The Effect of Camera Calibration Space on Visual Pose's Precision

Abstract

In this paper, the relationship between the camera calibration space and measurement error is analyzed in order to enhance the precision of a model based monocular vision pose estimation system. We proved that the calibration error of camera intrinsic parameters can be reduced when the calibration space is designed in a full field of view no matter how small the imaging range of the measuring area is, thus obtaining better precision of pose estimation. This conclusion provides important guidance for engineering application of the visual pose measurement system.