雷达信号交互式分选方法  

A Visual Analytics Approach for Radar Signal Clustering

作　　者：赵颖[1] 钱万云 李影[1] 张蓉[1] 吴青 陈斌[4] 周芳芳[1] Zhao Ying;Qian Wanyun;Li Ying;Zhang Rong;Wu Qing;Chen Bin;Zhou Fangfang(School of Computer Science and Engineering, Central South University, Changsha 410083;School of Automation, Central South University, Changsha 410083;School of Information and Network Center, Central South University, Changsha 410083;Southwest Electric & Telecom Engineering Institute, Shanghai 200434)

机构地区：[1]中南大学计算机学院,长沙410083 [2]中南大学自动化学院,长沙410083 [3]中南大学信息与网络中心,长沙410083 [4]西南电子电信技术研究所上海分所,上海200434

出　　处：《计算机辅助设计与图形学学报》2019年第10期1653-1665,共13页Journal of Computer-Aided Design & Computer Graphics

基　　金：国家科技基础资源调查专项(2018FY10090002);国家重点研发计划(2018YFB1700403);国家自然科学基金(61672538,61872388)

摘　　要：雷达信号分选指从混叠雷达信号数据中分选出属于各单部雷达辐射源发射的雷达信号.随着雷达技术快速发展,雷达核心属性分布不均、重叠度高、变换复杂,而且数据采集过程不稳定,导致传统自动化分选方法准确率不高.文中介绍了一种自动化分选和交互式分选结合的雷达信号监控数据可视分析方法;提出了一种改进的雷达信号自动化分选方法,结合核密度估计、密度聚类和排列熵技术对数据进行分区式聚类和量化结果评价;设计了一个可视分析原型系统,提供了可协同分析的多可视化视图和简单易用的交互手段,以直观地展示雷达信号分布情况和自动化分选结果,帮助用户参与整个信号分选过程.Radar signal clustering seeks to identify radar signals that belong to each radar radiation source from the radar signal data. With the rapid development of radar technology, the unevenly distributed and highly overlapping attributes of radar signals often result in unsatisfactory results by means of traditional radio signal clustering methods. This paper proposes an interactive clustering approach for radar signal data. Firstly, we propose an improved clustering method that integrates a kernel density estimation, density clustering and permutation entropy technology to process the radar signal data. Then, we design and implement a visual analysis system that presents the radar signal data and automatic clustering results from multiple perspectives. The system also provides a set of interactions to involve users in the entire process of radar signal clustering.

关 键 词：可视化 可视分析 雷达监控 雷达信号 信号分选 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]