基于支撑向量机的地杂波虚警抑制技术

由于杂波的非均匀特性,传统杂波抑制技术的杂波滤除能力有限。在强杂波环境下,处理后回波仍残留大量虚警。文中提出了新的点迹滤波技术,用于抑制杂波虚警。针对地面动目标检测体制雷达,通过提取目标和杂波的五种分类特征和五种先验特征,使用主成分分析和支撑向量机方法过滤虚假点迹。试验结果表明:文中提取的特征和设计的分类器, 可在较少目标损失下,有效地抑制地杂波虚警。

Filtering Ground-clutter-false-alarms Using Supported Vector Machine

For the heterogeneity of clutter, traditional methods are limited in clutter suppression. In strong clutter environment, a large number of false alarms still exist in the processed radar echo. With respect to ground-based moving-target-detection radar, five classification-features and five prior-features of target/clutter plots are extracted in this paper. Then, techniques of principal component analysis and supported vector machine are adopted for plot classifying. Experimental results show that the extracted features and designed classifier can effectively filter ground-clutter-false-alarms while few target plots are lost.