Research On Camera Calibration Method

From the two-dimensional image to three-dimensional information, the correspondence between the three-dimensional and two-dimensional image is first solved, so the camera needs to be calibration. Camera calibration technique is need to be solved the in the first place in the close shot photogrammetry and computer vision to widely used in the relevant fields laid a foundation. Existing camera calibration method consists of the traditional algorithm,self-calibration algorithm and active vision calibration method. Due to the existence of the camera lens distortion and for high precision measurement applications, distortion need to be corrected. Large amount of calculation and complex model are existed in the traditional calibration.On the basis of analysis of principle of camera calibration, the article main research content:First, the camera model is described in detail, and existing camera calibration methods are classified and researched.Second, the required for camera calibration images need to be processed, including image gray, smooth, enhance operation, then the image needs to be feature point extraction,At the same time, Harris corner algorithm is improved by joining the limiting factors to improve the effect of feature point extraction.Third, In view of the false convergence problem of particle swarm optimization, the particle swarm algorithm is improved, the fuzzy adaptive simulated annealing particle swarm optimization algorithm is proposed, the simulated annealing algorithm is introduced into the particle swarm algorithm, and the fuzzy theory is used to improve inertia weight.Experiments show that the improved algorithm has obviously improved on the convergence speed and accuracy.Fourth, using the improved particle swarm algorithm and using the real image for camera calibration, the experiment showed that the improved calibration method, comparing the results with ilf method show that the calibration method is feasible.