地面低速目标与背景模式分类和识别方法研究

Research on the Pattern Classification and Recognition about Slow Target and Its Background

识别地面低速目标是无线电近炸引信信号处理研究的难点之一。研究了基于连续波多普勒体制无线电近炸引信的地面低速目标与地面背景的分类问题。在分析目标与背景时域与频域特性的基础上,利用线性分类器和支持向量机进行分类识别实验,结果表明所提出的方法是可行的。

Recognizing slow target on ground is one of difficult research topic in radio proximity fuze signal processing field. The problem of classifying slow target and its background of radio proximity fuze based on continuous Doppler mechanism is studied in this paper. After analyzing their charicteristics in time and frequency domains, two classifying experiments were performed by means of linear classifier and support vector machine, respectively. The results show the method is efficient.