【摘要】 基于机器视觉原理的自动光学表面缺陷检测技术是当今工业生产中在线检测表面缺陷的一种新的技术方法,是精密制造与组装工业过程中保证零部件表面质量的重要检测手段。以液晶面板TFT阵列表面缺陷自动光学检测为例,介绍了表面缺陷自动光学检测的基本组成原理,阐述了周期纹理背景表面上的表面缺陷检测方法、缺陷信息处理的基本过程与实用算法。针对表面缺陷检测图像处理技术难题,详细论述了表面缺陷扫描图像中的周期纹理背景傅里叶变换频域滤波方法、缺陷分割双阈值统计控制法,并用实验结果给出了例证。更多还原

【Abstract】 Flat plates are widely used in industrial products.The defect on their surfaces is an important factor which can affect the quality of the surface.The on-line inspection technique for the surface defects is an important means to ensure product quality.An on-line automated optical surface defect inspection technique is investigated.The structure configuration of the optical inspection system and the defect extraction algorithms for the surface with periodic textures are illustrated,and exemplified by the surface defect inspection of TFT-LCD display module in flat panel display industry.The algorithm involves the basic principle of surface defect extraction,defect segmentation.In the algorithm,the canned surface images are firstly transformed from time domain to frequency domain by Fourier transform.The periodic texture background information on the surface is removed by using filtering method in the frequency domain.An one-dimensional transform Haar wavelet transform is used to extract defect information.Then,a dual-threshold statistical control method is used to separate the defects from the surface background.更多还原