using the soil latent heat flux. The estimation of the surface heat flux was based on the soil and vegetation canopy temperatures. The results show that the Priestley–Taylor formula method provided more accurate estimates of the latent heat flux than the linear relation method, and the reliability and precision were improved. The root-mean-squares error of the former method decreased by 38.8% compared with the latter method. The TSEB model was used to estimate the surface heat flux, and it was feasible for monitoring drought in typical drought-prone regions.

Remote sensing technology is an effective method of monitoring chlorophyll content, an important parameter for vegetation health. The chlorophyll information based on spectral information needs to consider the vertical characteristics of plants. Hyperspectral features [spectral reflectance (SR), spectral indices, and wavelet coefficients (WC)] were first selected to construct the cost function in the PROSPECT model-optimized inversion to improve the accuracy and efficiency of leaf chlorophyll content (LCC) inversion. Second, the sensitivity of LCC to leaf SR, vegetation index (VI), and WC were analyzed. Finally, LCC was inverted by the PROSPECT model using the iterative inversion algorithm, and the chlorophyll content of the vertical profile in maize was monitored. The following results were obtained. (1) According to the extended Fourier amplitude sensitivity test (EFAST) used to construct the cost-function inversion of LCC, the normalized difference vegetation index canste (NDVI_canste) and WC based on EFAST method which was used to construct the cost-function inversion of LCC, yielded higher accuracy than other spectral features. These two methods, combined with sensitivity analysis can provide accurate inversion results by weakening the influence of other parameters on spectral changes and eliminating the interference information between bands. (2) The cost function based on NDVI_canste in the LCC optimization exhibited overestimation, whereas WC can solve the problem efficiently. The continuous wavelet transform can extract weak information among the spectra, whereas the other two methods of SR and VI cannot easily obtain this information. (3) The vertical distribution of LCC in different maize varieties and treatments is in accordance with the parabola law, but the chlorophyll content of the different leaf positions on the vertical profile is somewhat different. The results reflect the vertical distribution of chlorophyll content through the radiation transfer of leaves, thus providing a theoretical basis for their further combination.

The Visible Infrared Imaging Radiometer Suite (VIIRS) imaging instrument on the polar-orbiting Suomi-National Polar-orbiting Partnership and National Oceanic and Atmospheric Administration-20 satellite platforms has infrared window bands but no carbon dioxide and water vapor absorption bands; the latter are essential for accurately deriving atmospheric variables (such as temperature and moisture profiles, cloud properties). An approach for fusing a high spatial resolution imager (e.g., VIIRS) and a high spectral resolution sounder [e.g., Cross-track Infrared Sounder (CrIS)] offers the opportunity to construct new spectral band radiances (radiance fusion) as well as sounder retrieval products (product fusion) at high spatial resolution. Furthermore, temperature and humidity profiles can also be provided at a high temporal resolution when geostationary imager [Advanced Baseline Imager (ABI)] data are used in combination with polar orbiting sounder (CrIS) data in a fusion approach. Promising results from VIIRS/CrIS and ABI/CrIS fusion are presented.

In the past, the Defense Meteorological Satellite Program—Operational Linescan System (DSMP/OLS) data have been successfully adjusted based on the presence of vegetation in each pixel. We propose a further adjustment based on radar data obtained by the SeaWinds scatterometer. It is shown that the adjustment results in substantial accuracy gains, measured as increased correlation with the more accurate Visible Infrared Imaging Radiometer Suite data as well as with the existing instances of radiance calibrated DSMP/OLS data. The proposed adjustment is easy to implement. The improvement of the proposed method is particularly significant in arid regions where vegetation is relatively sparse.

During construction of the Uma Oya Multipurpose Development Project, excavation of a headrace tunnel through highly tectonized metamorphic rock was slowed by repeated major water ingress events. In total, these events resulted in an estimated aquifer depletion of over 9  million m³, causing widespread surface subsidence. Persistent scatterer (PS) time series analysis of interferometric synthetic aperture radar (InSAR) scenes acquired by Sentinel-1 between July 28, 2016, and January 7, 2018, has revealed an ∼15  km² area of subsidence near reported water ingress locations, with a maximum line-of-sight subsidence value of 4.3 cm. A four-dimensional (4-D) filtering method based on averaging spatial and temporal pixel neighbors was applied to the unwrapped PS pixels, significantly reducing noise from sources that appear randomly in spatial and temporal domains and increasing the ability to accurately identify the onset of subsidence and track surface volume loss over time. A maximum average subsidence rate of 0.15  mm  /  day coincides with maximum tunnel discharge measurements, indicating that rock mass consolidation occurred as a result of rapid pore-pressure drawdown within an interconnected discontinuity network and led to fracture closure. The observed subsidence is punctuated by localized maxima and extends in a northwest–southeast orientation, suggesting that the extent of subsidence is governed by previously mapped structural trends of similar orientations. These results highlight the value of the 4-D filtering approach in processing InSAR PS data to help determine the magnitude and extent of surface subsidence.

Polarimetric SAR (PolSAR) can comprehensively describe the differences of targets combining the polarimetric scattering power and the relative phase information among the polarimetric channels. For the traditional image classification methods, a single classifier cannot fully solve the classification problem, in which certain inevitable bias between the classification results and the actual condition exists. To improve the classification performance, ensemble learning theory has been introduced into PolSAR classification, in which a number of different learners are integrated to get the final results. However, a larger number of learners may also lead to some negative effects. The concept of selective ensemble learning (SEL) is then introduced, in which some better learners are selected from a group of individual learners according to the selection strategy and then integrated to obtain a generalized classifier. Based on this idea, a SEL-based PolSAR image classification method is proposed, in which the representative features are extracted based on the characteristics of PolSAR image meanwhile high-performing learners are selected using genetic algorithm optimization. In order to verify the effectiveness and application of the proposed SEL-based PolSAR image classification method, the tests are implemented using UAVSAR L-band PolSAR data, besides the other two newly acquired experimental data of Chinese airborne and spaceborne system. The quantitative and qualitative experimental results confirm that the proposed method has excellent performance in classification.

Extracting roads from remote sensing images is an important task in the remote sensing field. We propose an approach of designing a light encoder–decoder network for road extraction. We analyze the relationship between road features and the receptive field of encoder–decoder networks and point out that the light encoder–decoder network can be achieved by controlling its receptive field. Based on this, we design road extraction networks on the architecture of a general encoder–decoder network, according to data specifications. In addition, we propose an adaptive weighted binary cross-entropy loss function to solve the problem of data imbalance in the training process. We validate our approach on the Massachusetts roads dataset and the DeepGlobe road extraction dataset. The experimental results show that our method reduces 98% and 94% of the parameters, respectively, compared with general encoder–decoder networks, whereas the performance of road extraction keeps well. Our approach has fewer parameters and good performance, so it is easier to deploy on mobile platforms.

