Comparative Analysis of Fabric Fault Detection Using Hybrid Approach

摘要

This paper focuses on the fabric fault detection for variable sized textile images collected from textile industry. This paper presents the comparative analysis of Fabric Detection using hybrid approach where GLCM, Gabor Wavelet technique is used for image extraction and Random Forest Decision technique is used for image classification. The texture is observed as one of the utmost significant feature in the process of analysis of image and recognition of patterns. The incorporation of GLCM and Gabor Wavelet is being applied in order to obtain the best feature images of fabrics. The co-occurrence matrix has better processing effect for global region of images. Similarly, in attaining several level scales. Several level directional and native information in frequency domain Gabor Wavelet results are found excellent in performing the work. To categorize the defective and non-defective images into defective or non-defectiveness of the intended fabric image and in detecting the same the classification phase involves the Random forest classifier involved.