基于稀疏表示和道路辅助的单幅SAR图像运动目标检测方法

本文提出一种利用单幅SAR(Synthetic Aperture Radar)图像实现运动目标检测的方法.首先提出一种基于压缩感知的SAR图像道路检测算法:根据SAR图像中道路的特点,使用模糊C均值方法将图像进行模糊分类,获得大致的道路区域,然后利用Hough变换域的稀疏性,用压缩感知精确定位图像中的道路信息.其次利用图像稀疏表示的方法对运动目标进行检测:不同速度运动目标的散焦量和距离单元跨越不同,由此生成样本图像,继而构造超完备字典.将待测图像分块,并计算子图像在字典下的稀疏系数,检测并匹配出运动目标的速度参数.最后,结合已检测出的道路辅助信息,消除多普勒模糊影响,剔除虚假的运动目标,并对运动目标速度参数进行校正.实验结果证明了所提方法的有效性.

Moving Targets Indication Method in Single SAR Imagery Based on Sparse Representation and Road Information

A method of moving target indication in single SAR(Synthetic Aperture Radar)imagery is proposed.First, a road detection method based on compressive sensing for single SAR image is presented, the fuzzy C mean method is used to classify the SAR image according to the road characteristics of the SAR images, and the road pixels are extracted.Then it shows how compressive sensing can be used to find lines in images, by exploiting sparseness in the Hough transform domain.Secondly, the moving targets are detected by a indication method based on sparse representation.In SAR image, different velocities of moving target lead to different defocuses and range cell migration.Based on this character, the over-complete dictionary of targets sample images with different speeds is constructed.Then the test SAR images are blocked into sub-images and the corresponding coefficients are calculated with the dictionary.According to the coefficients, moving target can be detected and the motion parameters can be estimated.Finally, the effects of Doppler ambiguity on motion parameters estimation are eliminated, and the false target and calibrate motion parameters are excluded.The results of experiments indicate the effectiveness of the proposed method.