Mejora de la detección de defectos superficiales en paneles solares con modelos VGG basados en IA

Naima El Yanboiy; Mohamed Khala; Ismail Elabbassi; Nourddine Elhajrat; Omar Eloutassi; Youssef El Hassouani; Choukri Messaoudi

doi:10.56294/dm202381

Original

Published: 2023-12-28

DOI: https://doi.org/10.56294/dm202381

Enhancing Surface Defect Detection in Solar Panels with AI-Driven VGG Models

Abstract

In recent years, the demand for solar energy has increased considerably. This growing demand has created a corresponding need for solar panel systems that not only demonstrate efficiency, but also guarantee reliability. However, The per-formance and durability of solar panels can be significantly affected by diverse faults such as surface defects, cracks, hot spots and accumulations of dust. Thus, early detection is crucial to ensure optimal operation of solar panels. In this study, we propose an intelligent system for detecting surface defects on solar panels us-ing the Visual Geometry Group (VGG) models. A camera is utilized to capture images of solar panels in both normal and defective states, these images are sub-sequently fed into the trained VGG model, which analyzes and processes them to identify defects on the surface of the solar panel. The experimental results show that the VGG19 model outperforms the VGG16 model in detecting faulty solar panels. VGG19 achieved a precision of 80%, a recall of 1, and an F1 score of 89%, while VGG16 achieved a precision of 79%, a recall of 92%, and an F1 score of 85%. Furthermore, the system demonstrated a high accuracy for the VGG19 in detecting surface panels in their normal state, while for the VGG16 it only achieved 90%. The results demonstrate the ability of the VGG19 model to detect surface defects on solar panels based on visual analysis.

Keywords:

Smart detection,

Surface defects,

Anomaly detection,

Solar panel,

VGG model,

How to Cite

El Yanboiy N, Khala M, Elabbassi I, Elhajrat N, Eloutassi O, El Hassouani Y, Messaoudi C. Enhancing Surface Defect Detection in Solar Panels with AI-Driven VGG Models. Data and Metadata [Internet]. 2023 Dec. 28 [cited 2024 Jul. 27];2:81. Available from: https://dm.saludcyt.ar/index.php/dm/article/view/81

Copyright Notice

The article is distributed under the Creative Commons Attribution 4.0 License. Unless otherwise stated, associated published material is distributed under the same licence.

Article metrics

Google scholar: See link

Metrics

Metrics Loading ...

Vol. 2 (2023)

See full issue

Revistas / Journals

Issue

About

Submissions