利用最优Gabor滤波器织物疵点检测算法

Fabric Fault Detection Algorithm Using Optimal Gabor Filter

针对织物疵点自动检测计算量大和准确度低的问题,文中提出一种快速有效的纺织品疵点检测算法,首先通过遗传算法调整Gabor滤波器,使其与无疵点织物图像的纹理信息相匹配,得到最匹配的Gabor滤波器参数;其次采用调整后的最佳Gabor滤波器对待检测织物图像进行疵点检测;最后对卷积图像进行自适应二值化处理,检测织物是否存在疵点。结果表明,利用遗传算法建立最优的Gabor滤波器,具有较高检测率且运算速度较快,正检率较同类算法提高了2.8%,达到了98.5%。

For the problem of large calculation and low accuracy,a fast and efficient textile fault detection algorithm is proposed to adjust the Gabor filter to match the texture information of the fabric image and obtain the most matching Gabor filter parameters and the best Gabor filter to binary the convolution image to detect the fault.The results show that the optimal Gabor filter using genetic algorithm has high detection rate and fast operation speed,and the positive detection rate increases by 2.8%to 98.5%.