基于数据挖掘技术的激光雷达硬件故障数据识别方法  

Recognition Method of Lidar Hardware Fault Data Based on Data Mining Technology

作　　者：李晓东 Li Xiaodong(School of Information and Electromechanical Engineering,Zhengzhou Business University,Zhengzhou,Henan 451200,China)

机构地区：[1]郑州商学院信息与机电工程学院,河南郑州451200

出　　处：《应用激光》2022年第7期87-93,共7页Applied Laser

基　　金：河南省高等学校青年骨干教师培养计划(2018GGJS192);河南省科技攻关项目(182102210548)。

摘　　要：采用目前方法识别激光雷达硬件故障数据时,未分析比较不同天气情况下的激光雷达回波数据特征,导致出现误判率高、召回率与调和指标低的问题,为此提出基于数据挖掘技术的激光雷达硬件故障数据识别方法,首先采用EMD和小波阈值联合法去除激光雷达回波数据中的杂波,然后通过模糊C均值聚类对去噪后的数据聚类处理,挖掘其数据特征,最后建立不同天气影响情况下的正常回波信号数据和故障回波信号数据的特征概率分布模型,完成故障数据的识别。试验结果表明,所提方法能有效地降低误判率,提高召回率与调和指标。When the current method is used to identify lidar hardware failure data, the characteristics of Lidar echo data under different weather conditions are not analyzed and compared, which leads to the problems of high false positive rate, low recall rate and harmonic index. A method for identifying lidar hardware fault data based on data mining technology is proposed. Firstly, the combined method of EMD and wavelet threshold is used to remove clutter in the lidar echo data, and then the denoised data is clustered by fuzzy C-means clustering. Moreover, its data characteristics are mined, and finally the characteristic probability distribution model of normal echo signal data and fault echo signal data under different weather influences is developed to complete the identification of fault data. Experimental results show that the proposed method can effectively reduce the misjudgment rate and improve the recall rate and the coordination index.

关 键 词：数据挖掘 激光雷达 硬件故障 数据识别 模糊决策树 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]