Research And Implementation Of Mobile Augmented Reality Technology

Mobile augmented reality is one of the research hotspots in recent years.It has broad application prospect and great economic value in a wide spectrum of application domains,such as medical research,industrial maintenance,monument restoration,digital cultural heritage protection,tourism exhibition,and games entertainment.The process of mobile augmented reality requires real-time response.When a user triggers the system,a real-time response should be given.Either low recognition accuracy or high response latency will seriously downgrade user experience.Existing mobile augmented reality systems mainly rely on mobile phone sensors or computer vision techniques.To accurately recognize the target with reasonable response latency,most of these systems generally require the location information from the GPS on mobile devices,and exploit this location information to locate a candidate set from the database for further searching and matching.Such treatment avoids the high latency problem that would be encountered by searching the whole database.However,it may still suffer from the latency issue because the coarse-grained property of the GPS(say an error of about 5m in outdoor environments),would lead to a candidate set of large size.In addition,these systems may fail to well accommodate indoor environments as the GPS is usually not available indoors.Although there are some works studying specific search methods without relying on the GPS information,they rely on extremely complex algorithms and thus have quite limited practicality.In light of the shortcomings of existing mobile augmented reality techniques,this paper studies and designs a new mobile augmented reality system.The proposed system creatively and cleverly fuses multiple cross-domain techniques such as mobile phone camera,mobile phone sensors,and computer vision,and provides detailed design on multiple functional modules,including target-user distance acquisition,target image extraction,target location acquisition,and search and matching.To achieve low latency,our system creatively adopts a geometric perception based search approach that builds on a distance-measurement model.It can accurately infer the relative positions of targets,establish the spatio-temporal relationship between targets,and locate a reliable and small candidate set from the database.Thus,the time consumed by searching the target can be significantly reduced,providing good user experience.The proposed mobile augmented reality system has the properties of high object recognition accuracy,low response latency,and robustness.It has little dependence on the external environments,and can be applied in indoor environments.The modules in the system are integrated in a reasonable and flexible way,which can provide users with quality service of mobile augmented reality.The experiments conducted in a real-world environment show that the proposed system can well accommodate indoor environments,and is easy to use with good stability.