Safety Monitoring And Implementation Of Pedestrian Recognition And Locomotive In Run For Unmanned Transportation Of Molten Iron

With the overall promotion of intelligent manufacturing strategy and the improvement of equipment and technology in China’s iron and steel industry,steel plants will put forward higher requirements on the efficiency of transportation of molten iron.The level of automation and intelligentization of the transportation of molten iron will be the key factor to promote the efficiency of production.Unmanned transformation of the transportation of molten iron can improve the efficiency of production,reduce the participation of staff and reduce the costs.However,compared with the traditional manned transportation of molten iron,the unmanned transportation of molten iron also brings problems of pedestrian safety and locomotive safety.Based on the unmanned transformation project of molten iron transportation in a domestic steel plant,this paper studies the technology of safety monitoring of pedestrian recognition and locomotive in run according to the process and characteristics of molten iron transportation system.In the aspect of person safety monitoring,this paper uses the methods of perspective transformation and sliding window to realize the detection of railway tracks in images,and divides the railway encroachment area.Then,the YOLOv4 detection model is used to train on public data set and self-built data set,and the effect of accurate pedestrian recognition is achieved.Finally,whether to send an alarm and brake the locomotive or not is decided on the position relationship between pedestrians and the railway encroachment area.In the aspect of locomotive safety monitoring,this paper uses Kalman filter to process the locomotive position signal.The railway lines for the transportation of molten iron are discretized and segmented topologically.A mapping rule between locomotive and railway is proposed to match the position information of locomotive to the railway lines.According to the matching position in the railway and driving plan of the locomotive,the railway section reservation algorithm and collision avoidance mechanism are designed to realize the safety control of multiple locomotives in different scenarios.Based on B/S framework,this paper divides the safety monitoring subsystem into three layers: the underlying data layer,the data processing layer and the front display layer.This paper completes the design of the function modules of pedestrian recognition and locomotive safety monitoring,and carries out the selection and design of database and data transmission.Finally,the safety monitoring subsystem of pedestrian recognition and locomotive in run is completed to ensure the safety of the unmanned transportation system of molten iron.