基于机器视觉技术和D-S证据理论数字显示仪表自动读数.pdfVIP

Scientific Journal of Information Engineering December 2013, Volume 3, Issue 6, PP.122-129 Reading Digital Display Instrument Automatically Based on Machine Vision and D-S Evidence Theory † Zhonghong Shen , Qunxing Liu, Chunxu Jiang, Haozhi Zhao The fifth research institute of MIIT, P. R. China, Guangzhou Guangdong Province 510610, P.R. China †Email: szh@ Abstract The automatic reading of digital display instrument is achieved by means of the application of machine vision technology and D-S evidence theory; in the accomplishment of which color space selection threshold binarization and other machine vision techniques are applied at first to preprocess images of the digital display instrument; then the characteristics of pixel distribution in the targeted areas are extracted and listed; further, integrated and assessed by D-S evidence theory to realize the automatic reading. It is demonstrated by a number of experiments that the algorithm characterized by high accuracy and recognition in real time can be applicable to practice after improvement . Keywords: Machine Vision; D-S Evidence Theory; Digital Display Instrument; Reading A utomatically 基于机器视觉技术和D-S 证据理论的数字显示仪 表自动读数 申中鸿，刘群兴，蒋春旭，赵浩之 工业和信息化部电子第五研究所，广东 广州 510610 摘 要：利用机器视觉技术与 D-S 证据理论，实现数字显示仪表的自动读数。首先，利用机器视觉中色彩空间选择，阈 值二值化等技术对数显仪表图像进行预处理；之后，统计并提取目标区域像素分布特征；最后，利用 D-S 证据理论，对 所提取的目标区域特征进行融合判断，实现数字自动读取。经过大量实验验证，该算法具有较高的识别准确度与良好的 实时性，研究改进后可推广应用于实际工业场所的数显仪表读数。 关键字：机器视觉；D-S 证据理论；数字显示仪表；自动读数 引言 对现代工业而言，如何提高生产效率并降低生产成本是众多企业追求的目标，于是，工业生产自动化应 运而生。工业自动化的显著特点是较少人工参与，并具有良好的精确性、及时性。机器视觉以其良好的精确 性、及时性、便捷性、自动性和控制性，成为现代工业自动化应用的一个主要技术手段。由于数显仪表直观 的读数特性，而被广泛的应用于现代工业各个行业，为入们读取测量数据提