The Study Of Oil Spill Detection Using Satellite Multi-spectral Data

qOil spill, as one of the most serious pollutions, has great impact to marine ecolo- gical environment. As FY and other earth resource satellite have been in orbit, oil spill detection by satellite remote sensing based on multi-spectral data show wide prospect and practical significance.In this thesis, the status of oil spill detection by satellite remote sensing based on multi-spectral data is introduced, including the physical principle of oil spill detection by satellite remote sensing based on multi-spectral data, the function of QUAC(Quick Atmospheric Correction) in ENVI,the method of oil spill information extraction based on multi-spectral data. The "Deepwater Horizon" drilling platform accident which ha- ppened in April 2010 in U.S. Gulf of Mexico and the "Montara" drilling platform acc- ident which happened in Sep 2009 in Australia are taken as an example, to study the relationship between the sun glint reflectance and oil spill detection. The results show that the sun glint reflectance plays a key role in oil slick detection for visible and near -ifrared bands, and a suggested threshold value 10^-4 is also provided in this thesis. The positive contrast is observed when the sun glint reflectance is higher in the oil spill ar- ea than that in the clean seawater. Contrarily, the negative contrast is showed when the sun glint reflectance is lower in the oil spill area than that in the clean seawater. Ana- lysis of the FY-3/MERSI image about "Deepwater Horizon" drilling platform accident suggests that the contrast changes when sun glint reflectance is between 0.04 and 0.05. FY-3/MERSI and MODIS data are used to study the oil spill detection on the basis of the Quick Atmospheric Correction module (QUAC) in ENVI software by removing the atmospheric effect. In this thesis, the FY-3/MERSI data of the "Deepwater Hori- zon" drilling platform accident which was captured at 16:30 on April 29, 2010 (Uni- versal Time) with the coverage of 92oW-85oW and 27oN-32oN is taken as an example to study the influence of QUAC in oil spill detection. The result indicates that the con- trast between oil and water will be improved by 30% -120% after atmospheric correc- tion and the near-infrared band is considered as the best channel for oil spill detection using FY-3/MERSI data. And also the "Lake Maracaibo" oil spill accident which hap- pened in January 2003 in Venezuela is taken as an example to study the influence of QUAC in oil spill detection. The result indicates that the contrast between oil and wa- ter will be improved after QUAC. Additionally, RGB image and band-radio image provide enough oil spill information.