Calibration-free and convenient monocular lens image radial distortion correction method

Abstract

A calibration-free and convenient method for correcting radial distortion of monocular lens images, which is characterized in that it is divided into the following steps: 1) For a single image or a certain frame image in the video acquired by the monocular camera, mark the image on the image 3 correction mark points are selected; 2) According to the distortion characteristics of the image, a radial distortion correction calculation model is constructed; 3) Data optimization of the correction mark points is performed to eliminate errors and improve the robustness of the correction algorithm; 4) Apply the optimized Correction mark points, numerically solve the distortion correction calculation model, and reversely find the radial distortion parameters; 5) Perform image correction based on the radial distortion parameters, that is, calculate the pixel value of each point on the corrected picture. In the present invention, users do not need any auxiliary calibration of images, nor do they need to predict or input any camera parameters. As long as three points are interactively input, image correction can be realized in real time.

Description

technical field [0001] The invention relates to a post-processing technology of a camera image in the field of image processing, in particular to a calibration-free and convenient monocular lens image radial distortion correction method. Background technique [0002] Because the camera lens does not conform to the ideal pinhole imaging model, the captured images generally have barrel distortion, pincushion distortion and fisheye distortion. Barrel distortion and pincushion distortion are collectively referred to as radial distortion. Barrel distortion is generally caused by wide-angle lenses. The magnification of the area far away from the optical axis of the lens is lower than that near the optical axis, and a convex scene appears in the image plane. The pincushion distortion is generally formed by the telephoto lens. The magnification of the edge of the field of view is higher than that of the center of the optical axis, and the rectangular object is shot into an image wi...

AI Technical Summary

Problems solved by technology

The biggest defect of this method is: In many practical applications, it is impossible to use a camera to shoot and sample the calibration template, and it is impossible to correct it.

The disadvantage of this type of method is that multiple calibration images taken by multiple cameras are required, and the calculation process is cumbersome.

Images