CAJ | 학술논문

Velodyne 64线雷达被广泛应用于自主车的环境感知功能中。为了在该雷达产生的大量数据中识别出障碍物信息，本文提出了一种新的基于分离地面点的障碍物识别算法。首先从原始数据集中提取出特征点集并建立地面模型；利用地面模型识别出地面点并将其从数据集中滤除；将地面点分离后的数据集投影至平面栅格中，并利用区域生长法聚类。该算法能够准确、高效地识别出障碍物。
Velodyne 64선뢰체피엄범응용우자주차적배경감지공능중。위료재해뢰체산생적대량수거중식별출장애물신식，본문제출료일충신적기우분리지면점적장애물식별산법。수선종원시수거집중제취출특정점집병건입지면모형；이용지면모형식별출지면점병장기종수거집중려제；장지면점분리후적수거집투영지평면책격중，병이용구역생장법취류。해산법능구준학、고효지식별출장애물。
The Velodyne 64E 3D Radar is widely used in situation awareness of autonomous vehicle. For the purpose of identifying obstacle in large amounts of data generated by the Radar, a new method is presented based on ground point segmentation. First, feature points are extracted from the original data set, and ground model is built base on these points. We use the ground model to identify ground points and remove them from the original data set, then map the processed data set to grid, and use the region labeling to cluster non-ground points. This approach can detect obstacle accurately and efficiently.