Skip to content

Latest commit

 

History

History
48 lines (34 loc) · 2.98 KB

README.md

File metadata and controls

48 lines (34 loc) · 2.98 KB
Raw

Conditional GANs for Crop Growth Simulations

Teaser image Picture: Generate images of different crop mixture growth stages under simulated different treatments.

Data-driven Crop Growth Simulation on Time-varying Generated Images using Multi-conditional Generative Adversarial Networks (arXiv 23)

Requirements

A suitable conda environment can be installed from the provided package file environment.yml

conda env create -f environment.yaml
conda activate crop-growth-cgan

Data

All datasets used in this work are publicly available: Arabidopsis (re-calibrated): abbreviated abdc, MixedCrop-CKA: mix, MixedCrop-WG: mix-wg, GrowliFlower: grf

Data preprocessing scripts and process-based simulation files will be available soon. Please contact the author for further information.

Training

There are train scripts depending on which conditions are to be considered for the multi-conditional GAN. The conditions are input images (img), the time or growth stage (t), treatment information as class variables (cls) and other influencing factors (if), such as in this case daily simulated biomass. A training can be started after setting the requirered parameter in the corresponding config file in ./configs/ (descriptions in the file).

python train_img_t_cls.py

Thereby the settings from ./configs/config_train_img_t_cls.py are used. For other train scripts accordingly. To resume a previous training, please enter the experiment name there and ensure, that architecture parameters are consistent.

Testing

To test a model trained with the above command

python test_img_t_cls.py

You need to specify at least the log dir and the experiment name which you want to evaluate in ./configs/config_test_img_t_cls.py. Beyond that per default all parameters used to train the model are also used for testing. But you can also try to change e.g. the dataset e.g. from mix to mix-wg. Weights of the growth estimations models will be provided soon and should be saved in the folder ./eval_model_weights.

Citation

If you use this code for your research, please cite our paper.

@Article{drees2023datadriven,
  author  = {Lukas Drees and Dereje T. Demie and Madhuri R. Paul and Johannes Leonhardt and Sabine J. Seidel and Thomas F. D{\"o}ring and Ribana Roscher},
  title   = {Data-driven Crop Growth Simulation on Time-varying Generated Images using Multi-conditional Generative Adversarial Networks},
  journal = {arXiv preprint arXiv:2312.03443},
  year    = {2023},
  doi     = {10.48550/arXiv.2312.03443},
}