Water–land classification is a basis for water depth calculation or suspended sediment concentration inversion through airborne LiDAR bathymetry (ALB). Traditional classification methods using ALB waveform data offer high accuracy but exhibit low efficiency and convenience in engineering applications. The three-dimensional (3-D) point cloud data of ALB are easier to analyze and utilize than waveform data. Therefore, we propose a water–land classification method that uses the 3-D point cloud data of ALB based on the threshold intervals of water surface points. First, a random sample consensus algorithm is applied to rough water–land classification using the 3-D point cloud data derived by an infrared laser of ALB. Second, the water surface points derived from rough classification are used to determine the means, standard deviations, and threshold intervals. Finally, accurate water–land classification is achieved on the basis of the threshold intervals of the water surface points. The proposed method is applied to a practical ALB measurement using Optech coastal zone mapping and imaging LiDAR and achieves 98.26% accuracy in water–land classification.

In some practical classification applications, our focus may be just a specific crop (e.g., rice, corn, or soybean). Therefore, how to distinguish the crop of interest from other classes accurately and quickly is an issue worth studying. We propose an approach that combines whitening transformation (WT)/interest-class-based whitening transformation (ICWT) and one-class classification (OCC) methods to identify the crop of interest. The image created by applying WT/ICWT to the original image is referred to as the totally/partially whitened image, respectively. Three typical OCC classifiers—one-class support vector machine (OCSVM), positive and unlabeled learning (PUL), and maximum entropy (MaxEnt)—are trained to identify the crop of interest, using a totally/partially whitened image. The research results demonstrate that (1) both ICWT and WT enabled OCSVM to improve the producer’s accuracy of the crop of interest by more than 20%, and the performance of the proposed approach reached or surpassed that of conventional supervised methods; (2) PUL and MaxEnt performed better than OCSVM, especially in identifying local major crops using different phased data; and (3) the proposed approach is efficient in single crop identification with remote sensing imagery and is worth recommending for practical applications.

As hyperspectral LiDAR (HSL) combines the advantages of hyperspectral remote sensing and LiDAR, it has the potential of estimating vegetation biochemical contents at any three-dimensional (3-D) location. We investigate the capability of HSL to monitor leaf chlorophyll and nitrogen contents at a distance of 7.5 m. Using full-waveform LiDAR data obtained by HSL, the performance of the partial least square regression (PLSR) model using three strategies (only wavelength bands, only vegetation indices, and both wavelength bands and vegetation indices) is explored based on the correlation analysis of 16 wavelength bands and 16 vegetation indices. The result shows that the PLSR model can make full use of various wavelength bands and vegetation indices, having a strong ability to monitor leaf chlorophyll and nitrogen contents using full-waveform LiDAR data. The analysis can be applied for predicting leaf biochemical components of other vegetation and provides the basis for monitoring 3-D distribution of biochemical contents using HSL in the future.

Soils in arid countries predominantly suffer from salinity and drought related to environmental problems that can lead to crop stress and low productivity. In this study, true-color aerial images from an unmanned aerial vehicle were used to assess the effect of soil and water salinity on date palm growth. Random soil samples (n  =  75) and irrigation water samples were collected from five sites and chemically analyzed. The custom algorithms were developed in ENVI^® and MATLAB.^® Green leaf index (GLI) was implemented to determine crop canopy attributes. Two segmentation methods namely between-class variance and foreground pixels were used to recognize the vegetation cover from other image pixels. The image analysis demonstrated that the mean value of GLI increased as the salinity levels decreased, R  =  0.96 and 0.92 for soil electrical conductivity (EC) and water EC, respectively. The percentage of area covered with vegetation was correlated to soil EC and water EC with about 70% accuracy. On the other hand, the percentage of area covered with palm trees only was used accurately to evaluate the soil EC by R²  =  0.89 and the water EC by R²  =  0.86. The findings of this research can set foundations for the development of aerial color imaging on salinity stressed date palm monitoring, providing useful information for decision makers on salinity management.

Narrowband normalized difference spectral indices (SI) have found wide application in vegetation studies. Consequently, several studies have investigated the utility of optimized SI for targeted applications. The objective of this study is to statistically develop optimized two-band normalized difference SI from a subset of hyperspectral bands derived using the Boruta wrapper algorithm. These indices are applied to model Fusarium circinatum stress in Pinus radiata seedlings. The performance of our developed optimized indices was compared with a selection of widely used existing SI (n  =  111) noted in the literature. Analyses were undertaken within a univariate (using the Jeffries–Matusita distance) and a multivariate (using the random forest algorithm) framework. Our results clearly demonstrate the improved accuracies using optimized SI (overall accuracy ranged from 76% to 96%) compared with using existing indices (overall accuracy ranged from 83% to 90%). Additionally, our results show that a multivariate approach yields superior results compared with a univariate approach. Overall, the results demonstrate the operational potential of optimized two-band normalized difference SI within a multivariate paradigm.

Multisensor image fusion has gained tremendous significance due to various satellites operating in different parts of the electromagnetic spectrum. We present a hybrid fusion approach to integrate information from synthetic aperture radar (SAR) and multispectral (MS) imagery to improve land use land cover (LULC) classification. The major concern in SAR and optical fusion is the spectral distortion in the fused image, which is significantly less in pansharpening algorithms. The primary objective of our work is to inject unique spatial information from the SAR image into MS images, deriving enhanced data. The proposed approach is based on the integration of principal component analysis and wavelet decomposition to reduce spectral distortion in the fused image. Fused images are evaluated visually and statistically. Results are compared with conventional fusion approaches. In order to explore the effectiveness of the proposed technique, LULC classification is performed on the fused and original data. The LULC classification results are analytically compared with the standard thematic map to derive classification accuracy. A comparative analysis with other approaches conclusively proves that the proposed hybrid approach is superior to conventional approaches.

