Deep learning models for classification of 15 common weeds in the southern U.S. cotton production systems.
- pytorch
- torchsummary
- tensorboard
- PIL
- Scikit-learn
- The weed image dataset is publicly available at https://www.kaggle.com/yuzhenlu/cottonweedid15
- To prepare your own dataset, you can run
python common/partition_imgs_Ubuntu.py
- To train the models, just specify the name of the models, and then run
python train.py
. - To test the images, just specify the name of the models, and then run
python test.py
. - To eval new data, just specify the name of the models, and then run
python eval.py
. - To visualize the training, run
tensorboard --logdir=runs
Detailed documentation of deep transfer learning for weed classification of the cotton weed dataset is given in our paper: https://www.sciencedirect.com/science/article/pii/S0168169922004082. If you use the dataset or models in a publication, please cite this paper.
@article{chen2022performance,
title={Performance evaluation of deep transfer learning on multi-class identification of common weed species in cotton production systems},
author={Chen, Dong and Lu, Yuzhen and Li, Zhaojian and Young, Sierra},
journal={Computers and Electronics in Agriculture},
volume={198},
pages={107091},
year={2022},
publisher={Elsevier}
}