Construction of a Precision Thinning Framework for Cunninghamia Lanceolata Plantations Driven by Point Cloud Fusion and Multi-Criteria Weighting

Abstract

Thinning is a cornerstone of sustainable forest management (SFM), yet decisions are often constrained by subjective experience and a lack of quantitative spatial data. To address this, we developed a precision thinning framework for Cunninghamia lanceolata plantations by fusing terrestrial and aerial LiDAR data. Instead of relying on manual selection, this study established a quantitative mechanism for identifying harvesting targets. We employed a layer-stacking seed-point algorithm to accurately segment individual trees and extracted key parameters (tree height, DBH, and crown width) with high precision (Overall R2 ≥ 0.85). Furthermore, an objective multi-criteria weighting method (AHP – CRITIC) was constructed to prioritize thinning targets based on stand stability and growth status. Results indicated that the proposed framework achieved an individual tree segmentation F1-score of 89.24%. Simulated thinning based on the calculated weights significantly optimized the stand spatial structure: canopy openness increased by 11.8%, and spatial indices converged toward optimal ranges. Compared with conventional practices, the proposed approach effectively reduced growth dispersion and enhanced population coordination. These findings demonstrate that integrating multi-platform LiDAR with objective weighting algorithms offers a scientifically rigorous pathway for precision forestry, promoting both productivity and the long-term sustainability of plantation ecosystems.

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.