Pesticide residue detection by hyperspectral imaging sensors

This paper presents a new application of hyperspectral imaging to detection of pesticide residues. We conducted a series of base-line experiments which are particularly designed for pesticides commonly used in agriculture in Taiwan. In order to detect different levels of pesticide residue concentration, an unsupervised subpixel detection algorithm is developed for this purpose. It firsts applies Automatic Target Generation Process (ATGP) to find potential target signal sources and then followed by Constrained Energy Minimization (CEM) to perform subpixel detection. The conducted experiments demonstrate that using hyperspectral imaging processing techniques can effectively detect pesticide residues which cannot be accomplished by traditional image processing methods.