Warehouse Smoke Detection And Alarm System Based On Machine Vision

In recent years,warehouse fires happen frequently,one of the reasons is that the warehouse management staff can not be at the warehouse site 24 hours for supervision;On the other hand,the warehouse is prone to smouldering,and the traditional smoke sensor cannot quickly generate smoke,which eventually leads to the spread of fire.There are all kinds of goods stored in the warehouse.The fire accident caused by the warehouse will damage people’s property and even cause casualties.Therefore,based on this background,through the collection and display of various parameters in the warehouse and the cloud and local storage of data,this study uses an image recognition program to detect the smoke in the early stage of the fire in the warehouse,and sends out an alarm at the early stage of the fire.A fully automatic warehouse smoke detection and alarm system with less manpower and centralized management has been built to make contributions to reducing warehouse fire accidents.This paper firstly analyzes the design of warehouse safety system at home and abroad,and according to the "Code for Design of automatic fire Alarm System" determines the various kinds of sensors needed in the whole system and their setting requirements;Secondly,choose to use PLC as a reliable logic control,through the sensor collected all kinds of data,PLC according to the designed ladder diagram to control the alarm system;Secondly,in view of the smoke produced in the early stage of fire,the image recognition program is used to detect the smoke,which is integrated with the traditional sensor detection.Finally,a host computer program written by Python is used to display and store all the collected data locally.Through the connection with the cloud server,the function of data uploading is realized,and the web page provided by the cloud can be used for real-time detection of the situation in the warehouse.This set of detection and alarm system in the process of operation can achieve the whole process of unmanned automatic operation,image recognition program can replace the human eye for the early fire smoke detection,completely solve the warehouse management personnel due to fatigue and other reasons,failed to find smoke caused by fire accidents.The work of this study shows that the smoke detection and alarm system based on image recognition can effectively supervise the warehouse,and the upper computer program can realize the data display,local storage,data upload and other functions.The image recognition program can detect whether there is smoke in the warehouse 24 hours a day without interruption;In the event of an accident,PLC will automatically alarm.The whole system can help managers to supervise the warehouse and ensure the safety of the warehouse.