Research On Technique Of Camera Calibration

Camera calibration is the first step of computer vision and one of the most active research fields nowdays. With an accurate calibrated camera we can carry out three-dimension reconstruction, motion analysis and precise measurement.In this paper, we first introduce the relative 'knowledge and technique about the theory of camera calibration. Compared with the classic camera calibration by calibrated objects, we present a new algorithm using one dimension object, which makes the method of calibration based on calibrated object more perfect and have solved problems in distributed camera system. In the case of self calibration, we modify the algorithm to meet the requirement of certain applications. They are self-calibration algorithm in case of invariant focal length, and new algorithm with provided motion parameters. The results of simulation and experiment proved that all algorithms are robust and applicable. The whole paper can be divided into the following:In the first chapter.we give the main content of camera calibration and introduce the key method presented in this paper.In the second chapter, a new flexible technique of camera calibration is developed using one-dimension objects, in which the configuration of camera is proposed in detail.In the third chapter, we present a review and classification of camera self-calibration techniques developed in recent years. After introducing the background of all methods in detail, we propose the comment on those calibration techniques.In the fourth chapter, two novel self-calibration algorithm are proposed according to certain application envirenment.Finally the previous work is summerized, and the prospect of camera calibration is laid out in this chapter.