SECTION C: ENGINEERING
This research presents an algorithm for self-calibration of the extrinsic parameters of
a stereoscopic vision system by using both road infrastructure and a metaheuristic
algorithm, to estimate the pitch, roll and yaw angles, and the height. This algorithm
is constituted by three stages: The first one is a non-linear equation of the height and
angles. The second one is the extraction of visual information from the road infrastructure
captured by the stereo vision system, in this case, the lines of the road. Finally, the
implementation of a metaheuristic based on the ant colony optimization algorithm
is carried out to solve the mathematical model which represents the world camera. Experimental results show the correct estimation of the four parameters, obtaining an
overall error of 8% and 9% in scenarios of laboratory and real road, respectively.