A method that extracts segments in 2D range images from LIDAR (light detection and ranging) for mo-bile robotic localization and mapping was presented. The method included two stages, splitting and merging. In the first stage, a set of points was split recursively with IEPF, and then the end points of all the sets were reconfigured by using the principle of LT segmentation method. In the second stage, two situations are corresponding to the number of points included in two adjacent sets of points, and two adaptive threshold merging methods were used accordingly. The merging stage was also a recursive process. Experimental results show that the method presented in this paper well resolved over-splitting and over-merging problems in exacting segments. The end points of segments are denoted more accurately.%提出一种从2D激光雷达距离图像中提取线段特征的方法,以满足移动机器人定位和地图创建的需要.该方法包括分割和合并两个阶段.采用IEPF(iterative end point fit)分割算法对输入点集进行递归分割,利用LT (line tracking)分割算法对已分割点集的端点进行重新分配.根据两个相邻点集所包含的点数将合并过程分为两种情况考虑,分别采用不同的自适应阈值合并算法对两点集进行合并.对比实验结果表明,该方法能解决线段提取中过分割和过合并问题,对线段端点表示较为精确.
展开▼
