Research On Intelligent Maintenance Method Of Complex Equipment Based On Augmented Reality

The traditional equipment maintenance support activities have many inconveniences during operation,limited by factors such as environmental interference,paper or electronic technical manual distractions,and often have the problems of long maintenance inspection time and low maintenance efficiency.Augmented reality(AR)maintenance assistance technology can integrate and superimpose virtual auxiliary information with real scenes,and push multi-source information to the real world through new interactive methods,thereby improving the maintenance efficiency of maintenance personnel in the maintenance guarantee of an open environment.Aiming at the problems of traditional maintenance auxiliary tasks,this paper takes the complex industrial equipment model as the research object,researches the AR maintenance method based on object recognition and real-time tracking,and builds a prototype system of AR intelligent maintenance.The main research work of this paper is as follows:(1)A method of target recognition and 3D tracking registration based on high precision RGB-D camera is proposed.Aiming at the problems such as an unstructured environment and numerous scene parts in maintenance assistance,a multi-perspective surveillance database was constructed by using the 3D model of the equipment,and real-time tracking of the target was realized by combining transfer learning and neural network.The target region of interest was mapped to the three-dimensional point cloud space,the initial pose guidance of the target was calculated by combining the sampling consistency algorithm and ICP registration algorithm,and the KLD particle filter algorithm was improved to complete the registration and tracking of the complex equipment.(2)A method for editing augmented reality auxiliary information based on IETM interactive electronic manual is proposed.According to the S1000 D standard,the IETM electronic manual is analyzed,and the unified conversion and description mechanism of multi-source auxiliary information is established.Information such as the 3D model and its geometric constraints,part layout,component explosion diagram and tool annotation are acquired and analyzed in a process.The HSG rendering engine for helmet end is developed,and the virtual prototype,3D animation,video,voice,and other multi-source maintenance process information is made into AR packets suitable for helmet end,which provides guidance content of augmented reality for maintenance assistance.(3)A prototype system of AR intelligent maintenance for complex devices was built.Firstly,the portable hardware equipment is built and debugged.Based on the architecture and analysis of the software system,the prototype system of AR intelligent maintenance for complex equipment is established.Finally,the feasibility and effectiveness of the prototype system are verified by several typical cases of augmented reality assisted maintenance.