Extract Tree Stem(TLS)

Functional Overview

This function is used to extract stems from terrestrial laser scanning point cloud data. Point clouds belonging to the same stem will be given a unique tree ID. This function is mainly aimed at the point cloud data obtained by high-precision station scanning to solve the complex problem of forest land vegetation, and the DBH of the extracted stem can reach a minimum of 0.1m. This algorithm is not suitable for point clouds acquired by airborne (ALS), backpacks and other acquisition methods, because the normal consistency of the stem surface of these point clouds were not strong.

Usage

Click TLS Forest> Extract Tree Stem(TLS).

Parameters Settings

• Input Data: The input data must be normalized point cloud data. For normalization method, please refer to normalization or normalization according to ground points. The input file can be a single data file or a point cloud data set; the file(s) to be processed must be opened the in LiDAR360 software.
• Height Above Ground (m) (default value: "0.5"): The point with z value lower than this will be regarded as the low understory vegetation and will not participate in the stem extraction. This parameter is used to reduce the influence of low understory vegetation on the effect of individual tree segmentation. 0.5-0.8m is appropriate for most TLS data.
• Max Stem Height (m) (default value: "4.0"): Maximum height of a stem. Points above this value will not be involved in the stem extraction. If this value is set too large, it will affect the accuracy of stem extraction and reduce the calculation efficiency. Taking rainforest data as an example, the height difference between trees may reach 20 to 30 meters, so only point clouds of common part height (less than 4.0 meters) are taken for stem extraction, and DBH and other forest parameters are calculated.
• Min Stem Height (m) (default value: "1.5"): Minimum height of a stem. The height of the shortest tree in the forest land. This height mainly affects the detection results of low stems. This value must be greater than "Height above ground point" and less than "Maximum stem height". Low stems with a height of 0.8 times this value will be detected.
• Buffer Size (m) (default value: "2.0"): Buffer area for getting data is used to solve the edge connection problem of adjacent data blocks in the algorithm. This value needs to be set to be slightly larger than the maximum diameter at breast height (DBH) of the trees in the scene. If this value is set too large, the calculation efficiency will be reduced, and if it is set too small, it will result in over-segmentation of part of the stems. 2.0 is suitable for most TLS data.
• Max Stem Inclination (degree) (default value: "15"): The stem forms the maximum inclination with the vertical direction. Stem points with the inclination angle less than this value will participate in subsequent calculation as candidate points. The larger this value is set, the more stem points are extracted, but the accuracy is reduced. On the contrary, the smaller this value is set, the fewer the number of tree stem points are and the higher the precision is. 10~15° is suitable for most data.
• Voxel Sampling (m) (default value: "0.02"): The grid size of Voxel thinning. Raw TLS data needs voxel thinning to reduce the computational burden. When this threshold is greater than the average point density of the point cloud (calculated automatically within the algorithm), grid thinning will be performed according to this threshold, otherwise grid thinning will not be performed. If the data has been subject to resampling, this value can be set as the default value.

> @inproceedings{ author={Shengli Tao, Nicolas Labrière, Kim Calders ,Fabian Jörg Fischer, E‑Ping Rau,Laetitia Plaisance, Jérôme Chave}, title={Mapping forest disturbances across the Southwestern Amazon: tradeoffs between open-source, Landsat-based algorithms}, booktitle={Environmental Research Communications, 3(9):091001(13pp)}, year={2021} }