key words: machine learning ; deep learning ; SAT-6 ; SAT-4 ; ARID-5 ; Remote Sensing ; Classfication ; mgrForest
a machine learning method improved by deep forest
achieved SOTA on SAT-4 & SAT-6
In addition, we released a dataset named ARID-5, which is used for arid area classification on 28*28 pixel image.
https://doi.org/10.1016/j.cageo.2020.104557
if you used our work or code, please google our work and cite:
@article{weng2020land,
title={Land use/land cover recognition in arid zone using A multi-dimensional multi-grained residual Forest☆},
author={Weng, Liguo and Qian, Ming and Xia, Min and Xu, Yiqing and Li, Chunzheng},
journal={Computers & Geosciences},
volume={144},
pages={104557},
year={2020},
publisher={Elsevier}
}
The dataset:https://pan.baidu.com/s/1RjF3BSOzIDhmq0oAi7-xnA, passwd:dpcu