论文部分内容阅读
光栅式双目立体视觉传感器的难点之一在于立体匹配问题,为此,提出了一种基于极线约束和空间点最小距离搜索的立体匹配方法。该方法将光栅式双目立体视觉传感器看作两个光栅结构光传感器,分别标定后可测定光条中心点关于某个结构光模型的三维坐标,若两点匹配,则其三维坐标间的距离理论上为零。引入极线约束,在左摄像机成像光条上找一个特征点,在右摄像机所成像中便可计算出一条极线与之对应,在极线与各光条中心的交点中寻找匹配点。该方法在三维空间进行匹配,计算量小,能够实现点与点的唯一匹配。仿真实验表明了该方法的有效性。
One of the difficulties of grating binocular stereo vision sensor lies in the problem of stereo matching. For this reason, a stereo matching method based on polar line constraints and minimum distance search of space points is proposed. The method takes the raster binocular stereo vision sensor as two grating optical sensors and respectively calibrates the three-dimensional coordinates of the center of the light beam with respect to a structured light model. If the two points match, the distance between the three-dimensional coordinates Theoretically zero. By introducing the polar constraint, a feature point is found on the left camera imaging light bar, an epipolar line can be calculated in the imaging of the right camera, and a matching point can be found in the intersection of the epipolar line and each light bar center. The method matches in three-dimensional space, the calculation is small, and can achieve the only match between point and point. Simulation results show the effectiveness of this method.