A low-cost image-based monitoring system for automatic rock displacement measurement using YOLO

Accurate monitoring of rock movement is essential for understanding rock slope instability and early failure mechanisms. However, conventional monitoring techniques often rely on manual measurements or expensive instrumentation, limiting their practicality for continuous and large-scale deployment. This study presents a low-cost image-based monitoring system that enables fully automatic measurement of rock displacement using a YOLOv8 pose estimation model. With a total cost of approximately EUR 200, the system integrates a mechanical crackmeter, a low-cost digital wildlife camera, and an AI-based image-processing pipeline. By detecting the dial indicator and estimating pointer positions from images, the system enables automatic reading of rock movements from the crackmeter, which has a resolution of 0.01 mm. The proposed system was evaluated at the Miroir d’Argentine rock wall, a limestone rock flake in the Swiss Alps. To further assess its practical applicability, additional images acquired at different time periods were used for independent validation. The results demonstrate that the proposed approach can reliably measure rock movement with high accuracy and strong generalization capability under real-world conditions. In addition, the robustness of the method was evaluated under simulated fog and blur degradations. The results show that the system maintains stable performance under light to moderate visual degradation, while performance decreases under severe fog and strong motion blur due to reduced geometric visibility. Perspective-related distortion was quantitatively evaluated and found to be limited across the dataset, while the potential thermal deformation remained small relative to the observed displacement variations. Overall, the proposed method provides an effective, low-cost, and practical solution for continuous rock movement monitoring and shows strong potential for long-term deployment in challenging alpine environments.