基于双目视觉的玻璃瓶三维重建研究

Research on 3 D reconstruction of glass bottles based on binocular vision

三维重建技术作为机器视觉和图形学的重要分支一直都是研究的热点,在工业自动化的生产进程中,通常需要运用机器视觉技术完成不同物体的三维重建。其中,透明物体的三维重建一直是个难点。文章选用玻璃瓶为研究对象,使用改进的立体标定算法对双目RGB相机进行标定,对采集的图像进行校正、预处理以及立体匹配。该实验在研究OpenCV与SGM的基础上,对SGM算法进行了改进,对比原算法和改进算法获取的三维重建图像,改进算法获得的图像更清晰。

3D reconstruction technology,as an important branch of machine vision and graphics,has always been a research hotspot.In the production process of industrial automation,it is usually necessary to use machine vision technology to complete the 3D reconstruction of different objects,among which the 3D reconstruction of transparent objects has always been a difficulty.This article selects glass bottles as the research object and uses an improved stereo calibration algorithm to calibrate a binocular RGB camera.The collected images are corrected,preprocessed and stereo matched.On the basis of studying OpenCV and SGM,this experiment has improved the SGM algorithm.Compared with the original algorithm and the improved algorithm,the 3D reconstruction images obtained by the improved algorithm are clearer.