Research On Calibration Technology Of Zoom Camera

The imaging mechanism of zoom lens is to adjust the viewing angle width,image resolution and scene size,and to achieve auto-focus,zoom and other functions,through continuously adjusting the focal length of the optical imaging system in a certain range.At present,zoom imaging technology is widely used in photography,camera,depth reconstruction,security monitoring,visual tracking,industrial detection and many other fields.The main contents of this paper are as follows :(1)In this paper,the principle of imaging and the zoom camera calibration method is discussed.What Two kinds of distortion model(multiplication and division model model)have impact on the calibration results is analyzed by experiment;and summarizing the evaluation criteria for camera calibration.(2)The calibration methods of zoom lens with fixed and variable intrinsic parameter are studied.This paper proposed a new self-calibration method based on two vanishing points and a new active camera self-calibration method rotating around the camera optical center.The precision feature matching between two zoom images is researched,and zoom camera internal parameters and external parameters,and the displacement and zoom center can be calculated in this paper.(3)To carry the camera calibration technique out image correction,image mosaic and depth reconstruction,the experimental results show that the proposed calibration method on the premise of ensuring real-time,improves the accuracy and speed of operation,enhance the practicability of the system.Innovation point of this paper:(1)In this paper,a new sub-pixel corner detection method based on Harris operator is proposed,improving the traditional Harris corner detection algorithm.The results show that the more accurate sub-pixel corner points can be obtained.(2)This paper presents a camera self-calibration method based on the two vanishing points.The camera's interior parameters can be solved linearly only four images taked from different viewpoint,containing rectangular pattern in the scene.Compared with the results of Zhang Zhengyou and machine vision software Halcon,the results show that the proposed methods are simple and efficient,and verify the robustness and practicability of the algorithm.(3)This paper presents an active self-calibration method of camera rotating around the center.This method does not require the known information calibration target,is suitable for any scene and a fully automatic calibration process.(4)This paper presents a precise matching algorithm used for zoom images.After the initial matching based on SIFT algorithm,the geometric properties of zoom center,the epipolar geometry constraint,the RANSAC algorithm are used for further optimization of the initial matching.Through the thick lens model of zoom lens,optical center displacement can be calculated according to the zoom camera external parameters have been calculated.