Research On Key Technology For Rail Transit Equipment IETM Embedded With AR Function

At present,China's high-speed railway,subway and other rail transit are in a period of rapid development.The maintenance and management of related equipment are of great significance for the normal operation of rail transit.According to the characteristics of rail transit equipment maintenance work,the Interactive Electronic Technical Manual(IETM)technology and Augment Reality(AR)technology are introduced into the maintenance work.By adding maintenance data information management,virtual reality overlay,human-computer interaction and other functions to the daily maintenance,management and malfunction repair of the equipment,some problems would be solved,such as,high technical content and complicated structure of rail transit equipment,massive maintenance information management difficulties and limited technical level of maintenance personnel,which provides a new information-based means for the maintenance,repair and management of rail transit equipment,and also provides an idea for solving the problem of complex equipment maintenance support in other industries.Firstly,based on the analysis of IETM research and development at home and abroad,the S1000 D specification involved in the IETM construction process is studied and combed,and the information management methods of equipment technical are analyzed,which use Common Source Data Base and Data Module.By analyzing the characteristics of rail transit equipment maintenance work,the requirements analysis of the IETM software and hardware system embedded with AR function is carried out,the description of the maintenance process is realized by using Petri net modeling technology,and the management of multimedia prompt information is realized by using XML technology.Secondly,aiming at the maintenance scene and object detection problem in the AR system,the HHA image is obtained by encoding the depth image,the SSD target detection algorithm using the lightweight network MobileNet as the base network is adopted,and the RGB image and the HHA image are input into the network for training by data fusion,so that the network can simultaneously learn color features and spatial information features,and the accuracy of the target detection algorithm is improved.Aiming at the problem of target tracking and 3D registration,an algorithm combining target tracking and target repetitive detection is proposed.By sampling the video image of the maintenance site,Kalman filtering is performed by target repeated detection to predict the target motion trajectory.When target lost,the tracker would be re-initialized to improve tracking accuracy.Using the point cloud matching method,the position and pose of the maintenance object in the camera coordinate system are calculated,and the virtual information overlay is realized.Finally,on the basis of the above theoretical research,taking the ZD6 switch machine as an example,an IETM prototype system embedded with AR function is realized and applied to the representation circuit fault maintenance of the switch machine.The evaluation of the above system is carried out by setting the AR system evaluation index,which confirms the superiority of the system in the management,retrieval and expression of equipment technical data.