Self-calibrating technique for 3D displacement measurement using monocular vision and planar marker

Abstract

Although the vision-based displacement measurement of structures has been widely studied for decades, it still poses challenges for practical applications, including the issues of calibration and automatic long-term monitoring. This paper introduces a monocular-vision based self-calibrating technique for accurately measuring the 3D structural displacement. The technique utilizes a tailor-made planar marker with unique graph patterns, enabling precise image recognition and PnP resolution. The proposed technique automatically estimates the 3D position and 3D orientation of the marker. The limitations of marker-free techniques in terms of long-term monitoring have also been discussed. The accuracy and robustness of the proposed technique are systematically examined by Monte-Carlo simulation, and then verified by experiments. Results show that accuracy of the proposed technique for 3D displacement measurement is 0.049 mm, which indicates the effectiveness of the technique for automatically measuring and calibrating stereo displacement of structures using monocular vision.