Achieving the accurate and efficient monitoring of forests at the tree level can provide detailed information for precise and scientific forest management. However, the detection of individual trees under planted forests characterized by dense distribution, serious overlap, and complicated background information is still a challenge. A new deep learning network, YOLO-DCAM, has been developed to effectively promote individual tree detection amidst complex scenes. The YOLO-DCAM is constructed by leveraging the YOLOv5 network as the basis and further enhancing the network’s capability of extracting features by reasonably incorporating deformable convolutional layers into the backbone. Additionally, an efficient multi-scale attention module is integrated into the neck to enable the network to prioritize the tree crown features and reduce the interference of background information. The combination of these two modules can greatly enhance detection performance. The YOLO-DCAM achieved an impressive performance for the detection of Chinese fir instances within a comprehensive dataset comprising 978 images across four typical planted forest scenes, with model evaluation metrics of precision (96.1%), recall (93.0%), F1-score (94.5%), and AP@0.5 (97.3%), respectively. The comparative test showed that YOLO-DCAM has a good balance between model accuracy and efficiency compared with YOLOv5 and advanced detection models. Specifically, the precision increased by 2.6%, recall increased by 1.6%, F1-score increased by 2.1%, and AP@0.5 increased by 1.4% compared to YOLOv5. Across three supplementary plots, YOLO-DCAM consistently demonstrates strong robustness. These results illustrate the effectiveness of YOLO-DCAM for detecting individual trees in complex plantation environments. This study can serve as a reference for utilizing UAV-based RGB imagery to precisely detect individual trees, offering valuable implications for forest practical applications.

Urban tree canopy (UTC) area is an important index for evaluating the urban ecological environment; the very high resolution (VHR) images are essential for improving urban tree canopy survey efficiency. However, the traditional image classification methods often show low robustness when extracting complex objects from VHR images, with insufficient feature learning, object edge blur and noise. Our objective was to develop a repeatable method—superpixel-enhanced deep neural forests (SDNF)—to detect the UTC distribution from VHR images. Eight data expansion methods was used to construct the UTC training sample sets, four sample size gradients were set to test the optimal sample size selection of SDNF method, and the best training times with the shortest model convergence and time-consumption was selected. The accuracy performance of SDNF was tested by three indexes: F1 score (F1), intersection over union (IoU) and overall accuracy (OA). To compare the detection accuracy of SDNF, the random forest (RF) was used to conduct a control experiment with synchronization. Compared with the RF model, SDNF always performed better in OA under the same training sample size. SDNF had more epoch times than RF, converged at the 200 and 160 epoch, respectively. When SDNF and RF are kept in a convergence state, the training accuracy is 95.16% and 83.16%, and the verification accuracy is 94.87% and 87.73%, respectively. The OA of SDNF improved 10.00%, reaching 89.00% compared with the RF model. This study proves the effectiveness of SDNF in UTC detection based on VHR images. It can provide a more accurate solution for UTC detection in urban environmental monitoring, urban forest resource survey, and national forest city assessment.

Since tree morphological structure is strongly influenced by internal genetic and external environmental factors, accurate simulation of individual morphological–structural changes in trees is the premise of forest management and 3D simulation. However, existing studies have few descriptions, and the research on the impact of growth environments and stand spatial structures on tree morphological structure and growth is still limited. In our study, we constructed a comprehensive grade model of spatial structure (CGMSS) to comprehensively evaluate individual tree growth states of the stands and grade them from 0 to 10 correspondingly. In addition, we developed a Chinese fir morphological structure growth model based on CGMSS, and dynamically simulate the growth variations of Chinese fir stands. The results showed that the overall stand prediction accuracy of CGMSS-based Chinese fir diameter at breast height, tree height, crown width and under-living branch height growth models was more than 94%. According to the analysis of the comprehensive grade of spatial structure (CGSS) of trees in the stand, except for the prediction accuracy and systematic error of the under-living branch height growth model at the CGSS = 3–5 levels, the systematic error of the Chinese fir growth model at each level was lower than 21.2%, and the prediction accuracy was greater than 73%. Compared with the spatial structural unit (SSU)-based Chinese fir growth model proposed by Ma et al., all growth models fit better at all levels, except for the CGMSS-based Chinese fir tree height and under-living branch height growth models that fit significantly lower than the SSU-based Chinese fir growth model at CGSS = 3–5 levels. In this study, the main conclusion is that the simulation results of CGMSS’s Chinese fir morphological structure growth model are closer to the real growth state of trees, achieving accurate simulation of differential growth of trees in different growth dominance degrees and spatial structure states in forest stands, making visualized forest management more effective and realistic.

