Research On Online Quality Control Method And System For Small And Precise Machines Based On Machine Vision

The industrial technology in today's society is moving into the industrial 4.0 era of information physical system.As one of the important criteria of for measuring a country's manufacturing industry,precision machinery products play an increasingly important role in the development of industrial technology.At present,most of the quality testing methods for precision mechanical products are traditional manual sampling inspection or mechanical repetitive teaching inspection.These testing methods are inefficient and slow and has become less and less consistent with the rapid pace of contemporary production.In recent years,mechanical product detection based on machine vision has become the focus of many scholars because of its good stability,high reliability,fast speed,etc.It is also one of the important directions for the development and research of the future detection field.In this paper,hydraulic torque converter is used as the research object,which aims to solve the object identification and the route planning mechanical arm of screw hole in the process of quality testing of hydraulic torque converter.Firstly,the research status of machine vision technology is expounded.Based on the current vision inspection technology and the deep learning theory framework model,a deep learning machine vision detection method based on Fast R-CNN model is constructed.Secondly,on the basis of the product type and the position of the screw hole which are obtained through the visual detection,the application objects and the advantages and disadvantages of the traditional path planning method and the intelligent path planning method are analyzed,a method of dynamic detection path optimization based on Dijkstra's algorithm method is proposed,and the method is verified in two-dimensional plane and three-dimensional space respectively.Finally,according to the guidance of the above theoretical methods,the software and hardware architecture is analyzed,an on-line quality control system for small precision mechanical products based on machine vision is designed and developed,the system architecture and operation logic are elaborated,and the effectiveness of this system is verified,which provides more extensive solving ideas and technical support for the quality testing of small precision mechanical products.