远距离采集图像损伤区域三维重建仿真研究

摘要

It is difficult to eliminate noise and deformation of the shape base property, and realize the 3D reconstruction.The traditional reconstruction methods cannot effectively repair damaged areas.This paper proposes a 3D reconstruction method of image damage region based on volume pixel interpolation.The image deformation in remote images acquisition is irregular.In this paper, the law of singularity generation was used to implement singular point interpolation compensation.After the non-homogeneous matrix of attribute point of projection images was obtained, the paper compensated the singularity of the damage region by using body pixel, and the singularity was eliminated by edge conflict principle and a more realistic 3D reconstruction image could be obtained.The experimental results shown that the proposed method could enhance the information of the image distortion area in 3D reconstruction.%远距离采集图像的损伤区域形状基属性形变和噪音难以消除,三维重建难度大,传统重建方法不能有效修复损伤区域.提出基于体像素插值的图像损伤区域三维重建方法,在远距离图像采集中形变不规则的前提下,利用产生奇异点的规律进行奇异点插值补偿,提取到图像仿真投射映像属性点的非齐次矩阵后,利用体像素补偿损伤区域奇异点,并通过边缘冲突原理进一步消除奇异点,获取到更加真实的远距离采集三维重建图像.实验结果证明:所提方法增强了远距离采集图像损伤区域三维重建的信息量.