Hyperspectral Manipulation for the Water Stress Evaluation of Plants

There are high demands for water content estimation in vegetation, e.g. water-stress control for sweet crops, forest disease monitoring and drought monitoring. In this paper, normalized difference-based and ratio-based water stress indices by means of hyperspectral information from NIR to SWIR, spectral ranges of InGaAs sensor, are introduced to facilitate realizing simple measurement system at reasonable cost. Regardless of the simple definition, sufficient estimation accuracies are realized in the proposed indices under the condition of laboratory observation. The experimental results based on airborne hyperspectral forest images showed that the water-stress indices are useful to detect oak wilt areas. Keyword: Hyperspectral data, water stress, vegetation, normalized difference index, ratio-base index.