基于神经网络的视觉检测系统的研究

Neural Network-based Vision Inspection System

随着电子学和计算机技术的日趋成熟,视觉检测系统取得了突飞猛进的发展并在许多领域得到了广泛地应用。本文提出了一种基于神经网络的结构光中心线提取的方法,实验结果表明神经网络方法可达到亚像素级精度,从而能有效地提高视觉检测的精度。

Vision inspection system has become an important technique and been utilized in many fields, following the increasingly maturity of electronics and computer technology. In this paper, a neural network was proposed and implemented for detecting the laser stripe center from the image plane. The accuracy of the measurement was improved by raising the resolution to the subpixel level in detecting the laser stripe center.