基于相对坐标ICP的室内场景三维重建算法

摘要

Aimed at the problem that the coordinate centers in multiframe of point cloud data viewed at a different angle are noncoincident, a 3-D reconstruction algorithm of indoor scene is presented based on relative coordinate ICP(iterative closest point).Firstly, in the proposed method, the colored image and depth image of the scene are obtained with Kinect sensor and the transformation matrix of adjacent frames is calculated by using pairs of SIFT(scale-invariant feature transform) matching feature points.Then the ICP algorithm is used to establish the relative coordinate relation with initial point-cloud data as its center and conduct the interframe matching.Finally, the registered point cloud data is made fused with the corresponding information of color texture on the basis of the registration of upper-lower layer to get a visual 3D reconstruction model of indoor scene.Experimental results show that the method proposed in this article has a good effect of 3D reconstruction of indoor scene, having a certain theoretical and practical value.%针对不同视角下多帧点云数据的坐标中心不统一的问题,提出了一种基于相对坐标ICP(iterative closest point)的室内场景三维重建算法.该方法首先基于Kinect传感器获取场景的彩色图像和深度图像,并利用SIFT(scale-invariant feature transform)特征匹配点对计算相邻帧的坐标变换矩阵;然后由ICP算法建立以初始点云数据为中心的相对坐标关系进行帧间匹配;最终基于上下层配准点云数据,将配准后的点云数据与对应的颜色纹理信息融合,得到可视化的室内场景三维重建模型.实验结果表明,本文方法对于室内场景三维重建有较好的效果,具有一定的理论和实际应用价值.