This repository contains a case study for the work developed by Malof, J. et al. in Mapping solar array location, size, and capacity using deep learning and overhead imagery [1]. Here is shown the transfer learning process for photovoltaic panels segmentation with the DeepLab V3 architecture trained for ImageNet.
Run the file main.py. If you want tho check the transfer learning process, see the segmentation_model/deep_lab_v3.py file. To download the dataset and the trained model, go to this link: https://drive.google.com/drive/folders/1HF9cHprGdPsjziL-NKyU9dU2R6fUEUGu?usp=sharing
See SETUP.md.
To use the repo run the main.py file that has the following arguments:
- train_dataset: Path to train dataset (default: 'dataset/train').
- test_dataset: Path to test dataset (default: 'dataset/test').
- train_model: Set True to train a new model (default: False).
- im_resize: Pixels to resize the images (default: 500).
- batch_size: Batch size for training process (default: 8).
- backbone: Model backbone (default: 'RESNET101').
- optimizer: Optimizer for training process (default: 'Adam').
- lr: Learning rate for training process (default: 0.001).
- epochs: Number of epochs for training process (default: 50).
- trained_model: Path to trained model (default: 'trained_models/deeplab_v3_RESNET101_model.pt').
- output_results: Path to save the inference outputs (default: 'results/').
Example:
$ python main.py --train False@article{gaviria_machine_2022,
title = {Machine learning in photovoltaic systems: A review},
issn = {0960-1481},
url = {https://www.sciencedirect.com/science/article/pii/S0960148122009454},
doi = {10.1016/j.renene.2022.06.105},
shorttitle = {Machine learning in photovoltaic systems},
abstract = {This paper presents a review of up-to-date Machine Learning ({ML}) techniques applied to photovoltaic ({PV}) systems, with a special focus on deep learning. It examines the use of {ML} applied to control, islanding detection, management, fault detection and diagnosis, forecasting irradiance and power generation, sizing, and site adaptation in {PV} systems. The contribution of this work is three fold: first, we review more than 100 research articles, most of them from the last five years, that applied state-of-the-art {ML} techniques in {PV} systems; second, we review resources where researchers can find open data-sets, source code, and simulation environments that can be used to test {ML} algorithms; third, we provide a case study for each of one of the topics with open-source code and data to facilitate researchers interested in learning about these topics to introduce themselves to implementations of up-to-date {ML} techniques applied to {PV} systems. Also, we provide some directions, insights, and possibilities for future development.},
journaltitle = {Renewable Energy},
shortjournal = {Renewable Energy},
author = {Gaviria, Jorge Felipe and Narváez, Gabriel and Guillen, Camilo and Giraldo, Luis Felipe and Bressan, Michael},
urldate = {2022-07-03},
date = {2022-07-01},
langid = {english},
keywords = {Deep learning, Machine learning, Neural networks, Photovoltaic systems, Reinforcement learning, Review},
file = {ScienceDirect Snapshot:C\:\\Users\\jfgf1\\Zotero\\storage\\G96H46L2\\S0960148122009454.html:text/html},
},
@article{malof2019mapping,
title={Mapping solar array location, size, and capacity using deep learning and overhead imagery},
author={Malof, Jordan M and Li, Boning and Huang, Bohao and Bradbury, Kyle and Stretslov, Artem},
journal={arXiv preprint arXiv:1902.10895},
year={2019}
}[1] Jorge Felipe Gaviria, Gabriel Narváez, Camilo Guillen, Luis Felipe Giraldo, and Michael Bressan. Machine learning in photovoltaic systems: A review. ISSN 0960-1481. doi: 10.1016/j.renene.2022.06.105. URL https://www.sciencedirect.com/science/article/pii/S0960148122009454?via%3Dihub
[2] Malof, J. M., Li, B., Huang, B., Bradbury, K., & Stretslov, A. (2019). Mapping solar array location, size, and capacity using deep learning and overhead imagery. arXiv preprint arXiv:1902.10895.
The software is licensed under an MIT License. A copy of the license has been included in the repository and can be found here.