基于机器学习的非平衡环境下多目标智能检测算法  

Multi-target intelligent detection algorithm in unbalanced environment based on machine learning

作　　者：沈文杰[1] SHEN Wenjie(Information Engineering Department,Fujian Vocational College of Agriculture f Fujian Fuzhou 350119)

机构地区：[1]福建农业职业技术学院信息工程学院,福建福州350119

出　　处：《宁夏师范学院学报》2023年第1期83-89,112,共8页Journal of Ningxia Normal University

摘　　要：在非平衡环境下多目标检测受到目标回波信息的干扰,导致检测精度不高,为了提高对多目标的智能跟踪检测能力,提出基于机器学习的非平衡环境下的多目标智能检测算法.首先,建立多目标跟踪识别的非平衡数据采集模型,通过近场源回波检测方法进行多传感网络下非平衡数据多目标特征检测,然后,利用空间网格聚类方法进行非平衡数据多目标特征量分类挖掘,最后,结合机器学习方法进行非平衡数据多目标智能检测过程中的寻优控制,实现非平衡环境下多目标智能检测算法的优化设计.实验结果表明,采用该方法进行非平衡环境下多目标检测的智能性较好,检测精度较高,检测时间开销较短.The multi-target detection method in the non-equilibrium environment is disturbed by the target echo information,resulting in low detection accuracy.In order to improve the intelligent tracking and detection ability of multi-target,a multi-target intelligent detection algorithm based on machine learning in non-equilibrium environment is proposed.Firstly,a non-equilibrium data acquisition model of multi-target tracking recognition method is established.Secondly,the multi-target features of unbalanced data under multi-sensor network are detected by neai^field source echo detection method.The spatial grid clustering method is used for multi-target feature amount classification mining of non-equilibrium data.Then the optimization control of multi-target intelligent detection process of non-equilibrium data is combined with machine learning method to realize the non-equilibrium environment.Finally*the optimization design of multi-target intelligent detection algorithm in unbalanced environment is realized by combining machine learning methods to optimize the control strategy in the process of multi-target intelligent detection of unbalanced data.The experimental results show that the method has better intelligence,higher detection accuracy,and shorter detection time overhead for multi-objective detection in non-equilibrium environment.

关 键 词：机器学习 非平衡数据 多目标 智能检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]