Detection and identification of oil spill species based on polarization information.

Abstract

Aiming at the problem of poor oil identification accuracy in existing oil spill detection technologies, the polarization degree model of oil spill on rough sea surface under different azimuths and zenith angles was established based on Fresnel theory. The analytical expressions of visible light polarization degree in calm and fluctuating water surface were derived respectively, and the polarization degree model of oil spill in reflection space was constructed. The effectiveness of the method and its influence on the polarization distribution of oil spill were analyzed by simulation. A portable turntable was designed to test the polarization characteristics of the experiment, and the visible light polarization detection experiment was carried out. The visible light polarization images of five typical oil spills at different observation azimuth and zenth angles were obtained. The differences in the polarization degrees of different oil species were analyzed, and the correctness of the theoretical model was proved by experiments. The polarization detection experiment of visible light pBRDF was completed, which more intuitively showed the variation law of the polarization characteristics of light reflected by different oil spills in different spatial positions. Using polarization information to distinguish oil species is a useful supplement to the traditional oil spill detection method and has important significance to improve the marine pollution control ability.