Multivision Sensor Extrinsic Calibration Method With Non-Overlapping Fields of View Using Encoded 1D Target

Abstract

Multiple vision sensors (multi-VSs) are difficult to calibrate in large fields of view vision measurement owing to the lack of overlapping areas. Thus, this study proposes a novel method for the external parameter calibration of multi-VSs by using an encoded 1D target. The quick response code is used in this study to uniquely identify the feature circle. The Hough transform method and ellipse center solving method are jointly applied to optimize the feature points and further improve the collinearity of the feature points. As the cross-ratio of the collinearity feature points involves a projection invariance, the rotation and translation vectors between the multi-VSs are solved by the inverse process of projection and rigid transformation. The accuracy and advantages of the proposed method are verified by simulation and real experiments. The proposed target has a simple structure, the operation is convenient and flexible, and the method can be applied to the global calibration of multi-VSs with non-overlapping fields of view.