提出一种基于边缘检测的快速溢油信息提取方法, 首先对溢油图像进行ROA(radio of average)边缘检测, 根据检测结果进行AOI(area of interest)提取, 然后使用改进的Weibull-CFAR检测算法对AOI进行溢油检测, 并与全局CFAR检测结果进行对比。实验结果证明, 所提出的方法对于非均匀灰度SAR图像溢油检测准确性较好、效率较高, 特别适用于大图像的快速溢油检测。

The side-look imaging of SAR (Synthetic Aperture Radar) may result in the pockety density, so we can not get a satisfying result by using CFAR (constant false alarm rate) detection on the whole image. To improve the oil detection capability, we propose a new oil detection method based on edge detection in this paper. First, the AOI (area of interest) of oil is detected through radio edge detection, then an improved Weibull-CFAR detector is applied to the AOI. Comparing with the whole CFAR detection results, this detection method demonstrates better accuracy and efficiency on SAR images with pockety density, especially suitable for the oil detection of big images.