曲波变换结合Sobel算子的SAR图像边缘检测方法.pdf

文档介绍：323 Vol. 41 No. 5
总第 期nce of conventional gradient edge detection operators and false edges. This paper pro⁃
，
poses a SAR image noise suppression method based on curvelet transform before using the Sobel operator to detect edges which re⁃
，
duces the effect of speckle noise on edge extraction and the extracted edges are accurate and detailed. Compared the method in this
，
paper with the Sobel operator after wavelet hard threshold denoising and the Sobel operator based on median filtering the experi⁃
，
ment proves that this method is superior to the other two methods in extracting edges.
SAR images curvelet transform Sobel operator edge detection
Key Words ， ， ，
TB391
Class Number
1 引言 提取的边缘图像上会呈现出纷乱的纹理细节，导致
边缘信息体现不够明显。
SAR
图像具有复杂的边缘纹理结构以及特有 针对上述问题，本文提出将曲波变换作为边缘
的相干斑噪声，