I. Introduction
Fibers and yarns create a stable relationship through specific crosses, knots, or links to form the fabric. In the production of the textile industry, fabrics are not only a primary material for the production of clothing but also other decorative and industrial categories of raw materials for reprocessing[1], Among them, the textile raw material class of solid waste identification and classification of the textile industry has great significance. At present, the traditional recognition method of solid waste of raw textile materials is carried out manually. However, as the time cost increases, the errors generated manually will rapidly increase, and the results are subjective. In addition, limited by the physiological characteristics of the human eye, recognition is slow, time-consuming, and challenging to meet the rapid requirements of modern textile industry need. Manual annotation has disadvantages, such as being time-consuming, fatigue- prone, and subj ective[2]. Therefore, the automatic identification and classification method of images based on machine learning is introduced into the image classification of textile raw material type solid waste to achieve the automatic recognition of textile raw material type solid waste[3].