A novel calibration method based on fixed multi-plane targets for cameras with large field of view

Abstract

For camera calibration with large field of view (FOV), the accuracy is impacted by the size of the calibration targets. Considering the challenges associated with producing and carrying large targets, existing methods resort to using small targets to achieve calibration without high accuracy. In order to solve this problem, a photogrammetric calibration method based on the fixed multi-plane targets (FMPT) is proposed in this paper for cameras with large FOV. The FMPT is composed by multiple identical small planar targets (PTs), with fixed pose transformation relationships among PTs. The proposed calibration method involves the following main steps. Firstly, the camera is moved several times to capture a series of images of FMPT. Secondly, the fixed pose transformation matrices of different small PTs in FMPT are calculated through the initial parameters of camera. Finally, calibration based on photogrammetry is conducted, with a continuous optimization of camera intrinsic and extrinsic parameters based on the multiple 2D and 3D constraint in the FMPT. Simulation and real data experiments show that the calibration accuracy using the proposed method is much higher than that using a small target, and sightly higher than that using a large target. Furthermore, the experiments demonstrate the robust stability of this method, maintaining high calibration accuracy even in the presence of increased noise and error in target production.