基于非标定图像的犁胫热态锻件三维重构  

3D reconstruction of hot forging piece of plough shank based on the uncalibrated images

作　　者：邢作常[1] 田素博[1] 刘思瑶[1] 林静[1] 张祖立[1] 

机构地区：[1]沈阳农业大学工程学院,沈阳市110161

出　　处：《中国农机化学报》2016年第5期49-53,共5页Journal of Chinese Agricultural Mechanization

基　　金：国家自然科学基金(51275318);公益性行业(农业)科研专项(201503116)

摘　　要：热态锻件尺寸的实时测量是长期困扰锻造工艺的重要问题。机器视觉测量是有望解决这一难题的重要途径。现场标定是机器视觉测量最为繁琐的步骤之一。建立基于非标定图像的三维重构模型,可以避免现场标定,提高机器视觉的测量效率和使用的方便性。提出基于RANSAC自动估计两幅图像之间基本矩阵的方法,研究基于基础矩阵估计的三维重构过程,给出基于非标定图像的三维重建结果。实验表明,非标定三维重建能为犁胫热态锻件的计算机视觉测量提供一定精度的三维描述。Real time measurement of the size of hot parts is important problem in forging process for a long time.Machine vision measurement is an important way to solve this problem.Field calibration is one of the most complicated steps of machine vision measurement.The establishment of 3Dmodel based on the uncalibrated images can avoid the field calibration,improve efficiency and convenience of the use of machine vision.A method of estimating the fundamental matrix between two images based on RANSAC is proposed,and the 3Dreconstruction process based on the fundamental matrix estimation is studied.The experiment shows that the noncalibration 3Dreconstruction can provide a certain accuracy for machine vision measurement.

关 键 词：非标定三维重建 犁胫 热态锻件 基础矩阵 RANSAC 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]