基于Kinect的物体表面重建方法研究  

Study on Kinect-based surface reconstruction

作　　者：徐祯琦[1] 龚光红[1] 

机构地区：[1]北京航空航天大学自动化科学与电气工程学院,北京100191

出　　处：《中国体视学与图像分析》2013年第3期255-262,共8页Chinese Journal of Stereology and Image Analysis

基　　金：仿真预研项目

摘　　要：随着测量技术的发展,基于深度图像的物体表面重建技术得到了越来越多的应用。本文实现了一种基于Kinect的物体表面重建方法。通过Kinect采集原始数据并经过感兴趣剔除、原始点云滤波、不同帧点云之间的配准、物体表面重建以及网格平滑等步骤,构造出了物体表面的模型。与传统的基于深度图像的重建方法相比,本文方法简便实用,能够获得较好的重建效果。With the development of measurement technology, depth-image based surface reconstruction is becoming more and more popular. In this paper, we implement the task of Kinect-based surface reconstruction. By acquiring data from Kinect, filtering region of interest, filtering original point clouds, registering point clouds from different frames, reconstructing surface and smoothing the mesh, we get the surface model of object concerned. Compared with traditional depth-image based method, our method is more applicable and can get a good result.

关 键 词：KINECT 表面重建 点云配准 网格平滑 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]