Leaf equivalent water thickness (LEWT) is an important parameter in ecological and environmental monitoring. Our study proposes two new indices, the normalized differential projection index (NDPI) and the distance projection ratio index (DPRI), by considering the angle of βSWIR1 and the length, projection, and relationships among the random forest feature bands at 721, 1466, and 2061 nm, corresponding to NIR, SWIR1, and SWIR2 wavebands, respectively. Single-factor analysis of NDPI and DPRI and multifactor analysis of other vegetation indices are used to build a LEWT linear model whose effectiveness is evaluated using 10-fold cross validation. Despite their simple structure, the NDPI and DPRI can explain the majority of the variation in LEWT. After single-factor analysis, results from DPRI are superior to those from previous vegetation indices or equations with a higher coefficient of determination (R²  =  0.79) and lower relative root-mean-square error (rRMSE  =  9.79  %  ). Multifactor analysis shows that the model (R²  =  0.81, rRMSE  =  9.45  %  ) built using DPRI and the water index is the most accurate. The proposed use of NDPI and DPRI as parameter bands to build such models provides a method for further study of the vegetation water content inversion.

Superpixel-level image classification methods take advantage of contextual information of pixels and reduce the time cost in training and test processes. However, extracting a superpixel-level feature is a challenging task because each superpixel has irregular size and shape. A superpixel-level polarimetric feature extraction (SLPFE) based on circular loop spatial pyramid pooling (CLSPP) (SLPFE_CLSPP) is proposed for polarimetric synthetic aperture radar (PolSAR) image classification. The main idea is that the CLSPP layer divides the feature maps of each superpixel into the same number of circular loops and produces consistent feature size for each superpixel, which then feeds into the next processing step, such as classification. SLPFE_CLSPP not only makes it possible to generate the same size feature but also to utilize the spatial information of pixels and reduce the running time during SLPFE process. Compared with the existing methods, the proposed method makes a balance between overall accuracy and running time. Experimental results on real PolSAR dataset demonstrated the superiority of the proposed approach.

Image fusion is an important technique in remote sensing to improve visual interpretation and classification. Pansharpening is the procedure of fusing panchromatic (PAN) and multispectral images to produce high spatial and spectral resolution images. Synthesized pansharpening is performed on Linear Imaging Self-Scanning Sensor III and Advanced Wide Field Sensor data, which are freely available and provided by the National Remote Sensing Center. The Adaptive Neuro-Fuzzy Inference System (ANFIS) in multiscale transform domain for multisensor image fusion application is evaluated. The state-of-the-art method has been evaluated by various quality metrics. The computational cost of ANFIS with wavelet, contourlet, shearlet, and curvelet transform is investigated. This study proves that curvelet with ANFIS-based fusion technique outperformed state-of-the-art techniques. The application will be used to incorporate the missing spectral information in the high spatial resolution PAN image to identify objects, highlighting the regions clearly.

To assess the potential of machine learning methods for predicting and mapping soil salinity in highly vegetated and slightly salt-affected croplands, a random forest (RF) regression model was constructed based on the measured soil salinity and spectral indices of remote sensing data, Landsat 8 Operational Land Imager (OLI), and Moderate Resolution Imaging Spectroradiometer (MODIS). The model was evaluated under different soil salinity conditions and with different resolutions of imagery data (Landsat 8 OLI and MODIS) to compare it with the multivariate adaptive regression splines (MARS) model. The results indicated that a higher accuracy estimation can be achieved from the RF model under conditions of poor correlations between the soil salinity and the spectral indices in the highly vegetated croplands, with a cross-validation determination coefficient   (  R²  )    =  0.86, root-mean-square error   (  RMSE  )    =  1.83, and ratio of performance to deviation   (  RPD  )    =  2.7; these were better than the MARS model (R²  =  0.81, RMSE  =  4.8, and RPD  =  1.96). The performance of the MARS model shows sensitivity to the variety of salinity levels. Estimated total salinity maps of Jiashi County were produced using the RF model based on OLI and Moderate Resolution Imaging Spectroradiometer (MODIS) data for September 2015 and April 2016, respectively. Acceptable mapping accuracy was achieved from the cross validation of the salinity mapping results for September 2015 (R²  =  0.84 and 0.81 for OLI and MODIS data, respectively) and April 2016 (R²  =  0.85 and 0.82 for OLI and MODIS data, respectively). The results indicated that the method developed is reliable and accurate for digital soil salinity mapping of slightly and moderately salt-affected croplands.

We exploit a bag of visual words approach with random point patterns to remote sensing image classification. The image is divided into patches of equal size, and the local maximums of its histogram are determined. For every intensity in the image, a point pattern is formed based on patch centers and according to local histogram maximums. Further, we investigate the type of the generated random point pattern and mark corresponding fragments with a label. The created point patterns can be of three types: cluster, regular, or random. Therefore, image patches are labeled as a cluster, regular, random, cluster-regular, cluster-random, regular-random, cluster-regular-random, or none when patch’s center does not belong to any point pattern. We form a representation in which the frequencies of patch types are used to differentiate images between the classes. Also, a spatial extension, which considers the relative arrangement of patch types based on their co-occurrences, is used. In addition, a median frequencies vector of the detected patch intensity maximums as the image spectral feature description is added. The introduction of such features significantly reduces the size of a codebook while delivering a classification accuracy comparable to matching the known methods.

The use of ground truth (GT) data in the learning and/or assessment of classification algorithms is essential. Using a biased or simplified GT attached to a remote sensing image to partition does not allow a rigorous explanation of the physical phenomena reflected by such images. Unfortunately, this scientific problem is not always treated carefully and is generally neglected in the relevant literature. Furthermore, the impacts of obtained classification results for decision-making are negative. This is inconsistent when considering investments in both the development of sophisticated sensors and the design of objective classification algorithms. Any GT must be validated according to a rigorous protocol before utilization, which is unfortunately not always the case. The evidence of this problem is provided, using two popular hyperspectral images (Indian Pine and Pavia University) that misleadingly are frequently used without care by the remote sensing community since the associated GTs are not accurate. The heterogeneity of the spectral signatures of some GT classes was proven using a semisupervised and an unsupervised classification method. Through this critical analysis, we propose a general framework for a minimum objective assessment and validation of the GT accuracy, before exploiting them in a classification method.

We report the results of our long-term monitoring study of cirrus clouds using our locally developed lidar sytem located in Hanoi, Vietnam (21.03°N, 105.85°E). Our data, spanning the period from 2011 to 2017, show that the cirrus clouds above Hanoi are generally found between 6.7 and 18.0 km with a mean height of 14.0 km. The cloud thickness is estimated to vary between 0.2 and 7.0 km with an average thickness of 2.3  ±  1.2  km. We also derive an average optical depth of 0.32  ±  0.22 and an average lidar ratio 30  ±  18. We also compare the geometrical and optical properties of cirrus clouds above Hanoi, where until now lidar study is still lacking, with that obtained for other locations in Asia and around the world.

