顾及点云邻域特征的道路里程推算方法实现

Realization of Highway Mileage Calculation Method Considering the Neighborhood Characteristics of Point Clouds

提出车载测量系统获取点云进行公路里程推算方法,基于布料模拟滤波算法分类路面、非路面点云,以点云强度信息为标量并顾及点云邻域几何特征分割地面点云,采用八叉树连通聚类方法提取特征标线点云并以此构造路面中心线,制作ArcToolbox里程桩点提取工具来推算道路里程。实验表明本方法可快速分割标线特征点云并准确提取路面中心线,相比于传统的方法可更加准确地实现道路里程推算和桩点坐标提取。

In view of the problems such as the total mileage of expressway, the serious phenomenon of broken chain, and the difficulty of precise positioning management, we proposed a highway mileage calculation method based on point clouds obtained by the vehicle mobile measurement system. We used the cloth simulation filtering algorithm to classify the road and non road point clouds, and taking the point clouds intensity information as scalar, segmented the ground point clouds considering geometric characteristics of the point clouds neighborhood. And then, we used the octree connected clustering method to extract the feature line point clouds, and constructed the road centerline. Finally, we made the ArcToolbox mileage point extraction tool to calculate the highway mileage. The experimental results show that this method can quickly segment the feature point cloud of feature line, and accurately extract the center line of the road surface. Compared with the traditional method, the method can more accurately calculate the highway mileage and extract the coordinates of pile points.