Real-time removal method of calibration plate in augmented reality

Abstract

The invention discloses a real-time removing method for an augmented reality calibration plate. Picture holes existing after the calibration plate is removed are patched in real time through an image restoring method in an existing augmented reality system based on a calibration plate. The restoring method comprises the following steps that the position of the calibration plate in a video picture is obtained through camera calibration, and an area to be restored is determined; a linear feather image is computed and is combined with a current video frame image to be a texture image; a texture piece sample library is built for the non-patched area of the texture image; for the boundary pixels of the area to be restored, a neighborhood texture piece is used for conducting matching in the texture piece sample library, and the center point color of a texture piece obtained through matching is assigned to be the current boundary pixels; whether the restoring of all areas to be restored is achieved or not is judged, the restoring is over if yes, and if not, the restoring steps are conducted on the boundary pixels of the rest areas to be restored; finally, a previous frame restoring result and a current frame restoring result are used as the neighborhood, and bilateral filtering is conducted on the restoring result of the current frame.

Description

technical field [0001] The invention relates to a method for repairing a video frame, in particular to a method for real-time repairing of a video frame after removing a calibration plate in augmented reality. Background technique [0002] Augmented reality is a technology that enhances the observer's visual perception of the real world through the virtual information provided by the computer system. It uses related computer technology to superimpose the virtual scene into the video screen of the real environment in real time. In order to ensure the spatial consistency between the superimposed virtual scene and the real environment, it is necessary to use the camera calibration method to calculate the internal and external parameters of the camera. [0003] The current camera calibration methods are mainly divided into two categories: methods based on pure computer vision and methods based on other physical hardware. The calibration method based on pure computer vision need...

AI Technical Summary

Problems solved by technology

[0005] However, this method requires that the calibration board must be placed in the video screen, which blocks some real scenes and affects the authenticity and aesthetics of the augmented reality screen, thus greatly limiting the application of this technology in the industry

Images