PCA、ICA和Gabor小波决策融合的SAR目标识别

提出了一种基于主成分分析(PCA)、独立分量分析(ICA)和Gabor小波决策融合的合成孔径雷达SAR(Synthetic Aperture Radar)图像目标识别方法。首先用PCA、ICA和Gabor小波变换分别对SAR目标图像提取特征向量,再用3个支持向量机分类器分别对3种方法提取得到的特征向量分类,通过基于等级的决策融合方法对3个支持向量机分类器的输出进行决策融合,得到最终类别决策结果。采用MSTAR数据库中3个目标进行识别实验,实验结果表明,PCA、ICA和Gabor小波决策融合后得到的识别率高于单独用其中任何一个特征得到的识别率。因此,该方法可提高目标的正确识别率,是一种有效的SAR图像目标识别方法。

SAR target recognition using PCA, ICA andGabor wavelet decision fusion

A method for Synthetic Aperture Radar (SAR) image target recognition based on Principal Component Analysis (PCA),Independent Component Analysis (ICA) and Gabor wavelet decision fusion is presented in this paper. PCA, ICA and Gaborwavelet transformation were used to extract feature vectors from SAR target images, respectively. Three Support Vector Machine(SVM) classifiers were applied to classify the feature vectors extracted via three algorithms, respectively. Ranking based decisionfusion algorithm was then used to fuse the outputs of three classifiers. The final classification decision result was obtained fromthe output of the fuser. Experiments were implemented with three military targets in MSTAR database. The experimental resultsshow that the probability of correct classification obtained by PCA, ICA and Gabor wavelet decision fusion is better than that attainedby any of the individual feature. Therefore, it is concluded that the method proposed in this paper advances the probabilityof correct classification and can be an effective approach for SAR image target recognition.