Uncalibrated Stereo Image Reconstruction Based on RANSAC Algorithm

Abstract

Nowadays, Computer vision is a very coveted discipline by many engineers. It plays an important role in the current applications in the fields of medicine, education, robotics, weapons, even in the smallest home appliances etc. The main and essential element of these applications is the camera which by dint of its affordable price, its ease to be embedded and its lower consumption of energy makes it possible to work wonderfully in computer vision systems. However, setting up a camera requires an engineer who must know most of the parameters that highlight a camera. Former is not easy especially when one don't have any information about of the internal parameters of camera used to take images. In this paper, the use of a camera which an internal parameters are unknown has been investigated to reconstruct an uncalibrated stereo image.