Research On Rapid Augmented Reality Visualization Of Large-scale Geo-spatial Social Media In A Distributed Mobile System

Augmented reality(AR)is widely known as the next-gen human-computer interface.The i OS and Android platforms have launched their respective AR APIs,ARKit and ARCore,making billions of smartphones become AR devices.In 2016,the AR game Pokemon Go became popular all over the world as soon as it was launched and received 130 million downloads within a month in more than 70 countries.Current AR applications mostly focus on 3D model animations.The user experience of simple social multimedia data,such as text,pictures,audio,and video,in AR environment is not good enough,but social media data is the easiest to access or produce for common users.Studying the visualization of social media data in the AR environment is of great significance.In the era of big data,when the scale of social media data is getting larger and larger,in a distributed mobile network environment,rapid AR visualization of largescale geo-spatial social media data(LGSM)is extremely important.And data access model,high-performance data processing are the primary issues for the rapid visualization of AR.AR visualization is presented in real 3D environments,which have distinct geospatial attributes,thus should be called geospatial social media.One factor in the AR visualization of LGSM is to use geospatial information effectively.Based on the previous research,this paper focus on rapid AR visualization of LGSM:(1)A geospatial social media data specification is proposed which is suitable for large-scale network data transmission,called Geo ARMedia.The Geo ARMedia data specification complying with gl TF and Geo JSON data standards,defines a data model of common geospatial social media,and specifies the JSON storage expression for common multimedia data types(text,pictures,audio,video,etc.).(2)In order to facilitate LGSM access in a distributed environment,a steady-state Z curve algorithm that facilitates database smooth reading and writing is constructed,and a common space query based on the steady-state Z curve of the data is studied.A distributed data division strategy for the data based on a steady-state Z curve,is proposed,to optimize the storage distribution in the database and improve the distributed access efficiency.(3)An AR visual zone and three-layer spatial interaction architecture on smart phone devices is constructed.A spherical model of AR visualization of spatial social media data is proposed.This paper also developed a spatio-temporal coupling algorithm for large-scale social picture retrieval that can greatly improve the efficiency of AR recognition map hitting with geographic information.(4)In order to quickly respond to LGSM loading requests in a distributed environment and ensure the efficiency of front-end AR visualization,we propose a multi-level caching mechanism for spatial social media data augmented visualization.The spatial proximity of media data is used to develop the AR visualization front-end cache pre-scheduling mechanism,and cloud detection and distributed cache scheduling mechanism,to spot geographical location-based hot data effectively,ensuring the rapid visualization and high concurrent response requirements.(5)LGSM AR visualization is accompanied by massive computing tasks.By using big data technology and high-performance computing technology,this study developed a parallel processing computing framework for LGSM,and proposes a algebraic method of parallel processing of LGSM which can describe parallel processing tasks concisely and effectively,organize and schedule multiple parallel computing tasks.(6)Combining current distributed database technology,big data technology,scalable network service architecture,mobile front-end 3D engine technology,and relying on the above research results,this paper designs and develops a LGSM rapid AR visualization prototype system—Retinar.Retinar is discusses in detail the AR interaction and visual design of common social media data,implemented front-end and back-end AR services,developed a mobile-end social media data AR editor,which reduced the cost of AR social media data content production.Finally,this article takes the surrounding area of Nanjing University Xianlin campus as an example to carry out application experiments.A large number of tests were conducted,and the results shew that Retinar can respond to tens of millions of data scales within 1 seconds,reaching an ideal experimental expectation.