Deep learning techniques have been widely applied to classify tree species and segment tree structures. However, most recent studies have focused on the canopy and trunk segmentation, neglecting the branch segmentation. In this study, we proposed a new approach involving the use of the PointNet++ model for segmenting the canopy, trunk, and branches of trees. We introduced a preprocessing method for training LiDAR point cloud data specific to trees and identified an optimal learning environment for the PointNet++ model. We created two learning environments with varying numbers of representative points (between 2048 and 8192) for the PointNet++ model. To validate the performance of our approach, we empirically evaluated the model using LiDAR point cloud data obtained from 435 tree samples scanned by terrestrial LiDAR. These tree samples comprised Korean red pine, Korean pine, and Japanese larch species. When segmenting the canopy, trunk, and branches using the PointNet++ model, we found that resampling 25,000–30,000 points was suitable. The best performance was achieved when the number of representative points was set to 4096.

In recent years, light detection and ranging (LiDAR) has been increasingly utilized to estimate forest resources. This study was conducted to identify the applicability of a LiDAR sensor for such estimations by comparing data on a tree’s position, height, and diameter at breast height (DBH) obtained using the sensor with those by existing forest inventory methods for a Cryptomeria japonica forest in Jeju Island, South Korea. For this purpose, a backpack personal laser scanning device (BPLS, Greenvalley International, Model D50) was employed in a protected forest, where cutting is not allowed, as a non-invasive means, simultaneously assessing the device’s field applicability. The data collected by the sensor were divided into seven different pathway variations, or “patterns” to consider the density of the sample plots and enhance the efficiency. The accuracy of estimating the variables of each tree was then assessed. The time spent acquiring and processing real-time data was also analyzed for each method, as well as total time and the time required for each measurement. The findings showed that the rate of detection of standing trees by LiDAR was 100%. Additionally, a high statistical accuracy was observed in pattern 5 (DBH: RMSE 1.22 cm, bias—0.90 cm, Height: RMSE 1.66 m, bias—1.18 m) and pattern 7 (DBH: RMSE 1.22 cm, bias—0.92 cm, Height: RMSE 1.48 m, bias—1.23 m) compared to the results from the typical inventory method. A range of 115–162.5 min/ha was required to process the data using the LiDAR, while 322.5–567.5 min was required for the typical inventory method. Thus, the application of a backpack personal LiDAR can lead to higher efficiency when conducting a forest resource inventory in a coniferous plantation with understory vegetation. Further research in various stands is necessary to confirm the efficiency of using backpack personal laser scanning.

In Republic of Korea, the digital transformation of forest data has emerged as a critical priority at the governmental level. To support this effort, numerous case studies have been conducted to collect and analyze forest data. This study evaluated the accuracy of forest resource assessment methods using terrestrial laser scanning (TLS) and backpack personal laser scanning (BPLS) under Leaf-on and Leaf-off conditions in the Gari Mountain Forest Management Complex, Hongcheon, Republic of Korea. The research was conducted across six sample plots representing low, medium, and high stand densities, dominated by Larix kaempferi and Pinus koraiensis. Conventional field survey methods and LiDAR technologies were used to compare key forest attributes such as tree height and volume. The results revealed that Leaf-off LiDAR data exhibited higher accuracy in capturing tree height and canopy structures, particularly in high-density plots. In contrast, during the Leaf-on season, measurements of understory vegetation and lower canopy were hindered by foliage obstruction, reducing precision. Seasonal differences significantly impacted LiDAR measurement accuracy, with Leaf-off data providing a clearer and more reliable representation of forest structures. This study underscores the necessity of considering seasonal conditions to improve the accuracy of LiDAR-derived metrics. It offers valuable insights for enhancing forest inventory practices and advancing the application of remote sensing technologies in forest management.

New technologies are utilized to improve forest management, but they require the collection and analysis of digital data. This study assessed the applicability of using light detection and ranging (LiDAR) devices for the examination of forest resource inventories to obtain digital forest resource information. Two terrestrial LiDARs, a backpack laser scanner (BPLS) and a terrestrial laser scanner (TLS) were used and compared with the traditional method to identify which was optimal. The findings showed that the TLS single-scan method was least accurate due to occlusion. The TLS multi-scan method and the BPLS showed high levels of accuracy for the height and diameter at breast height (DBH) estimates in most sample plots. However, the BPLS underestimated height to a greater extent than the other methods in a sloped plot (24°). Nevertheless, the efficiency of the BPLS was 2.8 times higher than that of the TLS when considering the amount of time spent on completing all indoor and outdoor tasks. Thus, these results prove that the utilization of LiDAR devices increases the efficiency of data collection and overcomes the limitations of existing methods. Moreover, they provide accurate information that will be a basis for forest management plans, the estimation of biomass, and the transaction of forest products.

Accurate estimation of stem profiles is critical for forest management, timber yield prediction, and ecological modeling. However, traditional taper equations often fail to capture species-specific growth variability and exhibit significant biases, particularly in the upper stem regions. Machine learning regression models were applied to estimate Quercus mongolica stem profiles across South Korea, and performance was compared with that of a traditional taper equation. A total of 2503 sample trees were used to train and validate Random Forest (RF), XGBoost (XGB), Artificial Neural Network (ANN), and Support Vector Regression (SVR) models. Predictive performance was evaluated using root mean square error, mean absolute error, and coefficient of determination metrics, and performance differences were validated statistically. The ANN model exhibited the highest predictive accuracy and stability across all diameter classes, maintaining smooth and consistent stem profiles even in the upper stem regions where the traditional taper model exhibited significant errors. RF and XGB models had moderate performance but exhibited localized fluctuations, whereas the Kozak taper equation tended to overestimate basal diameters and underestimate crown-top diameters. Machine learning models, particularly ANN, offer a robust alternative to fixed-form taper equations, contributing substantially to forest resource inventory, carbon stock assessment, and climate-adaptive forest management planning.