基于自主研发的点阵结构光投射系统并设计了一套从影像到点云的详细流程,对立体相机检校之后将左右片提取出的特征点通过核线约束等多种约束规则实现特征点的匹配,实验结果表明可以得到较高的匹配成功率,最终获取到物体表面轮廓的点云。
This paper designed a detailed process from images to point cloud based on the lattice structural light pro-jection system.After calibrating stereo camera,extracted feature points and matched these points by epipolar constraint and other constraints.The experimental results show that this matching method can get high matching success rate.Finally get the point cloud of object surface.