Open AccessA novel method of extracting geometric features of ships based on GrabCut algorithm
Author(s) -
Jia-Le Zha,
Huaixin Chen,
Chao Ren,
Chenggang Wang,
Siqi Li
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1693/1/012092
Subject(s) - rectangle , cut , segmentation , artificial intelligence , robustness (evolution) , computer science , computer vision , image segmentation , gaussian , minification , algorithm , pattern recognition (psychology) , mathematics , geometry , biochemistry , chemistry , physics , quantum mechanics , gene , programming language
Using remote sensing images to accurately identify ships is of great significance in both military reconnaissance and civil surveillance. This paper presents a method of extracting geometric features of ships based on GrabCut image segmentation. Firstly, by establishing gaussian mixture model of image area, the problem of segmentation of suspected target area is transformed into graph minimum cut, then the suspected target is segmented by iteration energy minimization, and the moment method is used to extract the minimum enclosing rectangle of ship target, so as to obtain the geometric features of ship target. Experimental data analysis shows that compared with the traditional algorithms, the proposed method can quickly extract the geometric features of ships in different directions in remote sensing images, and has strong robustness, which can meet the requirements of target detection engineering.