Based on the publicly opened SAR dataset of ships, methods for ship classification have been presented in this paper. For comparison, a joint feature based method for ship classification for SAR is described first. In this method, features for SAR ship classification are concluded, in which density of RCS and main-structure feature have been proposed to discriminate ships. Afterwards, a modified LeNet based method has been presented for SAR ship classification, by restricting the size of convolutional window and layers according to the properties of SAR. Experiments are conducted on the real measured data to show the effectiveness of the methods above. And by comparing the methods, the proper method for SAR ship classification has been concluded, as well.
Ship Classification Methods for Sentinel-1 SAR Images
Lect. Notes Electrical Eng.
International Conference in Communications, Signal Processing, and Systems ; 2019 ; Urumqi, China July 20, 2019 - July 22, 2019
2020-04-04
11 pages
Article/Chapter (Book)
Electronic Resource
English
Sentinel-1 Ship Monitoring Applications
VDE-Verlag | 2008
|Sentinel-2 Images and Finnish Corine Land Cover Classification
British Library Conference Proceedings | 2012
|