CORTEXA
← Browse
crossrefForests2023-10-31Cited by 2

PointDMS: An Improved Deep Learning Neural Network via Multi-Feature Aggregation for Large-Scale Point Cloud Segmentation in Smart Applications of Urban Forestry Management

Jiang Li, Jinhao Liu

Background: The development of laser measurement techniques is of great significance in forestry monitoring and park management in smart cities. It provides many conveniences for improving landscape planning efficiency and strengthening digital construction. However, capturing 3D point clouds in large-scale landscape environments is a complex task that generates massive amounts of unstructured data with characteristics such as randomness, rotational invariance, sparsity, and serious barriers. Methods: To improve the processing efficiency of intelligent devices for massive point clouds, we propose a novel deep learning neural network based on a multi-feature aggregation strategy. This network is designed to divide 3D laser point clouds in complex large-scale scenarios. Firstly, we utilize multiple terrestrial laser sensors to collect a large amount of data in open scenes such as parks, streets, and forests in urban environments. These data are integrated into a practical database called DMSdataset, which contains different information variables, densities, and dimensions. Then, an automatic block integrated with a multi-feature extractor is constructed to pre-process the unstructured point cloud data and standardize the datasets. Finally, a novel semantic segmentation framework called PointDMS is designed using 3D convolutional deep networks. PointDMS achieves a better segmentation performance of point clouds with a lightweight parameter structure. Here, “D” stands for deep network, “M” stands for multi-feature, and “S” stands for segmentation. Results: Extensive experiments on self-built datasets show that the proposed PointDMS achieves similar or better performance in point cloud segmentation compared to other methods. The overall identification accuracy of the proposed model is up to 93.5%, which is a 14% increase. Particularly for living wood objects, the average identification accuracy is up to 88.7%, which is, at least, an 8.2% increase. These results effectively prove that PointDMS is beneficial for 3D point cloud processing, division, and mining applications in urban forest environments. It demonstrates good robustness and generalization.

View free PDFSource page

Related papers

crossrefForests2026-07-24

Estimating Individual Leaf Area of Dendrocalamus brandisii: Allometric vs. Regression vs. Machine Learning

Bin Wu, John L. Innes, Shixin Deng, Renyi Gui

Accurate estimation of individual (culm-level) leaf area (LA) is essential to quantifying canopy structure, productivity, and carbon dynamics in bamboo plantations, yet it is often constrained by limited field data and high measurement costs. Using destructive sampling data from…

View free PDFSource page
crossrefForests2026-02-21Cited by 4

Spatial Prediction of Forest Fire Occurrence Integrating Human Proximity: A Machine Learning Approach for Korea’s Eastern Coast

Jeman Lee, Sujung Ahn, Sangjun Im

Forest fire occurrence prediction remains challenging despite advances in operational fire danger rating systems. In South Korea, the Korea Forest Fire Danger Rating Index (KFDRI) incorporates meteorological conditions, terrain (elevation, aspect), and forest type to assess regio…

View free PDFSource page
crossrefForests2026-02-08

Terrain–Climate–Human Couplings of Net Primary Productivity in the Chengdu–Chongqing Economic Circle Revealed by Optimal GeoDetector and Explainable Machine Learning

Sijie Zhuo, Bin Yang, Pan Jiang, Yingchao Sha, Yuxi Wang, Xinchen Gu, et al.

Terrestrial net primary productivity (NPP) integrates vegetation responses to climate, terrain, and human activities, yet their combined effects in mountainous–basin regions remain unclear. Focusing on the Chengdu–Chongqing Economic Circle (CCEC) in southwest China, we build a fr…

View free PDFSource page
crossrefForests2026-01-14Cited by 1

Computer Vision, Machine Learning, and Deep Learning for Wood and Timber Products: A Scopus-Based Bibliometric and Systematic Mapping Review (1983–2026, Early Access)

Gianmarco Goycochea Casas, Zool Hilmi Ismail, Helio Garcia Leite

This systematic mapping review and bibliometric analysis examines Scopus-indexed research on computer vision, image processing, and deep learning applied to wood and timber materials and products. A rule-based Scopus search (TITLE-ABS-KEY, 9 December 2025), combining wood and tim…

View free PDFSource page
crossrefForests2025-07-17

Analyzing Forest Leisure and Recreation Consumption Patterns Using Deep and Machine Learning

Jeongjae Kim, Jinhae Chae, Seonghak Kim

Globally, forest leisure and recreation (FLR) activities are widely recognized not only for their environmental and social benefits but also for their economic contributions. To better understand these economic contributions, it is vital to examine how the regional economic level…

View free PDFSource page
crossrefForests2025-06-19Cited by 2

A Study on Airborne Hyperspectral Tree Species Classification Based on the Synergistic Integration of Machine Learning and Deep Learning

Dabing Yang, Jinxiu Song, Chaohua Huang, Fengxin Yang, Yiming Han, Ruirui Wang

Against the backdrop of global climate change and increasing ecological pressure, the refined monitoring of forest resources and accurate tree species identification have become essential tasks for sustainable forest management. Hyperspectral remote sensing, with its high spectra…

View free PDFSource page