论文部分内容阅读
工程上常用静态雷达散射截面(RCS)统计特性进行目标识别,但其可分测度小,正确识别率较低。文中在精确获取目标动态RCS序列的基础上,提出了一种基于离散小波能量的特征提取方法,对典型飞机目标进行分类识别。首先,根据空气动力学和运动学方程设定五种典型飞机目标的飞行航迹并解算其实时飞行坐标,从而获取时变的雷达视线姿态角;其次,应用多层快速多极子电磁计算方法仿真各型目标的动态RCS数据;然后,再基于动态RCS序列,计算其位置、分布等统计特征,并进行小波分解和重构,提取各型目标的统计特征和小波能量特征;最后,采用基于距离的类间距离判据,比较两种特征量的分类识别效果。仿真计算结果表明:相对传统的统计特征,离散小波能量特征能完整地体现目标的特征,且可分性测度更大,识别效果更为理想。
Commonly used static radar cross section (RCS) statistical characteristics of the target identification, but the separable measurement is small, the correct recognition rate is low. Based on the accurate acquisition of target dynamic RCS sequences, a feature extraction method based on discrete wavelet energy is proposed to classify and identify the typical aircraft targets. Firstly, the flight path of five typical aircraft targets is set according to the aerodynamic and kinematic equations and the real-time flight coordinates are calculated to obtain the time-varying radar gaze attitude. Secondly, the multi-layer fast multipole electromagnetic calculation Method to simulate the dynamic RCS data of all kinds of targets. Then, based on the dynamic RCS sequences, the statistical characteristics such as position and distribution are calculated, and the wavelet decomposition and reconstruction are used to extract the statistical features and wavelet energy features of each type of targets. Finally, Based on the distance between the class distance criteria, the classification of recognition of the two features compared. The simulation results show that, compared with the traditional statistical features, the discrete wavelet energy features can completely reflect the characteristics of the target, and the separability measure is larger and the recognition effect is more ideal.