基于OpenCV的水银体温计示数实时检测

Reading Inspection of Mercury Thermometer Based on OpenCV

为了实现对水银体温计产品的在线非接触式检测,设计一种示数检测系统;基于特殊的机械运载平台设计配合线阵相机对体温计产品进行图像采集,基于OpenCV开源视觉库对图像进行处理,提出一种基于模板匹配和纵向扫描端点查找的图像搜索算法,通过确定刻度线和水银柱升降端在图像中的相对位置关系计算体温计示数;不借助外部测量仪器对产品示数进行实时检测,可有效取代传统的人工操作方式;实验结果表明,该系统运行稳定且具有较好的实时性,检测精度较高,满足测量不确定度小于0.1℃的检测要求。

In order to realize the non-contact inspection of mercury thermometer products in the assembly line, a kind of reading inspection system was designed. Images were acquired with line-scan digital camera based on the mechanical vehicle platform, the image was processed based on OpenCV open-source vision library, a kind of image algorithm based on template matching and longitudinal scanning for the endpoint lookup, thermometer reading was calculated through the interested points in the image, which are located in scale line and mercury lift side; The system can realize real-time inspection without external measuring instrument, which can effectively replace the traditional manual operation mode. The experimental results showed that the system has stable operation,good real-time performance and high precision, the measurement uncertainty is less than 0.1 ℃.