基于Kinect的三维重建-信号与信息处理专业论文.docxVIP

摘要 三维重建是计算机视觉的一个重要目标，可以帮助人们快速精确地将日常生 活中的物体数字化，并有着广泛的应用前景。本文提出了一种成本低廉、快速且 操作简便的三维重建方法。借助于微软公司的 Kinect 体感传感器作为采集深度图 像和彩色图像的输入设备，通过对原始深度图像的去噪、平滑、表面重建等一系 列方法，最终可以获得在三维空间中的点云模型。 在完成本文的过程中，本人做了以下工作：提出了一种新的基于均值滤波和 中值滤波的帧间滤波算法，对从 Kinect 中获取的原始深度图像进行该滤波处理； 使用 ICP(Iterative closest point)算法来计算两组点云数据间的变换矩阵；借助于 PCL(Point Cloud Library)进行三维点云数据的配准；实验结果表明本文算法具有一 定的三维重建效果。 通过本文的研究工作，能够使普通用户不花费高昂费用购买深度摄像机，也 能在自己家里使用电脑和 Kinect 对日常物品进行三维重建。可以从最后重建的结 果看出，尽管由于设备精度的原因，三维模型的细节不够精确，但基本轮廓清晰， 可以满足日常应用需要。 关键词：三维重建 Kinect 深度图像滤波 ICP PCL Abstract Three-dimensional reconstruction is an important target in computer vision, which can help people to digitalize the object in the real world and is widely used in different fields. This paper introduces a cheap, fast and easy way for 3D reconstruction. With the help of Microsofts Kinect as a depth sensor to get depth images and RGB color images, users can finally obtain 3D point clouds model through the process of denoising, smoothing introduced in this paper. This paper have completed the following tasks: a new inter-frame filter method based on mean filtering and median filtering is proposed, which can filter the raw depth images from the Kinect; by using the iterative cloud point (ICP) algorithm, the transformation matrix of two sets of point clouds data can be calculated; the point cloud library (PCL) is used for 3D point clouds registration. The experiment shows that the method can achieve the desired results of 3D reconstruction. According to the work in this paper, users can reconstruct the real objects without expensive equipment, what they need are just Kinect and PC. As the results shown, although the 3D model is not accurate enough because of lower accuracy, the general outline is clear enough to meet the need in daily application. Keyword: 3D reconstruction , Kinect, Depth images filtering, ICP, PCL 目录 HYPERLINK \l _bookmark0 摘要...............................................................................................................