Design And Implementation Of A Fast Video Splicing System Under Multiple Cameras

Video stitching is the process of stitching multiple overlapping areas of video into a single,wider viewing angle.The spliced video image is broader and can display more content than video from an ordinary viewing video.The video angle of view taken by the single-lens camera is not broad enough.In the future,there will have plenty of scenes that require a large viewing angle of video.Video stitching will be more widely used in future applications,such as driving,virtual reality,security monitoring,multiple projectors screening,map three-dimensional panoramic display,etc.This paper is mainly about design and implementation of video splicing system under muti-camera,moreover analyzing the knowledge and technology needed in the system.Implement multiple algorithm in the system,and obtain the experimental results,analyzing the simulation results of the algorithm.Under the guarantee of quality and speed,choose areasonable algorithms to reslize a smooth and clear video stitching system.The main work of this paper is in the following aspects:(1)The development status of image and video splicing at home and abroad is studied,and the feature point detection and matching in splicing,image fusion and parallel computing technology in computer image processing are deeply understood.In the case of image matching,it is proposed to scale the image according to a certain scale of downsampling reduction to improve the running speed of the system.By comparing the experimental simulation results,an appropriate proportion coefficient of downsampling is selected.In the system,the poisson fusion algorithm,the direct average fusion algorithm and the gradual in and out fusion algorithm are realized,and the experimental simulation of the three algorithms and the comparative analysis of the results are carried out.(2)When the images in the video stream are fused,in order to ensure that the video fusion fast enough,offline is proposed to calculate the camera parameters design,calculates the parameters of good video stitching before video fusion,when streaming video fusion,direct use of these parameters,can greatly reduce the streaming video fusion time,meet the needs of users,can play the video after the fusion of fluent and clear.CUDA parallel programming is used to improve the speed of video stream image fusion.