从二维图像序列获取三维轮廓的算法及其实现

立体信息的获取是图像分析、计算机视觉中的重要任务之一,国内外学者在这一领域中作了大量的工作。目前获取立体信息的方法主要有接触法、立体视差法、飞行时间测距、结构光测距等,论文提出的新算法是根据摄像机或照相机连续拍到的被测物体图像序列,对每一幅图像进行水平切线分割,找到图像中的切点即图像中的某个水平面与背景的两个交点,依次对二维图像序列中的每一幅图像进行同样的操作,可以获得同一水平面上的若干点,对每个点赋予深度信息,然后将水平面上的点按顺序连接,从而获得图像序列所表示的三维物体的水平截面轮廓。实验结果表明论文提出的算法确实可行、有效。

An Algorithm Implementation of 3D Cross Section Derived from 2D Image Sequences

It is one of the important studies of image analysis and computer vision that the three-dimensional data are acquired.There are many researches in the field.The main means of three-dimensional data acquired are three-dimensional parallex,distance measurement with flying time,distance measurement with structural ray,etc.A new algo-rithm is proposed in the paper,people can obtain the image sequence of measured solid acquired by pickup camera and divide up the three dimensional objective from the background with a gradual approach,then people can get the aclinic edge of the three-dimensional image.The algorithm's proposion and implementation are important to three dimension reconstruction of solid surface.