基于激光雷达点云数据的树种分类

Classification of Tree Species Based on LiDAR Point Cloud Data

以杭州钱江新城森林公园和新疆维吾尔自治区阿克苏市红旗坡农场的水杉、柳树、女贞、竹子和苹果树为研究对象,基于机载LiDAR获取高分辨率点云数据,结合支持向量机分类器,提出了多种树木特征,如结构特征参数、纹理特征参数和冠形特征参数等,以实现树种分类。实验结果表明,5种树木分类的整体准确率达85%,Kappa系数为0.81。所提分类方法不仅从LiDAR数据中获得了更有前景的单株树特征,还展示了一个可用于提高树种分类性能的有效框架。

This study involved the Metasequoia glyptostroboides,Salix babylonica,Ligustrum lucidum,bamboo,and Malus pumila Mill.from the Qianjiang new town forest park of the Hangzhou city and the Hongqipo farm of the Aksu city in the Xinjiang Uygur Autonomous Region.The structural,textural,and crown features were proposed based on high-resolution point cloud data acquired by the airborne LiDAR and a support vector machine classifier.The experimental results demonstrate that the overall accuracy of the classification is 85%,with a Kappa coefficient of 0.81.The proposed method derives promising features for a tree based on the LiDAR data and demonstrates an effective framework for improving the classification performance of the tree species.