Fabric defect detection plays a crucial role in the quality control of manufacturing in the textile industry. Focusing on detection of fabric defects, the paper proposes a algorithm for detecting fabric defects based on choosing the optimal Gabor filter. Firstly, based on sixteen Gabor filters for four scales and four directions, it calculates amount of informations of the image with defects which is filtered by direction masks of five by five to determine two optimal directions. Secondly, optimal Gabor filters are selected for each of the eight filters in the two directions by Signal to Noise Ratio and a cost function. Then the image with defects which is filtered by the optimal Gabor filters can be processed in the step of image fusion and image binaryzation to achieve fabric defect detection. From the results of experiment, show that the algorithm can effectively detect the fabric defects.%织物的疵点检测在纺织工业生产的质量控制中起着至关重要的作用。针对织物疵点，论文提出一种基于最优Gabor滤波器的织物疵点检测算法。首先，针对4尺度4方向的16个Gabor滤波器，借助的方向模板对疵点图像滤波后的信息量确定2个最优方向；其次，综合信噪比和损失评价函数在2个方向的8个滤波器中各自选取1个最优滤波器；然后，对滤波后的疵点图像做融合和二值化，实现了疵点检测。从实验结果看，该算法能够有效地检测出织物疵点。
展开▼
