Fabric Texture Analysis Using Computer Vision Techniques

• Published in: IEEE transactions on instrumentation and measurement Vol. 60; no. 1; pp. 44 - 56
• Main Authors: Xin Wang, Georganas, N D, Petriu, E M
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.01.2011; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Computer vision; Counting; Fabrics; Fractal analysis; fractal dimension; Fractals; fuzzy c-means clustering (FCM); grey level cooccurrence matrix (GLCM); Pattern recognition; Principal component analysis; principal component analysis (PCA); Rough surfaces; Roughness; Studies; Surface layer; Surface roughness; Textile industry; Texture; texture analysis; Weave; Weaving; woven fabric; Yarns
• ISSN: 0018-9456; 1557-9662
• DOI: 10.1109/TIM.2010.2069850

This paper presents inexpensive computer vision techniques allowing to measure the texture characteristics of woven fabric, such as weave repeat and yarn counts, and the surface roughness. First, we discuss the automatic recognition of weave pattern and the accurate measurement of yarn counts by analyzing fabric sample images. We propose a surface roughness indicator FD FFT , which is the 3-D surface fractal dimension measurement calculated from the 2-D fast Fourier transform of high-resolution 3-D surface scan. The proposed weave pattern recognition method was validated by using computer-simulated woven samples and real woven fabric images. All weave patterns of the tested fabric samples were successfully recognized, and computed yarn counts were consistent with the manual counts. The rotation invariance and scale invariance of FD FFT were validated with fractal Brownian images. Moreover, to evaluate the correctness of FD FFT , we provide a method of calculating standard roughness parameters from the 3-D fabric surface. According to the test results, we demonstrated that FD FFT is a fast and reliable parameter for fabric roughness measurement based on 3-D surface data.