雷达目标三维特征的提取与识别研究

以逆合成孔径雷达(ISAR)成像技术和计算机视觉理论为基础，提出了一套新的从动态目标ISAR成像序列中提取目标散射点三维结构信息，以此作为目标特征的识别方法。这一研究方法包含了4个重要环节：动态目标的ISAR像序列的获得；“散射点像元”质心的检测、跟踪和匹配；基于光流分析的目标散射点三维结构特征的提取；目标三维特征的识别。由于经过了时间和空间上信息的积累，目标的散射点的三维结构特征具有稳定和直观的特点。作为一种新的目标识别的依据是很有效的，且只需要较少的训练样本就可以获得较高的识别率。

Research on Target Identification with ISAR Image Sequence

Basing on ISAR imaging technology and computer vision theory, a novel target recognition method which could extract the 3-dimensional feature of targets by analyzing the 2-dimensional ISAR image sequence of the targets is presented in this paper. Four important phases are included in the method: first, obtaining of ISAR images sequence of the targets. Second, tracking and detecting 'the imaging cells of scattering points'. Third, rebuilding the 3D feature of scattering points of the target basing on the method of optical flow analysis. Finally, identifying targets by 3D feature of the target. With the theory and simulation experiment results, the obvious conclusion is that the 3D feature of the scattering points is steady because of the accumulation of the target's information in time and space. It is effective as a novel method of target recognition, and could get better recognition result with littler stylebooks.