Publication: AI-based pv module extraction and hotspot fault detection using image processing technique
| datacite.subject.fos | oecd::Engineering and technology::Electrical engineering, Electronic engineering, Information engineering::Electrical and electronic engineering | |
| dc.contributor.author | Soh, Jing Cheng | |
| dc.date.accessioned | 2025-03-19T07:17:51Z | |
| dc.date.available | 2025-03-19T07:17:51Z | |
| dc.date.issued | 2023-07 | |
| dc.description.abstract | The rise in awareness of the shortage of non-renewable energy has led to the drastic development of renewables, including solar energy. The wide application of PV modules to generate energy brings out another problem that needs to be solved: efficiency and maintenance. Thus, three objectives were established in this study: building a PV module extraction model, detecting faults using thermal imaging, and creating a GUI to visualize the results. After reading a few related articles, valuable parts are learned and utilized, such as involving AI to achieve the objectives, imposing image processing techniques on the images, etc. Some terminologies were studied to understand the principle of the technique used in this study. This paper presents a solution with three stages: module extraction model stage involving the YOLOv5 algorithm, hotspot detection stage involving statistical analysis, and creation of a GUI stage using Python language. Consequently, the stated objectives were successfully achieved. A module extraction model was created based on the YOLOv5 algorithm, which achieved an accuracy of 98% when tested with 44 samples. In addition, the statistical analysis achieved an accuracy of 85% for hotspot detection. A GUI is built, assimilating the results of previous stages and visualizing them for a better user experience. Manual operation is also available in case of extraction failure. | |
| dc.identifier.uri | https://erepo.usm.my/handle/123456789/21327 | |
| dc.language.iso | en | |
| dc.title | AI-based pv module extraction and hotspot fault detection using image processing technique | |
| dc.type | Resource Types::text::report::technical report | |
| dspace.entity.type | Publication | |
| oairecerif.author.affiliation | Universiti Sains Malaysia |