The objective of this paper is to present the methods of oil spill detection and segregation of look a-likes using SAR images of India's first Radar Imaging Satel-lite (RISAT-1) that are based on Hybrid Polarity SAR architecture. The Hybrid polarity SAR will transmit pulses in circular polarization and receive echoes in linear polarizations. Detection of oil spills on sea sur-face is a complex process, because some of the natural phenomena, which resemble oil spills ( look a-likes ) frequently occur, particularly at low wind conditions. Rain cells, wind front areas, organic film, grease ice are some examples of look a-likes. The images of the Oil, spilled for experimental reasons on ocean surface near Norway were acquired by RISAT-1 and the Single Look Complex images are obtained. Stokes parame-ters S1,S2,S3,S4 are derived for the images. From these Stokes parameters, the relative phase, orientation, ellipticity, and polarization angle are calculated. Fur-ther, four decomposition methods viz;, m-chi (elliptici-ty parameter), m-delta (relative phase parameter), and the new m-psi (orientation parameter), m-alpha (polari-zation angle), are applied to detect the oil slick and also to isolate look a-likes. The relative phase, orienta-tion angle, ellipticity, and polarization angle parameters will change their characteristics with respect to dielec-tric constant of the sea surface material. Since the di-electric constant of the oil spill is different from that of a the sea and look a-likes, the possibility to separate the oil spill area from the sea and look a-likes is ex-plored using all the four decomposition methods. The preliminary results are encouraging and the viability of unsupervised classification based on decomposition methods for identification of oil spill regions by Hybrid Polarity SAR images is presented in this paper....
Authors: LINGENAHALLI JAYADEV VIJAYA KUMAR, J.K.KISHORE, P KESAVA RAO.