| With the rapid development of railway informationization,the demand of automatic data acquisition and analysis system is increasing.The research of locomotive surveillance video automatic dump and analysis system is one of the important directions of Railway Informationization research at present.Railway locomotive safety protection system is a necessary on-board protection system for every locomotive,in which the automatic video monitoring and recording subsystem of locomotive records the key monitoring video information.But at present,video information is stored in the 6A host of locomotive,and can only be acquired and analyzed manually.According to the current situation that the locomotive monitoring video data is transferred and analyzed manually,the locomotive monitoring video automatic dump and analysis system is completed through the following research.Firstly,a special equipment for video acquisition of locomotive monitoring is designed and implemented.The equipment is used to acquire the monitoring video data of locomotive 6A host computer.By studying the USB data interface of video data monitored by 6A host,the development board based on ARM processor,with USB OTG data interface system and using Linux operating system is selected as the device motherboard.Zero W and Firefly RK3128,which are representative of the development of raspberry pie,were studied and developed respectively.The automatic data acquisition program in the device can connect the USB OTG data interface of the control device with the USB data interface of the 6A host computer by controlling the USB driver module program in the Linux kernel.It can automatically acquire the locomotive monitoring video data and store it in the on-board automatic dump industrial control computer to prepare for the subsequent automatic dump.Then,the client program and the server program of the ground server are designed to realize the automatic dump of the locomotive monitoring video when the locomotive enters the locomotive depot for preparation.In order to solve the problems of real-time and reliability of message communication in automatic dump,RabbitMQ message bus is used as the medium of connecting client program and server program.The communication protocol based on JSON format is designed to realize the conversation between client program and server program,disconnect and reconnect,overtime and retransmit,so as to ensure the reliability of the automatic dump system.Video files are transmitted using HTTP transport protocol to realize the function of breakpoint continuation.MD5 data validation is used to ensure the correctness of video data after transmission.Secondly,in order to solve the bottleneck problem of wireless network speed in data transmission stage,a multi-channel parallel transmission gateway system is designed and implemented to improve the transmission speed of video data.With the rapid development of mobile devices,the speed and quality of wireless networks are increasingly demanded by industrial applications.However,the current development of wireless networks is slow,far from meeting the actual needs.Therefore,aiming at the bottleneck of transmission speed and quality in wireless networks,a parallel gateway system using multi-channel wireless transmission is proposed.The system uses a pair of parallel gateways to establish a multi-channel wireless connection to realize the parallel transmission of Ethernet frame data,so as to improve the transmission speed of wireless network.In addition,UDT protocol is used to transmit data between parallel gateways to improve the quality of data transmission.Finally,a video behavior analysis software is designed and implemented.After the completion of video data transmission,the driver can automatically determine whether there is a violation of the driving seat during the driving process.Video from the surveillance room is directed at the driver’s seat,and the driver is detected whether he appears in the video to determine whether the driver leaves the seat or not.Using image similarity comparison technology,the driver is detected in the frame image.Detecting the first frame image of every second video,generating the decision value p,p value 0-1 binary data,p 1 means that the frame image is unmanned.When P is 1 and lasts for 15 video frames,it is judged that the driver has violated the rule of leaving the driver’s seat.The four modules of locomotive monitoring video capture,automatic dump,multichannel parallel transmission and video behavior analysis constitute the locomotive monitoring video automatic dump and analysis system. |
PDF Full Text Request