【摘要】 提出一种基于有效背景重构和对比度增强的Mura缺陷检测方法。首先,提出了一种新的基于缺陷区域预剔除的背景重构方法,能够有效地重构背景图像,消除图像亮度不均等干扰。然后,引入了基于Otsu的双γ分段指数变换法对差分图像进行增强处理,能够有效解决背景残余问题且极大增强了Mura区域的对比度和轮廓度。最后,直接运用动态阈值分割方法,可以快速准确地将Mura缺陷分离出来。实验结果表明,与传统的多项式曲面拟合方法以及离散余弦变换法相比,本文方法对各种类型的Mura缺陷检测效果稳定,且检出率和无误报率均达到了97%以上。更多还原

【Abstract】 A Mura defect detection method based on effective background reconstruction and contrast enhancement is proposed.Firstly,a new background reconstruction method based on defect region pre-elimination is proposed,which can effectively reconstruct the background image and eliminate the interference of uneven brightness.Then,the dual-γpiecewise exponential transform method based on Otsu is introduced to enhance the difference image,which can effectively solve the problem of background residual and better enhance the contrast and contour of Mura region.Finally,the Mura defects can be separated quickly and accurately by using the dynamic threshold segmentation method.The experimental results show that,compared with the traditional polynomial surface fitting method and the discrete cosine transform method,the detection effect of this method for various types of Mura defects is stable,and the detection rate and no false alarm rate are more than 97%.更多还原