目的 为了解决OLED显示屏表面周期性纹理背景和缺陷边界模糊、对比度低的特征导致其表面缺陷检测困难的问题，开展OLED显示屏表面缺陷自动检测方法研究。方法 对OLED显示屏图像进行奇异值分解，选择前2个较大的奇异值重构图像纹理背景，对原图像和重构图像进行差分运算，获得残差图像。将残差图像像素随机赋予初始隶属度值，采用模糊C均值聚类法获得像素最终隶属度值。根据隶属度大小，将残差图像像素聚成2类，并从残差图像中准确地分割缺陷。结果 选取较大的2个奇异值可以有效地重构OLED显示屏的周期性纹理背景;模糊C均值聚类法分割缺陷获得的区域灰度一致性(U)平均值为0.9846。结论 基于奇异值分解的背景重构方法可以有效地检测OLED显示屏表面缺陷;与分水岭法和Otsu方法相比，模糊C均值聚类可以准确地分割模糊边界的缺陷区域。

Surface defect detection of OLED display is difficult due to the characteristics of periodic texture background, fuzzy defect boundary and low contrast. In order to solve this problem, some investigations are carried out for automatic surface defect detection for the OLED display in this paper. The singular value decomposition (SVD) method is implemented on the OLED image, and the first two larger singular values are selected to reconstruct the image texture background. The differential operation between the original image and the reconstructed image is carried out to obtain the residual image. The initial membership value is randomly set to every pixel of the residual image, and the final membership value is obtained by the fuzzy c-means (FCM) clustering method. Based on the final membership value of every pixel, the pixels are grouped into two categories to segment the defects accurately from the residual image. The periodic texture background of OLED display screen can be reconstructed effectively by selecting two larger singular values, and the average U value obtained by the FCM method is 0.9846. The method of background reconstruction based on SVD can effectively detect the surface defects of OLED display. Compared with the watershed method and the Otsu method, the FCM method can accurately segment the defect areas of fuzzy boundary.