基于模糊关联空间的激光雷达三维扫描数据过滤研究  

A Study on 3D Scanning Data Filtering LiDAR Based on Fuzzy Correlation Space

作　　者：陈改霞 叶萧然 CHEN Gaixia;YE Xiaoran(Hebi Institute of Engineering and Technology,Henan Polytechnic University,Hebi Henan 458030,China)

机构地区：[1]河南理工大学鹤壁工程技术学院,河南鹤壁458030

出　　处：《现代雷达》2023年第12期102-108,共7页Modern Radar

基　　金：2021年度河南省高等学校重点科研资助项目(22B520018,22B470007)。

摘　　要：激光雷达三维扫描数据中存在大量无效点和有效点,但是两者之间的空间关联性阈值处在不断变动中,以固定关联规则的过滤方式很难准确区分,造成实际过滤应用效果较差。为此,提出基于模糊关联空间的激光雷达三维扫描无效数据过滤方法。采集激光雷达三维扫描样本数据构建无效数据识别规则库,将无效数据识别规则库中的无效数据集合转换为模糊集合,并利用聚类算法填补无效数据造成的数据空格,以避免出现数据过滤误差。填补后运用支持向量机分类器计算激光雷达三维扫描数据之间的模糊关联度,并对其进行分类。结合分类结果利用网闸过滤组织对激光雷达三维扫描无效数据过滤。实验结果表明:这种方法过滤精度高,具有较好的过滤效果,时间复杂度低、过滤耗时短,实际应用效果较好。There are a large number of invalid and effective points in the 3D scanning data of LiDAR,but the spatial correlation threshold between the two is constantly changing,making it difficult to accurately distinguish using fixed association rules for filtering,resulting in poor practical filtering applications.To this end,a method for filtering invalid data in LiDAR 3D scanning based on fuzzy correlation space is proposed,which converts the invalid data set in the invalid data recognition rule library into a fuzzy set,and uses clustering algorithms to fill in the data gaps caused by invalid data to avoid data filtering errors.After filling,support vector machine classifier is used to calculate the fuzzy correlation between the 3D scanning data of the LiDAR and classify it.Combined with the classification results,the invalid data of LiDAR 3D scanning are filtered using a mesh filter organization.The experimental results show that this method has high filtering accuracy,good filtering effect,low time complexity,short filtering time,and good practical application effect.

关 键 词：激光雷达 三维扫描数据 模糊关联空间 数据过滤 聚类算法 支持向量机分类器 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]