Description: Low-cost unmanned airborne vehicles (UAVs) are emerging as a promising platform for remote-sensing data acquisition to satisfy the needs of wide range of applications. Utilizing UAVs, which are equipped with directly georeferenced RGB-frame cameras and hyperspectral push-broom scanners, for precision agriculture and high-throughput phenotyping is an important application that is gaining significant attention from researchers in the mapping and plant science fields. The advantages of UAVs as mobile-mapping platforms include low cost, ease of storage and deployment, ability to fly lower and collect high-resolution data, and filling an important gap between wheel-based and manned-airborne platforms. However, limited endurance and payload are the main disadvantages of consumer-grade UAVs. These limitations lead to the adoption of low-quality direct georeferencing and imaging systems, which in turn will impact the quality of the delivered products. Thanks to recent advances in sensor calibration and automated triangulation, accurate mapping using low-cost frame imaging systems equipped with consumer-grade georeferencing units is feasible. Unfortunately, the quality of derived geospatial information from push-broom scanners is quite sensitive to the performance of the implemented direct georeferencing unit. This paper presents an approach for improving the orthorectification of hyperspectral push-broom scanner imagery with the help of generated orthophotos from frame cameras using tie point and linear features, while modeling the impact of residual artifacts in the direct georeferencing information. The performance of the proposed approach has been verified through real datasets that have been collected by quadcopter and fixed-wing UAVs over an agricultural field.

Description: Segmentation is a key issue in the processing of multidimensional images such as those in the field of remote sensing. Most of the segmentation algorithms developed for multidimensional images begin by reducing the dimensionality of the images, thus loosing information that could be relevant in the segmentation process. Evolutionary cellular automata segmentation (ECAS-II) is an evolutionary approach that provides cellular automata-based segmenters considering all the spectral information contained in a hyperspectral image without applying any technique for dimensionality reduction. This paper presents an efficient graphics processor unit implementation of the type of segmenters produced by ECAS-II for land cover hyperspectral images. The method is evaluated over remote sensing hyperspectral images, introducing it on a spectral–spatial classification scheme based on extreme learning machines. Experiments have shown that the proposed approach achieves better accuracy results for land cover purposes than other spectral–spatial classification techniques based on segmentation.

Description: The soil moisture active passive (SMAP) L-band synthetic aperture radar (SAR) could continuously provide global km scale ocean surface wind observations, which had a better coverage than other SARs and a higher spatial resolution than scatterometers. This paper investigates SMAP normalized radar cross sections (NRCS) dependence on wind vectors using more than 5 million matchups consisting of Defense Meteorological Satellite Program F17 Special Sensor Microwave Image/Sounder wind speed, National Center for Environmental Predication wind direction and SMAP L-band NRCS. An L-band geophysical model function (GMF) is proposed for SMAP wind retrieval on the basis of these matchups, and it indicates wind speed and direction dependence of SMAP L-band NRCS for about 40° incidence angle and 0–25 m/s wind speed range in both HH and VV polarization. The wind speed dependence increases rapidly with wind speed, and HH-polarized one is greater than VV polarization. The upwind–downwind difference for HH polarization is greater than that for VV polarization. A negative upwind–crosswind asymmetry occurs for HH- and VV-polarized backscatter at lower wind speeds. The retrieved SMAP wind speed using the proposed GMF is validated by using National Data Buoy Center buoy winds. The root mean square differences and biases are 1.77 and 0.19 m/s, respectively. The accuracies of SMAP wind speeds at 0–10 m/s range are better than those at higher wind speed range. In addition, SMAP wind speeds in upwind and downwind directions are relatively more accurate than those in crosswind directions.

Description: Many studies have demonstrated the efficient extraction of the spatial extent of urban areas from Defense Meteorological Satellite Program/Operational Linescan System imagery using a fixed thresholding technique. These studies may underestimate and overestimate the extents of small and large cities, respectively. To overcome this problem, a new intensity gradient (IG) and vegetation fractional coverage (VFC) method is developed for identifying cities or towns, principally based on the assumption that there is a border around a city at which the nighttime light intensity decreases sharply. Using this method, the spatial extents of urban areas for two of the biggest countries in the world, namely China and the United States, were extracted in 2010. The urban areas thus identified are compared with the urban areas interpreted from Landsat Thematic Mapper imagery, and the results show that there is a significant linear relationship between the former and latter areas. This demonstrates that the IG/VFC model is effective for efficiently extracting the extent of urban areas from nighttime light imagery.

