Search

Hyperspectral satellite imagery has developed rapidly over the last decade because of its high spectral resolution and strong material recognition capability. Nonetheless, the spatial resolution of available hyperspectral imagery is inferior, severely affecting the accuracy of ground object identification. In the paper, we propose an adaptively optimized pulse-coupled neural network (PCNN) model to sharpen the spatial resolution of the hyperspectral imagery to the scale of the multispectral imagery. Firstly, a SAM-CC strategy is designed to assign hyperspectral bands to the multispectral bands. Subsequently, an improved PCNN (IPCNN) is proposed, which considers the differences of the neighboring neurons. Furthermore, the Chameleon Swarm Optimization (CSA) optimization is adopted to generate the optimum fusion parameters for IPCNN. Hence, the injected spatial details are acquired in the irregular regions generated by the IPCNN. Extensive experiments are carried out to validate the superiority of the proposed model, which confirms that our method can realize hyperspectral imagery with high spatial resolution, yielding the best spatial details and spectral information among the state-of-the-art approaches. Several ablation studies further corroborate the efficiency of our method. [ABSTRACT FROM AUTHOR]

Heterogeneity, including the inhomogeneity of landscapes and surface variables, significantly affects the accuracy of evapotranspiration (ET) (or latent heat flux, LE) estimated from remote sensing satellite data. However, most of the current research uses statistical methods in the mixed pixel to correct the ET or LE estimation error, and there is a lack of research from the perspective of the remote sensing model. The method of using frequency distributions or generalized probability density functions (PDFs), which is called the "statistical-dynamical" approach to describe the heterogeneity of land surface characteristics, is a good way to solve the problem. However, in attempting to produce an efficient PDF-based parameterization of remotely sensed ET or LE, first and foremost, it is necessary to systematically understand the variables that are most consistent with the heterogeneity (i.e., variability for a fixed target area or landscape, where the variation in the surface parameter value is primarily concerned with the PDF-based model) of surface turbulence flux. However, the use of PDF alone does not facilitate direct comparisons of the spatial variability of surface variables. To address this issue, the objective of this study is to find an indicator based on PDF to express variability of surface variables. We select the dimensionless or dimensional consistent coefficient of variation (CV), Gini coefficient and entropy to express variability. Based on the analysis of simulated data and field experimental data, we find that entropy is more stable and accurate than the CV and Gini coefficient for expressing the variability of surface variables. In addition, the results of the three methods show that the variability of the leaf area index (LAI) is greater than that of the land surface temperature (LST). Our results provide a suitable method for comparing the variability of different variables. [ABSTRACT FROM AUTHOR]

Hydro-agricultural applications often require surface soil moisture (SM) information at high spatial resolutions. In this study, daily spatial patterns of SM at a spatial resolution of 1 km over the Babao River Basin in northwestern China were mapped using a Bayesian-based upscaling algorithm, which upscaled point-scale measurements to the grid-scale (1 km) by retrieving SM information using Moderate Resolution Imaging Spectroradiometer (MODIS)-derived land surface temperature (LST) and topography data (including aspect and elevation data) and in situ measurements from a wireless sensor network (WSN). First, the time series of pixel-scale (1 km) representative SM information was retrieved from in situ measurements of SM, topography data, and LST. Second, Bayesian linear regression was used to calibrate the relationship between the representative SM and the WSN measurements. Last, the calibrated relationship was used to upscale a network of in situ measured SM to map spatially continuous SM at a high resolution. The upscaled SM data were evaluated against ground-based SM measurements with satisfactory accuracy—the overall correlation coefficient (r), slope, and unbiased root mean square difference (ubRMSD) values were 0.82, 0.61, and 0.025 m3/m3, respectively. Moreover, when accounting for topography, the proposed upscaling algorithm outperformed the algorithm based only on SM derived from LST (r = 0.80, slope = 0.31, and ubRMSD = 0.033 m3/m3). Notably, the proposed upscaling algorithm was able to capture the dynamics of SM under extreme dry and wet conditions. In conclusion, the proposed upscaled method can provide accurate high-resolution SM estimates for hydro-agricultural applications. [ABSTRACT FROM AUTHOR]