基于非下采样轮廓波变换的SAR图像变化检测  

SAR Image Change Detection Based on Nonsubsampled Contourlet Transform

作　　者：黄凌霄 HUANG Lingxiao(College of Intelligent Engineering,Fuzhou Polytechnic,Fuzhou Fujian 350100,China)

机构地区：[1]福州职业技术学院智能工程学院,福建福州350100

出　　处：《曲靖师范学院学报》2024年第3期61-67,共7页Journal of Qujing Normal University

基　　金：福建省自然科学基金“基于视觉特征及深度核极限学习机的浮选精矿品位软测量方法”(2019J01224);福建省教育厅科技项目“NSCT域浮选气泡双模态多尺度特征提取及优化的量子神经网络状态识别”(JAT210817).

摘　　要：针对传统方法对合成孔径雷达(Synthetic Aperture Radar,SAR)图像变化检测精度不稳定等问题,提出了一种将非下采样轮廓波变换(NSCT)、显著性检测和差异图融合相结合的变化检测方法.采用Frost滤波器对SAR图像进行预处理;构造预处理图像的均值比和领域对数比两类差异图;通过NSCT变换将两类差异图分解为低频图像和高频子带图像,并采用SR算法提取高频子带的显著信息;然后采用自适应模糊逻辑算法和模值取大算法进行图像融合并得到NSCT反变换图像;最后经过k均值聚类算法得到SAR变化检测结果图和评价数据.实验结果表明,与传统方法相比,本方法对SAR图像变化检测的稳定性较高,精度在0.65到0.8之间浮动,验证了该方法的有效性.A change detection method combining Nonsubsampled Contourlet Transform(NSCT),saliency detection,and difference map fusion was proposed to solve the problem of unstable accuracy in SAR(Synthetic Aperture Radar)image change detection in traditional methods.Firstly,SAR images were pre-treated by Frost filters and two types of difference maps for preprocessed images with mean ratio and domain logarithmic ratio being constructed.Secondly,the two types of difference maps were decomposed into low-frequency images and high-frequency sub band images through NSCT transformation,and SR algorithm was used to extract salient information of high-frequency sub bands.Then the adaptive fuzzy logic algorithm and the maximum modulus algorithm were used for image fusion to obtain the NSCT inverse transform image.Finally,the SAR change detection result map and evaluation data were obtained through the k-means clustering algorithm.The experimental results show that the proposed method,compared with traditional methods,has higher stability in SAR image change detection,with an accuracy ranging from 0.65 to 0.8,verifying the effectiveness of the method.

关 键 词：NSCT SAR图像 SR算法 图像融合 图像变化检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]