Research on the classification for MODIS images of oil slick on the sea

Moderate-resolution Imaging Spectroradiometer (MODIS) is regarded as the important detection instrument onboard the TERRA and AQUA satellites, and its data is multi-spectral. Compared with NOAA-AVHRR and other Landsats, besides more frequently it transits, more extensive area with lower cost, MODIS has higher spatial resolution and spectral resolution. These characteristics are helpful to recognize and real-time monitor the type of surface. Based on the character of MODIS data, this paper focuses on the research of marine oil slick classification. Firstly, obtain each band data of oil slick area to create oil slick characteristic vector and spectral library, secondly, set up the threshold according to the spectral library, finally, classify the oil slick area. Because of the different character of oil slick area in each band and the distortion at random, this paper researches the problem of band selection. The experimental results show that the proposed method has well effect on marine oil slick area classification.