摘要
提出一种新的基于双树复小波变换的SAR图像边缘检测算法.算法在复小波子带上求取图像直方图方向梯度矩阵,并基于复小波变换的多尺度性质和方向选择性求取全局的梯度矩阵.通过对这一矩阵阈值化实现边缘检测.该算法能有效检测出SAR图像上的显著边缘,并对SAR图像中存在的相干斑噪声、灰度不均匀性和边缘模糊等现象具有一定的鲁棒性.实验结果证明了该算法的有效性.
We propose an edge detection algorithm of SAR images based on dual-tree complex wavelet transform. We use the oriented gradient of histogram method to calculate the gradient matrix on all complex wavelet subbands, and acquire the global matrix based on the directional selectivity and muhi-scale property of dual-tree complex wavelet transform. The edges are extracted by thresholding the matrix. The proposed algorithm effectively detects significant edges, and it is robust to the speckle noise, intensity inhomogeneity, and blurred boundaries of SAR images. The experimental results show effectiveness of the proposed algorithm.
出处
《中国科学院大学学报（中英文）》
CAS
CSCD
北大核心
2014年第2期238-242,248,共6页
Journal of University of Chinese Academy of Sciences
基金
国家自然科学基金面上基金(61272317)
安徽省自然科学基金面上基金(1208085MF90)资助
关键词
SAR图像
边缘检测
双树复小波变换
多尺度
方向选择性
直方图方向梯度
SAR image
edge detection
dual-tree complex wavelet transform
multi-scale
directional selectivity
oriented gradient of histogram
