摘要
对于海洋背景下低可观测目标检测问题,可通过选取能够有效区分纯海杂波与目标单元回波的观测量作为主要特征,最终采用机器学习法实现综合分析和自动分类。本文基于海杂波的统计特性和分形特性,将去相关时间及FRFT域Hurst指数作为特征矢量,采用支持向量机(Support Vector Machine,SVM)实现海洋背景下的低可观测目标检测。经四级海况下IPIX雷达(X波段)实测海杂波数据验证,表明该算法可以有效检测到海面弱目标,且在低信杂比条件下仍表现出良好的检测性能。
For the low observable targets detection in sea clutter, we can select the observed values which can distinguish targets signal from sea clutter effectively to constitute the feature vectors. Machine learning is applied to carry on comprehensive analysis and automatic detection finally. Based on sea clutter's statistical and fraetal characteristics, this paper used de-correlative time and Hurst exponents in FRFT domain as feature vectors, and enforced targets detection by using support vector machine (SVM). IPIX real-life sea clutter in four-level Douglas sea states was used for verification, and the results showed that the algorithm can detect weak target in sea clutter effectively. Or rather, it showed good performance in low signal-to-clutter ratio condition.
出处
《中国海洋大学学报(自然科学版)》
CAS
CSCD
北大核心
2013年第7期104-109,共6页
Periodical of Ocean University of China
基金
国防基础科研计划项目(B1120110005)资助