Real-Time Intelligent Monitoring of Rockfall in the Complex Environment

Rockfall is a global geological hazard characterized by wide distribution, strong randomness, and inapparent nature. The difficulties associated with deploying contact monitoring methods, recovering equipment, and high costs have led to the proposal of a non-contact monitoring method based on computer vision. This research study focuses on optimizing this method to improve its efficiency and effectiveness. Initially, image preprocessing techniques, such as gray transform, image enhancement, and image smoothing, are employed to facilitate rapid and accurate feature extraction. These optimizations aim to enhance the quality of the images obtained for further analysis. Next, the feasibility and generalization ability of the optical flow method are verified through dynamic target monitoring experiment. Finally, the software is developed to visualize the motion state and trajectory of rockfall. The experimental results show that: (1) By leveraging image processing and recognition technology, the optical flow method can enables remote and non-destructive monitoring of rockfall; (2) This technology is capable of capturing the motion characteristics of rockfall in various settings, and the algorithm’s processing efficiency allows for real-time detection; (3) Additionally, software for dynamic target detection can be developed and designed to provide a clear and intuitive display of the detection results and the movement trajectory of rockfall.