Description: Features are of great importance for synthetic aperture radar (SAR) imagery terrain classification, but low-level features usually readily suffer from the speckle noise and they are incapable or inaccurate to capture some complex and irregular texture structure. In this paper, a novel feature learning framework is proposed to address this problem, in which some mid-level and high-level features are simultaneously learned by exploiting the spatial context constraints and sparse priors. More specifically, the mid-level features served as the intermediates are extracted from several initialized low-level features by the spatial constraints to reduce the influence of the speckle noise. Then, more abstract and discriminative high-level features are learned with an effective dictionary learning algorithm so as to represent the complex structures in SAR imagery. Finally, both artificial synthesis and real SAR imagery are utilized to verify the effectiveness of the proposed framework. It is demonstrated from both quantitative evaluations and visual results that the proposed algorithm performs better than other compared algorithms and the learned high-level feature is robust to the speckle noise and can improve the classification performance.

Description: Rice crops are important in global food economy and are monitored by precise agricultural methods, in which crop morphology in high spatial resolution becomes the point of interest. Synthetic aperture radar (SAR) technology is being used for such agricultural purposes. Using polarimetric SAR (PolSAR) data, plant morphology dependent electromagnetic scattering models can be used to approximate the backscattering behaviors of the crops. However, the inversion of such models for the morphology estimation is complex, ill-posed, and computationally expensive. Here, a metamodel-based probabilistic inversion algorithm is proposed to invert the morphology-based scattering model for the crop biophysical parameter mainly focusing on the crop height estimation. The accuracy of the proposed approach is tested with ground measured biophysical parameters on rice fields in two different bands (X and C) and several channel combinations. Results show that in C-band the combination of the HH and VV channels has the highest overall accuracy through the crop growth cycle. Finally, the proposed metamodel-based probabilistic biophysical parameter retrieval algorithm allows estimation of rice crop height using PolSAR data with high accuracy and low computation cost. This research provides a new perspective on the use of PolSAR data in modern precise agriculture studies.

Description: This paper proposes a novel class allocation strategy in units of object (UOO) for soft-then-hard super-resolution mapping (STHSRM). STHSRM involves two processes: 1) subpixel sharpening and 2) class allocation. The UOO is implemented in the second process by integrating the object boundaries as an additional structural constraint. First, UOO obtains the object boundaries from remote-sensing images by image segmentation. The number of subpixels within an object is then calculated for each class to meet the coherence constraint of fractional images imposed by soft classification. Finally, a linear optimization model is built for each object to obtain the optimal hard class labels of subpixels. A synthetic image and two real remote-sensing images are used to evaluate the effectiveness of UOO. The results are compared visually and quantitatively with two existing class allocation methods: 1) the highest attribute values first (HAVF) and 2) units of class (UOC). The experimental results show that UOO performs better than these two methods. UOO can better reduce the salt and pepper effect in resultant maps than both HAVF and UOC when dealing with real remote-sensing images. Moreover, UOO can better maintain the structure of land-cover patches, with smoother boundaries as compared with the two methods.

Description: Open ocean and coastal area monitoring requires multispectral satellite images with a middle spatial resolution $({sim 300 {text{m}}})$ and a high temporal repeatability $({sim 1 {text{h}}})$ . As no current satellite sensors have such features, the aim of this study is to propose a fusion method to merge images delivered by a low earth orbit (LEO) sensor with images delivered by a geostationary earth orbit (GEO) sensor. This fusion method, called spatial spectral temporal fusion (SSTF), is applied to the future sensors—Ocean and Land Color Instrument (OLCI) (on Sentinel-3) and Flexible Combined Imager (FCI) (on Meteosat Third Generation) whose images were simulated. The OLCI bands, acquired at t 0 , are divided by the oversampled corresponding FCI band acquired at t 0 and multiplied by the FCI bands acquired at t 1 . The fusion product is used for the next fusion at t 1 and so on. The high temporal resolution of FCI allows its signal-to-noise ratio (SNR) to be enhanced by the means of temporal filtering. The fusion quality indicator ERGAS computed between SSTF fusion products and reference images is around 0.75, once the FCI images are filtered from the noise and 1.08 before filtering. We also compared the estimation of chlorophyll (Chl ) , suspended particulate matter (SPM), and colored dissolved organic matter (CDOM) maps from the fusion products with the input simulation maps. The comparison shows an average relative error s on Chl, SPM, and CDOM, respectively, of 64.6%, 6.2%, and 9.5% with the SSTF method. The SSTF method was also compared with an existing fusion method called the spati- l and temporal adaptive reflectance fusion model (STARFM).