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.
作者
王兆庆
牛朝一
佘维
宰光军
梁波
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)
出处
《郑州大学学报(理学版)》
CAS
北大核心
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)。
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
3D reconstruction
passive vision
monocular vision
binocular vision
multi-vision
struc-ture from motion
deep learning