Oil Spill Detection Using Hyperspectral Infrared Camera

Oil spill pollution is a severe environmental problem that persists in the marine environment and in inland water systems around the world. Remote sensing is an important part of oil spill response. The hyperspectral images can not only provide the space information but also the spectral information. Pixels of interests generally incorporate information from disparate component that requires quantitative decomposition of these pixels to extract desired information. Oil spill detection can be implemented by applying hyperspectral camera which can collect the hyperspectral data of the oil. By extracting desired spectral signature from hundreds of band information, one can detect and identify oil spill area in vast geographical regions. There are now numerous hyperspectral image processing algorithms developed for target detection. In this paper, we investigate several most widely used target detection algorithm for the identification of surface oil spills in ocean environment. In the experiments, we applied a hyperspectral camera to collect the real life oil spill. The experimental results shows the feasibility of oil spill detection using hyperspectral imaging and the performance of hyperspectral image processing algorithms were also validated.