Research On The Method Of Camera Self-Calibration Based On Image Sequences

Camera calibration is an important step for obtaining3D world information from2D images, which is widely applied to3D reconstruction, navigation of autonomousvehicles, visual supervision etc.Thereinto,camera self-calibration is the direction ofdevelopment of the calibration.Compared to traditional laboratory calibration, cameraself-calibration is closer to the actual application and can be used in-linethree-dimensional reconstruction, so it has broad application prospects.The camera self-calibration method has been carried out now, but it is essentiallybased on a constant intrinsic parameter,and in practical application,the camera intrinsicparameters can not be regarded as constant when the object is near the cameraregardless of camera movement or moving objects,therefore,the variable intrinsicparameters self-calibration makes more sense.I have done some exploring in the fieldof variable intrinsic parameters self-calibration after learning the basic theories relatedto calibration.The main work is as follows:(1) Described the research background,significance of the research situation, andtheory value and application prospect of self-calibration.(2) Introduced the theoretical basis of self-calibration about the camera model andmulti-view geometry, and frequently-used self-calibration methods.(3) This chapter presents an method of self-calibration of varying internal cameraparameters that based on image of dual absolute quadric transformation based onAbsolute quadric-based self-calibration method. Absolute dual quadric elements haveso large differences in magnitude that solutions are extremely sensitive tonoise.Through the transformation, all the elements of the image of dual absolutequadric are transformed into the same magnitude, So the solutions become more stable.The theorical analysis and experiments with real data demonstrate that thisself-calibration method can lead to an enormous improvement on the stability androbustness of the results without increasing computation.(4) This chapter presents an method of self-calibration of varying internal cameraparameters that based on quasi-affine reconstruction.In a stratified approach toself-calibration,a projective reconstruction is obtained first and this is successivelyrefined first to an affine and then to a Euclidean reconstruction.It has been observedthat the difficult step is to obtain the affine reconstruction,or equivalently to locate the plane at infinity in the projective coordinate frame.So,a quasi-affine reconstruction isobtained first by image sequences,then we can obtain the infinite plane,or equivalentlyto affine reconstruction.Then the infinite homography matrix can be calculated throughthe affine reconstruction,and then using the infinite homography matrix and constraintsof the image of absolute conic to calculate the camera internal parameters matrix,andfurther to measure the metric rreconstruction.This method does not require a specialscene constraints(such as prapllel、perpendicular) information,and also the cameramovement informations(such as pure translation or orthogonal movement),to achievethe goal of self-calibration. The theoretics analysis and experiments with real datademonstrate that this self-calibration method is available、stable and robust.