The comprehensive indexes approach based on stand structure parameters is mainly used to select trees for harvest. However, these indexes do not consider the comprehensive impact of horizontal and vertical structures, leading to an incomplete analysis of the forest structure and an inaccurate selection of trees for harvest. To solve this problem, we constructed a comprehensive decision index of logging (CDIL), integrating horizontal and vertical structure parameters which can identify harvest trees more scientifically. In this study, we took the Shanxia Forest Farm in the Jiangxi Province of China as the experimental area and used mixed broadleaf/conifer forests at different ages as our experimental sample. We selected eight horizontal and vertical spatial structure parameters to establish an efficient, objective, and accurate comprehensive decision index of logging. We combined 3D visualization technology to realize the dynamic visualization simulation of the index at different intensities of tending and felling management. The results indicated that the proposed CDIL-index could effectively optimize the forest spatial structure. From the perspective of stand structure adjustment, the optimal thinning intensity was 20%. The average CDIL in each plot decreased by more than 80% after logging, while the change range of each plot was between 30% and 70% after the F index was applied to implement tending and logging. The CDIL was 11.4% more accurate in selecting trees for harvesting than the F index. In this study, the main conclusion is that the CDIL would enable forest managers to more accurately choose trees for harvesting, leading to forest adjustment that would reduce the competition pressure among trees and improve the distribution and health of trees, possibly making the forest structure more stable.

Complex forest structures, interspecies similarities, and intraspecies variations constrain the acquisition of species-specific tree phenotypes. This study develops a scalable framework for extracting species-specific structural parameters at the individual tree level. Leveraging ultrahigh-resolution UAV-based RGB and LiDAR data, we propose a novel self-attention-guided spectral–structural multimodal fusion transformer (SAMFormer). Key components include: (1) an adaptive feature enhancement module (AFEM) that employs spatial and channel attention to selectively highlight canopy features while suppressing background noise; (2) a cross-modal fusion module (CMFM) that captures intra- and inter-modal dependencies through the cross-attention mechanism, generating highly discriminative representations. SAMFormer achieves fine-grained tree identification in complex forest environments, relieving issues of blurred canopy segmentation and species misclassification. K-fold cross-validation demonstrates robust performance across diverse scenes, achieving 86.3 % F1-score and 88.0 % mAP@0.5, significantly outperforming single-modal inputs and mainstream instance segmentation models. We generate large-scale species-specific maps of tree structural parameters based on SAMFormer outputs, allometric equations, and a sliding window strategy. Subsequently, these parameters are utilized to map carbon stock. Ecological analysis reveals a coupling relationship between tree competition and structural parameters/carbon stock: competition intensity exhibits a significant negative correlation with both (p < 0.001). Trees adapt by adjusting growth strategies (e.g., reducing radial growth and limiting canopy expansion), ultimately lowering biomass accumulation and carbon stock. Additionally, species mixing enhances carbon stock, as mixed forests store more carbon than monocultures. This work provides a high-throughput, non-destructive pathway for forest phenotyping, supporting precision forestry and climate-adaptive management practices.

Environmental factors substantially influence the growth of trees. The current studies on tree growth simulation have mainly focused on the effect of environmental factors on diameter at breast height and tree height. However, the influence of environmental factors, especially light, on canopy morphology has not been considered, hindering the accurate understanding of the range of characteristics of tree morphology that occur due to environmental changes. To solve this problem, this study investigated the influence of light on the changes in canopy morphology and constructed a coupled canopy–light model (CCLM) to visually simulate the polymorphism of fir morphology. Using the Huangfengqiao Forestry Farm in You County, Hunan Province, China, as the study area, we selected a typical sample plot. Field surveys of the fir trees in the sample plot were conducted for three consecutive years to obtain longitudinal data of fir tree canopy shape. We constructed the canopy curves using a cubic uniform B-spline to construct 3D models of the fir trees in different years. The topographic and spatial location distribution data of the fir trees were used to construct a 3D scene of the sample plot in the UE4 3D engine, and the light distribution for each part of the canopy was calculated in a 3D scene by using the annual average photosynthetically active radiation (PAR) as the light parameter, which we combined with the ray-tracing algorithm. This study constructed the CCLM from the fir diameter using the breast-height growth model (BDGM) and the height–diameter curve model (HDCM), the fir trees’ canopy shape description from two years, and the light distribution data. We compared the canopy data obtained from canopy simulations using the CCLM with those obtained using a growth model based on spatial structure (GMBOSS) and those obtained from field surveys to identify any difference in the effectiveness of the canopy simulations using the CCLM and GMBOSS. Based on the BDGM and HDCM, we constructed the CCLM of firs with a determination coefficient (R²) of 0.829, combining data on canopy shape descriptions obtained from two years of field surveys and the light distribution data of each part of the canopy obtained through the ray-tracing algorithm. The Euclidean distance between the canopy description data obtained using the CCLM and the canopy description data obtained from the field survey was 15.561; that between the GMBOSS and the field survey was 23.944. A virtual forest stand environment was constructed from the survey data, combining ray-tracing algorithms to construct the CCLM model of fir in a virtual forest stand environment for growth visualization and simulation. Compared with the canopy description data obtained using the GMBOSS, the canopy description data obtained using the CCLM better fit the canopy description data obtained from the field survey, and the Euclidean distance decreased from 23.944 to 15.561.