摘要
为减少经典目标识别算法中的计算量与模板存储量,本文将雷达目标散射中心作为目标识别的特征,通过支持向量机(SVM)方法分类目标。首先,采用基于传播算子的多重信号特征法(PM-MUSIC)提取雷达目标散射中心参数;其次,计算散射中心的中心矩以建立统一标准的识别特征,并采用SVM方法对目标进行分类。最后,通过仿真实验比较了该算法与散射中心联合自适应高斯分类(AGC)算法用于5类战斗机识别的结果,说明了散射中心目标SVM识别方法提高了目标的识别性能。
To reduce the computational complexity and template storage, the scattering centers on the target are utilized in the radar target recognition as the features, and the support vector machine （SVM） is used to identify radar targets. Firstly, the scattering centers are extracted from the MUSIC based on propagator method. Secondly, the scale and translational-invariant features are obtained by the central moments from the scattering centers, and the support vector machine （SVM） is utilized to classify the targets. Finally, SVM and adaptive Gaussian classifier （AGC） are compared in recognizing five kinds of fighter planes by 1-D scattering center features. Simulations show that the SVM recognition method based on scattering centers can improve the recognition performance effectively.
出处
《数据采集与处理》
CSCD
北大核心
2011年第5期521-525,共5页
Journal of Data Acquisition and Processing
