基于机器视觉的光性智能检定系统设计与实现

The Design and Implementation of a Computer-vision Based Intelligent System for Optical Property Identification

基于机器视觉技术,深度融合了光机电一体化系统,利用K均值(KMeans)聚类分析函数,提出了适用于光学仪表读数的图像处理算法,研发了一种具备自动观测、自动识别功能的光性智能检定系统。该系统由折光仪,底座,样品旋转模块,加液模块,视频采集模块,识别分析模块和总控模块等七大模块组成,可以准确检定折光仪读数,智能判别介质光性,提高了测试精度和工作效率。系统内置的多种工作模式可满足教学示范、快速无损检测、光性研究、晶体精细加工等不同需求。校验结果表明,该系统平行测定结果的绝对差值<0. 000 2,达到国标要求,可广泛应用于各类专业实验室和教学实验室中。

In this paper,an intelligent system for optical property identification is developed with cutting-edge technologies,such as opto-mechatronics,computer vision,K-Means arithmetic.The system is composed of seven parts,i.e.,refractometer,base,sample rotating apparatus,liquid injection device,video capture device,image recognition software,and general control module.To meet the needs of teaching classroom demonstration,rapid identification of crystals,study of optical axis characteristics and orientation processing of gemstones,four working modes are designated.The inspection results show that the absolute error of the parallel measurements of unknowns is less than 0.000 2,which meets the requirements of the Chinese national standard.Therefore,this system can be widely used in various scientific and teaching laboratories.