So Yongguk, Kim Jongchol, Hwang Hyok
J Opt Soc Am A Opt Image Sci Vis. 2020 Jul 1;37(7):1229-1235. doi: 10.1364/JOSAA.391317.
Recently, fabric defect inspection techniques have received attention in textile production procedures, since demands for various textile fabrics are growing. However, visual inspection for fabric defect detection is a very difficult problem because of the complexity of the fabric pattern and various defects. In this paper, we propose a method to detect the defects in fabric surfaces using the hybrid Particle Swarm Optimization-Gravitational Search Algorithm (PSO-GSA) and ellipse Gabor filter (EGF). In the proposed method, the hybrid PSO-GSA been employed to optimize the parameters of the EGF. Gabor filter parameters for the texture of the nondefective fabric images adjusted via the hybrid PSO-GSA with good convergence and solution characteristics. The defective fabric image is convoluted with the selected optimal Gabor filter, and we generate binary images by thresholding processing. The proposed method uses only one optimal filter, so fabric defect inspection is faster and more cost effective. Experimental results show that the proposed method is robust and achieves accurate detection of fabric defects.
近年来，由于对各种纺织面料的需求不断增长，织物缺陷检测技术在纺织品生产过程中受到了关注。然而，由于织物图案的复杂性和各种缺陷，通过目视检查来检测织物缺陷是一个非常困难的问题。在本文中，我们提出了一种使用混合粒子群优化-引力搜索算法（PSO-GSA）和椭圆伽柏滤波器（EGF）来检测织物表面缺陷的方法。在所提出的方法中，混合PSO-GSA被用于优化EGF的参数。通过具有良好收敛性和求解特性的混合PSO-GSA来调整无缺陷织物图像纹理的伽柏滤波器参数。将有缺陷的织物图像与选定的最优伽柏滤波器进行卷积，并通过阈值处理生成二值图像。所提出的方法仅使用一个最优滤波器，因此织物缺陷检测速度更快且更具成本效益。实验结果表明，所提出的方法具有鲁棒性，能够准确检测织物缺陷。
