Design And Implementation Of Automatic Crack Identification System For Subway Tunnel Surface

In recent years,with the rapid growth of urban subway operation mileage and routes,a large number of infrastructure,especially subway tunnels,have been transferred from the construction period to the curing period,and the facilities are affected by diseases such as aging of their own structures and damage from external factors.For the subway tunnel,the crack on the surface of the tunnel has become one of the hidden dangers of subway operation safety.The detection of subway tunnel diseases mainly relies on manual inspection and visual judgment.The test results are greatly affected by human factors,resulting in insufficient and inefficient results.Therefore,computer technology and image processing technology are used automatically,effectively and at high speed.Identifying cracks in subway tunnels has become one of the hot issues in the modernization of subway testing.In this context,the paper designs an automatic identification system for surface cracks in subway tunnels.The system is developed in an object-oriented,highly portable Java language,and adopts the popular SSH(Struts+Spring+Hibernate)open source framework.Using the small and fast MySQL relational database,the detection and identification of crack information is based on convolutional neural network features in the convolutional architecture(Caffe)framework with high recognition and fast recognition.The region-recommended(Faster R-CNN)algorithm is used for crack identification.The main work of the thesis is as follows:(1)Firstly,the paper introduces the development background of the system and the research status at home and abroad,and expounds the techniques used in the development of the system.(2)According to the functional requirements of the tunnel surface crack automatic identification system,the paper determines the overall architecture of the system,including the software architecture and technical architecture of the system.Furthermore,the overall functional structure and database structure of the system are also designed.(3)The paper elaborates on the detailed design,implementation process and testing process of the system.In this part,the training of the network model of the fast-rcnn algorithm and the design of the business processing logic of each functional module are introduced in detail.The application of the thesis shows that the recognition efficiency and recognition accuracy of the automatic crack recognition system for subway tunnels designed in this paper reach a good level.In addition,the system realizes the automatic detection and location of cracks in subway tunnels by applying the latest target detection technology Faster R-CNN,and establishes a tunnel crack disease distribution database to analyze the deformation and crack of the tunnel through historical data comparison.Development trend,timely guidance to maintenance personnel on crack repair and maintenance.