基于Kinect的3D全景图像扫描重建技术  

Panoramic Image Scanning and Reconstruction Technology based on Kinect

作　　者：黄君君[1] HUANG Junjun(School of Information Engineering,Fujian Agricultural Vocational and Technical College,Fuzhou 350007,China)

机构地区：[1]福建农业职业技术学院信息工程学院,福州350007

出　　处：《成都工业学院学报》2023年第6期47-51,共5页Journal of Chengdu Technological University

摘　　要：为解决传统的迭代最近点(ICP)算法耗时长的问题,基于K-D树改进ICP算法,通过Kinect对室内场景进行三维重建。利用加速鲁棒性特征(SURF)算法提取特征点,分别使用传统的ICP和基于K-D树改进ICP算法完成多帧点云数据的配准,对比单一场景和多场景下的模型重建效果。结果表明,在面积为30 m^(2)的房间三维模型重建中,传统ICP算法处理时间分别为1.31 min(单一场景)和8.06 min(多场景),而改进ICP算法处理时间为0.67 min(单一场景)和5.23 min(多场景)。改进后的ICP算法三维重建速度较快,没有明显的物品位置错乱等情况,能满足日常需要。In order to solve the time-consuming problem of the traditional iterative closest point(ICP)algorithm,the ICP algorithm was improved based on K-D tree,and the 3D reconstruction of indoor scenes was performed through Kinect.The Speeded Up Robust Features(SURF)algorithm was used to extract feature points.The traditional ICP algorithm and the improved ICP algorithm based on K-D tree were respectively used to complete the registration of multi frame point cloud data,and the model reconstruction effects under single scene and multi scene were compared.The results show that in the 3D model reconstruction with 30 m^(2) room,the processing time of the traditional ICP algorithm was 1.31 min(single scene)and 8.06 min(multiple scenes)respectively,while the processing time of the improved ICP algorithm was 0.67 min(single scene)and 5.23 min(multiple scenes).The improved ICP algorithm has a fast 3D reconstruction speed,and no obvious dislocations of items,so it can meet daily needs.

关 键 词：三维重建 KINECT 迭代最近点算法 K-D树 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]