Skin Detection over Vegetation Backgrounds using Hyperspectral Imaging

Hyperspectral imaging or imaging spectroscopy can be used for the detection of people over cluttered backgrounds in applications such as defense, security, and search and rescue operations.  The spectral signature information collected at high spectral resolution by hyperspectral imagers over the VIS/SWIR region of the electromagnetic spectrum can be used to detect skin signatures over vegetation and other types of backgrounds. This poster will present results of our work in studying hyperspectral detection of skin. Data collection using hyperspectral imagers and spectrometers for algorithm testing and validation will be described. Then experimental results comparing different algorithms based on band ratio and standard signature-based detection and classification algorithms for hyperspectral images will be presented. Spectral angle detector show good results for detecting skin spectra over vegetation backgrounds.