Defect detection and identification in textile fabrics using Multi Resolution Combined Statistical and Spatial Frequency Method

In textile industry, reliable and accurate quality control and inspection becomes an important element. Presently, this is still accomplished by human experience, which is more time consuming and is also prone to errors. Hence automated visual inspection systems become mandatory in textile industries. This Paper presents a novel algorithm of fabric defect detection by making use of Multi Resolution Combined Statistical and Spatial Frequency Method. Defect detection consists of two phases, first is the training and next is the testing phase. In the training phase, the reference fabric images are cropped into non-overlapping sub-windows. By applying MRCSF the features of the textile fabrics are extracted and stored in the database. During the testing phase the same procedure is applied for test fabric and the features are compared with database information. Based on the comparison results, each sub-window is categorized as defective or non-defective. The classification rate obtained by the process of simulation using MATLAB was found to be 99%.