This repository is part of the codes used in this paper. Specifically, this code implements the feature extraction of pre-trained deep-learning models using Caffe framework. This code, which is very easy to use, was created following this tutorial.
The networks implemented in this code and that can be used to extract deep features of any dataset are:
The Brazilian Coffee Scenes Dataset exploited in this work can be found here.
This dataset is a composition of scenes taken by SPOT sensor in 2005 over four counties in the State of Minas Gerais, Brazil: Arceburgo, Guaranesia, Guaxupé and Monte Santo. It has many intraclass variances caused by different crop management techniques. Also, coffee is an evergreen culture and the South of Minas Gerais is a mountainous region, which means that this dataset includes scenes with different plant ages and/or with spectral distortions caused by shadows.
Some of the features exploited in this work can be downloaded here:
If you use this code or features in your research, please consider citing:
@article{nogueira2017towards,
title={Towards Better Exploiting Convolutional Neural Networks for Remote Sensing Scene Classification},
author={Nogueira, Keiller and Penatti, Ot{\'a}vio AB and Santos, Jefersson A dos},
journal=pr,
volume={61},
number={1},
pages={539--556},
year={2017},
publisher={Elsevier}
}
@inproceedings{penatti2015deep,
title={Do deep features generalize from everyday objects to remote sensing and aerial scenes domains?},
author={Penatti, Otavio and Nogueira, Keiller and dos Santos, Jefersson},
booktitle=cvprw,
pages={44--51},
year={2015}
}