Tree Species Classification Using Ground-Based LiDAR Data by Various Point Cloud Deep Learning Methods

Abstract

Tree species information is an important factor in forest resource surveys, and light detection and ranging (LiDAR), as a new technical tool for forest resource surveys, can quickly obtain the 3D structural information of trees. In particular, the rapid and accurate classification and identification of tree species information from individual tree point clouds using deep learning methods is a new development direction for LiDAR technology in forest applications. In this study, mobile laser scanning (MLS) data collected in the field are first pre-processed to extract individual tree point clouds. Two downsampling methods, non-uniform grid and farthest point sampling, are combined to process the point cloud data, and the obtained sample data are more conducive to the deep learning model for extracting classification features. Finally, four different types of point cloud deep learning models, including pointwise multi-layer perceptron (MLP) (PointNet, PointNet++, PointMLP), convolution-based (PointConv), graph-based (DGCNN), and attention-based (PCT) models, are used to classify and identify the individual tree point clouds of eight tree species. The results show that the classification accuracy of all models (except for PointNet) exceeded 0.90, where the PointConv model achieved the highest classification accuracy for tree species classification. The streamlined PointMLP model can still achieve high classification accuracy, while the PCT model did not achieve good accuracy in the tree species classification experiment, likely due to the small sample size. We compare the training process and final classification accuracy of the different types of point cloud deep learning models in tree species classification experiments, further demonstrating the advantages of deep learning techniques in tree species recognition and providing experimental reference for related research and technological development.

Internal Tree Trunk Decay Detection Using Close-Range Remote Sensing Data and the PointNet Deep Learning Method

“…The number of points defines how many points will be plotted on a 3D mesh, creating a point cloud. PointNet does not seem to perform differently when exceeding 2048 points, so this value was used even in this study [34,39].…”

“…Nevertheless, the task of object classification is common. It is applied to many more or less complicated objects in various fields of research, including forestry [38,39,[51][52][53], providing 92.5% accuracy in the task of tree segmentation from a complex scene [35] or tree species classification from 3D point clouds with more than 90% accuracy [34]. For this reason, the proposed study correctly assumed that deep learning classification of decayed or healthy tree stems is possible.…”

“…A new trend in data evaluation is the use of deep learning. Usually, images or 3D point clouds are often evaluated by neural networks, and exciting results are reported in object detection or classification-specifically, in the detection of individual trees or tree species classification [32][33][34][35], but it is appropriate for the detection of any feature with distinctive patterns [36]. As mentioned before, it is known that some fungi species cause specific changes to the external shape of a tree stem base, and stem decay is commonly related to the presence of bulges, cancer, and swelling.…”

Semi- and weak-supervised learning for Norwegian tree species detection

“…Other deep learning approaches utilize the point cloud itself for classifying tree species. However, these are rarely applicable across large areas due to the need for intensive data processing and often aim for single-tree classification [16,14]. Overall, related published deep learning models are not directly comparable to a model that uses rasterized medium-resolution lidar-derived data products such as Digital Surface and Digital Terrain models.…”

TSE-UNet: Temporal and Spatial Feature-Enhanced Point Cloud Super-Resolution Model for Mechanical LiDAR

“…To overcome the shortcomings of mechanical liDAR in practical applications, we delve into recent advances in point cloud [5][6][7], image-based super-resolution technology [8,9], and deep-learning-based modeling [10,11], which can provide new ideas and methods to solve the limitations of mechanical LiDAR. First of all, as the main data form of LiDAR environment perception, the point cloud has accurate 3D information.…”