车载LiDAR点云数据中杆状地物自动提取与分类

针对城市道路场景中车载LiDAR点云数据质量差、各类地物相互遮掩的情况,提出杆状地物自动提取与分类算法。先通过改进数学形态学算法移除点云数据中的地面点,再根据杆状地物的形态特征,使用纵向格网模板初步提取杆状地物,然后对提取的疑似杆状地物进行点云数据规则化并通过统计分析移除噪声点,最后根据预先建立的杆状地物样本训练SVM分类模型,对提取的杆状地物进行分类。试验表明,本文方法能够在数据质量欠佳的情况下有效提取城市道路场景中的杆状地物,并对提取的杆状地物进行高精度分类。

Automatic extraction and classification of pole-like objects from vehicle LiDAR point cloud

Aiming at the poor quality of vehicle LiDAR point cloud data and the mutual concealment of various ground objects in urban road scenes,an automatic extraction and classification algorithm for pole-like objects was proposed.Firstly,ground points in point cloud data were removed by improving the mathematical morphology algorithm.According to the morphological characteristics of the pole-like objects,preliminary extraction of pole-like objects was carried out through the longitudinal grid template.Secondly,the extracted suspected pole-like objects were regularized with point cloud data and some noise was removed by statistical analysis.Finally,SVM classification model was trained according to the previously established pole-like object samples to classify the extracted pole-like objects.The experimental results showed that the method could effectively extract the pole-like objects in urban road scenes under the condition of poor data quality,and classified the extracted pole-like objects with high precision.