Design And Implementation Of Intelligent Inspection Software For Workpiece Defects

Under the trend of global industry 4.0,many advanced manufacturing technologies are becoming mature,especially the landing application of robotics,Internet of things,cloud computing,big data and other technologies,providing people with more and more high-quality products and services,but at the same time,the requirements for product quality are getting higher and higher.More and more manufacturers put product quality into a more important position,and the workpiece defect detection efficiency and accuracy is an important factor restricting the overall quality of the product.In use of workpiece products,workpiece defect problems seriously affect the system and the working state of the equipment as well as the life,while the industrial CT(Computed Tomography)imaging technology can ensure the workpiece structure on the premise of not damaged,show the inside of the work piece image details effectively,namely to nondestructive testing of workpiece internal defects,on the basis of the condition by the staff through the study of the artificial recognition of CT image so as to achieve the purpose of the extraction and classification of internal defects in the workpiece.In this paper,based on the requirements of the project,the intelligent detection software of the workpiece defect is implemented,which can assist the manual to realize the efficient extraction and accurate classification of the workpiece defect.The main contents of this paper are as follows:(1)Through the practical problems encountered in the work,translate the requirements and determine the intelligent detection target function of the software.The main functions of this software include image preprocessing,defect extraction,defect feature selection,defect sample collection and classifier training,and realization of defect classification function.(2)Based on the morphological processing of CT image and the pre-processing of noise removal points,the adaptive threshold segmentation method was used to extract defects from the image.(3)Support Vector Machine(SVM)is used to identify defects.It can better solve common problems such as small samples,nonlinearity,high dimension and local minima.The one-to-one multi-classification defect recognition method is adopted to realize defect classification.(4)Program each functional module of the software,build the software interface based on QT5.7 on the platform of Microsoft Visual Studio 2015,and implement the image related processing algorithm based on Open CV.The Python programming language is used to train the classification algorithm,and the software is called to realize the intelligent detection and classification function of the workpiece defects.