<![CDATA[This paper proposes the development of an electronic device that integrates image processing algorithms for quality control of woven fabrics by detecting three defects: holes, bumps, and stains. Although classic and deep learning-based image processing methodologies for detecting fabric defects exist in the literature, these proposals lack hardware integration with the software. The proposed device overcomes this limitation by integrating digital image processing algorithms implemented in Python on a single-board computer, along with control devices for the camera, lighting, and motor, to enable real-time fabric quality control. This adaptable architecture meets the needs of the textile industry and reduces the risks associated with manual inspection. The device was validated with 100 fabric samples evaluated by a textile engineering specialist, yielding a Cohen's Kappa index of 0.959 for light-colored fabrics and 0.869 for dark-colored fabrics, as well as an average inspection speed of 15.68 m/min.]]>
Copyrights © 2026
