The Method And Implement Of Edge Extraction Of Oil Spill In SAR Image

The frequent occurrence of oil spills give a serious threat to the ecology. environment and sustainable economic development of the coastal city, and attract the growing attention of the international community. When oil spill occurs, in order to take appropriate emergency measures actively and effectively, it is necessary to get the information of oil spills such as location, distribution scope and volume, while edge extraction of oil spills image is the key that determines the accuracy. At present, aviation remote sensing combined with aerospace remote sensing is still the most commonly used method for oil spill monitoring in most countries. Because synthetic aperture radar (SAR) can work in all conditions, it wins a wide range of concerns in marine oil spill monitoring. The objective edge extraction of SAR image is also the focus of researchers from all the countries.At present, the methods used for SAR image segmentation are mostly classical algorithms, such as threshold and so on. The new method like level set is limited in the theoretical study. The remote sensing processing software we commonly used such as ENVI, ERDAS have problems like lack of specific algorithms and complicated operates, so it can not extract the border region of oil directly. Therefore, the design of algorithms that targets oil border region extract is meaningful.This study is supported by the Ministry of Science and Technology, aiming to study the marine oil spill information extraction by space-borne SAR image. First, the paper studies the format of SAR data, and read, display SAR information by programming. Then the paper studies problems included during the SAR image preprocessing, especially selection of filter methods and filter sizes. To further enhance the accuracy of edge extraction, the paper use gray-balanced operation to enlarge the difference of pixels between regional in the pre-processing stage. For the image after preprocessing, paper presents three methods of edge extraction: (1) Edge operator, with the use of five kinds of edge operators on getting the edge of the image after preprocessing; (2) Seed filling, the use of manual and automatic extraction of a combination to achieve a more accurate edge Extraction, and get the area data of oil spill; (3) Level set method, we improve the Chan-Vese model for its level set function affected by the shortcomings of the noise to achieve more accurate boundary extraction. Finally, combined with algorithms that paper discussed earlier, the writer uses VC + + language design and develop a software which could use space borne SAR image to monitor the oil spill. This software can implement basic functions of SAR image processing, and gain the data we are more concerned during the oil spill more precisely (including area, perimeter, etc.)...