摘要
无人机集群协同工作成为当下无人机技术的发展趋势,为了实现对天空中无人机集群目标进行检测和跟踪,提出了一种基于局部特征的无人机集群目标的检测跟踪方法。首先,根据目标在图像中的局部特征,利用改进的多级梯度检测方法,筛选出候选目标点,在候选目标点上进行局部对比度检测得到目标区域,再通过连通域检测得到无人机集群目标在图像中的位置;然后,综合考虑空间距离和目标区域框交并比等因素,结合卡尔曼滤波器和匈牙利算法,实现检测目标与目标轨迹之间的数据关联;最后,利用轨迹建立、轨迹保持和轨迹删除的策略管理无人机目标的运动轨迹。实验仿真结果表明,该方法能对复杂天空背景中的无人机集群目标进行检测,准确率可达97%,并在出现虚警、漏检和目标交叠的情况时也能有效跟踪。
Collaborative work of UAV swarm has become the popular trend in UAV technology filed.In order to detect and track UAV swarm in the sky,a detecting and tracking method of UAV swarm based on Local features are proposed.First,according to the Local feature of the targets in the image,the candidate target points are screened by an improved multistage gradient detection method,and local contrast detection is carried out on the candidate target points to get the target region.Then,use connected component analysis to get the location of the UAV swarm in the image.Further,considering the spatial distance and the intersection over union of the targets,combining Kalman filter and Hungarian algorithm to realize the association of the data between the detection targets and the trajectories.Finally,use the strategy of trajectory to manage the establishment,retention and deletion of the trajectories of UAV swarm.Results of experimental simulation show that the proposed method can detect UAV swarm in the complex sky background with an accuracy of 97%,and it can also track UAV swarm effectively when there are false alarms,missed detections and targets overlapping.
作者
赵瑜
孙宏海
高文
Zhao Yu;Sun Honghai;Gao Wen(Changchun Institute of Optics,Fine Mechanics and Physics,Chinese Academy of Sciences,Changchun 130033,China;University of Chinese Academy of Sciences,Beijing 100049,China)
出处
《国外电子测量技术》
北大核心
2023年第8期183-189,共7页
Foreign Electronic Measurement Technology
基金
吉林省科技发展计划重点研发项目(2022021146GX)资助。
作者简介
赵瑜,硕士研究生,主要研究方向为目标识别、目标跟踪。E-mail:zhaoyu971104@163.com;通信作者:孙宏海,副研究员,主要研究方向为先进光电成像技术及高速数字图像实时处理系统。E-mail:sunhh@ciomp.ac.cn;高文,副研究员,主要研究方向为视频图像跟踪、目标识别等。E-mail:gaowen@ciomp.ac.cn。
