基于被动视觉的三维重建技术研究进展  

Progress of Study on 3D Reconstruction Technology Based on Passive Vision

作　　者：王兆庆 牛朝一 佘维[1,3] 宰光军 梁波[4] 易建锋 李英豪 WANG Zhaoqing;NIU Chaoyi;SHE Wei;ZAI Guangjun;LIANG Bo;YI Jianfeng;LI Yinghao(School of Cyber Science and Engineering,Zhengzhou University,Zhengzhou 450002,China;XJ Electric Co.,Ltd,Xuchang 461000,China;Zhengzhou Key Laboratory of Blockchain and Data Intelligence,Zhengzhou 450002,China;Information Management Center,Zhongyuan Oilfield Branch of SINOPEC,Puyang 457001,China)

机构地区：[1]郑州大学网络空间安全学院,河南郑州450002 [2]许继电气股份有限公司,河南许昌461000 [3]郑州市区块链与数据智能重点实验室,河南郑州450002 [4]中国石油化工股份有限公司中原油田分公司信息化管理中心,河南濮阳457001

出　　处：《郑州大学学报（理学版）》2024年第5期13-19,共7页Journal of Zhengzhou University:Natural Science Edition

基　　金：国家自然科学基金项目(62206252);河南省重点研发与推广专项(212102310039)。

摘　　要：基于被动视觉的三维重建技术方法多样、应用广泛。按照不同的分类方法,对基于被动视觉的三维重建技术研究进展进行了分析总结。首先,根据采集装置的数量进行分类,介绍了基于单目视觉、双目视觉和多目视觉的三维重建技术,并对各种方法的优缺点进行比较。其次,根据不同应用方法进行分类,对运动恢复结构法和深度学习法的研究进展进行了阐述。最后,对基于被动视觉的三维重建方法进行了综合对比分析,并对三维重建的应用和发展进行了展望。The methods of 3D reconstruction technology based on passive vision are diverse and widely used.Most of studies of 3D reconstruction technology based on passive vision were analyzed and summa-rized based on different classification methods.Firstly,according to the different number of acquisition devices,the 3D reconstruction techniques were classified into monocular vision,binocular vision and multi-version,and the advantages and disadvantages of each method were compared.Secondly,the re-search progress of structure from motion and deep learning method were described according to the classi-fication of different application methods.Finally,the 3D reconstruction technology based on passive vi-sion were compared and analyzed comprehensively,and the application and development of 3D recon-struction were explored.

关 键 词：三维重建 被动视觉 单目视觉 双目视觉 多目视觉 运动恢复结构 深度学习 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]