摘要： 对基于计算机视觉的指针式仪表自动检测系统设计中所涉及到的数字图像处理关键技术进行了研究。利用指针图像的先验特征,构造了基于空间相关和灰度相关的邻域平滑算子。在常规图像分割理论的基础上,进行了图像多灰度等级分割的理论分析,提出了多灰度等级分割算法。根据仪表图像的噪声特征,构造了阈值减影算法,以实现指针的检出。其研究结果已应用于“基于计算机视觉的指针式仪表自动检测系统”的设计开发中,实现了指针式仪表质量检测的自动化。

关键词: 计算机视觉, 仪表, 检测

Abstract: Some key technologies about image processing for design of index meter examination system with computer vision are studied, including neighbhood smoothing algorithm with gray and spatial correctivity, image segmentation with multiple gray levels, threshold subtraction algorithm. Based on the results above, index meter examination system with computer vision is developed, which makes it possible for the quality examination and control of index meter to be finished automatically and intelligently.

中图分类号: 

• TP274.4