Adaptive wide-angle image correction method and device based on single-grid image segmentation mapping

Abstract

The invention discloses a self-adaptive wide-angle image correction method and device based on segmented mapping of a single grid map, which has the characteristics of easy operation and strong practicability, not only realizes the automatic extraction of control point pairs, but also can adapt itself according to the distortion of the lens. The optimal segmentation correction is carried out to achieve the automatic correction processing and improve the local and global correction accuracy. The method includes: preprocessing a distorted grid template image to obtain a single-pixel binary distorted grid template image; extracting control point pairs in the distorted grid template image by using the neighborhood features of grid intersections; establishing radial distortion The minimum segmented optimization model of the Pixel interpolation is performed on a corrected result image to restore its gray value to obtain a complete second corrected result image.

Description

technical field [0001] The present invention relates to the technical field of wide-angle image correction, and in particular, to an adaptive wide-angle image correction method and device based on single-grid image segmentation mapping. Background technique [0002] Compared with ordinary lenses, wide-angle lenses can obtain images containing more target information in a wider field of view due to their short focal length, large field of view, and long depth of field, which greatly reduces visual blind spots. Therefore, it is widely used in industrial monitoring, video surveillance, assisted driving, robot navigation, medical endoscopy, virtual reality and many other visual fields. However, due to the manufacturing defects and nonlinear imaging mechanism of the wide-angle lens, there is a relatively serious barrel distortion in the captured images. Wide-angle barrel distortion mainly includes thin prism distortion, tangential distortion and radial distortion. This nonlinea...

AI Technical Summary

Problems solved by technology

For a large field of view lens with severe distortion, these two types of models can only effectively correct the area around the distortion center, but for the four corners of the image, the correction accuracy is low and the effect is very poor

For example, patent application No. 201410424349.8 "A Method of Image Correction after Fisheye Lens Calibration", similar to Zhang's calibration method, uses the checkerboard grid calibration method to obtain the parameters of the high-order polynomial model, due to the integrity requirements of the grid blocks of the checkerboard template diagram , the control points in this type of template image cannot cover the entire viewing angle area, resulting in the inability to calibrate the corner area of ​​the image, so all areas of the distorted image cannot be effectively restored

No. 201510514238.0 patent application "A Method for Accurately Correcting Image Distortion of Ultra-Wide-Angle Camera" has been improved on this basis, and the correction accuracy and the range of correction angle of view have been improved to a certain extent, but the overall correction effect is not good

Images