Based On Marine Oil Spill Monitoring And Evaluation Methods SAR Images

The ocean is an important ingredient of the global life system. With the advancement of the global technology and the expansion of industrialization development, more and more pollutants were discharged into the sea, which has led a fatal damage to the marine environment. Oil spill, as one of the marine pollution, has a significance to the comprehensive assessment for the losses that caused by oil spilling.While the losses are always divided into three parts:Clean-up costs, social economic losses and natural environmental hazards. The traditional assessment methods require individual assessment of these three parts, then superimposed on each other to form the comprehensive damage assessment results. But that made the process becoming complicated and longer assessment cycle, as well as more factors for calculation, and most part for those factors were so difficult to calculate or quantify that result in a great uncertainty of the results. So there still exist a significant limitation between the operability of the traditional assessment methods and practical application. All of the data which was used in the research is bases on the historical oil spill case database of "The International Oil Pollution Compensation Funds, IOPCF". Using multiple regression analysis on the historical data, investigating the oil spill loss, analyzing the intrinsic relationship between oil spill and natural environment features, was the main content of this paper. Contrasting different type of self-variable composition,combining with significance test and goodness of fit test, to find the multiple regression model which used the oil spills--the sea breeze--and the distance from the coast as the independent variable, establish the appropriate multiple regression equation for the oil spill damaging assessment. This paper use Penglai19-3oil spill as an example to evaluate the synthesize loss of oil spill.Research and conclusions of this paper:2. Oil spill extended model which was corrected by Xiao-Dan Wu is used to extract the oil spill thickness. And it was showed that the accuracy of the model can reach more than90%so it was good enough to be applied in this research.3. Combining with texture analysis, SAR Image was used to calculate the oil spill area through Neural Network Classification. While in the process of texture analysis, extract oil-spill area with different parameters. We find the best results comes when Contrast, dissimilarity, Angular Second Moment, correlation acting as a combination of texture features,21×21as the window size,3as a texture spacing,0°as the angle related to the correlation and angular second moment texture features statistics,45°as the angle about the dissimilarity texture features statistics,135°as the angle of contrast texture features statistics.4. The analyst based on historical oil-spill data has built the relationship between the loss and all the factors including of oil spills, sea-surface wind speed, the distance to the coast. We input the feature information of Penglai19-3oil-spill accident which were extracted from those data supported by remote sensing, and finally get to know that the loss is about CNY508,000,000.