Spider Mite Detection and Canopy Component Mapping in Cotton Using Hyperspectral Imagery and Spectral Mixture Analysis

摘要

Spectral mixture analysis and hyperspectral remote sensing are analytical and hardware tools new to precision agriculture. They can allow detection and identification of various crop stresses and other plant and canopy characteristics through analysisof their spectral signatures. One stressor in cotton, the strawberry spider mite (Tetranychus turkestani U.N.), feeds on plants causing leaf puckering and reddish discoloration in early stages of infestation and leaf drop later. To determine the feasibility of detecting the damage caused by this pest at the field level, AVIRIS imagery was collected from USDA-ARS cotton research fields at Shafter, CA on 4 dates in 1999. Additionally, cotton plants and soil were imaged in situ in 10 nm increments from 450 to 1050 nm with a liquid-crystal tunable-filter camera system. Mite-damaged areas on leaves, healthy leaves, tilled shaded soil, and tilled sunlit soil were chosen as reference endmembers and used in a constrained linear spectral mixture analysis to unmix the AVIRIS data producing fractional abundance maps. The procedure successfully distinguished between adjacent mite-free and mite-infested cotton fields although shading due to sun angle differences between dates was a complicating factor. The resulting healthy plant, soil, mite-damaged, and shade fraction maps showed the distribution and relative abundance of these endmembers in the fields. These hardware and software technologies can identify the position, spatial extent, and severity of crop stresses for use in precision agriculture.