基于图像的焊件深度信息提取  

Depth Information Extraction of Welding Piece on Binocular

作　　者：马国红[1] 朱政强[1] 汪鎏[1] 黄凌翔[1] 刘国权[1] 

机构地区：[1]南昌大学机器人研究所,南昌330031

出　　处：《上海交通大学学报》2008年第S1期110-111,118,共3页Journal of Shanghai Jiaotong University

基　　金：国家自然科学基金(50705041);江西省自然科学基金(2007GQC1823);江西省教育厅基金([2007]8)项目资助

摘　　要：阐述了一种简化的基于竞争匹配机制的双目立体图像匹配算法.通过对摄像机安装位置进行配置,使得双目视觉传感器的模型大为简化.使用竞争匹配机制对图像中的对应点进行匹配.然后把各个点在图像上的坐标系代入简化后的传感器模型,计算出三维焊缝的空间坐标.最后在VC++平台上对所介绍的方法编程实现,并通过双目测距实验证明该算法的实时性和可靠性.This paper described a simplified binocular images matching algorithm based on competition and match mechanism.By setting down the position of CCD sensor,a model of binocular sensor was simplified.With this match way,the corresponding point of image was processed and three dimension space-coordinate was calculated after different point positions were sent to that simplified model.The experiment with this algorithm shows that this algorithm has good reliability and works well.It can extract three-dimensional...

关 键 词：双目视觉 竞争匹配 深度提取 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]