Vision-based structural displacement measurement: System performance evaluation and influence factor analysis

• A vision-based method is proposed for noncontact, multi-point and real-time structural displacement measurement.
• The system performance of the developed vision-based system is evaluated through different shaking table tests.
• The main influencing factors on the measurement accuracy of the developed vision-based system are analyzed.
• Experimental study shows that the illumination and vapor may significantly affect the test results.

In the past decade, the emerging machine vision-based measurement technology has gained great concerns among civil engineers due to its overwhelming merits of non-contact, long-distance, and high-resolution. A critical issue regarding to the measurement performance and accuracy of the vision-based system is how to identify and eliminate the systematic and unsystematic error sources. In this paper, a vision-based structural displacement measurement system integrated with a digital image processing approach is developed. The performance of the developed vision-based system is evaluated by comparing the results simultaneously obtained by the vision-based system and those measured by the magnetostrictive displacement sensor (MDS). A series of experiments are conducted on a shaking table to examine the influence factors which will affect the accuracy and stability of the vision-based system. It is demonstrated that illumination and vapor have a critical effect on the measurement results of the vision-based system.