UAV-Based Detection of Rock Mass Instabilities Using DBSCAN and 3D Region Growing Algorithms

This study presents an unmanned aerial vehicle (UAV) photogrammetry system that integrates DBSCAN clustering and 3D region growing algorithms to detect unstable rocks on high-angle slopes (50°–80°), addressing critical safety challenges in geotechnical monitoring. The system was applied at China’s Y...

Full description

Saved in:
Bibliographic Details
Published in:Geotechnical and geological engineering Vol. 43; no. 6; p. 295
Main Authors: Chen, Na, Hu, Min, Xiao, Henglin, Li, Yahu, Hou, Zeng
Format: Journal Article
Language:English
Published: Cham Springer International Publishing 01.08.2025
Springer Nature B.V
Subjects:
Accuracy
Aerial photography
Algorithms
Alpha shapes
Automation
Blasting
Civil Engineering
Clustering
Data collection
Data processing
Deep learning
Earth and Environmental Science
Earth Sciences
Efficiency
Geology
Geotechnical Engineering & Applied Earth Sciences
Hydroelectric power
Hydroelectric power stations
Hydrogeology
Lithology
Measurement techniques
Monitoring
Original Paper
Photogrammetry
Quantitative analysis
Risk reduction
Rock
Rock masses
Rocks
Safety factors
Simulation
Slope stability
Terrestrial Pollution
Unmanned aerial vehicles
Waste Management/Waste Technology
Intelligent identification
Point cloud
Unmanned aerial vehicle
3D-region growing algorithm
Unstable rock masses
ISSN:0960-3182, 1573-1529
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This study presents an unmanned aerial vehicle (UAV) photogrammetry system that integrates DBSCAN clustering and 3D region growing algorithms to detect unstable rocks on high-angle slopes (50°–80°), addressing critical safety challenges in geotechnical monitoring. The system was applied at China’s Yulong Kashi Hydropower Station, where it combined Alpha-shape algorithms with polyhedral projection integration to quantitatively analyze 3D rock mass characteristics, including backwall inclination (28.54°–55.2°), area (96.01–17,892.27 m 2 ), volume (320.49–174,964.55 m 3 ), and elevation difference (3.71–27.88 m). Stability assessments based on safety factor (K-value) and volumetric classification identified one relatively stable rock mass (LA6) and nine unstable ones, with LB1 being the most severe due to its exceptionally large volume (174,964.55 m 3 ) requiring urgent controlled blasting. The system demonstrated significant efficiency improvements, achieving a 71.8% reduction in processing time compared to manual methods (2.8–5.0 s vs. 8.9–20.2 s per case) and 83% lower time variability, particularly excelling in complex scenarios (e.g., saving 16.9 s for LA14). By combining advanced algorithms with 3D analytics, this approach provides a robust, rapid tool for risk mitigation on steep slopes, offering a new standard for geotechnical monitoring in challenging environments.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0960-3182
1573-1529
DOI:10.1007/s10706-025-03226-8