Automatic tree parameter extraction by a Mobile LiDAR System in an urban context

Area of Science:

• Urban forestry
• Geospatial analysis
• Remote sensing

Background:

• Accurate urban tree data is crucial for city planning, hazard assessment, and environmental monitoring.
• Existing methods for urban tree parameter estimation can be labor-intensive and time-consuming.

Purpose of the Study:

• To develop an innovative, automated methodology for estimating individual urban tree structural parameters using Mobile LiDAR System data.
• To validate the accuracy and scalability of the proposed methodology for city-level urban tree analysis.

Main Methods:

• Diameter at Breast Height (DBH) estimation using RANSAC circle fitting and extreme point distance for non-circular trees.
• Tree size extraction via connectivity analysis, Crown Base Height calculation using voxelization.
• Canopy Volume estimation with mesh generation and α-shape methods, and tree location determination via Principal Component Analysis.

Main Results:

• High correlation (R²=0.92 for DBH in 20 cm height bin) and coefficients (>0.91 for other parameters) achieved against field measurements.
• Validation performed on 29 trees and testing on a larger dataset of 58 trees demonstrated methodology feasibility.
• Investigated the influence of point density and height bins on DBH estimation accuracy.

Conclusions:

• The proposed automated methodology is feasible and scalable for comprehensive urban tree analysis.
• This approach provides accurate structural parameter estimation, supporting efficient urban forest management.
• Mobile LiDAR System data combined with advanced algorithms offers a powerful tool for urban tree inventory.