基于中值滤波的三维重建研究  

Research on Three Dimensional Reconstruction Based on Median Filtering

作　　者：李晓明[1] 仝明磊[1] 

机构地区：[1]上海电力学院电子与信息工程学院,上海200090

出　　处：《仪表技术》2016年第3期19-22,共4页Instrumentation Technology

摘　　要：探讨如何通过单目摄像机重建变形曲面三维结构。模板已知,单幅图像重建变形曲面三维结构由于透视投影的影响难以确定。在两个三维点的关系中,可由约束求出点的深度的最大值,创新点在于考虑到变形曲面的平滑性,加入中值滤波,对初始得到的最大位深度发生突变的特征点进行处理,代替迭代求最优深度最大值的过程,可以缩短计算时间。最后由仿真实验验证了该方法的可行性,并由真实实验验证了该方法的效果。This paper mainly discusses how to reconstruct the 3D structure of the deformed surface by single camera. The template is known. Due to the effect of perspective projection,this problem is underdetermined. In the relationship between the two 3D points,the maximum value of the depth of the point can be determined by the constraints. The innovation of this paper is adding median filtering in the consideration of the smoothness of the surface deformation to process the initial maximum depth of the feature points instead of the iterations to get the optimal depth maximum value. This method can reduce the computing time. The feasibility of this method is verified by simulation experiments. And the effect of the method is verified by real experiments.

关 键 词：三维重建 单目 中值滤波 模板 特征点 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]