利用极线约束方法实现图像特征点的匹配

Image token correspondence with constraint of epipolar line

点匹配一直是计算机视觉和模式识别领域中的一个重要问题。通过对双目立体视觉测量数学模型的研究,在2副图像中特征点数相等的情况下,基于极线约束等理论,给出了一种解决图像特征点匹配的新方法,对三维空间中的离散点在2个视平面上的投影点进行了正确的匹配。实验结果表明,该方法效果良好,具有实用价值。

Point matching is an important problem in the field of computer vision and pattern recognition. Through the research on binocular stereo vision measurement system, based on the concept of epipolar constraint, a new method is proposed to resolve the corresponding problem between two images which have the same number of points.The projection on the view plane of the discrete points in threedimensional space are correctly matched. The experimental results indicate that the method is effective and is of great value in use.