Research And Implementation Of The Camera Calibration Methods

Human apperceives and understands the outer world mainly by vision. Now with the development of computer technology and the continual increase of the demanding of people on digital information, both computer vision and stereo vision have been developing greatly. As an important part of stereo vision, and a premise and base for computer vision obtaining the spatial information of 3D objects, research on the camera calibration methods has great important significance of theoretical study and practical value.In this thesis, we research the theory and methods on the camera calibration, and gain the parameters about the camera to construct the relation between the 3D object and the 2D image, and provide the reliable data and lay a good foundation for the next research on computer vision and stereo vision.In this thesis, we first introduce the foundational knowledge of camera calibration in detail, and then study some existing calibration algorithms in classic camera calibration and self-calibration. For the classic camera calibration, we present a new algorithm using a coplanar target which makes the camera calibration method based on calibrated object simpler, more flexible and accurate. This method only requires a coplanar target and without camera motion. When we require low accuracy, we can neglect the lens distortion and linearly solve all the parameters. When high accuracy is required, we can include Weng's lens distortion model and solve the distortion coefficient by nonlinear algorithm. The experiments of synthetic images and real images prove the proposed method is correct and effective. By applying the proposed algorithm to the vision inspection, we can see the practicability of the algorithm and the necessity of the camera calibration. Finally, the program of the proposed method is given by the combination of VC++and Matlab softwares.There are many camera self-calibration methods based on the Kruppa equation in the literature. We present a new improved self-calibration method based on the Kruppa equation in this thesis. In this improved method, we can not eliminate the unknown scale factors, but approximately estimate the unknown scale factors in Kruppa equation frist, and then optimize them by paticle swarm optimization algorithm, and linearly complete camera calibration finally. The results of the simulation experiment show that the improved method is quite effective and robust.