基于一维匹配滤波的SAR目标识别

为克服合成孔径雷达(SAR)目标位置和姿态不确定性对识别性能的影响,提出了一种基于一维匹配滤波的SAR目标识别方法。基于姿态估计对目标图像进行预处理,基于一维匹配滤波定义两样本图像间的相似度或距离,并采用最近邻分类器。在MSTAR公开数据集上进行了实验,已知和未知目标姿态情况下的识别率分别达到98.8%和98.2%。实验结果表明,所提出的方法性能优于支持向量机(SVM)方法。

SAR target recognition based on one dimensional matched filtering

In order to overcome the effects on the recognition performances caused by position and pose uncertainty of Synthetic Aperture Radar（SAR） targets,a new method for SAR target recognition was proposed based on one dimensional matched filtering.The target image was pre-processed based on the pose estimation,the similarity or distance between two sample images was defined based on the one dimensional matched filtering,and the nearest neighbor classifier was used.Experiments were conducted on the MSTAR public data set,and the recognition rates with and without knowledge of the target pose achieved 98.8% and 98.2% respectively.According to the experiment results,the proposed method outperforms the Support Vector Machine（SVM） method.