基于TLS辅助的长白落叶松一级枝条生物量模型构建

【目的】探究利用地基激光雷达(terrestrial laser scanning, TLS)点云数据估测枝条生物量的可行性，构建预测长白落叶松(黄花落叶松)枝条生物量的最优模型。【方法】以利用孟家岗林场26株长白落叶松点云数据提取出的733个一级枝条的特征因子枝长(L_BL)、弦长(L_BCL)、基径(d_B)、着枝角度(A_B)、弓高(H_BAH)、枝条基部断面积(S_BAB)、相对着枝深度(d_RDINC)]和对应的实测数据为数据源，分别建立枝条水平上的一级枝条生物量基础模型，通过对比基础模型之间的差异来分析利用TLS数据建立枝条生物量模型的可行性。最后利用TLS数据分别对比基础模型、混合效应模型和随机森林模型的预测效果。【结果】基础模型中最终选定的自变量为S_BAB和L_BCL。利用TLS数据建立的枝条生物量基础模型具有更好的预测精度。对比3种模型预测能力结果显示，随机森林模型无论在训练集还是测试集...

Constructing a biomass model of Larix olgensis primary branches based on TLS

【Objective】 To explore the feasibility of using terrestrial laser scanning (TLS) point cloud data to estimate branch biomass and find the optimum model to predict the branch biomass of Larix olgensis. 【Method】 The characteristic factors (branch length (L_BL), branch chord length (L_BCL), branch diameter (d_B), branch angle (A_B), branch arch height (H_BAH), basal area of branches (S_BAB), and relative depth within the crown (d_RDINC)) of 733 first-order branches extracted using point cloud data of 26 artificial larch plants and the corresponding measured data from Mengjiagang Forestry Farm were used as data sources to establish first-level branch biomass base models at the branch level. Additionally, the feasibility of using TLS data to establish branch biomass models by comparing the differences between base models were analyzed. Finally, the prediction effects of the base model, mixed-effects model, and random forest model were compared separately using TLS data. 【Result】 The independent variables selected in the base model were S_BAB and L_BCL. Using TLS data to build the branch biomass base model had a better prediction accuracy. Comparing the results of the three models showed that the random forest model showed the best results both on the training and test sets, in the following order: random forest model > mixed-effects model > base model. The R² of the random forest model improved by 1.32% and 4.89%, and the RMSE decreased by 11.23% and 13.60%, respectively, compared to the mixed model and the base model. 【Conclusion】 The results of this paper proved that the branch biomass can be estimated accurately based on TLS using the random forest algorithm, which not only laid a practical foundation for the application of the random forest algorithm on stand growth models, but also provided an important reference value for the application of TLS in the study of crown structure.