11 November 2010 Calibration target reconstruction for 3-D vision inspection system of large-scale engineering objects
Author Affiliations +
Abstract
It is usually difficult to calibrate the 3-D vision inspection system that may be employed to measure the large-scale engineering objects. One of the challenges is how to in-situ build-up a large and precise calibration target. In this paper, we present a calibration target reconstruction strategy to solve such a problem. First, we choose one of the engineering objects to be inspected as a calibration target, on which we paste coded marks on the object surface. Next, we locate and decode marks to get homologous points. From multiple camera images, the fundamental matrix between adjacent images can be estimated, and then the essential matrix can be derived with priori known camera intrinsic parameters and decomposed to obtain camera extrinsic parameters. Finally, we are able to obtain the initial 3D coordinates with binocular stereo vision reconstruction, and then optimize them with the bundle adjustment by considering the lens distortions, leading to a high-precision calibration target. This reconstruction strategy has been applied to the inspection of an industrial project, from which the proposed method is successfully validated.
© (2010) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Yongkai Yin, Xiang Peng, Yingjian Guan, Xiaoli Liu, Ameng Li, "Calibration target reconstruction for 3-D vision inspection system of large-scale engineering objects", Proc. SPIE 7855, Optical Metrology and Inspection for Industrial Applications, 78550V (11 November 2010); doi: 10.1117/12.870287; https://doi.org/10.1117/12.870287
PROCEEDINGS
6 PAGES


SHARE
Back to Top