A Hybrid Approach for Detection And Classification of Fabric Faults In Textile Industries Using Wavelet Transform And GLCM Technique

This paper presents a hybrid approach for the classification of faulty fabric, which is a challenging task in textile industries. In this paper, Discrete Wavelet Transform (DWT) is applied over the fabric image and texture features are extracted from the sub-band possessing maximum energy using GLCM (Gray Level Cooccurrence Matrix) technique. The DWT is used for decomposing the image into sub-bands because the features derived from the sub-band coefficients uniquely characterize the texture. Further, it reduces the computation time and resources required. GLCM method used for feature extraction here has proven to be a powerful basis for use in texture classification. Euclidean distance is used as metric to classify the fabric image as defect free or not. The performance of our proposed method SBCM (Sub-band domain Co-occurrence Matrix) is compared with that of SDCM (Spatial Domain Cooccurrence Matrix) method using classification rate as a measure.