Research And Implementation Of Surface Defect Detection Algorithm Of Sanitary Ware

Sanitary products are products that people often encounter in daily life,they are required to carry out strict quality testing during the production process.Surface defect detection in quality detection is an important aspect.The traditional detection method,namely the artificial detection method,has the disadvantages such as high time consumption,high labor cost,high subjectivity,lack of continuity and detection results.Unable to meet the high speed detection requirements of industry,machine vision instead of manual detection has become a new trend.This paper is based on the practical requirements of industrial site,explores the surface defect detection algorithm of sanitary ware,and constructs the surface defect detection system of sanitary ware.This article main research is as follows:Firstly,in order to achieve industrial testing strict requirements of high speed and high precision,at the same time for sanitary products in acquisition,the characters of surface defect images and build the overall design scheme of the system,and analyses the system hardware construction,for the later image processing provides a necessary material base.Secondly,For sanitary products in the industrial test of the product the change of the position,put forward in the standard image pop a characteristic information of the most abundant one small area,the area for subpixel contour extraction,the characteristics of the pixel outline and it as our template,at the same time can get its coordinates;Using the feature template to search in the image to be detected,find matching area,and obtain the coordinates of the matching area,according to the coordinate calculation of the two affine transformation matrix,to detect the image affine transformation,to implement the orientation between the two images.Experimental results show that this method effectively improves the function of image positioning.Thirdly,According to the characteristics of surface defects of sanitary products,different methods are studied in different aspects.Fast Fourier transform FFT detection method,MLP detection method and differential detection method are used in frequency domain space.The characteristic parameters of each method and the effect of these parameters on the detection effect are studied.At the same time,according to the characteristics of the sanitary products texture image,the difference method can't separate the dirty point features such as the size and location for effective identification,so of the processed image is difference and connected component method such as image segmentation,morphology and processing,finally through the stain defect parameters for judgment of the defects in the image of the circumference.Then use the circular box to mark the area.Compared with other methods,the experimental results show that the difference test achieved good detection effect,can satisfy the requirements of the industrial test on the speed and accuracy,and also to verify the effectiveness of the proposed algorithm in this paper.Finally,The surface defect detection system of sanitary products was set up,and the surface defect detection system of sanitary ware was constructed.Experiments show that the speed and precision of the system can meet the requirements of industrial inspection.