Research on Software and Algorithms for Displacement Monitoring System Based on Machine Vision

Shield tunnels excavated for urban subway construction experience complex defects since they are plagued by various factors such as rock properties, tunnel lining deformation, and erosion damage. In practical subway tunnel inspection, most measurements still hinge predominantly on traditional manual inspections and manual marking, supplemented by a limited number of inspection instruments, which is insufficient to meet modern testing requirements. To address the challenge of automated measurement during complex underground construction and subsequent operations, research was conducted on a multi-target displacement online monitoring system. The system design process is elaborated in detail from both software and image processing perspectives, encompassing the software design architecture, target capture program, image acquisition, and edge detection procedures. Considering potential target occlusion in actual monitoring environments, this study capitalizes on a triple-condition judgment method to determine target occlusion, thereby ensuring the system's field applicability and providing robust data support for the subsequent development of an automatic alarm system, a testament to a reduction in false alarm rates. Finally, the system was tested in an outdoor environment with multiple targets, and the results indicated a stable error of less than 0.05 mm, catering to field testing requirements. The research findings reveal the system's adaptability to the vibrating and dusty environments of underground construction sites, exhibiting high displacement measurement accuracy, real-time measurement, and a high degree of automation. It holds promising application prospects in displacement monitoring for complex underground engineering projects.