激光与光电子学进展, 2020, 57 (22): 221017, 网络出版: 2020-11-05
综合布料滤波与改进随机森林的点云分类算法 下载: 976次
Point Clouds Classification Algorithm Based on Cloth Filtering Algorithm and Improved Random Forest
图像处理 激光雷达 布料滤波算法 随机森林 点云归一化 最大互信息系数 image processing LiDAR cloth filtering algorithm random forest normalized point cloud maximal information coefficient
摘要
城区激光雷达点云建筑物提取技术是近年来发展的热点,如何准确区分植被、建筑物以及人造物,提高分类精度一直是研究难点。为此,针对分类精度较低的问题,提出一种基于随机森林的点云分类算法。首先使用改进布料滤波算法对点云数据进行地面滤波;其次,构建决策树并进行基于最大互信息系数的相关性分析,选出相关系数最小、精度最高的决策树,得到弱相关随机森林模型;最后,对决策结果进行加权投票处理,得到一种综合布料滤波和加权弱相关随机森林的点云分类算法,并通过Vaihingen城区数据集对算法进行验证。实验表明,与传统随机森林分类算法相比,本文算法提高了4.2%的分类精度,也提高了算法效率。
Abstract
Building extraction technology in urban areas has been a hot topic in recent years, but how to accurately distinguish vegetation, buildings, and man-made objects and improve classification accuracy has always been a difficult point. Aiming at the problem of low classification accuracy, we propose a point cloud classification algorithm based on random forest. First, the improved cloth filtering algorithm is used to perform ground filtering on the point cloud data. And a decision tree is constructed and the correlation analysis based on the largest mutual information coefficient is performed to select the decision tree with the smallest correlation coefficient and the highest accuracy to obtain a weakly correlated random forest model. The decision results are processed by weighted voting, and finally a point cloud classification algorithm combining cloth filtering and weighted weakly correlated random forest is obtained. Compared with the traditional random forest classification algorithm, the algorithm is verified by the Vaihingen urban dataset, and the classification accuracy is improved by 4.2%.
薛豆豆, 程英蕾, 释小松, 秦先详, 文沛. 综合布料滤波与改进随机森林的点云分类算法[J]. 激光与光电子学进展, 2020, 57(22): 221017. Doudou Xue, Yinglei Cheng, Xiaosong Shi, Xianxiang Qin, Pei Wen. Point Clouds Classification Algorithm Based on Cloth Filtering Algorithm and Improved Random Forest[J]. Laser & Optoelectronics Progress, 2020, 57(22): 221017.