基于多模标签多伯努利滤波-支持向量机的扇扫红外预警设备图像多目标跟踪算法  

Multi-Target Tracking Algorithm for Infrared Warning Equipment Image based on Label Multi-Bernoulli Support Vector Machine

作　　者：鲜勇 XIAN Yong(Wuhan Military Representative Bureau of Naval Armament Department,Wuhan 430223,China)

机构地区：[1]海装驻武汉地区第七军事代表室,湖北武汉430223

出　　处：《光学与光电技术》2022年第5期57-62,共6页Optics & Optoelectronic Technology

摘　　要：随着技术的不断发展,传统多目标跟踪算法在红外警戒设备应用上面临新的挑战与要求。讨论了基于多模标签多伯努利-支持向量机的扇扫红外警戒设备图像多目标跟踪问题。首先分析了传统多目标跟踪算法运用于扇扫红外警戒设备存在的不足之处,提出了基于多模标签多伯努利-支持向量机滤波算法的优势;分析了基于多模标签多伯努利滤波-支持向量机算法运用于扇扫红外警戒设备面临的问题,并提供了对应的解决方案。仿真结果表明,在典型复杂背景下,该算法的弱小目标航迹抗干扰成功率平均提升20%左右,对仅50%检测率的弱小目标的建航成功率提升45%以上,显著提高了弱小多目标跟踪能力。With the high-technique developing quickly,the traditional multi-target tracking algorithm is facing new challenges and requirements in the application of infrared warning equipment. In recent years,the approaches to MTT based on random finite set(RFS)have been achieved substantial results. This paper discusses MTT problem of sector scanning infrared warning equipment image based on multiple model label multi-Bernoulli support vector machine(MM-LMBSVM). Firstly,the problems existing in the application of traditional algorithms to sector scanning infrared warning equipment are discussed. The advantages of MM-LMB-SVM filtering algorithm are proposed. At the same time,the problems faced by the application of MM-LMB-SVM filtering to sector scanning infrared warning equipment are analyzed and the solution are provided. The simulation results shows that under the typical complex background,the track antijamming success rate of weak and small targets is improved by more than 20%,and the tracking success rate of weak and small targets with only 50% detection rate is improved by more than 45%. It significantly improves the weak and small multi-target tracking ability.

关 键 词：多模标签多伯努利 多目标跟踪 红外警戒 弱小目标 

分 类 号：TN216[电子电信—物理电子学]