We describe the response of synthetic aperture radar to precipitation normalized radar cross section using a forward scattering model. Based on the snow-layer, rainfall-layer, and zero-layer bright band, the attenuation models of volume scattering and surface scattering echoes are established, and a retrieval algorithm based on a three-layer structure is proposed. The information of the starting point and width of the rainfall area is obtained by data preanalysis, and then the precipitation intensity is inverted by combining the model-oriented statistical and Volterra integral equation methods. The retrieve results in three cases show that the algorithm is scientific and effective, reduces the hypothetical conditions of the model, reduces the error of retrieve of precipitation by bright band, and the retrieved data are accurate and reliable.

Monitoring grassland biomass throughout the growing season is of key importance in sustainable, site-specific management decisions. Precision agriculture applications can support these decisions. However, precision agriculture relies on timely and accurate information on plant parameters with a high spatial and temporal resolution. The use of structural and spectral features derived from unmanned aerial vehicle (UAV)-based image data from low-cost sensors is a promising nondestructive approach to assess plant traits such as above-ground biomass or plant height. Therefore, the main objectives were (1) to evaluate the potential of low-cost UAV-based canopy surface models to monitor sward height as an indicator of grassland biomass, (2) to evaluate the potential of vegetation indices from low-cost UAV-based red-green-blue (RGB) digital image data, and (3) to compare the mentioned methods with established methods for biomass monitoring such as rising plate meters and spectroradiometer-based narrowband vegetation indices over the growing season in 2017, including three cuts. We compared the accuracy of each single UAV-based height feature and vegetation index using a combined multivariate approach to estimate fresh and dry biomass. The heterogeneous sward structure with high spatiotemporal variability led to varying performance in biomass estimation depending on the growths (time between two cuts) and choice of predictor variable. The results showed that biomass prediction by height features provided moderate-to-good results (cross-validation R²  =  0.57 to 0.73 for dry biomass and 0.43 to 0.79 for fresh biomass), but reference measurements based on rising plate meters were more robust when estimating biomass. The spectral features (RGB-based vegetation indices and spectroradiometer-based vegetation indices) yielded varying accuracy and suitability for biomass prediction. Despite the variability, our findings indicate a promising approach for grassland biomass monitoring.

Knowing the thermodynamic phase of a cloud–whether it is composed of spherical water droplets or polyhedral ice crystals–is critical in remote sensing applications and in climate studies. Liquid water and ice have different absorptive properties in certain spectral bands that can be exploited to identify the phase of clouds using ground-based, passive remote sensing. Our simulations found that ground-based radiance measurements at three spectral channels (1.55, 1.64, and 1.70  μm) provide improved discrimination when analyzed in three spectral dimensions as opposed to previous approaches based in two-dimensional parameter space. Our simulations show that these bands provide good discrimination between liquid-water and ice clouds when the optical depth is large. We also show measurements from a ground-based spectrometer confirming the cloud-phase sensing ability of these three channels, with validation provided by a dual-polarization lidar system.

Hyperspectral imaging (HSI) continues to grow as a method for remote detection of vegetation, materials, minerals, and pure chemicals. We have used a longwave infrared (7.7 to 11.8  μm) imaging spectrometer in a static outdoor experiment to collect HSI data from 24 minerals and background materials to determine the efficacy with which HSI can remotely detect and distinguish both pure minerals and mineral mixtures at a 45-deg tilt angle relative to ground using two different backgrounds. Measurements were obtained separately for the minerals and materials mounted directly on both a bare plywood board and a board coated with aluminum foil: 19 powders (3 mixtures and 16 pure mineral powders) held in polyethylene bottle lids as well as five samples in rock form were taped directly to the boards. The primary goal of the experiment was to demonstrate that a longwave infrared library of solids and minerals collected as directional-hemispherical reflectance spectra in the laboratory could be used directly for HSI field identification along with simple algorithms for a rapid survey of the target materials. Prior to the experiment, all 24 mineral/inorganic samples were measured in the laboratory using a Fourier transform infrared spectrometer equipped with a gold-coated integrating sphere; the spectra were assimilated as part of a larger reference library of 21 pure minerals, 3 mixtures, and the polyethylene lid. Principal component analysis with mean-centering was used in an exploratory analysis of the HSI images and showed that, for the aluminum-coated board, the first principal component captured the difference between the signal that resembled a blackbody and the highly reflective aluminum background. In contrast, the second, third, and fourth principal components were able to discriminate the materials including phosphates, silicates, carbonates, and the mixtures. Results from generalized least squares target detection clearly showed that laboratory reference spectra of minerals could be utilized as targets with high fidelity for field detection.

Change detection using remote sensing imagery is a broad and highly active field of research that has produced many different technical approaches for multiple applications. The majority of these approaches have in common that they do not deliver any detailed information concerning the type, category, or class of the detected changes. With respect to the extraction of such information, recent research often suggests that a land use classification is required. This classification can be accomplished in an unsupervised or supervised way, whereas the practicability of both strategies is more or less limited by the usage of reference or training data. Moreover, expert knowledge is needed to arrive at meaningful land use classes. An approach is presented that overcomes these drawbacks. A time series of synthetic aperture radar amplitude images is considered, enabling the detection of so-called high activity objects in urban environments. Such objects represent the basis of the investigations and denote the input for unsupervised categorization and classification procedures. The method supports even the unexperienced user in learning the actual information content leading to the capability to define a suitable scheme for change classification. Tests carried out on two different datasets suggest that the method is both practical and robust.

Polarimetric synthetic aperture radar (PolSAR) image classification is a challenging task due to lack of effective feature representation approaches. However, when it comes to deep feature representation, there is little research that pays enough attention to make full use of existing expert knowledge. We propose a model for deep polarimetric feature extraction, and a superpixel map is used to integrate contextual information. The proposed model uses multiple polarimetric algebra operations, polarimetric target decomposition methods, and convolutional neural network (CNN) to extract deep polarimetric features. The core idea is to utilize expert knowledge of the target scattering mechanism interpretation to assist the CNN classifier in feature extraction and employ superpixel algorithm based on Wishart distribution to improve the final classification performance. The proposed method is able to get abstract and discriminative representation from initial features, achieving robust and improved performance for PolSAR images. Compared with other state-of-the-art methods, experiments on the classification of three real PolSAR datasets are performed to demonstrate the superiority of the proposed approach.

Forest carbon estimation currently largely relies on remote sensing techniques in combination with field measurement. High-resolution images, which are commonly utilized for carbon estimation, are not readily available, and their cost prohibits communities from reaping the benefits of maintaining their forest under the UN reducing emissions from deforestation and forest degradation program. Our study explores the combination of readily available and relatively cheaper unmanned aerial vehicle (UAV) (4-cm resolution) and multispectral Pleiades (50-cm resolution) images for species classification robustness in view for carbon estimation through object-based image analysis. The images are resampled and used to evaluate the effect of combining multispectral Pleiades image on the accuracies of segmenting UAV images for tree crown projection area (CPA) estimation and species classification. RGB images from a UAV platform are processed in a photogrametric software and combined with the near-infrared band of a Pleiades image to get a UAV-Pleiades image composite. The images are segmented using the ESP 2 tool and the segmentation accuracy compared using a paired t-test. The segmented tree crowns are classified using random trees (RT), support vector machines (SVM), and maximum likelihood (ML) classifiers, and the classification accuracies of the three classifiers are compared using the McNemar’s chi-squared test. Our study demonstrates a 93.5% accuracy of segmenting UAV-Pleiades image composite, which is significantly higher than the 84.8% accuracy of segmenting UAV images (p  <  0.05). Also an 84% classification accuracy of UAV-Pleiades image composite is significantly higher than the 54% classification accuracy of the UAV images (p  <  0.05). Of the three classifiers used, the classification accuracies of SVM and RT are significantly higher (p  <  0.05) than that of the ML classifier. Given the significantly high accuracies observed from this study for tree CPA extraction and tree species classification, carbon/above ground biomass modeling is possible with significantly high accuracy using the combination of multispectral Pleiades and UAV images.

With the increasing use of natural gas, gas emissions (predominantly methane) from natural gas infrastructure are exacerbating the greenhouse effect. Therefore, it is crucial to detect natural gas leaks in a timely manner. The remote sensing of plants offers the potential to identify leakage from the reflectance responses of the plants to elevated concentrations of natural gas in the soil. Solar-induced chlorophyll fluorescence (SIF) is directly related to photosynthesis and is thus proposed to track the impact of natural gas leakage on vegetation growth. In order to explore the feasibility and advantage of SIF for detection of natural gas leaks, a field experiment of natural gas microleakage is performed, and soybean and grass are chosen as the test plants. Sixteen plots, eight with soybean and eight with grass, are established, and four of each plant are chosen for the injection of natural gas into the soil. The other plots act as controls. We present weekly ground measurements of canopy spectra over plots of gassed and ungassed soybean and grass. SIF at 760 nm (F760) and other published vegetation indices are calculated and analyzed. The indices can identify grasses under natural gas microleakage stress throughout the whole experiment, whereas only F760 can identify soybeans under natural gas microleakage stress during the later stage of the experiment. When the canopy chlorophyll content of soybean is at high levels, F760 can still increase along with it, whereas other indices fail to increase because of saturation. The results suggest that the F760 index can stably and reliably identify plants under natural gas microleakage stress. Therefore, in the future, the location of leakage of natural gas can be detected by monitoring the F760 of plants with hyperspectral remote sensing.

A training sample refining method is proposed to improve the classification performance of very high-spatial resolution (VHR) remote sensing images. The proposed approach involves three major steps. First, for a given image, an initial sample set with a limited number for each class is prepared manually. Second, neighboring pixels around each available labeled pixel are gradually distinguished by an adaptive extension algorithm. When an iterative extension around the available pixel is terminated, the neighboring pixels that are within the extended region are taken into account as candidate training samples. The candidate training sample is then used to refine the signature of each initial sample. Third, when the whole available labeled pixels are scanned and processed pixel-by-pixel in the above manner, the revised training sample set is trained specially for a supervised classifier for classification. Three VHR remote sensing images with limited initial samples are used for evaluating different classifiers and advanced methods based on spatial–spectral features to investigate the feasibility and performance of the proposed approach. Higher classification performance and accuracies are obtained by our proposed approach with respect to the classification maps based on the initial training sample set and an existing method that improves the initial training set by a regular window.

Subpixel target detection in hyperspectral images (HSIs) is a problem that emerges in many important civil, military, and environmental applications. The widely used, simple in calculation, matched-filter detector is well-known to exhibit reduced performance when the target-hosting image comprises spectrally diverse, incoherent background regions. We propose to calculate and apply matched-filter subpixel target detection in image regions that exhibit both spectral and spatial coherence. Specifically, our contributions are as follows. (a) We propose the spatial–spectral expectation-maximization (SSEM) algorithm for the segmentation of a target-hosting HSI. (b) We conduct local matched filter computation and application in the SSEM-calculated image segments. The proposed method is accompanied by extensive experimental studies that corroborate its merits.

Satellite image fusion techniques have been used for more than two decades. Development of satellite sensors for very high-resolution satellite imagery monitoring contributed to the development of new image fusion techniques. We examine the quality of Ehlers, Brovey transform, modified intensity–hue–saturation (M-IHS), and high-filter resolution merge fusion methods on vegetation indices values combining high-resolution WorldView-4 and low-resolution Sentinel-2 imagery. For image fusion, four bands—blue, green, red, and near-infrared—were used. The effect of fusing Sentinel-2 and WorldView-4 imagery is tested on the various vegetation indices [normalized difference vegetation index (NDVI), blue normalized difference vegetation index (BNDVI), and green normalized difference vegetation index (GNDVI)]. M-IHS fusion showed the best result, with the least difference between indices calculated based on the original Sentinel-2 and fused bands. Difference between vegetation indices calculated using original Sentinel-2 and M-IHS fused bands for 30.08.2015, 30.09.2016, and 30.09.2017 is between −0.041 to 0.037. The second part of this research is an evaluation of three different vegetation classes (grassland, mixed vegetation, and forest) on three different years and three different months within a year. Difference between NDVI, GNDVI, and BNDVI calculated based on the original Sentinel-2, and M-IHS fused bands are ranging from −0.043 to 0.103, −0.041 to 0.050, and −0.052 to 0.030 for grassland, mixed vegetation, and forest, respectively. On the other hand, the difference between vegetation indices for grassland, mixed vegetation, and forest are from −0.110 to 0.050, −0.041 to 0.038, and −0.052 to 0.038, when observing August, September, and October of 2017, respectively.

Object detection is one of the most important research topics for remote sensing image (RSI) analysis. Because of the rapid development of satellite imaging technology, the resolution of RSI has increased dramatically, which brings a great challenge to the efficiency of RSI-based object detection. We present an object detection model, which greatly improves the detection speed by using learning-based hashing techniques. Specifically, first, a selective search method is used to generate many high quality object proposals that may contain objects. Then for each proposal, we learn rotation-invariant binary codes, which can handle the objects with arbitrary orientations in RSI, to quickly eliminate most nonobject proposals in Hamming space. And finally, the object detection task can be achieved by classifying the left (very limited amount of) proposals with more discriminating classification model. Experimental evaluations on a public very high-resolution remote sensing dataset show the superiority of the proposed method. Specifically, the rotation-invariant learning metric has been combined with three popular hashing methods, and performance improvements have been obtained for all the cases. Furthermore, comparisons with deep learning-based detectors show that hash-based methods fail to achieve state-of-the-art detection performance, but when the graphics processing unit is inaccessible, it achieves a good compromise between performance and runtime.

