The color classification of wall and floor tiles is the final but important procedure in the whole process. Due to the technology and some other reasons, this task is still done by human inspectors at present. And there exist series of problems in manual operation due to the subjectivity. Therefore developing an automated color classification algorithm is of great importance, which can effectively resolve the problems brought by manual activity, and can help to improve the full automation in the tile industry and advance the development of automated color recognition technology in other related industries.To acquire the accurate and stable tile images in real time is the basis for the whole classification. Starting from the task requirement and design principle, some key problems such as the design for the illuminating system and the selection of the CCD camera and frame grabber are explained and the system hardware configuration is given also. Then the development and realization of the whole classification algorithm, including the image noise analysis and denoising method, the extraction of the target image, the feature analysis of color texture images and the design of the classifier, is introduced from the viewpoint of the software. The main research work and important results are listed below:1. The system configuration is designed after an analysis of the system task and a comprehensive investigation are conducted. The whole task includes the design of the illuminating system, the selection of the CCD camera and framer grabber.2. The denoising of wall and floor tiles is analyzed and discussed from the viewpoint of theory and experiment. And an effective method to evaluate the performance of denoising method is proposed. The result of the edge detection conducted by the differential operator, in the combination of the evaluation by eye and performance index, is proved to work well in denoising performance evaluation.3. Two ways to extract the tile image from the background are given based on the analysis of tile image characteristics. The applications of edge extraction based on projection and area segmentation based on gray level thresholding to the tile image extraction are discussed. The difference projection is chosen for tile image extraction and some problems and measurements, which might be taken, is pointed out to be worthy of notice.4. The co-occurrence matrix and difference matrix are used to extract the color texture features of wall and floor tiles. The co-occurrence matrix depicts the spatial dependency between different gray levels, while the difference matrix, which is derived from the image based on the conception of the difference, depicts the texture features by the gray variation in the neighborhood of the pixel.