Generative Adversarial Network for Synthetic Imaging Data of Sub-optimal Conditions in Photovoltaic Panels

Efrén Jiménez-Delgado; A. Méndez-Porras; J. Alfaro-Velasco

doi:10.1007/978-3-030-96293-7_20

Generative Adversarial Network for Synthetic Imaging Data of Sub-optimal Conditions in Photovoltaic Panels

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

8 Scopus citations

Abstract

Image detection technology is currently used to perform inspection and detection of defects in solar panels. Deep learning models are typically used to process visual light images, thermal or infrared images. Building machine learning models takes miles or millions of images. Currently, there are few freely available or freely accessible image data sets that contain suboptimal conditions in solar panels. This paper proposed an Adversary Generative Network (GAN) model to build a data set of photovoltaic panels under suboptimal conditions for Deep Learning training. We use a data set created at the Tecnológico de Costa Rica to generate an adversary neural network model. In addition, we evaluate our model using a convolutional neural network that classifies whether or not a visual light image or thermal images contains suboptimal conditions.

Original language	English
Title of host publication	Information Technology and Systems - Proceedings of ICITS 2022
Editors	Álvaro Rocha, Carlos Ferrás, Efren Jimenez Delgado, Abel Méndez Porras
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	206-218
Number of pages	13
ISBN (Print)	9783030962920
DOIs	https://doi.org/10.1007/978-3-030-96293-7_20
State	Published - 2022
Event	International Conference on Information Technology and Systems, ICITS 2022 - San Carlos, Costa Rica Duration: 9 Feb 2022 → 11 Feb 2022

Publication series

Name	Lecture Notes in Networks and Systems
Volume	414 LNNS
ISSN (Print)	2367-3370
ISSN (Electronic)	2367-3389

Conference

Conference	International Conference on Information Technology and Systems, ICITS 2022
Country/Territory	Costa Rica
City	San Carlos
Period	9/02/22 → 11/02/22

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 7 Affordable and Clean Energy

Keywords

CNN
Deep learning
GAN
Generative adversarial network
Neural network
Photovoltaic panels
Suboptimal conditions
Synthetic data

Access to Document

10.1007/978-3-030-96293-7_20

Cite this

Jiménez-Delgado, E., Méndez-Porras, A., & Alfaro-Velasco, J. (2022). Generative Adversarial Network for Synthetic Imaging Data of Sub-optimal Conditions in Photovoltaic Panels. In Á. Rocha, C. Ferrás, E. J. Delgado, & A. M. Porras (Eds.), Information Technology and Systems - Proceedings of ICITS 2022 (pp. 206-218). (Lecture Notes in Networks and Systems; Vol. 414 LNNS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-96293-7_20

Jiménez-Delgado, Efrén ; Méndez-Porras, A. ; Alfaro-Velasco, J. / Generative Adversarial Network for Synthetic Imaging Data of Sub-optimal Conditions in Photovoltaic Panels. Information Technology and Systems - Proceedings of ICITS 2022. editor / Álvaro Rocha ; Carlos Ferrás ; Efren Jimenez Delgado ; Abel Méndez Porras. Springer Science and Business Media Deutschland GmbH, 2022. pp. 206-218 (Lecture Notes in Networks and Systems).

@inproceedings{11c1291220ee4af0940ec531058d16cf,

title = "Generative Adversarial Network for Synthetic Imaging Data of Sub-optimal Conditions in Photovoltaic Panels",

abstract = "Image detection technology is currently used to perform inspection and detection of defects in solar panels. Deep learning models are typically used to process visual light images, thermal or infrared images. Building machine learning models takes miles or millions of images. Currently, there are few freely available or freely accessible image data sets that contain suboptimal conditions in solar panels. This paper proposed an Adversary Generative Network (GAN) model to build a data set of photovoltaic panels under suboptimal conditions for Deep Learning training. We use a data set created at the Tecnol{\'o}gico de Costa Rica to generate an adversary neural network model. In addition, we evaluate our model using a convolutional neural network that classifies whether or not a visual light image or thermal images contains suboptimal conditions.",

keywords = "CNN, Deep learning, GAN, Generative adversarial network, Neural network, Photovoltaic panels, Suboptimal conditions, Synthetic data",

author = "Efr{\'e}n Jim{\'e}nez-Delgado and A. M{\'e}ndez-Porras and J. Alfaro-Velasco",

note = "Publisher Copyright: {\textcopyright} 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.; International Conference on Information Technology and Systems, ICITS 2022 ; Conference date: 09-02-2022 Through 11-02-2022",

year = "2022",

doi = "10.1007/978-3-030-96293-7\_20",

language = "Ingl{\'e}s",

isbn = "9783030962920",

series = "Lecture Notes in Networks and Systems",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "206--218",

