Scheme Design Of Train Positioning System In Non Electrified Section Of Special Railway Line Based On Video

The railway special line section separated from the scope of centralized electrical control is called "non electrified section of railway special line".This section lacks electrification transformation and cannot realize centralized dispatching like the conventional electrification section.However,due to the needs of safety and management,the relevant railway departments cannot simply rely on manual confirmation of train on track status as before.Therefore,it is urgent to find a technical solution to solve the problem of non visualization of trains in the non Electrification Section of the special line.Based on the above purposes,the research contents of this paper are as follows:By analyzing and comparing the existing train positioning methods,starting from the actual situation,under the dual drive of optimal economic benefits and accurate and effective data,this paper finally selects the video-based positioning method,and designs the overall scheme.Generally speaking,the whole train visual positioning system is divided into three modules: bottom algorithm module,background system module and front-end display module.The train positioning task is divided into two tasks: target detection and target tracking.By analyzing and comparing the advantages and disadvantages of various deep learning algorithms,according to the application scenario of the algorithm and combined with the actual needs,YOLOv5 algorithm and Deep SORT algorithm are used for target detection and target tracking respectively.Finally,the position of the train in the image and video is mapped to the real space to realize the positioning.According to the overall scheme design,a suitable example analysis site is selected.During the specific implementation,firstly,according to the site conditions and data acquisition requirements,formulate the panoramic coverage deployment scheme of cameras along the railway,collect the original video information of the train,and transmit the collected information to the positioning system.The positioning system realizes the real-time detection and tracking of train position in railway scene based on YOLOv5 algorithm and Deep SORT algorithm,and transmits the structured data to the background system for storage.The background system transmits the data to the front-end visualization system to realize the visual real-time display of train position.This paper provides a new idea for railway transportation enterprise managers to solve the problem of train positioning and visualization in the non electrified section of railway dedicated line.It has practical application value in theory,and can be further explored and implemented in the overall idea of this paper.