摘要
为满足高精度地图制作中对道路标识线提取需要,为充分挖掘车载点云数据三维空间特征及反射强度特征提出一种道路标识线提取算法。首先,采用数学形态学滤波算法完成地面数据提取;其次,基于法向量特征及聚类算法获取路面数据;然后,基于反射强度信息的分层模板法提取标识点云,完成道路标识线信息提取;最后,对实际道路的验证结果表明:该方法的准确率达到96%以上,完整率达到94%,检测质量指标达到94%,能提取比较完整的道路标识线。
In order to meet the needs of extracting traffic sign lines in high-precision map production,in order to fully explore the three-dimensional spatial characteristics and reflection intensity characteristics of vehicle-mounted point cloud data,a road sign line extraction algorithm is proposed.Firstly,mathematical morphology filtering algorithm is used to extract ground data;Secondly,obtain road surface data based on normal vector features and clustering algorithms;Then,based on the reflection intensity information,the layered template method is used to extract the identification point cloud and complete the extraction of road identification line information.The validation results on actual roads show that the accuracy and completeness of this method reached over 96%,94%,and the detection quality index reached 94%,which can extract relatively complete road identification lines.
作者
段永涛
DUAN Yongtao(Beijing Huaxing Exploration of New Technology Co.,Ltd.,Beijing 100000,China)
出处
《经纬天地》
2023年第3期48-51,共4页
Survey World
关键词
点云
道路标识线提取
反射强度
分层模板法
point cloud
road marking line extraction
reflection intensity
layered template method
