A Method of Image Texture Feature Analysis Base on Wavelet Decomposition and its Application

A method for analyzing the textural features of image is put forward by using the wavelet decomposition. That is used to measure color difference of ceramic tiles on-line. A linear array color CCD camera is used to grab the image of ceramic tiles, and then the image is pretreated. Two level wavelet decomposition is used on the ceramic tile image, and the energy feature of each sub-graph is extracted. The energy feature will be used as a feature matrix,and it can be carried out the classfication according to minimum distance classifier. It can reflect spatial relationship of color more effectively, when contrast with the color histogram, which is used as a traditional method. Experiments show that it is effective using wavelet analysis on image processing,and the sort results are approximately in accordance with the artificial detection.