Tree Species Identification Based on Point Cloud and Panoramic Images

Function Overview

Tree Species Identification identifies tree species for individual trees and writes the results back to the tree attribute file. The tool supports point-structure-based identification, panoramic-image-assisted identification, and LLM-based identification.

Usage

Click TLS Forest > Tree Species Identification to open the tool.

Parameter Settings

• Input Data: Input individual tree attribute file (*.treedb).

• Recognition Method: Method used for tree species identification.

Point Structure

• LiData File: Point cloud file corresponding to the tree attribute file.

• Tree Species Confidence: Confidence threshold for tree species identification.

• Prefer GPU: Use GPU acceleration when a supported GPU is available.

• Select Model: Select the tree species recognition model.

Panoramic Images

• Point Cloud File: Point cloud file corresponding to the tree attribute file.

• Panoramic Image Path: Folder containing panoramic images.

• Tree Species Confidence: Confidence threshold for tree species identification.

• Use LLM Optimization: Use LLM optimization for the panoramic-image-based result.

LLM

• Panorama Image: Optional panorama image path.

• Projected Coordinate System: Optional coordinate system used with panorama images. See Define Projection.

Output

The tool updates the input tree attribute file with tree species and confidence information.

Notes

• Point-structure-based identification requires the corresponding LiData point cloud file.
• Panoramic-image-based identification requires both the point cloud and the panoramic image path.
• For LLM-based identification, panorama images and coordinate system information are optional and depend on the selected model capability.