Open AccessSolar Panels Crack Detection using Overhead Images
Author(s) -
Saksham Checker
Publication year - 2021
Publication title -
international journal for research in applied science and engineering technology
Language(s) - English
Resource type - Journals
ISSN - 2321-9653
DOI - 10.22214/ijraset.2021.38532
Subject(s) - renewable energy , drone , overhead (engineering) , computer science , solar energy , deep learning , simulation , automotive engineering , artificial intelligence , reliability engineering , embedded system , electrical engineering , engineering , operating system , genetics , biology
Many countries like India have been working on plans to shift from conventional energy sources to renewable, and solar energy is one of them. The technology of Solar Panels faces defects on a large scale. These defects can be within the cells or on the panel as a whole. Management authorities require automated systems to detect the physical faults on the solar panels. These faults can either be detected by physically looking at the panels or through the energy supply, which requires Machine Learning models. This paper proposes a model with 95.34% accuracy that can be deployed on drones and automatically check for physical damage on the panels. We collected the dataset manually from the internet, cleaned and split it into training and validation parts. Data augmentation was performed to get a better view of the functioning of the model. Keywords: Solar Panels, Solar Energy, Renewable Energy, Deep Learning, Machine Learning.