宇航学报

宇航学报 ›› 2017, Vol. 38 ›› Issue (10): 1114-1123.doi: 10.3873/j.issn.1000-1328.2017.10.012

• 制导、导航、控制与电子 • 上一篇    下一篇

临近空间高超声速目标修正随机Hough变换TBD算法

李岳峰，王国宏，李林，张翔宇   

  1. 海军航空工程学院信息融合研究所，烟台 264001
  • 收稿日期:2017-07-20 修回日期:2017-08-25 出版日期:2017-10-15 发布日期:2017-10-25
  • 基金资助:

    国家自然科学基金（61731023,61372027,61501489,61671462,61701519）；“泰山学者”建设工程专项经费资助项目

A Modified Randomized Hough Transform TBD Algorithm for Hypersonic Target in Near Space

LI Yue feng, WANG Guo hong, LI Lin, ZHANG Xiang yu   

  1. Institute of Information Fusion, Naval Aeronautical and Astronautical University, Yantai 264001, China
  • Received:2017-07-20 Revised:2017-08-25 Online:2017-10-15 Published:2017-10-25

PDF

274

摘要：

针对临近空间高超声速目标的检测跟踪问题，提出一种修正的随机Hough变换检测前跟踪算法。首先，为尽可能克服远距离条件下角度误差带来的较大位置偏差，通过解耦的方式将量测点迹映射至精度较高的径向距离-时间平面进行检测；然后，为更合理地合并参数空间特征点并提升积累效率，构建检验统计量并与自适应门限进行比较，将特征点合并问题转换成两个正态总体均值差的自适应假设检验问题，并利用点数积累与能量积累相结合的双重积累方式进行积累检测；最后，为进一步降低虚假航迹数，引入运动约束和航迹合并措施，得到最终按时序关联的检测航迹。仿真结果表明，相比标准Hough变换检测前跟踪算法，本文算法在检测概率相差不大的情况下具有更少虚假航迹和更低运行时间。

关键词: 临近空间, 高超声速目标, 检测前跟踪, 随机Hough变换, 特征点合并

Abstract:

Aiming at the detection and tracking issue for near-space hypersonic target, a modified randomized Hough transform track-before-detect algorithm is proposed. Firstly, to overcome the large positional deviation from the angle error under long distance, measurements are mapped into the range-time plane with a higher accuracy by decoupling. Then, to merge the feature points in the parameter space more reasonably and improve the accumulation efficiency, the test statistics is built to compare with the adaptive threshold, so that the merging of the feature points is transformed into the adaptive mean difference hypothesis testing of two normal populations. Double integration means of noncoherent integration and binary integration is also adopted for the accumulation detection. Finally, to further reduce the false trajectories, motion constraints and trajectory merging are introduced for the final detected trajectory after sequence association. Simulation results show that, compared with the standard Hough transform track-before-detect algorithm, the proposed algorithm has fewer false trajectories and less run time with close detection probability.

Key words: Near space, Hypersonic target, Track before detect, Randomized Hough transform, Merging of feature points

中图分类号: