SAR图像舰船目标边缘检测

Warship Target Edge Detection Algorithm of SAR Images

提出了一种在正则化基础上,利用小波变化实现合成孔径雷达(SAR)图像舰船目标边缘检测的新方法。传统的利用小波变换实现图像边缘检测时,阈值需要人为设定。针对这一问题,文章引入正则化超分辨技术,从贝叶斯框架下的估计问题出发,采用非二次正则化,平滑图像,保护强散射点目标,实现对SAR图像进行去噪。利用小波变换的局部化特性和多尺度分析能力,检测突变信号,实现对舰船目标的边缘检测。该方法去噪效果好,边缘定位准确,仿真结果表明了算法的有效性。

A new ship edge detection algorithm in synthetic aperture radar（SAR）images based on regularization method and wavelet transform was proposed in this paper. When existing edge detection algorithms based on wavelet transform was adopted, threshold was needed to filter candidate edge points for edge detection. Aiming at this problem, super resolution technology based on regularization method was adopted. First, the estimation problem under the Bayesian framework was considered. The quadratic regularization method was adopted to smooth the SAR images and protect the strong scattering point targets so as to filter the noise. Second, wavelet transform was adopted to detect the singular points and take the edge of ship targets because of its＇ localization ability and multi-resolution wavelet decomposition of an image yields detailed coefficients that contain the high frequency content of the image. The experiment showed the effectiveness of this method.