Search

Mid-infrared (MIR) reflectance spectroscopy is commonly studied as a rapid and nondestructive method for predictive soil analysis under laboratory conditions. The first objective of this paper is to report an MIR spectral library based on 20,000+ soil samples collected from the United States. The second objective is to assess, using partial least squares regression (PLSR) and artificial neural networks (ANN), the performance of the library to predict 12 physical and chemical soil properties: organic carbon (OC), inorganic carbon (IC), total carbon (TC), total nitrogen (TN), clay, silt, sand, Mehlich-3 extractable phosphorus (P), NH4OAc extractable potassium (K), cation exchange capacity (CEC), total sulfur (TS), and pH. The third objective is to investigate whether the use of auxiliary variables of master horizon (HZ), taxonomic order (TAXON), and land use land cover (LULC) would improve MIR model performance. The results showed that OC, IC, TC, TN and TS were predicted most satisfactorily with R² > 0.95 and RPD (ratio of performance to deviation) > 5.5. Soil CEC, pH, clay, silt, and sand were also predicted satisfactorily with R² > 0.75 and RPD > 2.0. P and K were predicted poorly, with R² < 0.4 and RPD < 1.4. The ANN models generally outperformed PLSR models, except for clay, silt and sand. Using auxiliary variables (HZ, TAXON, and LULC) to develop stratified models generally improved model performance. The HZ-specific models showed the greatest improvements. Using an MIR spectral library for routine soil analysis would positively impact many modern applications where high spatial resolution, quantitative soil data are demanded. [ABSTRACT FROM AUTHOR]

Abstract: Encroachment of invasive shrubs into grassland areas on rangelands in the southwestern United States threatens the viability of livestock production and can severely alter hydrology and biodiversity. Novel remote sensing technologies may provide useful information for monitoring and remediating this threat. The objectives were to investigate multiple endmember spectral mixture analysis (MESMA) as an approach to map rangeland vegetation using hyperspectral remote sensing imagery and to test the sensitivity of MESMA to alternative image spatial resolutions and spectral waveband combinations. Data from two Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) overflights at the Jornada Experimental Range in southwestern New Mexico were used in the analysis. Endmember spectra were selected from a library of ground-based spectral observations collected with a field spectroradiometer. A 4-endmember MESMA was conducted for both AVIRIS images at their native spatial resolutions using 113 10-nm wavebands from 422 to 2339nm. Additional MESMAs were conducted at 10 multiples of the images'' native spatial resolution and for 6 alternative combinations of spectral waveband subregions. Maps of endmember fractional cover for green shrub vegetation, nonphotosynthetic grass vegetation, and bare soil were comparable to an earlier vegetation classification map of Jornada. MESMA fractional cover estimates for the green vegetation endmember were positively correlated with the normalized difference vegetation index (NDVI) with correlation coefficients (r) greater than 0.58. Correlation coefficients between the sum of the green and nonphotosynthetic vegetation endmembers and the cellulose absorption index (CAI) were greater than 0.59. Correlation coefficients between MESMA fractional green vegetation cover and NDVI from independent multispectral images were greater than 0.57. Despite losses in spatial detail at coarser image spatial resolutions, MESMA results for images with spatial resolution degraded by a factor of 10 (~150m) were similar to aggregated results for MESMA at the native spatial resolution (~15m). Additionally, MESMA results were shown to be substantially more sensitive to the spectral wavebands used in the analysis as compared to the spatial resolution of the images. Considered together, the MESMA results at Jornada indicated that fine spectral resolution with hyperspectral remote sensing was substantially more important than incremental changes in image spatial scale from 15m to 150m. [Copyright &y& Elsevier]

Abstract: Hyperspectral imaging is a growing area in remote sensing in which an imaging spectrometer collects hundreds of images (at different wavelength channels) for the same area on the surface of the Earth. Hyperspectral images are extremely high-dimensional, and require advanced on-board processing algorithms able to satisfy near real-time constraints in applications such as wildland fire monitoring, mapping of oil spills and chemical contamination, etc. One of the most widely used techniques for analyzing hyperspectral images is spectral unmixing, which allows for sub-pixel data characterization. This is particularly important since the available spatial resolution in hyperspectral images is typically of several meters, and therefore it is reasonable to assume that several spectrally pure substances (called endmembers in hyperspectral imaging terminology) can be found within each imaged pixel. In this paper we explore the role of hardware accelerators in hyperspectral remote sensing missions and further inter-compare two types of solutions: field programmable gate arrays (FPGAs) and graphics processing units (GPUs). A full spectral unmixing chain is implemented and tested in this work, using both types of accelerators, in the context of a real hyperspectral mapping application using hyperspectral data collected by NASA''s Airborne Visible Infra-Red Imaging Spectrometer (AVIRIS). The paper provides a thoughtful perspective on the potential and emerging challenges of applying these types of accelerators in hyperspectral remote sensing missions, indicating that the reconfigurability of FPGA systems (on the one hand) and the low cost of GPU systems (on the other) open many innovative perspectives toward fast on-board and on-the-ground processing of remotely sensed hyperspectral images. [Copyright &y& Elsevier]

Abstract: The Process Analytical Technologies (PAT) initiative of the US Food and Drug Administration (US FDA) has established a framework for the development of imaging techniques to determine the real-time distribution of mixture components during the production of solid dosage forms. This study, which is the first in a series of two parts, uses existing mixing indices and a new criterion called the “percentage of homogeneity” (H%) to assess image homogeneity. Image analysis techniques use feature extraction procedures to extract information from images subjected to treatments including colour segmentation and binarization. The surface distribution of components was determined by macropixel analysis, which splits an image into non-overlapping blocks of a preset size and calculates several statistical parameters for the resulting divisional structure. Such parameters were used to compute mixing indices. In this work, we explored the potential of image processing in combination with mixing indices and H% for assessing blending end-point and component distribution on images. As a simplified test, an arrangement of binary and ternary systems of coloured particles was mixed collecting at-line multispectral (MSI) and non-invasive RGB pictures at preset intervals. [Copyright &y& Elsevier]