The Research And Development Of Passenger Monitoring System For Smart Landscape Based On Video Analysis Technology

The management of tourists is an important part of landscape security management and plays an important role in passenger density control,public safety and services.With the improvement of people's living standards,there is a great increase of tourists in the landscape,which causes more and more difficulties for tourist management.At present,artificial intelligence(AI)has been widely used in many sectors,such as subway security,financial payment and warehouse management,etc.AI has greatly improved the efficiency of many scenarios.An AI-based passenger flow management system can help a lot in the management of landscape,especially for the improvement of efficiency and reduction of cost.Based on video analysis technology,this thesis develops an intelligent system for landscape passenger flow detection,which realizes the detection of passenger density,key area monitoring,crowdintensive alarm,and so on.The detection of passenger density detects the number of tourists entering and leaving each entrance in real time,and generates temporal statistical reports.When the number of tourists reaches the threshold,the system will alarm the operator to take necessary limiting measures.The key area monitoring monitors the dangerous areas in real time,such as lakeside,cliff edge,electrical equipment,etc.When tourists enter the monitoring areas,the system will sound an alarm,so that the operator can deal with the emergency as soon as possible to prevent the possible occurrences of dangerous accidents.Besides these measures,the system will record and preserve the videos before and after the occurrence of dangerous actions.These videos can be used to make subsequent investigations when necessary.Crowd intensive alarm detects crowd density in real time by detecting all the enabled video streams in the landscape.Once the threshold of crowd density is reached,the system will sound an alarm and the image will be taken simultaneously.In this system,Open CV is used to read frames from the video stream,and YOLOv3 target detection algorithm is used to analyze the metadata of tourists' confidence and location in each frame,then Deep SORT multi-target tracking algorithm is used to generate each tourist's moving track,and Django framework is used to develop the functionality of passenger flow management.The system can help landscape managers to quantify the number of tourists,grasp the real-time density of landscape tourists,improve the landscape's ability to prevent potential hazards,and collect the behavior data of tourists based on the visual analysis technology through artificial intelligence,which provides data support for landscape managers to carry out scientific management and decision-making.