Research On Marine Oil Spills Monitoring Using Hyper-spectral Romote Sensing

With the increasing oil spills at the sea, the marine environment is subjetc to the harm of human society more seriously, which cause not only a large area of marine oil pollution, but also more damage to the marine life. Therefore, the establishment of a sound oil spill contingency plan has become one of the focuses in the world. The remote sensing technology has become one of the most important means in the oil spill monitoring and analysis. Currently, scientists use the multi-band optical sensors and SAR carried on the airborne and spaceborne platforms to monitor the oil spill, and then distinguish the target based on the spectral characteristics, but it is also limited by climatic conditions, weather and false targets, then to detect the location, type and thickness issues accurately is difficult.Hyper-spectral with high spectral resolution and large amout of data has become the leading field of remote sensing technology, it can obtain a lot of image data which has a very narrow band interval. Hyper-spectral can make up for the deficiencies of existing sensors in the quantitativly monitoring oil spill type, size and so on, and its development and application has cause more and more attention.This paper uses the Hyperion data which is the only satellite in orbit to carry out the oil spill information. We innovatively attempt to use the non-image area targets as the dark and bright objects respectively to correct the image atmospherically based on the empirical line calibration method. The results show that the spectral characteristics of green vegetation are obvious. At the same time according to the characteristics of hyper-spectral data, we use the spectral angle mapper method for image classification, and the reference spectral establishes on the basis of historical data, the classification results is consistent with the actual situation, which has a certain value to the oil spill emergency response system.