Research On Online Detection Method Of Ceramic Tiles Surface Quality Based On Multi-feature Fusion

Chinese ceramics show a continuous development since very early in the history,during which Chinese ceramic technology has improved greatly.Today,China has become a world leading ceramic producer.In recent years,many Chinese ceramic manufacturers have started their business in international market.To increase the international competitiveness of the Chinese ceramic enterprises,it is crucial for those enterprises to improve the quality of their ceramic products.A very important step to assess the quality of ceramic tiles is to examine geometry,size and any defects of the surface.At present,most of the ceramic production enterprises still manually inspect the quality of ceramic tiles,while often times,it could be inefficient or inaccurate to do so,which leads to poor improvement over the quality of the ceramic tiles.In recent years,however,with the development of image capture technology and image process technology,machine vision inspection has dominated the ceramic industry.Machine vision inspection not only is flexible and steady,but is easily manageable and economical as well.Currently,Machine vision inspection is becoming most popular method to inspect the quality of ceramic tiles.Because of this trend in ceramic quality assessment,an online detection method,which uses the Machine vision inspection technology that targets ceramic tiles,to assess the surface quality of ceramic tiles with multi-feature fusion technology is proposed by this paper.To meet the criteria for parameter of ceramic tile assessment,the following parts of the research are proposed by this paper:(1)Construction of on-line inspection hardware test platform for ceramic tile surface qualityThe hardware platform to evaluate the quality of ceramic tile surfaces is created with the help of LED lighting,framing CCD camera and other associated facilities in this paper.This platform is capable of capturing images of the surface of ceramic tile surfaces,as well as processing the images captured by the platform.(2)Research and design of on-line detection algorithm for ceramic tiles Surface qualityThe design of the algorithm to evaluate the quality of the ceramic tile surfaces based on analyzing criteria and the parameters of ceramic tile surface evaluation is proposed by this paper.The design of the algorithm is completed through two sections – the algorithm to measure the surface dimensions of ceramic tiles,and the algorithm to detect the defects on the surfaces of ceramic tiles.The algorithm to measure the surface dimensions of ceramic tiles is designed with adapted Hough Transformation and TMOA Algorithm(Tuo's Max of Acme Algorithm),and is capable to quickly measure the length,thickness and diagonal line of a ceramic tile online.And for the algorithm to detect the defects on the surfaces of ceramic tiles,multi-feature fusion technology is the core.This algorithm utilizes defect detector with the adaptive fuzzy interference system to accurately tell if the ceramic tiles from the five most common defects.(3)Design and development of the online detection software to evaluation the quality of ceramic tilesIn the Visual Studio 2010 software development environment,the C++ language combined with MFC framework and multi-threading technology to complete the design of on-line inspection software system for ceramic tile surface quality is used in this paper.With the user interface design for the human computer interaction,the algorithms mentioned in previous paragraph,and data acquired from the current database,this online detection method is capable of measuring dimension and detecting defections of ceramic tiles,as well as memorizing and displaying the information of ceramic tiles.The success to measure the dimensions and to detect deflections of ceramic tile surfaces through detailing the creation of the online detection hardware platform,researches on algorithms for quality inspection,and development of the software is established by this paper.The result of the research shows that with the proposed online detection method,the measured dimensions are accurate within ±0.1mm and the five most common deflections can be recognized accurately,which satisfies the expectations for evaluation of ceramic tile surface.