摘要
针对2D激光雷达扫描数据异常值检测问题,提出了最邻近距离-局部异常因子检测算法。首先给出2D激光雷达扫描数据孤立点的定义,并依此检测扫描数据的孤立点;其次基于局部异常因子检测扫描数据的局部异常点;最后采用5组实际环境中的2D雷达扫描数据对算法进行验证。研究结果表明:最邻近距离-局部异常因子检测算法能够检测出2D激光雷达扫描数据的异常值,且算法的平均误报率为1.228%、平均计算时间为0.842s,可满足实际需求。
Aiming at the problem of outlier detection in 2D lidar scanning data,the nearest neighbor distance local outlier factor detection algorithm is proposed.Firstly,the definition of global outliers in 2D lidar scanning data is given,and then the global outliers of scanned data are detected;secondly,the local outliers of scanned data are detected by using local outlier factor;finally,the algorithm is verified by using five sets of 2D radar scanning data in actual environment.The results show that:the nearest neighbor distance local outlier factor detection algorithm can detect the abnormal value of 2D lidar scanning data,and the average false alarm rate of the algorithm is 1.228%,and the average calculation time is 0.842 seconds,which can meet the actual needs.
作者
刘延彬
姜媛媛
LIU Yan-bin;JIANG Yuan-yuan(School of Mechanics and Optoelectronic Physics,Anhui University of Science and Technology,Huainan 232001.China;School of Electrical and Information Engineering,Anhui University of Science and Technology,Huainan 232001,China)
出处
《激光与红外》
CAS
CSCD
北大核心
2021年第9期1243-1248,共6页
Laser & Infrared
基金
国家自然科学基金项目(No.51604011)
安徽省高校省级自然科学研究项目(No.KJ2019ZD12)
安徽理工大学芜湖研究院研究基金项目(No.ALW2020YF21)资助。