研究三维立体图像优化问题,实现高真实度的纹理图.由于立体图像重建过程产生累加误差,影响匹配精度.目前半自动和自动纹理贴图中三维扫描数据与高分辨率纹理图像对应点配准精度低、计算量大.为解决上述问题,在标准ICP(Iterative Closest Point)算法的基础上,提出一种改进的LM-ICP 2D和3D配准算法.通过法向量内积加权的最近点迭代,动态更新特征对应,减小误匹配点对配准精度的影响,并利用LM(Levenberg-Marquardt)算法优化投影矩阵.采用真实数据进行仿真.实验表明,提出的算法能得到精度高、真实性强的匹配图像效果,为设计提供参考.%Highly realistic texture mapping is a key factor in 3D reconstruction based on the 3D scanning device. For semi-automatic and aotomatic registration of 3D scanning data and high resolution texture image, the existing algorithm is of low precision and needs complex calculation. A modified LM-ICP algorithm applied to 2D and 3D registration was presented based on the standard ICP algorithm. Updating the corresponding relationship between 2D and 3D feature points by weighting the inner product of normal vector, it enhanced the robustness and reduces the effect of registration precision resulted from mismatching points. Finally, the LM algorithm was used to optimize the projection matrix. The experimental results on real data show that this method is feasible and can achieve highly realistic texture mapping.
展开▼
