Font Size: a A A

Research On Key Technology Of Multi-Channel Visual Simulation

Posted on:2020-01-03Degree:MasterType:Thesis
Country:ChinaCandidate:Y Q YaoFull Text:PDF
GTID:2428330572478183Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Multi-channel visual simulation usually uses a multi-machine distributed interconnection to provide users with a larger field of view and higher immersion.However,as the richness of the simulation scene increases,the data size of the simulation scene increases sharply,and the number of moving entities also increases,causing the problem of the scene picture to be stuck.Moreover,due to the difference in the processing power of each rendering node and the discontinuity of the simulated entity data,the problem of ghosting,tearing,and moving entity jitter of the visual scene occurs when the multi-channel dynamic scene is displayed.Based on the research of the “Full Visual Virtual Simulation Project” of the Graduate Innovation and Entrepreneurship Fund Project and the “Complete Mission Maritime Integrated Simulation System”,this paper studies the above issues.In this thesis,the cone partitioning method and multi-factor decision method are proposed to select the least redundant data block,combined with viewpoint prediction and multi-thread scheduling strategy to avoid the picture jam;and based on this,a motion entity smoothing method based on ?-? filtering is proposed to solve The problem of picture jitter is presented.The method is tantamount to solve the problem of picture jitter.At the same time,the bucket synchronization method based on bucket cycle adaptation is proposed to repair the ghosting and hearing problems of the screen.Finally,a comprehensive evaluation method based on time and space consistency is proposed to evaluate the synchronization effect in all aspects.Specific work includes the following:Data block selection and multi-thread scheduling strategy based on cone partitioning and multi-factor decision method.According to the user's attention to each partition scene,different clipping strength is selected to avoid loading of redundant data blocks.Then,for the multi-channel dynamic scene,a multi-factor data node comprehensive judgment method is proposed to appropriately reduce the data block display precision of the high-speed dynamic scene.Finally,data prefetching strategy is introduced,and the parallel processing method is used to improve the rendering and scheduling speed of the data block.Motion entity processing based on ?-? filtering is synchronized with the bucket period adaptive scene.Firstly,the filtering algorithm is utilized to smooth the motion entities to enhance data continuity and solve the problem of physical motion jitter.Then,a bucket synchronization method based on bucket cycle adaptation is proposed.The bucket period is dynamically adjusted according to the feedback of the rendering pressure of each channel,and combined with the motion entity synchronization control strategy,the rendering data of each channel is synchronized in real time in the bucket cycle to avoid tearing of the screen.A comprehensive evaluation method based on time and space consistency is proposed to solve the problem of difficult evaluation of optical synchronization.The simulation results show that the proposed scene data block selection and multi-thread scheduling strategy based on cone partitioning and multi-factor decision can significantly reduce the loading of data blocks and improve the smoothness of visual images.The proposed bucket cycle adaptive synchronization method can effectively solve the problem of dynamic scene unsynchronization.The proposed smoothing method based on ?-? filtering can eliminate motion jitter and improve the dynamic scene display effect.
Keywords/Search Tags:cone partitioning cropping, multi-factor node decision, viewpoint prediction, bucket period adaptive, ?-? filtering
PDF Full Text Request
Related items