Camera Calibration Method For Machine Vision

Camera calibration is the process that using the corresponding relations between theobject surface feature point space coordinates and image pixel coordinates the related tocalculate the parameters of camera system according to the mathe matical model of imagingsystem. The accuracy of the calibration results and the stability of the algorithm influencethe accuracy of the visual system directly. Therefore, the work of research cameracalibration method in machine vision is of important significance to theoretical andpractical.Feature points extraction based on image is the foundation for camera calibration, thepoint coordinates have great influence on accuracy to calibration, So this article putforward can realize the pixel level two feature points extraction algorithm on the basis ofthe principle of camera calibration and the basic method and double camera systemreviewed:(1)The mathematical combination of Harris and Forstner algorithm can makeextraction accuracy to the sub-pixel level;(2)The fitting method of optimization method isapplied to Harris algorithm to make the feature point coordinates extraction to thesub-pixel level.Tasi two-step method of camera calibration method based on the plane template canmeet the precision need mostly situation, is a more mature calibration algorithm, but thealgorithm need to assume the main point coordinates, and that increases the instability o fthe algorithm. This paper puts forward a kind of method that can find out main pointcoordinates only using two pictures to avoid the instability of the original Tasi two-stephypothesis main point coordinates.In this paper we finally realized the entire process of the camera calibration and thedual camera system simulation on the experimental platform of MATLAB, achieves thedesired results. The process is mainly divided into three parts to complete, namely imagepreprocessing, pixel feature point extraction, single camera system inside and outsideparameter calibration, double camera system parameter calibration.