【摘要】 三维型面非接触测量在工业领域有着广泛的应用前景。双目视觉是一种有效获取三维轮廓信息的方法 ,其中立体匹配是关键的一个步骤。本文提出了结合空间二进制编码与外极线斜率约束来实现立体精匹配的方案。首先利用空间编码将被测物空间划分成众多的不同编码区域 ,然后结合外极线斜率约束在左右两幅图像中具有相同编码的区域内寻找匹配点。该方法能够实现亚像素级的立体精匹配 ,改善了交向摆放姿态双目立体视觉系统的匹配质量。更多还原

【Abstract】 Non-contact measurement of 3D surface has been extensively applied in industry.The stereo vision is an effective way for 3D profile measurement,and stereo matching is a key step.In this paper,a stereo precise matching method by combining spatial binary code and epipolar-slope constraint is presented.In this method,structured strip patterns is used to encode the measured space into many small sections,and based on epipolar-slope constraint the corresponding points can be searched in the same section encoded.Therefore,the stereo matching at sub-pixel level for double camera of converging optical-axis will be completed successfully.This method can improve efficiently the matching quality of binocular stereo vision system.更多还原