Panoramic video stitching method based on spatial spherical surface and equipment thereof

Abstract

The invention aims at providing a panoramic video stitching method based on a spatial spherical surface and equipment thereof. The method is characterized by through mapping each plane image in one frame of video of the video to the spatial spherical surface, acquiring a corresponding spatial spherical surface image; determining at least two groups of characteristic point pairs corresponding to each two adjacent plane images in the one frame of video or the spatial spherical surface image, acquiring positions of the characteristic point pairs in the spatial spherical surface; and based on the positions of the characteristic point pairs mapped to the spatial spherical surface, registering each spatial spherical surface image in the one frame of video on the spatial spherical surface. Distortion which is introduced when the plane image is unfolded into a longitude-latitude plane image is avoided so that a registering result is accurate. Each registered spatial spherical surface image is unfolded into a corresponding longitude-latitude plane image. The longitude-latitude plane images are stitched so as to acquire a panoramic image corresponding to the one frame of video so that all the plane images in the one frame of video of the video can be accurately unfolded so as to stitch into the panoramic image.

Description

technical field [0001] The present application relates to the field of computers, and in particular to a spatial sphere-based panoramic video splicing method and device. Background technique [0002] In the prior art, there are mainly two methods for panoramic stitching of panoramic videos. The first method is to expand the plane image into a latitude and longitude map, and obtain the position registration information between the plane images by translating and rotating the latitude and longitude map. And based on the position registration information, the corresponding longitude and latitude maps are spliced; the second method is to obtain the corresponding homography matrix by calculating the affine transformation between the plane images, and map the plane image to the space through the homography matrix and obtain the registration information, and then stitch the spatial images corresponding to the planar images based on the registration information. Due to the distorti...

AI Technical Summary

Problems solved by technology

[0004] One purpose of the present application is to provide a panoramic video stitching method and device based on a space sphere, so as to solve the problems of large image registration errors and severe distortion caused by stitching panoramic videos using the panoramic stitching technology in the prior art

Images