editor = "{\'A}lvaro Rocha and Carlos Ferr{\'a}s and Delgado, \{Efren Jimenez\} and Porras, \{Abel M{\'e}ndez\}",

booktitle = "Information Technology and Systems - Proceedings of ICITS 2022",

}

Jiménez-Delgado, E , Méndez-Porras, A & Alfaro-Velasco, J 2022, Generative Adversarial Network for Synthetic Imaging Data of Sub-optimal Conditions in Photovoltaic Panels. in Á Rocha, C Ferrás, EJ Delgado & AM Porras (eds), Information Technology and Systems - Proceedings of ICITS 2022. Lecture Notes in Networks and Systems, vol. 414 LNNS, Springer Science and Business Media Deutschland GmbH, pp. 206-218, International Conference on Information Technology and Systems, ICITS 2022, San Carlos, Costa Rica, 9/02/22. https://doi.org/10.1007/978-3-030-96293-7_20

Generative Adversarial Network for Synthetic Imaging Data of Sub-optimal Conditions in Photovoltaic Panels. / Jiménez-Delgado, Efrén ; Méndez-Porras, A.; Alfaro-Velasco, J.
Information Technology and Systems - Proceedings of ICITS 2022. ed. / Álvaro Rocha; Carlos Ferrás; Efren Jimenez Delgado; Abel Méndez Porras. Springer Science and Business Media Deutschland GmbH, 2022. p. 206-218 (Lecture Notes in Networks and Systems; Vol. 414 LNNS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Generative Adversarial Network for Synthetic Imaging Data of Sub-optimal Conditions in Photovoltaic Panels

AU - Jiménez-Delgado, Efrén

AU - Méndez-Porras, A.

AU - Alfaro-Velasco, J.

PY - 2022

Y1 - 2022

N2 - Image detection technology is currently used to perform inspection and detection of defects in solar panels. Deep learning models are typically used to process visual light images, thermal or infrared images. Building machine learning models takes miles or millions of images. Currently, there are few freely available or freely accessible image data sets that contain suboptimal conditions in solar panels. This paper proposed an Adversary Generative Network (GAN) model to build a data set of photovoltaic panels under suboptimal conditions for Deep Learning training. We use a data set created at the Tecnológico de Costa Rica to generate an adversary neural network model. In addition, we evaluate our model using a convolutional neural network that classifies whether or not a visual light image or thermal images contains suboptimal conditions.

AB - Image detection technology is currently used to perform inspection and detection of defects in solar panels. Deep learning models are typically used to process visual light images, thermal or infrared images. Building machine learning models takes miles or millions of images. Currently, there are few freely available or freely accessible image data sets that contain suboptimal conditions in solar panels. This paper proposed an Adversary Generative Network (GAN) model to build a data set of photovoltaic panels under suboptimal conditions for Deep Learning training. We use a data set created at the Tecnológico de Costa Rica to generate an adversary neural network model. In addition, we evaluate our model using a convolutional neural network that classifies whether or not a visual light image or thermal images contains suboptimal conditions.

KW - CNN

KW - Deep learning

KW - GAN

KW - Generative adversarial network

KW - Neural network

KW - Photovoltaic panels

KW - Suboptimal conditions

KW - Synthetic data

UR - https://www.scopus.com/pages/publications/85126229197

U2 - 10.1007/978-3-030-96293-7_20

DO - 10.1007/978-3-030-96293-7_20

M3 - Contribución a la conferencia

AN - SCOPUS:85126229197

SN - 9783030962920

T3 - Lecture Notes in Networks and Systems

SP - 206

EP - 218

BT - Information Technology and Systems - Proceedings of ICITS 2022

A2 - Rocha, Álvaro

A2 - Ferrás, Carlos

A2 - Delgado, Efren Jimenez

A2 - Porras, Abel Méndez

PB - Springer Science and Business Media Deutschland GmbH

T2 - International Conference on Information Technology and Systems, ICITS 2022

Y2 - 9 February 2022 through 11 February 2022

ER -

Jiménez-Delgado E , Méndez-Porras A , Alfaro-Velasco J. Generative Adversarial Network for Synthetic Imaging Data of Sub-optimal Conditions in Photovoltaic Panels. In Rocha Á, Ferrás C, Delgado EJ, Porras AM, editors, Information Technology and Systems - Proceedings of ICITS 2022. Springer Science and Business Media Deutschland GmbH. 2022. p. 206-218. (Lecture Notes in Networks and Systems). doi: 10.1007/978-3-030-96293-7_20

Generative Adversarial Network for Synthetic Imaging Data of Sub-optimal Conditions in Photovoltaic Panels

Abstract

Publication series

Conference

UN SDGs

Keywords

Access to Document

Other files and links

Fingerprint

Cite this