基于NSCT和支持向量机的SAR图像识别

针对传统的合成孔径雷达图像（SAR）识别算法识别精度低，用时长等问题，提出一种基于非下采样轮廓波变换（NSCT）和支持向量机（SVM）的SAR图像识别算法。首先通过非下采样轮廓波变换将目标图像分解成不同的尺度，然后得到目标图像的低频分量和高频分量；接着在高频分量中提取方向梯度直方图特征（HOG），在低频分量中利用局部二值化算法（Local Binary Pattern，LBP）提取纹理特征；然后将提取的梯度方向直方图特征和局部二值化特征空间连结，并使用支持向量机（SVM）作为分类器；最后对算法进行了测试。实验结果表明，该方法不仅能够有效地提高了SAR图像目标分类的精度，在MSTAR数据库上的准确率达到90.7%，而且对相干斑的影响具有较高的鲁棒性。

SAR image recognition based on NSCT and SVM

In view of the low recognition accuracy and long time used of traditional synthetic aperture radar (SAR) image recognition algorithms, a SAR image recognition algorithm based on non-subsampled contourlet transform (NSCT) and support vector machine (SVM) was proposed. Firstly, the image is decomposed at multiple scales through NSCT to obtain the high-frequency and low-frequency components. Then histogram of oriented gradient (HOG) was extracted from the high-frequency component, and LBP (Local Binary Pattern) algorithm was used to extract texture features from the low-frequency component. After that the extracted high and low frequency features are combined and divided by support vector machine. Finally, the algorithm proposed is tested by testing set. Experimental results show that this method can not only effectively improve the SAR image recognition accuracy, the recognition rate on the MSTAR database reaches 90.7%, but also robust to the coherent speckle.