基于三维虚拟vr的历史街景重现技术

传统方法采集点云数据时容易丢失信息,无法保证历史街景重现的完整性。为此,提出一种新的基于三维虚拟vr的历史街景重现技术,对采集的0像素值点进行去噪处理,防止大量空洞产生。通过取交集对由三维采样点构成的集合进行多帧融合处理,求出各三维采样点的深度可信度,根据深度可信度排序对三维采样点进行无重复融合处理,获取历史街景完整三维点云。依据ICP法,通过不停查找对应点集及运算变换关系的过程,找到目标点集与参照点集间的旋转矩阵与平移向量,将感兴趣重建部分从整个场景中分离实现点云配准。将历史街景三维重建模型添加至Google Earth平台,在Google Earth平台实现历史街景重现。实验结果表明,所提方法能够有效实现历史街景重现,重建精度高、完整性强。

Reconstruction of historical streetscape based on 3D virtual VR

It is easy to lose information when collecting point cloud data by traditional methods, and it can not guarantee the integrity of historical streetscape. To solve this problem, a new technology based on 3D virtual VR is proposed, which is used to remove the noise from the 0 pixel value points and prevent a large number of holes from occurring. By collecting and intersecting the set of 3D sampling points, the frame is processed by multi frame fusion, and the depth confidence of each 3D sampling point is obtained. Based on the ranking of depth confidence, the 3D sampling points are processed without repeated fusion to get the complete 3D point cloud of historical street scenes. According to the ICP method, we find the rotation matrix and translation vector between the target point set and the reference point set by finding the corresponding point set and operation transformation relation continuously, and reconstruct the interest reconstruction part from the whole scene to achieve point cloud registration. The historical streetscape 3D reconstruction model is added to the Google Earth platform, and the historical streetscape is reproduced on the Google Earth platform. The experimental results show that the proposed method can effectively reconstruct the historic street scene, and the reconstruction accuracy is high and the integrity is strong.