论文部分内容阅读
为了更加精准地从合成孔径雷达(SAR)图像中提取水边线,提出基于G0分布的SAR图像水边线提取方法。首先,将SAR图像域划分为大小相等的子块,并假设每个子块内像素强度服从独立同一的G0分布;利用矩估计方法得到对应每个子块的粗糙度和散射性参数;根据设定的参数阈值,可划分出粗略的水域,并确定该区域的几何中心;以该几何中心为圆心,向四周做射线,保留经过陆地区域的射线;利用似然函数确定每条射线上的水-陆分界点,依次连接分界点,进而实现SAR图像水边线提取。采用本文方法,分别对模拟和真实SAR图像进行水边线提取实验。定性和定量结果表明:采用基于G0分布的SAR图像水边线提取方法能够有效地克服斑点噪声的影响。
In order to extract the waterfront more accurately from synthetic aperture radar (SAR) images, a waterfront extraction method based on G0 distribution is proposed. Firstly, the SAR image domain is divided into sub-blocks with the same size and the pixel intensity in each sub-block is assumed to be independent of the same G0 distribution. The roughness and scattering parameters corresponding to each sub-block are obtained by the moment estimation method. The threshold of the parameter can be divided into rough waters, and the geometric center of the area is determined. Taking the geometric center as the center, radiating around and keeping the rays passing through the land area; using the likelihood function to determine the water-land The demarcation point, in turn, connects the demarcation point, so as to extract the waterfront of SAR image. Using this method, experiments on the waterfront extraction of both simulated and real SAR images are carried out respectively. Qualitative and quantitative results show that the water edge detection method based on G0 distribution can effectively overcome the speckle noise.