The Detection Of Ceramic Tile Color Difference And Planncss Based On Improved Probabilistic Neural Network

As an important building ceramic material，Ceramic tile is widely used in the modern life. The method about improving the quality of ceramic tile products has became a major issue for the ceramic production enterprise. color difference and planness are the two important indexes to measure the quality of ceramic-tiles surface, but there has appeared no equipment or theory research which could detect these two kinds of index together. Therefore, searching for a scheme to suit automatic detecting of ceramic tiles color difference and planness inland actively, which have important practical significance.This paper has constructed algorithm model to test ceramic tiles color difference and planness by useing the knowledge system about improved probabilistic neural network (PNN). According to the reasonable design of detection algorithmmodel and the theory research for identification of ceramic tiles color difference and planness, as the particularity of ceramic tile testing field, through theoretical analysis and lots of experiments we achieved a effective probabilistic neural network structure, then we put forward the differential evolution algorithm to optimize the smoothing parameters, to improve the rate of detection correctness further. In the end, the paper presents a detection algorithm of the color difference and planness of ceramic tile based on improved PNN algorithm.