Research On Image Feature Based Camera Calibration Technology

Image Based Modeling (IBM) is an important research direction of computer vision and had broad prospect in application. Camera calibration is the most important step in the whole modeling process, because the projection matrix of camera may be used in the posterior steps of modeling. In light of these, it's necessary to research further in Camera calibration technologies. In this paper I study mainly on the technology of acquiring the intrinsic and extrinsic parameters of camera using feature of images.By analyzing lots of calibration schemes, this paper makes a conclusion of the method of calibration using the feature of images. This method can be divided into two steps. In the first step, we get the initial value of parameters by using the configuration in single image, then, by using matched feature points in stereo-pair, we can optimize the parameters of camera.1 The paper puts out a linearity imaging model. We do some predigestion on this model for an easier and more quickly calibrate process. By this model, we can distill the intrinsic parameters and extrinsic parameters, furthermore, we can also find out the projection matrix.2 In order to make camera calibration direct and robust, we give a convenient method. We can use the characteristics of 3-point perspective projective, and estimate the vanishing points of orthogonal parallel lines in 3D space. Then we get a set of equations and camera parameters can be solved with high accuracy.3 This paper adopts a method to optimize the parameters of camera, by using nonlinear optimization. Based on the result of the first step and prior researches, we adopt a nonlinear optimization to make optimization of camera calibration, and this optimization can lead to a satisfied result.4 Finally, we do some experiments to validate our method. Experiments show that this method is flexible and extensible, and that the calibration result wins great satisfaction form users...