A calibration method for binocular stereo vision sensor with short-baseline based on 3D flexible control field. (January 2020)

• Record Type: Journal Article
• Authors: Yang, Shoubo; Gao, Yang; Liu, Zhen; Zhang, Guangjun
• Is Part Of: Optics and lasers in engineering. Volume 124(2020)
• Publication Date: 2020-01
• Subjects: Binocular stereo vision -- Camera calibration -- Short baseline -- Large field-of-view -- 3D Flexible control field; Lasers in engineering -- Periodicals; Optical measurements -- Periodicals; Optics -- Periodicals; Lasers en ingénierie -- Périodiques; Mesures optiques -- Périodiques; Optique -- Périodiques; 621.36605
• DOI: 10.1016/j.optlaseng.2019.105817
• Languages: English
• ISSNs: 0143-8166

Abstract

Highlights: A novel calibration method based on a 3D flexible control field is proposed. This method requires a narrow calibration space and adapts to the varied FOV. It is simple to operate and the workload remains unchanged with FOV changing. Abstract: Binocular stereo vision sensor (BSVS) based on disparity is an important technique in 3D shape measurement. In the application of large field-of-view (FOV), a relative short baseline distance of BSVS has heightened the demand for calibration. In this paper, a calibration method based on a virtual 3D flexible control field (FCF) for BSVS with short-baseline is proposed. The FCF is constructed by a laser tracker, a precision two-axis turntable and a camera calibration reference target (CRT) which can be exchanged with a standard laser tracker target precisely. Firstly, the turntable coordinate system is built and the position of the fixed CRT is measured by the laser tracker. Then the turntable with BSVS on it rotates in two dimensions regularly, during which the cameras shoot the fixed CRT synchronously, and the 3D FCF is constructed. Finally, the minimum objective function is established according to the BSVS imaging model to realize the calibration of internal and external parameters integrally. The experimental results show that under the conditions that the baseline is 120 mm and the working distance is 3 m ∼ 5 m, the residual error RMS of spatial points is 0.3654 mm and the reconstructed distance RMS error is 1.0728 mm .