VIDEO PLAYING METHOD AND SYSTEM BASED ON AUGMENTED REALITY TECHNOLOGY AND MOBILE TERMINAL

摘要

Disclosed is a video playing method based on augmented reality technology. In combination with augmented reality technology, a rectangular plane 3D model which completely covers a target picture in a current scene image is drawn, and at the same time, after a video file related to the target picture is separated into an audio stream and a video stream, a video image frame in the video stream is taken as a texture and is mapped to the 3D model frame by frame, and graphic rendering is performed; and finally, a scene image captured by a video camera and the rendered 3D model are fused and output, and the audio stream data is output synchronously. Correspondingly, also disclosed are a video playing system based on augmented reality technology and mobile terminal. The picture information in newspapers and periodicals, magazines and other print media can be easily converted into video information, and the video completely fits the picture in the spatial position, so that a user obtains a new audio-visual experience of combining emotion with scenery; and newspapers and periodicals, magazines and other paper media can also realize diversified propagation of information by virtue of the technology of the present invention, and provide a precise fast video push service.

主权项

1. A method for playing a video based on augmented reality technology, comprising:
capturing, by the video camera, the current scene image containing the target picture; performing feature detection on the current scene image captured by the video camera, extracting image feature points thereof, and describing said image feature points, so as to obtain feature description data of a whole image; searching an image in an image database according to the feature description data of the current scene image, so as to obtain a reference image matching with the target picture in the current scene image and a video file corresponding to said reference image; performing audio-video split on and decoding to the video file corresponding to said reference image, in order to obtain a video image frame sequence and an audio data; registering the current scene image captured by a video camera with the reference image, calculating a posture of the video camera to obtain a homography matrix, and obtaining a rectangular plane 3D model which completely covers the target picture in the current scene image according to said homography matrix; mapping a video image frame, as a texture, in said video image frame sequence to said 3D model frame by frame, and performing graphic rendering; and fusing the current scene image captured by the video camera with the rendered 3D model, outputting and displaying the fusing result, and outputting said audio data synchronously.