Deformation monitoring of ancient buildings based on computer vision method

To monitor the global displacement and local deformation of ancient buildings, a deformation monitoring system of ancient buildings based on computer vision was proposed. Considering the ancient building deformation monitoring have several characteristics such as small amount and slow change of deformation as well as limited camera layout in the monitoring process, ZHANG Zhengyou’ s camera calibration method was introduced to eliminate the limitation that the optical axis of the camera needs to be vertical, realizes the transfer of arbitrary tilted images to the 3D coordinate system for analysis, and improves the monitoring accuracy by using sub-pixel interpolation technology, and the normalized accuracy of the test reaches 99. 5% . To verify the feasibility of the system for monitoring the deformation of ancient structures, a shaking table test of an ancient building model and a static monitoring test of the simplified model of Xianfu Palace well pavilion in Beijing Palace Museum were monitored using the proposed approach. The results show that the deformation monitoring system of ancient buildings based on computer vision method provides the technical possibility of non-destructive monitoring for the monitoring of structural deformation of ancient buildings under long-term load and accidental load, and supports the deformation monitoring of ancient building structures with and without targets. The error of deformation monitoring of structures with and without targets under accidental load is less than 8% . Under long-term load, the target structure deformation monitoring error is less than 4% . © 2023 Science Press. All rights reserved.