Research On Distortion Correction Algorithm Of Fish-eye Lens With Largefield Of View

Fish-eye lens has been widely used in video monitoring,panoramic shooting and other areas.But,in the area of Measurement,the inherent nonlinear distortion of the fish eye lens will produce large measuring error,which limits the value of its application severely.In order to obtain a good visual effect of the large field fish-eye lens,the commonly used technique in the image processing of the optical camera imaging system is to use the computer image vision technology for correction.This article on this basis,the fisheye lens distortion correction algorithms as the main research object,and a summary of previous calibration algorithm is improved and optimized,and basically completed the following contents: first,in order to improve the correction accuracy of fisheye lens,make its can be used in actual measurement,the precision of fisheye lens debugging and testing,ensure that the experiment of the lens used all kinds of optical parameters meet the design requirements;At the same time,the correction of centrifugal distortion was introduced.According to its physical meaning,the method was analyzed in the simulation of Code V software.Secondly,based on the optical lens imaging theory and according to the mathematical thinking of fish-eye lens imaging design theory used in the experiment,the equidistant correction model was constructed,and the line-by-line scan algorithm was adopted to illustrate its practicability from the aspects of validity,error range and calculation quantity.Thirdly,on the basis of MATLAB Camera Calibration toolkit,standard black and white checkerboard grid was used to calibrate the fish-eye lens applied in the experiment,and the results were analyzed to lay the foundation for the subsequent higher precision measurement.Fourth,in order to satisfy the fisheye image in high precision measurement,the application of the Harris corner algorithm to extract the target area and surface fitting,indeed in the polynomial model is established in this paper,on the basis of application of Hooke-Jeeves pattern search method to make the corner detection pixels achieve sub-pixel level,through the second calibration,finish high precision of the distortion image recovery.The algorithm model adopted in this dissertation is based on the optimization calculation method and the linear interpolation algorithm to complete the image reconstruction.For the analysis of the final result,with the purpose of increasing the number of fish-eye cameras of the same type to enlarge the samples taken,the distortion correction of the captured images was completed,and the relative distortion,distortion coefficient and image point offset were analyzed by randomly sampling the samples.The final correction results show that this algorithm can quickly complete the distortion correction of the distorted image taken by fish-eye lens,and the correction accuracy is high,and the algorithm is easy to realize.