基于KSVD和PCA的SAR图像目标特征提取

提出一种基于核的奇异值分解(KSVD)与主成分分析(PCA)相结合的SAR图像目标的组合特征提取方法。该方法首先利用核的奇异值分解得到图像非线性的代数特征,然后进一步经过PCA变换得到图像的最终分类特征。实验中,将本文提出的KSVD+PCA两步特征提取方法与PCA、SVD、KPCA、KSVD方法分别结合简单、快速的最近邻分类器在MSTAR坦克数据上进行了比较,实验结果表明,KSVD+PCA方法不仅有效地提高了目标的正确识别率,而且大大降低了对目标方位的敏感度,在目标方位信息未知的情况下,识别率可达到95.75%,是一种有效的SAR图像目标特征提取方法。

SAR image target feature extraction based on KSVD and PCA

In this paper a Synthetic Aperture Radar (SAR) image target extraction method based on Kernel Singular Value Decomposition (KSVD) and Principal Component Analysis (PCA) is proposed. First it acquires the nonlinear algebraic feature of SAR images by performing KSVD; then obtains the last discriminating feature using PCA; and finally the nearest neighbor classifier is used for recognition. The KSVD and PCA are carried out on MSTAR tank dataset in comparison with traditional PCA, SVD, KSVD and Kernel Principal Component Analysis (KPCA). Experiment results demonstrate that the KSVD and PCA method proposed in this paper is effective for SAR image target feature extraction. Not only the right recognition rate is higher of the new method but also it is not sensitive to target azimuth.