In many remote sensing (RS) applications, haze greatly affects the quality of optical RS images, but we do not always have the conditions to acquire multiple images in the same area for haze removal tasks. Therefore, the research on haze removal from a single RS image is necessary. Previous haze-removal methods introduce various prior knowledge to solve this problem, and thus, the quality of these methods largely depends on the reliability and validity of prior knowledge, which brings various limitations. We propose and validate a deep-learning-based model for haze removal, named haze removal fully convolutional network, to estimate transmission maps and generate corresponding haze-removed images via an atmospheric scattering model. Moreover, we propose an approximate method to produce hazy-and-clear image pairs as a dataset for training and validation. Experiments using this dataset demonstrated that the proposed model achieved the desired results in both visual effect and quantitative measurement.

The presence of shadows in remote sensing images leads to misinterpretation of objects and a wrong discrimination of the targets of interest, therefore, limiting the use of several imaging applications. An automatic area-based approach for shadow detection is proposed, which combines spatial and spectral features into a unified and flexible approach. Potential shadow-pixels candidates are identified using morphological-based operators, in particular, black-top-hat transformations as well as area injunction strategies as computed by the well-established normalized saturation-value difference index. The obtained output is a shadow mask, refined in the last step of our method in order to reduce misclassified pixels. Experiments over a large dataset formed by more than 200 scenes of very high-resolution images covering the metropolitan urban area of São Paulo city are performed, where the images are collected from the WorldView-2 (WV-2) and Pléiades-1B (PL-1B) sensors. As verified by an extensive battery of tests, the proposed method provides a good level of discrimination between shadow and nonshadow pixels, with an overall accuracy up to 94.2%, for WV-2, and 90.84%, for PL-1B. Comparative results also attested that the designed approach is very competitive against representative state-of-the-art methods and it can be used for further shadow removal-dependent applications.

In recent years, target detection has drawn increasing attention in the field of radar signal processing. We consider the problem of coherent integration for detecting a highly maneuvering target with complex motion, involving range migration (RM), linear Doppler frequency migration, and quadratic Doppler frequency migration within the coherent interval. Owing to the different sensitivity of the envelope and phase to the target motion, the parametric and nonparametric models are, respectively, established. A coarse-to-fine integration algorithm is correspondingly proposed. Specifically, the keystone transform and fold factor searching are first utilized to remove the parametric RM and to achieve coarse integration. Then, the embedded phase gradient autofocus is applied to estimate the nonparametric nonlinear Doppler phase and realize refined integration. Compared with several representative algorithms, the proposed method could achieve a good trade-off between computational complexity and detection performance. Moreover, this method can be easily extended to detect a high-order maneuvering target. Finally, both simulations and real data experiments are presented to demonstrate the effectiveness of the proposed method.

The ability to obtain reconstructed floor maps of acceptable quality with less samples speeds up the acquisition phase, which is of the highest relevance for most radio-frequency (RF)-based through-the-wall mapping (TWM) applications, particularly in the area of law enforcement and security. The key issue impacting the acquisition time is the quantity of necessary samples for acceptable floor reconstruction quality, which in turn depends on the intersample distance. We evaluate the impact of enlarging the intersample distance in the reconstruction quality of an algebraic method applied to RF-based TWM, under a parallel-beam acquisition geometry. Finite-element method (FEM) is used to define a physical model, allowing high accuracy in the simulation results. Two test floor maps are built in the FEM model, and the correspondent very-high frequency RF samples are generated in each case, composing the so-called sinograms. The sinograms are forwarded to the algebraic reconstruction method—simultaneous iterative reconstruction technique—to provide estimates of the original floor maps. In simulations, the use of sinogram upsampling to improve image resolution leads to floor maps of acceptable reconstruction quality in a parallel-beam RF-based TWM setup with intersample distances up to two wavelengths for an angular step of 4.5 deg.

Agricultural burning in Mexico and Central America is a common practice to remove crop residues. Characterization of the spatial and temporal distribution of burning emissions is essential to evaluate their environmental impacts. For this reason, smoke emissions from agricultural fires were calculated using two methods: direct and indirect. The direct method calculates the rate of smoke emission by multiplying the smoke emission coefficient by the fire radiative energy (FRE) release rate. The smoke emission coefficient was calculated from satellite measurements of aerosol optical depth and FRE made by the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument. The indirect method calculates the rate of smoke emissions by using FRE observations from MODIS to parameterize a diurnal fire cycle based on a Gaussian equation for each ecoregion. Fire emission estimates from each of these methods were compared with each other and with the Global Fire Emissions Database, version 4.1. A total of 482,506 and 233,660 fire pixels were detected by MODIS from 2003 to 2014 in Mexico and Central America, respectively. Active fires concentrate in the tropical forests with maximum frequencies from March to May. Agricultural fires account for 40% of total detected burns from 2003 to 2014. Considerable interannual variability is observed, with 2003, 2005, and 2011 the most active years and 2008 and 2014 the least active years. Our calculations show that agricultural smoke emissions estimated by the indirect method are lower than those estimated by the direct method by a factor of 1.4 to 4, depending on the ecoregion, although the total smoke emissions are similar between both methods. Finally, average annual smoke emissions in Mexico are the greatest in tropical humid forests and least in tropical dry forests, whereas in Central America, average annual smoke emissions are very similar for both humid and dry tropical forests.

Hyperspectral unmixing is one of the most important procedures for remote sensing image processing. The non-negative matrix factorization (NMF)-based method has been widely used for hyperspectral unmixing since it can get endmember and abundance matrices simultaneously. However, the inherent single-decomposition structure of NMF may not achieve good performance for highly mixed data. To solve this issue, we propose a spectral and spatial total-variation-regularized multilayer non-negative matrix factorization (SSTV-MLNMF) for hyperspectral unmixing. In SSTV-MLNMF, we designed an effective multilayer factorization process and combined spectral and spatial total variation as extra regularization. These could enhance the smoothness for spectral signatures and spatial fields, which could achieve better performance. Experiments on both synthetic and real datasets have validated the effectiveness of our method and have shown that it has outperformed several state-of-the-art approaches of hyperspectral unmixing.

