Performance Study of the Genetic Algorithm and the Levenverg-Marquardt Optimization Algorithm for the Calibration of a Rotating Camera

摘要

Calibrating the camera is one of the fundamental problems of computer vision. It aims to determine the parameters of the intrinsic and extrinsic camera. During the camera calibration process, the initial camera settings are acquired by solving linear equations. Then these parameters are refined by a nonlinear optimization procedure. In this paper, a comparative study between Levenberg Marquardt and the genetic algorithm for the optimization of nonlinear system parameters is presented. Our main objective is to show the merits of the optimization of Genetic Algorithms and to determine its suitability in the field of camera calibration. Experimental results demonstrate that the proposed method is effective.