Design And Implementation Of Surface Defect Detection Algorithm For Unfixed Shading Sanitary Products

Sanitary products are essential daily necessities for people in daily life.With the development of industrial automation,people are paying more and more attention to the quality of sanitary products.Since Sanitary products are generally mass-produced in machines,it is inevitable that some surface defects of indentations,stains,and wrinkles may occur during the production process.For the detection of surface defects of sanitary products,the method of manual visual inspection is currently adopted,which is slow,subjective and inefficient,and is increasingly unable to meet the needs of enterprises.With the introduction of machine vision technology in the 1980 s,it became possible to use machines instead of manual inspection of surface defects in sanitary products.This paper is based on the actual needs of industrial surface defects detection of sanitary products.With Haclon as the platform,a detection algorithm for detecting unfixed shading flexible sanitary products is designed to meet the real-time and rapidity of industrial surface defects detection.And high accuracy requirements.The main research contents of this paper are as follows:(1)Proposal of the overall plan.In order to meet the requirements of high speed and high precision in industrial inspection,the overall design of the system is constructed for the characteristics of images and surface defects of sanitary products.At the same time,the hardware structure of the system is analyzed,which provides the necessary hardware foundation for image processing in the future.(2)The design of the detection algorithm.Aiming at the characteristics that the shading of sanitary products is not fixed and the flexible materials are easily affected by deformation,a sub-area detection algorithm is proposed.In the detection algorithm,the collected sample image is first preprocessed,and then the sample image is divided into an edge region,a cotton pad region and a cotton core region by using a threshold segmentation method in the sample image,different mathematical regions are processed in different regions after segmentation.Then the segmented image processed by mathematical morphology is designed according to its different regional features to detect the defects of the image.Finally,the image defects are extracted by using the expansion,corrosion and connectivity of the region.The algorithm is programmed with the halcon operator of MVtec of Germany,which can well identify defects such as product wrinkles,edge stains,cotton pad abnormalities,cotton pad stains,cotton core anomalies,cotton core stains and the like.(3)System design and analysis of experimental results.With VS2012 as the platform,the interface of the surface defect detection system of sanitary products was designed to realize the on-line detection of surface defects of sanitary products.Finally,through a total of 500 sample images with a resolution of 2048x1415 for three sets of images,The feasibility and stability of the surface defect detection algorithm proposed in this scheme are analyzed.Finally,the feasibility and stability of the surface defect detection algorithm proposed by this scheme are analyzed by testing a total of 500 BMP images with a resolution of 2048x1415.In the experimental stage,the average accuracy of the surface defect detection algorithm of the sanitary product is over 95%,and the average time for detecting an image is less than 100 ms,which can meet the requirements of industrial real-time.