Feature Extraction Method of Indoor Structured Environment Based on Two-Dimensional LiDAR

摘要

In order to realize the geometric feature extraction of two-dimensional lidar (Light Detection and Ranging) point cloud data in an indoor structured environment. First, the raw data collected by the lidar sensor are subjected to outlier culling processing. Then, a region segmentation method of the point cloud data based on adaptively dynamic threshold is proposed. This proposed method can perform dynamic threshold adjustment according to the adjacent point cloud data from the lidar in real time. The simulation verification results show that the method can effectively detect the division points of the region segmentation and improve the accuracy of the division point detection. Finally, the feature extraction operation of the common line feature and corner feature in the indoor structured environment is carried out. The simulation results show that this method can effectively extract the common two-dimensional geometric features in the indoor structured environment.