College of Artificial Intelligence, Yango University, Fuzhou 350015, China.
Department of Electrical and Computer Engineering, Tamkang University, New Taipei City 251, Taiwan.
Sensors (Basel). 2023 Jan 28;23(3):1452. doi: 10.3390/s23031452.
In this paper, the paper cups were used as the research objects, and the machine vision detection technology was combined with different image processing techniques to investigate a non-contact optical automatic detection system to identify the defects of the manufactured paper cups. The combined ring light was used as the light source, an infrared (IR) LED matrix panel was used to provide the IR light to constantly highlight the outer edges of the detected objects, and a multi-grid pixel array was used as the image sensor. The image processing techniques, including the Gaussian filter, Sobel operator, Binarization process, and connected component, were used to enhance the inspection and recognition of the defects existing in the produced paper cups. There were three different detection processes for paper cups, which were divided into internal, external, and bottom image acquisition processes. The present study demonstrated that all the detection processes could clearly detect the surface defect features of the manufactured paper cups, such as dirt, burrs, holes, and uneven thickness. Our study also revealed that the average time for the investigated Automatic Optical Detection to detect the defects on the paper cups was only 0.3 s.
在本文中，以纸杯为研究对象，结合机器视觉检测技术和不同的图像处理技术，研究了一种非接触式光学自动检测系统，用于识别制造纸杯的缺陷。采用组合环形光作为光源，采用红外（IR）LED 矩阵面板提供 IR 光，不断突出检测对象的外边缘，采用多栅像素阵列作为图像传感器。图像处理技术包括高斯滤波器、Sobel 算子、二值化处理和连通分量，用于增强对生产纸杯存在缺陷的检测和识别。纸杯有三个不同的检测过程，分为内部、外部和底部图像采集过程。本研究表明，所有的检测过程都可以清晰地检测到制造纸杯的表面缺陷特征，如污垢、毛刺、孔和不均匀厚度。我们的研究还表明，所研究的自动光学检测检测纸杯缺陷的平均时间仅为 0.3 秒。
