In this paper, we propose a novel ship detection method based on multi-visual features after analyzing the characteristics of ship in the sea. According to the principal of the visual contrast, brightness and orientation saliency map of ship object are respectively generated, and then they are integrated to obtain the total saliency map. In addition to the brightness and orientation of the ship objects, the method doesn’t use other prior knowledge of them. In ship detection experiment, the experimental results prove our method can effectively concentrate on the ship objects regardless of their size and brightness, and thereby improve the capacity of visual attention in complex scene. Thus, the design idea of our method is verified.
Multi-Visual Features Based Ship Detection in Remote Sensing Images
Applied Mechanics and Materials ; 643 ; 228-232
2014-09-12
5 pages
Article (Journal)
Electronic Resource
English
Marine Vision-Based Situational Automatic Ship Detection Using Remote Sensing Images
Springer Verlag | 2024
|Ship object detection in remote sensing images using convolutional neural networks
British Library Online Contents | 2017
|