一种基于结构化环境的线性距离特征提取算法

A Structured Line Segment Feature Extraction Algorithm Based on Linear Distance

为解决室内环境中移动机器人的自主导航问题,提出了一种基于结构化环境的线性距离特征提取算法。首先通过建立机器人运动模型,对激光雷达获得的点云数据进行预处理。然后采用聚类算法对预处理后的数据进行分割和合并。采用正交拟合算法,估算特征线段的最大角度公差,并提取竖直和水平特征线进行误差纠正。仿真实验结果表明:算法可有效提取室内环境特征线段并建立特征地图。同时调用数据集与ICP(iterative closest point)算法进行对比测试,结果表明使用该算法构建环境地图,可见使用此算法可降低建图时间复杂度,同时提高地图匹配精度。

In order to solve the problem of autonomous navigation of mobile robots in indoor environment,a linear distance feature extraction algorithm based on structured environment is proposed.Firstly,a robot motion model was established,and the point cloud data obtained by the laser radar was pre-processed.Then the pre-processed data was divided and combined by a clustering algorithm.The maximum angle tolerance of feature line segment was estimated by orthogonal fitting algorithm,and the vertical and horizontal feature lines were extracted for error correction.The simulation results show that the algorithm can effectively extract the feature line of the indoor environment and establish the feature map.At the same time,the data set and the iterative closest point(ICP)algorithm are called to carry out the contrast test,and the result shows that the environment map can be constructed by using the algorithm,so that the time complexity of the building graph can be reduced,and the matching precision of the map can be improved.