Monocular vision-based displacement measurement system robust to angle and distance using homography

Jun Hwa Lee, Soojin Cho, Sung Han Sim

Research output: Contribution to journalConference articlepeer-review

7 Scopus citations

Abstract

Displacement of a civil structure contains important information about structural behavior that helps confirm safety and serviceability. Displacement measurement by conventional devices such as LVDT is still challenging due to issues in the cost-effectiveness and accessibility to measurement locations. This paper presents a monocular vision-based displacement measurement system (VDMS) that is cost-effective and convenient installation. The proposed system is composed of a low-cost aftermarket camera, a specially-designed marker, and a laptop computer. The motion of the marker attached to the structure is captured by the video camera, that is subsequently converted to the time history of a dynamic displacement response using planar homography. Planar homography is a transformation matrix mapping any two planes in the three dimensional space, and it can map an image plane into the real marker's plane which may be at an angle with respect to the camera axis. The system was validated by a lab-scale test using a shaking table. Experimental results show proposed system has a good agreement with LDV regardless of camera positions.

Keywords

  • Camera
  • Displacement measurement
  • Marker
  • Planar homography
  • Vision

Fingerprint

Dive into the research topics of 'Monocular vision-based displacement measurement system robust to angle and distance using homography'. Together they form a unique fingerprint.

Cite this