Design Of Remote Control And Name Plate Detection System For Waste Battery Storage Cabinet

Heavy metals and sulfuric acid waste liquor,two major components of waste batteries,pose a potential yet serious threat to natural environment and human physical health if improperly controlled during in storage.At present,the simple storage device universally adopted in the market has poor controllability,causing lower the storage efficiency and being difficult to complete real-time control.The information of used batteries simultaneously rests on so precise artificial cognition and recording that makes it difficult to fulfill machine detection and identification,in addition,the overall storage efficiency is alow.Aimed at the above problems,the remote control and nameplate detection system for the storage cabinets of used batteries is designed.Generally speaking,the remote control subsystem accomplishes the long-distance and local storage control functions of the storage cabinets as might have been expected.Moreover,the nameplate detection subsystem through adopting deep learning algorithm also fulfills the nameplate detection function.It is under these circumstances that implementation of each system function contributes to conveniently obtaining the crucial information of waste batteries to achieve the classification and statistics function,enhancing the efficiency of environmental protection supervision and acting as a driving force for the vigorous development of hazardous chemicals recycling industry.In the beginning,the overall system is designed and the key technologies are studied.The system is divided into remote control subsystem design and nameplate detection subsystem design.Then the remote-control subsystem is designed.This system is mainly composed of We Chat mini program,background server and equipment controller.The functions of storage control and image acquisition are realized by combining with MCU and other hardware.The client We Chat applet adopts HTTPS network transmission technology based on TCP protocol.Remote storage control includes "open door","view equipment number and address","get latest picture" functions.Local storage control includes "open door","on light","off light" and "inspection door" functions.Remote control subsystem realizes the overall control function of storage cabinet.Then,the name plate detection subsystem is designed.Aiming at the problems of low accuracy and poor real-time performance of traditional YOLOv3 target detection algorithm,the convolution unit layer network structure of Darknet-53 is improved,and the Inception-Darknet-53 network model is proposed.DIo U is replaced by Io U to make it more consistent with the target box regression mechanism.The improved YOLOv3 target detection algorithm is used to detect the three main information of "brand","model" and "voltage/capacity" in the nameplate.The comparative experiment is carried out by making data sets and using Keras framework to build the experimental environment.The experimental results show that the detection frame rate of the improved algorithm reaches 83FPS/s,which is5FPS/s higher than that of the original Darknet-53 network model,and the detection accuracy m AP is increased by 3.54%.The multi category average accuracy of 92.01% is achieved in the test set,which achieves good detection effect and meets the accuracy and real-time requirements of the detection system.Finally,the stability of the remote-control subsystem and the improved nameplate detection algorithm are verified by building the experimental device of the storage cabinet to test the system as a whole,and the remote-control and nameplate detection functions of the storage cabinet are achieved.