摘要
为了使船舶跟踪算法更适用于工程实践,降低杂波、风浪带来的影响,结合工程经验提出一种融合凝聚耦合度与航迹成长值的船舶跟踪算法。根据LSTM模型的“遗忘门”、“细胞状态”思想,提出“暂时航迹成长值”概念,用于计算暂时航迹的可靠性,同时利用多个波门过滤航迹,结合点迹凝聚的耦合度,量化出凝聚点迹质量,剔除虚假航迹,降低虚警率,实现航迹的正确起始。通过仿真结果表明,该算法跟踪效果良好,验证了其可行性和有效性。
In order to make the ship tracking algorithm more suitable for engineering practice and reduce the impact of clutter and wind waves,a ship tracking algorithm combining condensing coupling degree and track growth value is proposed based on engineering experience.According to the"Forgetting Gate"and"Cell State"thoughts of the LSTM model,the concept of"temporary track growth value"is proposed to calculate the reliability of the temporary track,while using multiple gates to filter the track,the degree of coupling of dot trace aggregation is combined.The degree of coupling quantifies the quality of the condensed dots,the false trajectories are eliminated,the false alarm rate is reduced,and the correct start of the track is achieved.The simulation results show that the algorithm has good tracking effect and its feasibility and effectiveness are verified.
作者
徐瑞龙
石琳
XU Ruilong;SHI Lin(School of Computer,Jiangsu University of Science and Technology,Zhenjiang 212003)
出处
《计算机与数字工程》
2021年第8期1574-1579,1665,共7页
Computer & Digital Engineering
