多视点图像中线特征的三维重建

3D Reconstruction of Line Features from Multiple Views

将多视点图像中的离散线段(称为线特征)用于场景的三维重建过程.首先将平均重投影几何误差作为线特征三角形法和光束平差法的目标函数,保证了重建结果的最优性;其次提出一种对含噪声的空间直线Plücker坐标进行双线性约束矫正的解析方法,并给出理论证明;最后推导出针对具有4自由度空间直线的最少参数化迭代方法,消除了优化过程中的过参数化问题,以避免由于内约束存在而导致迭代难于收敛的情况,提高了重建结果的精度.仿真数据以及真实图像的实验结果验证了该方法的有效性和精确性.

This paper uses line features as the basic elements for 3D reconstruction of real scenes. Firstly, the average geometric reprojection error is used as the objective function, assuring the optimization of final results; secondly, an analytic correlation method of bilinear constraint for noisy 3D lines represented by Pluecker coordinates is proposed, and both its theoretical proof and geometrical interpretation are provided; thirdly, a minimal representation for 3D lines with 4 degrees of freedom is used during the bundle adjustment, by which the over-parameterization problem is avoided, and the algorithm convergence is enhanced and the accuracy of the final results are substantially improved. Experiments with synthetic data and real images demonstrate the feasibility and accuracy of our proposed method.