Unsupervised model-free camera calibration algorithm for robotic applications

摘要

This paper presents an algorithm for camera calibration. The algorithm, inspired by work in the field of developmental robotic on the concept of space in naive agents, is particularly suitable for robotic applications: it is completely unsupervised, and it does not assume any model of the camera, making it applicable to many kinds of optical devices. Testing of the algorithm, in a simulated environment, shows very good results, outperforming the main unsupervised and model-free calibration algorithm in the literature.