High-resolution inverse synthetic aperture radar (ISAR) imaging of fast rotating targets is a complicated task. Targets’ fast rotational motion always introduces migration through range cells (MTRC) when we use the conventional range Doppler algorithm. We mainly focus on the ISAR imaging of a fast rotating target. In that case, the samples in the cross-range dimension are insufficient, which is the undersampling case. Compressed sensing-based ISAR imaging methods are generally faced with the problem of basis mismatch, which may degrade the ISAR image. A two-dimensional iterative reweighted super-resolution algorithm is proposed by iteratively decreasing a surrogate function. We also compared the performance of the proposed method with other state-of-the-art sparse recovery methods. Simulation results show that the proposed method can achieve the high-resolution ISAR imaging of fast rotating targets. Moreover, the MTRC and the basis mismatch problems can be effectively solved.

We present a semisupervised hyperspectral unmixing solution that incorporates the spatial information between neighbor pixels in the abundance estimation procedure. The proposed method is applied to a polynomial postnonlinear mixing model in which each pixel reflection is characterized by a nonlinear function of pure spectral signatures corrupted by additive white Gaussian noise. The image is partitioned into different classes containing similar materials with the same abundance vectors. We model the spatial correlation of pixels of each class by the Markov random field. A Bayesian framework is used to iteratively estimate each class and its corresponding abundance vector. Here, we propose the sparse Dirichlet prior for abundance vectors to demonstrate a semisupervised scenario. A Markov chain Monte Carlo algorithm is used to estimate abundance vectors. The major contribution of this work is based on combination of spatial correlation with nonlinear mixing models in a semisupervised scenario. The proposed approach is compared to linear mixing model, generalized bilinear mixing model, and the conventional polynomial postnonlinear mixing model algorithms. The results on both simulated and real data show the outperformance of the proposed algorithm by achieving lower errors in unmixing and reconstruction of hyperspectral images.

Pansharpening is a process used to extract high spatial resolution from panchromatic images and high spectral resolution from multispectral images at the same time. The intensity–hue–saturation (IHS) is the simplest and most efficient fusion method in current fusion methods, which could preserve spatial information well, but there is some spectral distortion. The main reason is that, in the IHS space, the estimated intensity image used to replace the original intensity image is not accurate enough. A method is proposed to overcome this defect, which aims to combine the fast generalized IHS (FGIHS) fusion technique based on spectral adjustment with a variational optimization model to estimate a more accurate intensity image. In the variational model, it is assumed that the down-sampled intensity image should be close to the original intensity image, retaining the spectral information effectively. The experimental results show that this method not only preserves the spectral information effectively but also improves the spatial quality of the fusion results in the view of the subjective and objective evaluations.

Linear mixed model (LMM) has been extensively applied for hyperspectral compressive sensing (CS) in recent years. However, the error introduced by LMM that limits the reconstruction performance has not been given full consideration. We propose an algorithm for hyperspectral CS based on LMM under the assumption of known endmembers. At the sampling stage, only spectral compressive sampling is carried out to keep the abundance information as much as possible. At the reconstruction stage, the proposed algorithm estimates abundance by using linear unmixing from the spectral observed data. Moreover, the model error introduced by LMM is explored; a joint convex optimization scheme for estimation of both abundance and model error is established and solved by the alternating iteration approach to achieve the optimal reconstruction. Experimental results on a real hyperspectral dataset demonstrate that the proposed algorithm significantly outperforms the other state-of-the-art hyperspectral CS algorithms.

In recent years, the geosynchronous satellite-based passive inverse synthetic aperture radar (ISAR) has been widely employed for target detection and automatic target recognition. However, the great distance between the satellite and the receiver causes low signal-to-noise ratio (SNR) of the target signal, which makes the system unable to obtain a well-focused image of the target. To obtain a well-focused ISAR image within a limited coherent processing interval, a method for improving the SNR is required. We propose a two-step noise suppression method to enhance the SNR by eliminating the noise components contained in the received data. The target windowing method is utilized first to reject the noise and clutter interference by windowing the target data from the range-Doppler map. Then, a denoising method is employed to further remove the noise components from the selected target data by threshold techniques. By using the two-step noise suppression, the SNR of the target signal is increased, and consequently a quality-improved ISAR image can be obtained. Simulation results validate the effectiveness of the proposed noise suppression method.

The ability to differentiate a non-native aquatic plant, Myriophyllum spicatum (Eurasian watermilfoil or EWM), from other submerged aquatic vegetation (SAV) using spectral data collected at multiple scales was investigated as a precursor to mapping of EWM. Spectral data were collected using spectroradiometers for SAV taken out of the water, from the side of a boat directly over areas of SAV and from a lightweight portable radiometer system flown from an unmanned aerial system (UAS). EWM was spectrally different from other SAV when using 651 spectral bands collected in ultraviolet to near-infrared range of 350 to 1000 nm but does not provide a practical system for EWM mapping because this exceeds the capabilities of available airborne hyperspectral imaging systems. Using only six spectral bands corresponding to an available multispectral camera or eight wetlands-centric bands did not reliably differentiate EWM from other SAV and assemblages. However, a modified version of the normalized difference vegetation index (mNDVI), using a ratio of red-edge to red light, was significantly different among dominant vegetation groups. Also, averaging the full range of spectral to 65 10-nm wide bands, similar to available hyperspectral imaging systems, provided the ability to identify EWM separately from other SAV. The UAS-collected spectral data had the lowest remote sensing reflectance versus the out-of-water and boatside data, emphasizing the need to collect optimized data. The spectral data collected for this study support that with relatively clear and calm water, hyperspectral data, and mNDVI, it is likely that UAS-based imaging can help with mapping and monitoring of EWM.

Aimed at solving the limitation of the traditional unsupervised segmentation method of remote sensing images that rely more on features and prior knowledge, an unsupervised remote sensing image segmentation method combined with the atrous convolution autoencoder and the bidirectional long- and short-time memory autoencoder autoencoder (BiLSTM) network is proposed. The multispectral remote senses images using the dimensional reduction capacity of the autoencoder to obtain the transverse low-dimensional spatial features of the image through the atrous convolution autoencoder, and they use the BiLSTM autoencoder to learn the longitudinal low-dimensional spectral features of the multiband sequence of the remote sensing image. The batch normalization layer is added to optimize the whole network, and the visualization method is used to dynamically train and adjust the model. The data clustering of two kinds of low-dimensional features is obtained by the softmax algorithm, and unsupervised image segmentation of the multispectral data is finally realized by fusing the clustering results. In the experiment, the original data of Landsat8 remote sensing images were clustered and segmented directly. The results show that the accuracy is up to 83.25% through the clustering of dual autoencoder network in the water information extraction, which is obviously better than the common clustering algorithm and can get better image segmentation results.

