Graph-Based Registration and Blending for Undersea Image Stitching

doi:10.1017/S0263574719000699

CORC > 自动化研究所 > 中国科学院自动化研究所 > 复杂系统管理与控制国家重点实验室 > 机器人应用与理论组

	Graph-Based Registration and Blending for Undersea Image Stitching
	Yang, Xu 2; Liu, Zhi-Yong 2; Qiao, Hong 2; Su, Jian-Hua 2; Ji, Da-Xiong 3; Zang, Ai-Yun 4; Huang, Hai 1
刊名	ROBOTICA
	2020-03-01
卷号	38 期号:3 页码:396-409
关键词	Undersea image stitching Feature correspondence Graph matching Energy minimization Nonsubmodular function
ISSN号	0263-5747
DOI	10.1017/S0263574719000699
通讯作者	Huang, Hai(haihus@163.com)
英文摘要	Image stitching is important for the perception and manipulation of undersea robots. In spite of a well-developed technique, it is still challenging for undersea images because of their inevitable appearance ambiguity caused by the limited light in the undersea environment, and local disturbance caused by moving objects, ocean current, etc. To get a clean and stable background panorama in the undersea environment, this paper proposes an undersea image-stitching method by introducing graph-based registration and blending procedures. Specifically, in the registration procedure, matching the features in each undersea image pair is formulated and solved by graph matching, to incorporate the structural information between features. In the blending procedure, an energy function on the indirect graph Markov random field is proposed, which takes both image consistency and neighboring consistency into consideration. Coincidentally, both graph matching and energy minimization can be mathematically formulated by integer quadratic programming problems with different constraints; the recently proposed graduated nonconvexity and concavity procedure is used to optimize both problems. Experiments on both synthetic images and real-world undersea images witness the effectiveness of the proposed method.
资助项目	National Natural Science Foundation (NSFC) of China[61633009] ; National Natural Science Foundation (NSFC) of China[61503383] ; National Natural Science Foundation (NSFC) of China[U1613213] ; National Natural Science Foundation (NSFC) of China[61627808] ; National Natural Science Foundation (NSFC) of China[91648205] ; National Natural Science Foundation (NSFC) of China[U1509212] ; National Key R&D Program of China[2016YFC0300801] ; National Key R&D Program of China[2017YFB1300202] ; Strategic Priority Research Program of Chinese Academy of Sciences[XDB32000000]
WOS关键词	ENERGY MINIMIZATION
WOS研究方向	Robotics
语种	英语
出版者	CAMBRIDGE UNIV PRESS
WOS记录号	WOS:000512402600002
资助机构	National Natural Science Foundation (NSFC) of China ; National Key R&D Program of China ; Strategic Priority Research Program of Chinese Academy of Sciences
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/28562]
专题	自动化研究所_复杂系统管理与控制国家重点实验室_机器人应用与理论组
通讯作者	Huang, Hai
作者单位	1.Harbin Engn Univ, Natl Key Lab Sci & Technol Underwater Vehicle, Harbin 150001, Peoples R China 2.Chinese Acad Sci, Inst Automat, State Key Lab Management & Control Complex Syst, Beijing 100190, Peoples R China 3.Zhejiang Univ, Ocean Coll, Zhoushan 316000, Peoples R China 4.Ocean Univ China, Coll Engn, Qingdao 266100, Peoples R China
推荐引用方式 GB/T 7714	Yang, Xu,Liu, Zhi-Yong,Qiao, Hong,et al. Graph-Based Registration and Blending for Undersea Image Stitching[J]. ROBOTICA,2020,38(3):396-409.
APA	Yang, Xu.,Liu, Zhi-Yong.,Qiao, Hong.,Su, Jian-Hua.,Ji, Da-Xiong.,...&Huang, Hai.(2020).Graph-Based Registration and Blending for Undersea Image Stitching.ROBOTICA,38(3),396-409.
MLA	Yang, Xu,et al."Graph-Based Registration and Blending for Undersea Image Stitching".ROBOTICA 38.3(2020):396-409.