Research On Intelligent Detection And Control System Of Three-dimensional Warehouse

With the development of the Internet,the logistics industry has also risen rapidly.Due to the limitation of space and labor,the storage of products has become an important issue.In recent years,the development of research fields such as programmable controller and stacker has promoted the development of automated warehouse and effectively solved the storage problem of products.In addition,a few enterprises in industrial production have realized the integration of product warehousing and intelligent machine vision inspection.The integrated assembly line operation can not only realize the storage and delivery of products,but also accurately and efficiently detect product defects,greatly improving work efficiency.This dissertation uses SFI4.0-MINI industry 4.0 equipment and STEP7 software and image processing software to design an intelligent detection and control system based on three-dimensional warehouse.The system is mainly divided into two parts: One is the design of automatic stereoscopic warehouse,the other is the detection of gear size and defect by intelligent camera detection system,and the use of deep learning to realize the target detection of gear defect.The main contents of this dissertation are as follows:1.Siemens S7-200 SMART ST30 is used as the core controller of the automated warehouse,and the relative addressing method is adopted to make the target accurately in place;Its stacking car module is composed of servo motor,reducer,X-axis and Z-axis linear modules and pneumatic gripper.The PLC is programmed by using the sequential function diagram method,and the operations of stacker’s delivery and warehousing are realized.2.The monitoring system of the upper computer is designed by using the configuration software of Kunlun General MCGS,and the stacker’s outbound and inbound operations are realized through the upper computer;Select manual and automatic control operation;Pneumatic gripper action mode and display operation status.3.Download the ladder diagram to PLC and the configuration interface to MCGS running environment,and debug the three-dimensional warehouse system as a whole.The goal of controlling the operation of the stacker by the upper computer has been achieved,and the expected requirements have been met.4.Acquisition of gear images.The tray loaded with gears is transported by the conveyor belt to the intelligent camera inspection station and the double telecentric vision inspection station.The camera,the double telecentric lens and the annular light source are used to take photos automatically to acquire gear images.5.Use Vision Bank visual inspection software to measure the size and detect the defects of gears.After preprocessing,feature positioning,circle positioning,circle detection,patch detection,counting and patch area of the collected gear images,the diameter of the top circle and the diameter of the bottom circle of the gear are measured.Finally,the template inspection tool is used to detect whether the gear has defects.6.Four classical neural network models,Alex Net,VGG-16,SSD and YOLOv3,are used to detect the defects of the collected gear images based on deep learning,and the four detection results are compared.The results show that YOLOv3 has higher detection accuracy for small targets.