改进泊松算法的图像三维重建点云模型网格化

发布时间：2018-04-05 01:08

本文选题：三维重建　切入点：泊松算法　出处：《测绘科学》2017年04期

【摘要】：针对泊松表面重建算法在点云数据的网格化中仍不能有效满足"细节保持与噪声平滑"的平衡问题,该文提出了一种基于高斯滤波的改进泊松算法。通过将高斯滤波引入到点云数据等值面的向量场估计中,一方面实现了对点云拓扑结构的更准确估计以及对点云噪声的有效平滑;另一方面通过调节高斯滤波中的标准差参数,实现了对点云模型网格化的细节保持与噪声平滑的细微控制。以福州大学张孤梅雕像为实验对象,图像三维重建技术获得的点云数据作为数据源,利用改进的泊松算法进行点云网格化。结果表明,改进的泊松算法提高了网格模型的精确性与完整性,且在视觉上更好地逼近真实模型的细节,验证了改进算法的有效性。
[Abstract]:Poisson surface reconstruction algorithm can not effectively meet the balance problem of "detail keeping and noise smoothing" in the mesh of point cloud data. In this paper, an improved Poisson algorithm based on Gao Si filter is proposed.By introducing Gao Si filter into vector field estimation of point cloud data isosurface, on the one hand, more accurate estimation of point cloud topology and effective smoothing of point cloud noise are realized; on the other hand, standard deviation parameters in Gao Si filter are adjusted.The detail maintenance and noise smoothing of point cloud model are realized.The point cloud data obtained by 3D image reconstruction are used as the data source, and the improved Poisson algorithm is used to mesh the point cloud.The results show that the improved Poisson algorithm improves the accuracy and integrity of the mesh model, and approaches the details of the real model more visually, and verifies the effectiveness of the improved algorithm.
【作者单位】：福州大学地理空间信息技术国家地方联合工程研究中心/福建省空间信息工程研究中心;
【基金】：国家科技支撑计划课题项目(2013BAH28F02)
【分类号】：TP301.6;TP391.41

【相似文献】