基于LabVIEW的双目视觉识别系统设计

Design of binocular vision recognition system based on LabVIEW

双目视觉能够获得物体的三维信息,基于双目视觉的测量方法是一种行之有效的测距手段。研究了双目测距原理、摄像机标定、图像处理等关键技术,使用MATLAB工具箱完成相机标定,利用虚拟仪器程序设计语言LabVIEW、NI IMAQ和NI Vision,构造双目视觉系统,完成图像采集、预处理、识别、深度恢复等功能,得到目标物体特征点的空间三维坐标。通过设计实验,验证该系统测量结果的相对误差小于1.15%。结果表明,该系统操作简单、界面友好,具有较好的实时性和稳定性。

Binocular vision can obtain three-dimensional information of objects.The measurement method based on binocular vision method is an effective means of distance measurement.This paper studies the principle of binocular distance measurement,camera calibration,image processing and other key technologies.The camera calibration is completed using the MATLAB toolbox.The binocular vision system was constructed using virtual instrument programming language LabVIEW,NI IMAQ and NI Vision.The functions of image acquisition,preprocessing,recognition and depth recovery were completed,and the spatial three-dimensional coordinates of the target object′s feature points was obtained.The relative error of the system measurement results is less than 1.15%,through the design experiment.Results show that the system was characterized by simple operation,friendly interface,with better real-time and stability.