Aiming at the problem that the line segment feature recognition of the existing LiDAR feature recognition algorithm is not accurate enough,an improved line segment feature recognition algorithm of LiDAR is proposed.Firstly,continuous points are clustered based on the improved slope difference algorithm to find all corners,breakpoints and scattered points,and the breakpoint is used as the feature to segment the point set.Secondly,false breakpoints are eliminated by continuous edge tracking algorithm.Then,the improved iterative end point fit(IEPF)algorithm is used to filter all corner points found based on the slope difference algorithm,and the real corner points are obtained.Finally,feature line segment is obtained by least square fitting of the feature point set after segmentation and filtering so as to obtain feature line segment.The experimental results show that the method is more accurate than the existing feature recognition algorithms in corner point and breakpoint recognition,so the length robustness of the extracted feature line segment is better.
关键词
激光雷达/特征识别/斜率差算法/迭代适应点算法
Key words
LiDAR/feature recognition/slope difference algorithm/iterative end point fit(IEPF)algorithm
维普
万方数据