基于多阈值判断的单线激光雷达草丛识别方法

本文基于单线激光雷达提出了一种对于草丛障碍物的识别方法。首先对草丛特征进行分析,阐述了草丛障碍物与其他连续体障碍物的特征区别,由特征分析提出了一种多阈值草丛特征识别算法,通过候选框对单线激光雷达点云识别保留,最后通过试验进行验证该多阈值草丛识别方法。试验结果表明该算法能够充分利用单线激光雷达点云数据量小、计算时间短、鲁棒性强等优点,弥补了特殊环境下视觉识别适应性差、识别失效等问题,更好地发挥了单线激光雷达的作用。

Research on Grass Recognition Method of Single-line Lidar Based on Multi-threshold Judgment

This paper proposes a recognition method for grass obstacles based on single-line LiDAR. Firstly, the grass features are analyzed, the feature difference between grass obstacles and other continuum obstacles is elaborated, a multi-threshold grass feature recognition algorithm is proposed by the feature analysis, the single-line LiDAR point cloud recognition is retained by the candidate frame, and finally the multi-threshold grass recognition method is verified by experiments. The experimental results show that the algorithm can make full use of the advantages of small amount of single-line LiDAR point cloud data, short computation time and robustness, and make up for the problems of poor adaptation of visual recognition and recognition failure in special environments, and better play the role of single-line LiDAR.