Journals Proceedings

Abstract

Camera calibration is one of the most important steps, recognized as a bottleneck, in 3D scanning devices that a very small error in the camera parameters can lead to imprecise 3D models. In the last two decades, different approaches have been taken and improved the precision of the calibration parameters; but still the desire for better precisions and more stability exists. In this work, we have used an easy-to-implement and cheap checkerboard as the calibration object. The initial data for calibration is obtained with only three images captured from the checkerboard. To optimize the non-linear camera equations, both numerical and Particle Swarm Optimization (PSO) algorithms together have been applied in order to increase the precision and reduce the sensitivity of the algorithm to noise and errors. Capability of combined numerical and PSO-based algorithm for camera calibration especially lens distortion is demonstrated using a criterion called re-projection error. This work is distinguished from the former works in the application of combined numerical and PSO-based optimizations together in camera calibration focusing on lens distortion. This paper shows that solving camera calibration equations using numerical and PSO algorithms in an integrated manner can lead to acceptable results.