Design Of Train State Diagram Recognition System Based On Multiprocess

Railway is currently the most widely covered means of transportation.As its application level becomes more and more extensive,its traffic safety is the primary consideration.The detection of train status is currently the most widely studied direction in safe transportation.The section integrated monitoring system is an important system that guarantees the safety and maintenance of railway operation.If it fails,it will cause immeasurable casualties and economic losses.Therefore,it is very important to ensure the normal operation and maintenance of the section integrated monitoring system,but the failure data is not directly provided by the railway bureaus.Manufacturers mainly rely on staff to collect data on site,which is inefficient and wastes human resources.Secondly,equipment manufacturers cannot obtain fault data through wiring,so it is very necessary to use machine vision and artificial intelligence to detect train status.The main research contents of this thesis are as follows:(1)If the occupancy loss or failure occupancy alarm occurs,it may be an interval fault or a device failure.However,each railways bureau does not directly provide the fault data to the equipment manufacturer.When problems need to be analyzed,they need to report to the Telecommunication section.The staff need to copy the data on the site with few people,which is inefficient and waste of resources.Therefore,the system improves the way of obtaining data,uses CTPN and CRNN algorithm to identify the collected picture text content,and uses unsupervised image segmentation,Hoff algorithm and HSV space model to identify the graphics,It can meet the requirements of server,CPU,communication station,occupancy logic check and different section state detection of train status,and can judge the complete state of train based on the test results.(2)Because the requirements of network security are met,equipment manufacturers can not directly connect data lines to obtain the collected field data,but to realize network security is a huge systematic project.There is no relevant regulations and plans for monitoring the manufacturer network by Railway Corporation and all the railway bureaus.Therefore,the system adopts contactless one-way way to obtain data.The network security of the cloud host is higher than that of the traditional server,and it can run stably for a long time.Therefore,the cloud host platform is built to collect and transmit the data.Meanwhile,the train operation status map is uploaded in real time and brought into the algorithm to obtain the recognition results and saved and sent to the user,secure remote communication is realized.Through the analysis and research of the above problems,different sub interfaces and comprehensive interfaces are designed to identify the train status for each part of the image.For the requirements of non-contact detection and data transmission,the website design is carried out by Django in Python,and the website of multi-process train status detection is designed,And it is attached to the cloud host.The main goal of the system is to detect the running state of the train in the block area,and to store and enquire about the detection status in real time.The system can meet the needs of the equipment manufacturers for the train status detection,and has a strong practical value and significance.