Ship detection in optical remote sensing image based on visual saliency and AdaBoost classifier

In this paper, firstly, target candidate regions are extracted by combining maximum symmetric surround saliency detection algorithm with a cellular automata dynamic evolution model. Secondly, an eigenvector independent of the ship target size is constructed by combining the shape feature with ship h...

Full description

Saved in:
Bibliographic Details
Published inOptoelectronics letters Vol. 13; no. 2; pp. 151 - 155
Main Author 王慧利 朱明 蔺春波 陈典兵
Format Journal Article
LanguageEnglish
Published Tianjin Tianjin University of Technology 01.03.2017
Springer Nature B.V
Subjects
AdaBoost
Cellular automata
Lasers
Optical Devices
Optics
Photonics
Physics
Physics and Astronomy
光学
分类器
形状特征
显著性
舰船检测
视觉
遥感图像
A
Online AccessGet full text

Cover

More Information
Summary:In this paper, firstly, target candidate regions are extracted by combining maximum symmetric surround saliency detection algorithm with a cellular automata dynamic evolution model. Secondly, an eigenvector independent of the ship target size is constructed by combining the shape feature with ship histogram of oriented gradient（S-HOG） feature, and the target can be recognized by Ada Boost classifier. As demonstrated in our experiments, the proposed method with the detection accuracy of over 96% outperforms the state-of-the-art method. efficiency switch and modulation.
Bibliography:classifier AdaBoost histogram automata symmetric pixel candidate similarity surround segmentation
12-1370/TN
In this paper, firstly, target candidate regions are extracted by combining maximum symmetric surround saliency detection algorithm with a cellular automata dynamic evolution model. Secondly, an eigenvector independent of the ship target size is constructed by combining the shape feature with ship histogram of oriented gradient（S-HOG） feature, and the target can be recognized by Ada Boost classifier. As demonstrated in our experiments, the proposed method with the detection accuracy of over 96% outperforms the state-of-the-art method. efficiency switch and modulation.
ISSN:1673-1905
1993-5013
DOI:10.1007/s11801-017-7014-9