基于双边滤波和Frangi滤波的织物疵点检测方法

A Fabric Defect Detection Method based on Bilateral Filtering and Frangi Filtering

为提高织物疵点检测效率,提出了一种基于双边滤波和多尺度线性滤波(Frangi)相结合的织物疵点检测方法。首先,采用双边滤波对织物图像进行处理,以减少织物背景纹理信息对检测结果的影响;然后,应用多尺度线性滤波器对经双边滤波后的织物图像进行二次滤波,增强织物疵点区域信息;最后,使用最大熵法对增强后的织物图像进行分割,定位出织物疵点区域。通过对油污、破洞、线头、断经、双纬、节纱、吊经织物图像进行实验,发现该方法能快速有效检测出疵点信息。

To enhance the efficiency of fabric defect detection,a novel method combining bilateral filtering and frangi filtering has been proposed.Initially,bilateral filtering is employed to process fabric images,reducing the influence of background texture information on the detection results.Subsequently,a frangi filter is applied for secondary filtering of the bilateral-filtered fabric images,thereby enhancing the information of defect areas.Finally,the maximum entropy method is utilized to segment the enhanced fabric images,accurately locating the defect regions.Experiments conducted on images featuring oil stains,holes,loose threads,broken warps,double wefts,irregular yarns,and hanging warps demonstrate that this method can swiftly and effectively detect defect information.