Classify Tree Species from Point Clouds

Function Overview

Point Cloud Tree Species Recognition uses a defined tree species recognition tool to automatically identify tree species for individual trees in TLS forest point cloud data. The function is based on deep learning models and can determine tree species from point cloud data.

The recognition tool can come from:

• The built-in Classify Tree Species from Point Clouds tool.
• A custom model trained with Custom Tree Species Identification Model Training, then configured as a recognition tool.

Usage

Open the tree species recognition function in the TLS Forest module.

• Point Cloud Mode:

Parameter Settings

• Input Data: Input point cloud data that has already been segmented into individual trees.

• Evaluation Parameters: Parameters used to control the model recognition process.

• Prefer GPU: When selected, GPU acceleration is used if the GPU configuration meets the requirements.

• Name Priority: Controls which name is written for the identified tree species. Latin name, Chinese name, common name, and genus name can be used. The detailed species information is managed in the tree species library.

• Continue Searching Lower-Priority Names When the Current Name Is Empty: For example, if the priority is common name > genus name > Chinese name > Latin name and the common name is empty while other names are available, selecting this option writes the genus name. If it is not selected, the result remains empty.

• Panoramic Mode: If panoramic mode was used during training, the interface is as follows:

For parameter descriptions, see Individual Tree Attributes Statistics.