To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
Ship Detection in Satellite Imagery by Multiple Classifier Network |
|
Author |
Ahsan Raza Siyal, Zuhaibuddin Bhutto, Kashif Saleem, Abdul Sattar Chan, Mudasar Latif Memon, Murtaza Hussain Shaikh, and Saleem Ahmed |
| Citation |
Vol. 19 No. 8 pp. 142-148 |
|
Abstract |
Automatic ships detection in low-resolution satellite imagery is a challenging task. In this paper, we compare state of the art classifiers like support vector machine, Random Forest, Linear discriminant analysis, K nearest neighbor and deep Convolutional neural network for ship detection and classification in satellite imagery. We proposed a novel method with Convolutional neural network which improves robustness of the system, accuracy for ship detection and reduces noise due to weather conditions and high waves. The open-source dataset of planet ship is used to test the results of the proposed scheme, containing preprocessed 2800 images. The dataset has two main classes “ship” and “no-ship”, within ship class 700 images of various type and sizes of ships having diverse atmospheric conditions. The class having 2100 images of no-ship with random land covers with the ocean, earth and cities land cover with ocean featuring no ship in them. The simulation results presented here show that the proposed scheme gives improved detection accuracy and robustness in low-resolution satellite images having bad weather conditions. |
|
Keywords |
satellite, state-of-the-art, convolutional neural network, robustness. |
|
URL |
