基于机器视觉的液晶屏裂痕自动检测方法研究

LCD crack automatic detection method based on machine vision research

随着液晶的普及,对液晶屏生产过程中质量控制提出更高的要求,液晶屏裂痕自动检测系统应运而生,它能真正实现高效率、高稳定性的实时检测。针对液晶屏裂痕缺陷中的"线缺陷"、"点缺陷"进行深入研究,把机器视觉、数字图像处理技术运用到液晶屏裂痕自动检测系统中,并以HALCON和VC++联合编程完成了裂痕自动检测系统的设计。在数字图像处理部分,根据指定尺寸对图像进行傅里叶变换、高斯滤波、迭代处理等一系列操作,为了得到更清晰的图像,突出裂痕区域,对图像进行灰度化、频域图像卷积、滤波去噪、形态学处理。选择出适合在线测量的各种算法,进行了大量实验测试和实时自动检测,结果表明,该方法在识别液晶屏裂痕的几何特征上效果和精度较好,识别速度达到在线要求。

With the popularity of LCD,quality control of LCD screen in the production process needs higher requirements,thus the LCD screen crack automatic detection system came into being,which can truly achieve high efficiency,high stability and real- time detection. This paper studies the crack defects of LCD screen such as line defect and point defect deeply. The machine vision and digital image processing technology are applied to the liquid crystal screen crack automatic detection system,and HALCON and VC ＋ ＋ programming are used to complete the crack automatic detection system design. In the part of digital image processing,according to specified dimensions of the image,a series of operations like Fourier transform,Gaussian filtering and iterative processing are taken out. In order to get a clearer image and highlight the rift zone,the operations including gradation,frequency domain image convolution,noise filtering and morphological processing are taken out. Various algorithms suitable for online measurement are selected for a large number of tests and real-time automatic inspection. The results show that on the geometric characteristics of the recognition of LCD screen cracks,effect and precision are better,and recognition speed meets online requirement.