Development of a low-cost microscopic vision-based real-time strain sensor using Raspberry Pi

Abstract

Strain is one of the key indicators for structural health monitoring. In this study, we developed a low-cost microscopic vision-based real-time strain sensor using Raspberry Pi (called MISS-Dym). By strategies for image processing accelerated and the specific running logic, the strain can be outputted at a frequency of more than 30 Hz in real time. The MISS-Dym integrates multiple functions including real-time strain calculations, temperature compensation, data storage, and wireless transmission. Comparative experiments were performed with fiber Bragg grating to assess the accuracy of the sensor. In the static experiments, the maximum mean squared error was 1.77 µε, while the maximum relative error was 5.5% in the dynamic experiments. Additionally, a 10-day monitoring was conducted by MISS-Dym. The results show that the sensor can effectively capture both the vehicle-induced and the temperature-induced strain of the concrete bridge. The MISS-Dym provides an efficient and low-cost method for monitoring the dynamic strain responses of concrete structures.