Intelligent Recognition System Of Safety Helmet And Vest Based On YOLOv4

In the safety management of the construction site,safety helmets and reflective vests can greatly reduce the safety risks in the construction process.Safety helmets can greatly reduce the potential safety hazards caused by falling objects.Reflective vests can also remind vehicles on the site and reduce traffic accidents.However,there are often dangerous behaviors that operators do not wear safety helmets and take off reflective vests on the site,which brings great potential safety hazards to the safety management of the construction site.It is difficult to reduce these behaviors only by relying on the safety officer to patrol on each construction site.Therefore,there is an urgent need for a method that can automatically identify dangerous behaviors to improve the safety management of the site.In terms of technical scheme,firstly,more than 1600 photos were collected in a project in Jinan,and VOC data sets containing four types of behaviors were constructed.The algorithm model with a correct rate of about 84%was trained based on yolov4 algorithm,which can meet the use of the system;Secondly,QT is used to build the identification client and optimize the algorithm at the business level.Compared with the traditional edge computing equipment,it not only meets the normal use of users,but also has lower requirements for system performance.In terms of system implementation,the background management system is built through spring cloud to receive and identify the violation information transmitted by the client,and the front-end management system is realized by using typescript and Vue,which is responsible for the display of front-end pages and data statistics and analysis.Finally,a cross platform mobile application is built through Cordova framework,which is responsible for receiving violation information and reminding security managers.On the basis of in-depth understanding of the needs of users,through the above technical solutions,a set of intelligent identification and monitoring system for helmets and reflective vests is provided for users,which can effectively help users improve the safety management of the project.