Automatic recognition of woven fabric pattern based on image processing and BP neural network

As there are error judgments of float type in the traditional method based on image processing, it is hard to determine the woven fabric pattern from the recognition results. To solve this problem, fuzzy C-means (FCM) algorithm was selected to classify the floats into two groups in the experiment, and BP neural network is chosen to recognize woven fabric pattern. White-black co-occurrence matrix is used to extract its texture features. The texture and structure features of the normal fabrics extracted from the classification are input into the neural network to complete the learning process. During the recognition process, the texture features of the fabric are extracted from the classification results with white-black co-occurrence matrix. The structure features are extracted simultaneously. These features are then input into BP neural network and woven fabric pattern would be output from the neural network. The experiment on actual fabrics proves that the method proposed in this study has fault tolerant ability, and it can recognize fabric patterns correctly.