A method for realizing 360-degree panoramic video

Abstract

The invention discloses a method for realizing 360-degree panoramic video, which relates to the technical field of video processing. It uses the existing video compression standard to compress and transmit the video content captured by each camera in parallel. The video shot by a single camera is a two-dimensional flat video, which is suitable for using the existing video compression standard, and records the camera in the parallel video stream. Marking information, at the receiving end, can decode and display all videos, or use the camera marking information to select the video captured by the corresponding camera according to the viewer's viewing area, splicing and merging them and displaying them. The invention has high efficiency in video data processing, transfers the splicing work from the encoding end to the decoding end, does not need the video splicing and projection transformation process before encoding, reduces the complexity of 360-degree video encoding, and improves the encoding speed.

Description

technical field [0001] The present invention relates to the technical field of video processing, in particular to a data processing flow for parallel encoding of 360-degree panoramic video and splicing and displaying at a decoding end. Background technique [0002] In virtual reality (VR) applications, 360-degree panoramic video is a video image that is seamlessly connected around 180 degrees up and down and 360 degrees horizontally with the human eye as the center point. Through the panoramic player, users can click the mouse, touch the screen, Gyroscope and other methods realize up and down, left and right, zoom in, zoom out without dead angle browsing or watch video mode through virtual reality helmet. Shoot videos from different angles at the same time through multiple cameras, output them later through dedicated software, or synchronize, splice, adjust, and output through special video stitching software, and finally get a panoramic video file with a ratio of 2:1, which...

AI Technical Summary

Problems solved by technology

[0003] Spherical projection is to expand each meridian into a straight line segment, and circle around the equator to form a cylinder, and then expand into a rectangle. This method is widely used in the process of panoramic video coding. Its advantage is that the rectangle is conducive to The existing video compression standard is used for compression, but during the conversion process, the polar regions are stretched too much, resulting in deformation and distortion

[0004] Cube projection is to project the content of the spherical surface onto the six faces of the circumscribed cube, so that the pixels are evenly distributed and the picture has no geometric distortion. After the transformation, the file size is reduced by 25% compared with the original sphere, but because it is mapped to the circumscribed cube of the ball, The pixel density is reduced, which brings the problem of blurred images

[0005] The square prism projection is to project the content of the spherical surface into the square prism embedded in the cube. The projected plane area is smaller than that of the sphere, which reduces the memory by 80% compared with the source file, but because the unit area on the sphere is projected to the square The area on the plane of the prism varies with the position, resulting in more obvious scaling, which will lead to a certain degree of loss of original image information and display distortion

[0006] To sum up, no matter what kind of projection transformation the current spherical video has, there are certain deficiencies, and the 360-degree video after the projection transformation has higher resolution than the traditional flat video, and requires more bit rates. Limited problems, based on this, it is necessary to change the original method of transmission and processing of 360-degree video, and design a method for realizing 360-degree panoramic video

Images