Share Email Print
cover

Proceedings Paper

Ship detection based on spatial partial features
Author(s): Hong-hu Wang; Fu-gen Zhou; Ting Jin
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

This paper mainly studies how to detect a wide variety of ships from the ship-borne infrared images in order to implement sea monitoring. Different types of ships have significant differences in their appearance. The traditional detection method which uses the global texture features of the object is not suitable to detect varied ships. This paper presents a novel detection algorithm which extracts spatial partial texture features trained by Adaboost to establish the ship model for detection. We first extract all the partial regions of the object through random traversal, and then extract the texture features by using the “Uniform LBP” operator. Compared to the traditional way, we save each partial feature individually as one feature vector, which not only reduces the vector dimension but also highlights the key regions when the partial regions with strong generality are selected by Adaboost at the second step. Finally, the selected partial features are boosted with weights to establish ship model for the ship detection. The proposed approach is efficient and robust in the infrared ship detection.

Paper Details

Date Published: 11 September 2013
PDF: 8 pages
Proc. SPIE 8907, International Symposium on Photoelectronic Detection and Imaging 2013: Infrared Imaging and Applications, 89073M (11 September 2013); doi: 10.1117/12.2034060
Show Author Affiliations
Hong-hu Wang, BeiHang Univ. (China)
Fu-gen Zhou, BeiHang Univ. (China)
Ting Jin, BeiHang Univ. (China)


Published in SPIE Proceedings Vol. 8907:
International Symposium on Photoelectronic Detection and Imaging 2013: Infrared Imaging and Applications
Haimei Gong; Zelin Shi; Qian Chen; Jin Lu, Editor(s)

© SPIE. Terms of Use
Back to Top