基于机器视觉系统的半透明薄膜缺陷检测及厚度测量研究

Defect inspection and thickness measurement for semi-transparent thin film based on machine vision system

提出了一种基于机器视觉系统的薄膜缺陷检测与厚度测量的方法。该方法通过建立CCD接收到的放置薄膜前后的图像灰度与厚度之间的关系,利用同样的图像采集硬件和不同的软件分析方法,在保证实现最小缺陷0.1mm×0.1mm检测的同时,完成单次66mm×50mm的大范围厚度测量。在基于透射光密度原理进行厚度测量时,为减小测量误差,采用标准密度值标定板对系统进行标定,对计算的理论光密度值进行偏移校正。以四氟薄膜为例,其厚度检测平均误差为5.7%,标准偏差为6.66%。仅需匹配简单的扫描装置,便可在普通的视觉系统上完成薄膜划痕、气泡等表面二维缺陷的识别以及三维轮廓的全场在线测量。

In order to meet the industrial requirements of owline quality evaluation for semi-transparent thin film,a method of defect inspection and thickness measurement is presented based on machine vision system is presented. The method can retrieve the correspondence between CCD grayscales and thin fihn thickness,and the CCD grayscales are recorded respectively before and after the placement of thin film. By using the same image acquisition hardware and different analysis methods, the system can inspect the surface defects larger than 0.1 mm×0.1 mm,and can measure the thickness of thin film in 66 min×50 mm area at each time. To reduce the measurement error of thin film thickness based on principle of transmission optical density,the density calibration by standard calibration sheet is exploited to correct the density drift of calculated density value. PEFT film was adopted to demonstrate the feasibility of the method. The results show that the average error of thickness measurement is 5.7 % with a standard deviation of 6.66%. Assembling with a simple scanning module, the two-dimensional surface defects, such as scratch and bubble, can be recognized, and three-dimensional full field profile measurement of thin film also can be performed on a general machine vision system.