Design And Implementation Of Intelligent Waste Collection Supervision System

In recent years,driven by the rapid development of science and technology and rapid economic improvement,the production of garbage produced by residents has gradually increased,and with it,the situation of environmental problems has become increasingly serious.The state has introduced a variety of policies to address the problem of low garbage recycling efficiency,aiming to improve the efficiency of garbage recycling,and cities across the country have responded.At present,the effect of garbage recycling treatment is to cure the symptoms but not the root cause: garbage sites are still smelly,supervision is not strong,citizens lack knowledge of classification,weak awareness of classification,and even residents do not classify due to lack of time.In order to better improve the efficiency of garbage collection,this paper uses modern advanced technology embedded system design,artificial intelligence(deep learning,machine vision),and web front-end to develop an intelligent AI garbage supervision system for the problems of equipment that cannot be monitored in real time,weak supervision,and insufficient data statistics,and the main work is as follows:1.In view of the problem that the changeable environment of the garbage classification site cannot be monitored in real time,AI intelligence is used to monitor personnel and garbage separately.Based on the concept of low cost and high efficiency,the design idea of garbage collection supervision system is proposed.Through the selection of webcams,switches,and 4G industrial routers,the collection of personnel garbage information is completed;With the support of the motherboard,switching power supply and other equipment,the effect of the system on the processing of information is achieved.Equipped with industrial electric actuators,supervision speakers,RFID readers and other equipment,the overall hardware construction of the system is realized.2.Design and develop QT-based control system on PC and port it to the embedded Linux platform for operation.The thread coding method is used to detect personnel and ground garbage and voice reminder,realizing the intelligent function of equipment AI.Through the supervision and management of the violations of personnel in the garbage delivery process,the effect of AI supervision is achieved.3.In terms of handling personnel and garbage detection difficulties,the YOLOv4 deep learning network model is proposed,and Mobile Netv2 is selected as the backbone feature extraction network,and the image dataset is processed to realize the identification of personnel,ground and garbage cans.4.In the process of garbage collection,there is a problem of insufficient data statistics,this paper aims at this problem,using the Web front-end progressive framework vue to build a data management center platform,you can view different levels of delivery accuracy,self-classification rate,delivery participation rate and other statistical data,convenient for administrators and supervisors to compare the number of garbage drops,accuracy rate,self-classification rate and other indicators in different communities,through the display of the change trend of indicators such as the number of community delivery,accuracy rate,etc.,showing the effect of garbage collection supervision.