Research On Camera Calibration Method Of Wide Field

Computer vision technology has been widey employed in the machine manufacturing, Security Monitoring System, navigation guidance, robotics, etc, it is one of the most rapid development discipline. Camera calibration establish the projection relations between 2D images and 3D space. Base on the existing research results, three kinds of camera calibration method as well as corresponding camera calibration tool are developed according to the specific needs of large field view application environment in practical.First, current situation of the development of the camera calibration technology is system analysised. The holes imaging model and lens distortion model are illustrated under the engineering background of this paper this paper. several concepts of projective geometry are introduced.Secondly, three camera calibration methods for wide field application are presented:(1) improved flexible planar camera calibration method for wide field of view. Flexible calibration target is designed to reduce the large target from the difficulty of producing and processing of uncertainty on the calibration accuracy of the characteristics of the starting points raised by the geometric constraints to construct the calibration point calibration method. By considering lens distortion, the method of constructing the feature points is improved, a lens distortion coefficients pre-calibratable flexible planar target is developed, the improved method can be used in the occasion with more distortion and lager field of view;(2) Proposed camera calibration method based on a small target large field mosaic. In the large field applications, the use of a single small calibration target can only get the Local optimal calibration results, put forward a different position of the small field of view Target spliced into the camera field of view cover most of the regional mosaic of small calibration target Methods. The method for small target is not strictly the subject of installation requirements, just a small target to meet all the conditions placed coplanar to, and effective solution to the large field of applications using a single small calibration target limitations;(3) Proposed camera calibration method based with auxiliary pattern. The advantages and disadvantages of the two methods above are analysised, a model "coarse calibration with big target + correction with small one" is rised. This method does not require to consider lens distortion in the construction of calibration points, nor requires a small target as requested to meet coplanarity constraints, it weakens the conditions of implemente the camera calibration method for large field, simplifies the calibration process. This article uses simulations and practical experiments of the above three methods to verify the feasibility and stability. From the experimental results, all of these three methods can reduce camera calibration difficulty and improve the calibration accuracy. The accuracy achieved by these method is near to the traditional method.Finally, according to the modular development thought, camera calibration software is developed with Qt4.7 & OpenCV2.1 under the Visual Studio 2008 the development environment, models of image processing and feature point extraction, feature point coordinates matching, video camera parameters calculation and camera calibration accuracy checking etc are realized.