Deep learning neural networks for knitted fabric defect identification and classification

摘要

Production and quality of many engineering industries were improved by using artificial neural network. In textile industries inspection of fabric takes more time and labours. In order to reduce the manpower and time, computerized inspection using image processing in deep learning is used. It gives the accurate result in less time, and increase the production. The convolutional neural network in deep learning is used for image processing. The high quality images is given as input, then the image features were used to train the neural network for automatic fabric defect detection. The image processing used to identify the cotton knitted fabric defects such as holes, patches, stain and seam joint. The sample images were collected from the Sky Cotex India Private Limited, Tirupur, India and were processed in Teachable Machine, a web-based tool that makes creating machine learning models fast, easy, and accessible to everyone. There is an input layer, "n" number of hidden layer and output layer. The trained neural network was used to test with the defective cotton knitted fabric sample images and the efficiency was calculated as 89.8%.