利用双目序列图像恢复3D物体形状的方法

A Method on Getting Three-dimensional Information by Processing Sequence Images

针对机器视觉在3D物体形状恢复方面的研究和发展,提出一种运用图像处理技术对双目序列图像进行处理,获取物体三维信息的方法。利用两个不同方位的摄像机对旋转物体进行拍摄,获得俯视方向的图像和从主视方向连续采集的序列图像,使用图像处理算法将两个视角的图像进行匹配和对应,并对物体外轮廓进行跟踪、定位及坐标运算,最后计算出外轮廓的三维信息,在计算机中重建其模型。此方法可应用于零件检测和逆向工程。

A new method is mentioned to get three-dimensional information of the selected object by processing sequence images,using digital image processing technology. Two digital vidicons are located at the top and obverse of the object,which is placed on the rolling plate. The image getting from the top one and the sequence ones from the obverse are processed in different ways to get the required data,and the integrated three -dimensional information of the object would be collected by calculating the data. Then the digital model can be built on the computer.