哈尔滨工程大学学报2026，Vol.47Issue(5)：968-976,9.DOI:10.11990/jheu.202601009
基于图像声呐的无人水下航行器编队小目标识别与跟踪方法
Forward-looking-sonar-based recognition and tracking method for small targets in UUV formation operations
摘要
Abstract
Aiming at the problem of insufficient continuous recognition and tracking capabilities of unmanned un-derwater vehicles(UUV)for small targets from a long distance to a near field in complex underwater environments,this paper proposes a UUV formation small target recognition and tracking method based on image sonar.This method uses multi-beam forward-looking sonar to obtain images in real time and combines a lightweight detection network to achieve efficient target recognition.And a filter-based enhanced tracking algorithm and a three-stage data association mechanism are introduced to improve the tracking robustness.For targets at different distances,a distance hierarchical tracking strategy is proposed;Dynamic obstacle avoidance is achieved by combining the im-proved artificial potential field method.Through multi-UUV formation tests conducted on the reservoir,the system verified targets such as simulated frogmen,unmanned underwater vehicles and simulated mines.The results showed that the recognition success rate of the system in tracking and obstacle avoidance tasks exceeded 90%,demonstrating good real-time performance and environmental adaptability.It can provide reliable technical support for underwater collaborative search,target following and safe navigation tasks.关键词
无人水下航行器/前视声呐/目标识别/目标跟踪/编队控制/卡尔曼滤波/动态避障/数据关联Key words
unmanned underwater vehicle/forward-looking sonar/target recognition/target tracking/formation control/Kalman filter/dynamic obstacle avoidance/data association分类
信息技术与安全科学引用本文复制引用
徐硕,尹雪峰,万家豪,罗圣源,王博,胡庆玉..基于图像声呐的无人水下航行器编队小目标识别与跟踪方法[J].哈尔滨工程大学学报,2026,47(5):968-976,9.基金项目
国家自然科学基金项目(52501383,52501382) （52501383,52501382）
黑龙江省博士后面上项目(LBH-Z24119). （LBH-Z24119）
