The Intelligent Detecting Technology Research In The Detect Of Floor Tile

China ceramics production enterprise is numerous, in national economy plays an important role. However, the current domestic tiles quality testing is still using artificial sorting way, this already limiting sorting accuracy, also influence the further reduce production cost, which seriously restrict the product class is improved and the ability to export the enhancement. Therefore, the development suitable for China's national conditions tiles defect detecting system, can satisfy our country to build TaoHangYe technical equipment modernization demand, and the China early become ceramic production power has important social meaning and practical value.This thesis based on image processing and pattern recognition technology and neural network technology, aiming at tiles and types of the characteristic, the major research the tiles achieve quality automatic defect inspection of the theory and method of specific algorithm. First, the image preprocessing, used image average method and median filtering method denoise target image, adopt threshold algorithm and gray-level projection algorithm extracte image effective area, eliminate image background, compared two algorithm advantages and disadvantages; Second, image feature extraction, respectively take three feature extraction algorithm, one kind is based on graylevel co-occurrence matrix feature extraction, based on difference matrix feature extraction, and put forward a new feature extraction algorithm, which based on gray symbiotic matrix and color features extraction algorithm, increased the accuracy of feature extraction, but increased time complexity; Finally, the application of advanced BP neural network method of tiles on the characteristics of the training and recognition.Finally using BP neural network trained and identification three kinds of feature extraction algorithm, through experiment results of data analysis, this paper based on graylevel co-occurrence matrix and color feature extraction algorithm in accuracy advantages. This algorithm in feature extraction and recognition process, the accuracy of the algorithm will be affected by the noise, light and other factors, those will reduced the robustness of the algorithms, so in the future research process, need the stability algorithms.