利用尺度不变量特征的ISAR二维像自动识别技术

Automatic target recognition of ISAR images based on SIFT features

在平移、旋转、投影和遮挡等条件下,逆合成孔径雷达对目标成像不理想,会导致逆合成孔径雷达二维像目标识别困难.针对这个问题,提出了一种新的逆合成孔径雷达二维像特征提取方法,可以在不同的目标姿态获得有效的识别结果.利用尺度不变量变换,提取逆合成孔径雷达二维像的尺度信息,并按照从大到小的顺序重新排列,称为顺序尺度.将顺序尺度截取相同长度作为不变量特征,利用基于支持矢量机的识别算法选择RBF核函数,对逆合成孔径雷达二维像进行目标识别.仿真结果表明,利用顺序尺度作为特征变量,可以对逆合成孔径雷达二维像进行有效识别.

Under conditions of translation, rotation, projection and shelter of the aircraft targets, undesired inverse synthetic aperture radar（ISAR） images will make target classification more difficult. A new feature extraction method is proposed to identify ISAR images from different flight attitudes of the target aircraft. The approach provides effective features for identification using the order scales extracted by scale invariant feature transform. Experimental results show the presented algorithm is quite effective.