Study On Distortion Correction Of Wide-angle Lens And Image Mosaic Algorithms Based On Opencv

In order to get more information of image, in the field of image we pay more attention on useing large field of wide-angle lens and image mosaic technology. Such algorithm has a strong practical. Panoramic image mosaic has a wide range of applications in processing earth satellite photos, vehicle information such as medical images, and quite a number of areas. The image information is not accurate because of the distortion of wide-angle lens, and splicing out of the image is meaningless. This paper makes a further study in large field of wide-angle lens for distortion correction and image mosaic.In this paper, summarise and conclud the correction algorithm of distored image and image mosaic technology, study on nonlinear image distortion correction method. Wide-angle lens have optical distortion in the imaging process. The distortion of the optical system has directly impact on the geometry persition, thus people will make mistakes of the location of the things. For example, two images with overlapping region for image mosaic, as the nonlinear distortion of lens caused the two images changed, the stitching does not match or even splice failure. In this issue, we use template method (checkerboard template method and concentric template method) for wide-angle lens distortion correction, based on the polar coordinate transformation, propose the application of OpenCV in VC platform and achieve polynomial fitting algorithm to calculate the distortion correction parameters. Calculate distortion correction coefficient and calculate parameters of camera to improve the accuracy of aberration correction, which applied to real-time distortion correction and image mosaic systems. To assess the accuracy of calibration, the TV distortion of some position of the lens field of view was caculated before correction and after correction. By comparing the amount of distortion, the effect of correction can be assessed.Study the image mosaic algorithms for corrected images. Typically, image mosaic algorithm can be divided into two categories:one is based on the region and the other is based on the characteristic. The advantage of the stitching algorithm based on the region is intuitive, the disadvantage is large operation and only applicable in simple image. In addition, as region-based image mosaic algorithm matching operations are based on pixels correlation in the overlapping area between spliced and reference image. Therefore, noise will caused significant impact on mosaic effect. In this paper, an image mosaic method that ratio method based on the wavelet transformation was presented. Compareing with the other image mosaic method in the spatial domain, the improved method in this paper could reduce adverse effect on image mosaic effect and get better visual result. At the end of this paper, we conclud the innovations and shortcomings, and prospect the future of the distortion correction and image mosaic technology.