Gross primary productivity (GPP) is a significant indicator of terrestrial system carbon cycle and is one of the most important parameters of global climate change. Accurate modeling of GPP is important for understanding the changes of carbon dioxide in the atmosphere. Although many studies have been conducted to explore methods for estimating GPP, large uncertainties still exist among different models. The combination of remote sensing data with flux data is a feasible approach to determine site-based GPP. Using multisource remote sensing data, meteorological data obtained from Google Earth Engine, and GPP estimates from eddy covariance measurement, we present a model, based on the random forest regression (RFR) algorithm, to predict site-scale GPP. First, we train and tune the RFR model for 16 global deciduous broadleaf forest (DBF) flux sites by remote sensing-based drivers. Then we evaluate the RFR model’s performance by predicting GPP for the other eight DBF sites. Our results show that, compared with the MOD17A2H GPP product, this GPP prediction method improves modeling of GPP, with an R² of 0.82 and an RMSE of 1.93  g C m^−  2 d^−  1. Contribution analysis shows that the enhanced vegetation index is the most important driver among input variables. Our study may provide a reliable method to predict GPP in global DBF, which is promising for different-scale evaluations of carbon sequestration.

The integration of multiple Global Navigation Satellite Systems (GNSSs), such as the Chinese BeiDou Navigation Satellite System (BDS) and the American Global Positioning System (GPS), makes it possible to improve the accuracy of single-point positioning (SPP). However, the current accuracy of GNSS SPP with code measurement is on the order of several meters. In order to further improve SPP accuracy, we develop a multistep weighted least squares (MWLS) estimation method based on the elevation-dependent weighted least squares (EWLS) method. In this approach, the weight of each visible satellite is determined by its elevation and azimuth, and the coordinate components are semi-independently and separately calculated by MWLS. We conduct open-sky and blocked static tests as well as a vehicle-based kinematic test using single-frequency receivers to assess the effectiveness of this approach. Comparing the positioning errors of GPS and GPS/BDS SPPs between the MWLS and EWLS methods shows that the former improved the positioning accuracy more than the customary latter. Specifically, MWLS achieves a horizontal accuracy of about 1 m and a positional accuracy of 2 to 3 m in clear observational environments, about 15% better than ELWS.

Land surface water is one of the most vital resources on Earth for human survival, and it is necessary to distinguish water bodies from nonwater features. Remote sensing techniques are among the most widely used approaches for monitoring water resources; many waterbody delineation methods have been proposed, and multiband spectral water indices are among the most popular. These methods utilize blue, green, near-infrared (NIR), middle-infrared, and shortwave infrared bands but do not involve the red band. A waterbody delineation index is introduced, i.e., the deeply clear waterbody delineation index (DCWDI), which is based on the reflectance of the red and NIR bands. In NIR-red spectral space, the distance between deeply clear water pixels and the coordinate origin, O, is less than that of other land cover types. This method can use the distance between any point E and the coordinate origin O to differentiate deeply clear water pixels from nonwater pixels, i.e., objects near point O are always deeply clear water bodies or extremely wet regions. The accuracy and robustness of the DCWDI are tested using Landsat 8 operational land imager images of Hongjiannao Lake, Qinghaihu Lake, and Lingao Reservoir. The performance of the DCWDI is compared with that of the normalized difference water index (NDWI), automated water extraction index (AWEI), and the modified NDWI (MNDWI). The net shoreline movement (NSM) and the area errors between delineated water areas and the “true” areas of water bodies are adopted to evaluate the accuracies of the four classifiers. The mean   |  NSM  |   values from the DCWDI of Hongjiannao Lake, Qinghaihu Lake, and Lingao Reservoir are 17.033, 75.108, and 11.021 m, respectively, which are smaller than those from the MNDWI (34.641, 149.308, and 19.647 m), NDWI (71.607, 164.503, and 22.151 m), and AWEI_nsh (19.957, 113.119, and 11.126 m). The average of   |  NSM  |   and area error from the DCWDI are smaller than those from the other three classifiers. These results show that the lake boundaries derived from the DCWDI are spatially similar to the “true” lake boundaries, and the DCWDI, which is obtained from NIR-red spectral space, can be applied to delineate and detect the changes in information on deeply clear water. The spatial location information contained within NSM can validate the accuracy of a classifiers’ ability to extract water bodies.

The deep convolution neural network (CNN) has been of great interest in hyperspectral image classification recently. Current CNN-based approaches adopt a two- or three-dimensional convolution network to extract spectral–spatial features from local pixel neighborhoods. High computation cost and overfitting problems should be handled carefully for these large-scale networks. A compact one-dimensional (1-D) CNN-based method is proposed to explore the joint spectral–spatial features by employing the pixel coordinates as the spatial information. Furthermore, two kinds of CNN architecture for spectral–spatial feature fusion are compared. To be specific, the shallow fusion model (SF-CNN) concatenates the spatial information to the spectral features before they are fed into the final fully connected layer, whereas the deep fusion model (DF-CNN) combines the spectral and the spatial information in an early stage and extracts the high-level spectral-spatial features by the 1-D convolution layers. The experimental results demonstrate that the DF-CNN method provides competitive classification results to state-of-the-art methods. Moreover, attributed to the concise spatial information and the effective feature fusion structure, the proposed method is economical in terms of computation cost when compared with the current CNN-based spectral-spatial methods.

Considering data reliability, the optimized connections of dual residues for phase discontinuity reconstruction provide a more reliable scheme and produce more robust unwrapped results. However, their practical implementation usually involves a time-consuming iterative global operation, which is not suitable for application to the phase unwrapping (PU) of big blocks of interferometric synthetic aperture radar (InSAR) phase data. A parallel PU method based on local minimum reliability dual expanding is proposed. With given quality weight maps, the dual reliability is defined based on residues, and the minimum reliability residue pairs are introduced to represent possible discontinuous boundaries. We provide a dynamic expanding method for the pairs with local minimum reliability searching and dual merging. The final minimum balanced trees obtained are used for path integrating of PU with the aid of reliability maps. The calculation of reliability maps, the residue pair searching, and the dynamic expanding are designed to be carried out in parallel. We employ the interface propagation scheme based on the eikonal equation and flood-filling for the parallel implement. Two big blocks of airborne InSAR data were processed by the proposed method, the experimental results and analysis verified its robustness and efficiency for the large-scale PU problem.