Research On The Appearance Detection Algorithm Of Terracotta Based On Machine Vision

The surface quality of terracotta products is one of the important components of its product quality.The quality of its surface has a great impact on its commercial and practical value.Traditional ceramic tile surface quality inspection mainly relies on manual inspection,but manual inspection has many disadvantageous factors,which is not good for the long-term development of ceramic tile production industry.With the continuous development and improvement of industrial automatic detection technology and detection theory,non-contact automatic detection technology based on machine vision has shown great advantages in the field of detection,which has been successfully applied to many areas of industrial production automatic detection.The use of machine vision inspection technology to automatically detect the appearance defects of ceramic tiles is a difficult problem to be solved in the industrial upgrading of the ceramic tile industry.In this paper,a systematic research is carried out on the defect detection algorithms for the appearance of ceramic tiles based on machine vision.Among them,the three types of defects appearing on the surface of ceramic tiles,namely the detection algorithms of lack of glaze,bulging,and cracks,are studied.The main tasks of this article are as follows:1)First of all,the background and significance of the topics studied in this article are described,and the current development status of machine vision inspection technology at home and abroad is discussed.2)Secondly,the methods of pre-processing and feature extraction of terracotta images involved in the detection of terracotta appearance defects are studied.In this paper,based on the actual appearance of terracotta defects,Gaussian filtering is used to remove the terracotta image noise;the project area is used to locate the target of the detection area;the color space conversion method is used to improve the contrast of the defects and the terracotta color classification;the extraction uses the geometric feature method.3)Furthermore,for different types of ceramic tile appearance defects,it is difficult to detect all types of terracotta defects by one detection method.Therefore,this paper designs corresponding detection methods for the glaze defects,bulge defects and crack defects on the surface of terracotta.For the detection of terracotta bulge defects based on a combination of custom sliding filtering and interpolation low-pass filtering is proposed;A glazed glazing defect based on color space conversion and iteration method combined with Otsu method is proposed for glazed glazing defects Detection algorithm;for multi-scale Hessian matrix filtering,double-window filter filtering,and sub-region detection for tiling crack defects,tiling crack detection is implemented.In the end,the experimental results show that the average detection accuracy of terracotta defects is over 93%,which can effectively recognize the appearance defects of terracotta tiles.