Research On Fusion Method Of Multi-Monitoring Video And 3D Virtual Scene

Under the requirements of deepening the security construction and improving the threedimensional prevention and control system of social security,the monitoring system,as an effective security scheme of grid management,is developing in an all-round and ultra-high definition direction.With the increasing number of surveillance cameras,the traditional multi-screen switching playback method has gradually exposed the shortcomings of discrete content and weak correlation.This adds the cognitive burden to spatio-temporal information for users,and can not meet the multi-dimensional visualization needs of intelligent monitoring.This paper focuses on the urgent problem of how to realize the effective integration and intuitive visualization of a large number of surveillance videos.Combined with augmented virtual reality technology,through the virtual and real fusion of multi-channel surveillance videos and three-dimensional virtual scenes,as well as the view-dependent video scheduling,the visualization of omni-directional three-dimensional real-time dynamic monitoring is realized.The research work of this paper is as follows:(1)In order to meet the three-dimensional visualization requirements of the monitoring system,this paper designs and construts a web-side virtual-real fusion monitoring visualization system,and studies the camera registration method based on incremental motion recovery structure.The system consists of four modules.The first is the video acquisition module,which is used to obtain the video data stream of the simulation platform or the real scene.The second is the preprocessing module,which is used to generate the configuration files that the system needs to read in advance.The third is the video transmission module,which is used for encoding,packaging and pushing data stream,and the terminal pulling video data stream.The fourth is the fusion display module,which is used to display the virtual-real fusion effect of real-time monitoring video.At the same time,aiming at the camera registration problem involved in the preprocessing module,a camera registration algorithm based on incremental motion recovery structure is implemented,and the effectiveness of the algorithm is verified by experiments.(2)Aiming at the problem that a single virtual-real fusion display method cannot take into account the visualization effect and rendering performance,this paper studies a display method based on a hybrid strategy for virtual-real fusion of surveillance video and 3D scenes.Firstly,three visualization algorithms based on texture mapping,shadow projection and depth information are analyzed,the related principles and implementation processes of each algorithm are described,and the horizontal comparison experiment and analysis are carried out.Then,a fusion display method based on hybrid strategy is proposed.Combined with the registered pose of the camera and the flatness of the projection environment,the appropriate display method is selected according to the applicable conditions of different algorithms to achieve the balance between visual effect and rendering performance.(3)Aiming at the problems of reduced rendering performance and redundant resource occupancy in multi-channel video stream fusion display,this paper studies a view-dependent video data scheduling method.This method consists of three parts.First,the visual domain node clipping based on quadtree,which realizes the pre-division of video nodes in threedimensional space and the effective clipping of visible nodes at runtime.Second,video data scheduling based on multi-resolution level of detail,which realizes detail level judgment for visible nodes based on the screen space error,so as to select the video data stream with appropriate resolution.Third,multi-video stitching and rendering after transparency detection to realize the visual optimization effect of natural transition between adjacent overlapping videos.Finally,through experiments in a variety of experimental scenes,the virtual-real fusion effect of multi-camera video streams and 3D virtual scene is displayed,and the effectiveness and fluency of the algorithm are verified.