基于独立分量分析的ISAR图像特征提取方法

摘要

ISAR图像特征是利用ISAR图像进行目标识别的关键,根据目标区ISAR图像由强散射点组成这一特点,将独立分量分析的方法应用于目标ISAR图像的特征提取.所提取的ISAR图像的独立分量特征,反映了目标回波距离单元内散射点的位置分布,能够很好地表征一定角度范围内ISAR图像的特征.将目标ISAR图像投影到提取的独立分量特征空间上,得到投影系数,通过比较投影系数与目标样本之间的余弦测度,实现对目标的正确识别.仿真空间目标数据实验表明,该算法是有效的.%Inverse Synthetic Aperture Radar (ISAR) image feature is the key to target recognition based on ISAR images. In this paper Independent Component Analysis (ICA) algorithm is used to extracting features from ISAR images based on the characteristic that target area ISAR image is composed of strong scattering points. The independent features of ISAR images elucidate the distribution of scatting points in distance unit to the received wave of target and will show typical features of targets in a certain rotatable angle. Cosine measure between the projection coefficient of unknown ISAR image on independent features and samples are identified as the parameters, which will be used to find a exact sample that matching to the uncertain target. The result tested from simulated data of space objects is effective.