Research On Detecting Method Of Surface Defects In Tiles

With the improvement of living standards,people's requirements for home decoration are getting higher and higher and the requirements of the living environment have promoted from the most basics to meeting the needs of comfort and quality.It needs to be designed well,beautiful and set practicality and aesthetic in one.Tiles in the home decoration occupies a very important position.Not only the floor but also the wall will use tiles to decorate,so the quality of the tile itself and the quality of the paste will directly affect the quality of housing,overall beauty and the function.The size of tile surface flaw and how much directly affect the appearance of tiles,at the same time it determines the quality of tile and tile quality classification of an important indicator.In this paper,the tiles surface defect detection system can be used for on-line monitoring of tile production enterprises,so that enterprises can monitor the production of tiles in real time.This paper studies the tile surface defect detection system based on the theory of digital image processing and pattern recognition.In this paper,the implementation of the algorithm mainly includes image preprocessing algorithm,template image and tile image difference algorithm,tile defect identification algorithm and the calculation of the size of the algorithm.Through the introduction and comparison of the algorithm in each image preprocessing step,the image preprocessing algorithm of the subject system is designed.After the linear transformation and the piecewise linear transformation are discussed,the RGB color tile image acquired by the CCD industrial camera is grayed out by using the piecewise linear transformation in the preprocessing of the image,and then the k-means clustering algorithm which is more suitable for large target location is acquired to locate target and then segment it by OTSU threshold segmentation algorithm.By comparing the grayscale stretching and histogram equalization,the final use of grayscale stretching is to overcome the light changes.The algorithm of OTSU threshold segmentation is used for further deal with the tile image and prepare for the median filter.By discussing the commonly used mean filtering and median filtering,combined with the characteristics of tile image,the median filter is chosen to filter the denoising of the tile image.In this paper,the traditional "matching-difference algorithm" is improved.After preprocessing the template tile image and importing the tile image,and the tile image to be measured is directly determine whether the tile to be tested is defective and the location of the defect according to the output obtained after the difference and then identify the type of defect by the ratio of the perimeter and the area of the defect.Finally,the parameters of the number of defects,perimeter,area and coordinates of the center of the tiles are calculated by the connected domain analysis algorithm,and the tiles are classified according to the tiles classification standard.In the graphical user interface,by testing the tile image about the surface defect detection process,we can be very intuitive in the interface to observe the detection results.In the study of this subject,there are two main factors influencing the quality grade of the tested tiles:the number of flaws in the tiles and the size of the flaws in the tiles.Through using a large number of experimental samples,combining with the actual requirements of industrial pipelines,we optimize the relevant algorithm.Finally,the improved "fast-difference algorithm" improves the correct rate and detection rate of the surface flaw detection of the system.In this paper,the graphical user interface with complete test results is given in the system,which verifies the feasibility and efficiency of the system.At the end of this paper,the paper summarizes the shortcomings of the tile surface defect detection system,and puts forward the prospect of developing the next generation of tile surface defect detection system.