MS-DGCNN++: A Multi-Scale Fusion Dynamic Graph Neural Network with Biological Knowledge Integration for LiDAR Tree Species Classification

2025-07-16

AI Summary

To address insufficient semantic hierarchical modeling for tree species classification in terrestrial LiDAR point clouds under complex, multi-scale forest structures, this paper proposes a Hierarchical Multi-Scale Dynamic Graph Convolutional Neural Network (H-MSDGCNN). Unlike conventional parallel multi-scale approaches, H-MSDGCNN is biologically inspired by tree morphology and introduces scale-specific feature engineering coupled with hierarchical graph downsampling, enabling semantic alignment and cross-scale information propagation across local, branch-level, and canopy-level scales. It integrates geometric features, normalized relative vectors, and distance-aware representations to enhance structural awareness. On the STPCTLS dataset, H-MSDGCNN achieves 94.96% accuracy—significantly outperforming DGCNN, MS-DGCNN, and PPT. On FOR-species20K, it attains 67.25% accuracy, a 6.1-percentage-point improvement. Moreover, it maintains competitive performance on ModelNet40/10 while using fewer parameters, demonstrating suitability for edge deployment.

Technology Category

Machine Learning: Multi-class/Multi-label Learning & Extreme ClassificationComputer Vision: Multi-modal VisionNatural Language Processing: Language Grounding & Multi-modal NLP

Application Category

Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphsSemantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMsSearch and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

Tree species classification from terrestrial LiDAR point clouds is challenging because of the complex multi-scale geometric structures in forest environments. Existing approaches using multi-scale dynamic graph convolutional neural networks (MS-DGCNN) employ parallel multi-scale processing, which fails to capture the semantic relationships between the hierarchical levels of the tree architecture. We present MS-DGCNN++, a hierarchical multiscale fusion dynamic graph convolutional network that uses semantically meaningful feature extraction at local, branch, and canopy scales with cross-scale information propagation. Our method employs scale-specific feature engineering, including standard geometric features for the local scale, normalized relative vectors for the branch scale, and distance information for the canopy scale. This hierarchical approach replaces uniform parallel processing with semantically differentiated representations that are aligned with the natural tree structure. Under the same proposed tree species data augmentation strategy for all experiments, MS-DGCNN++ achieved an accuracy of 94.96 % on STPCTLS, outperforming DGCNN, MS-DGCNN, and the state-of-the-art model PPT. On FOR-species20K, it achieves 67.25% accuracy (6.1% improvement compared to MS-DGCNN). For standard 3D object recognition, our method outperformed DGCNN and MS-DGCNN with overall accuracies of 93.15% on ModelNet40 and 94.05% on ModelNet10. With lower parameters and reduced complexity compared to state-of-the-art transformer approaches, our method is suitable for resource-constrained applications while maintaining a competitive accuracy. Beyond tree classification, the method generalizes to standard 3D object recognition, establishing it as a versatile solution for diverse point cloud processing applications. The implementation code is publicly available at https://github.com/said-ohamouddou/MS-DGCNN2.

Problem

• Classify tree species from LiDAR point clouds
• Improve multi-scale feature fusion in tree structures
• Enhance 3D object recognition with reduced complexity

Innovation

• Hierarchical multiscale fusion dynamic graph network
• Semantic feature extraction at multiple scales
• Scale-specific feature engineering for trees