0
Powered by :
Indian Institute of Engineering Science and Technology, an India-based engineering institute, has developed a Deep Learning system for detecting faulty solar-panel cells. The system uses thermal images and Semantic Segmentation to support solar plant maintenance while reducing human involvement in inspections. Researchers trained and tested the proposed model using a publicly available thermal image dataset for faulty-cell identification. The lightweight architecture combines DensNet121 as its encoder with part of UNet serving as the decoder. This model identifies faulty solar panels from thermal imagery and was assessed against previously published methods and alternative architectures. The authors reported that the proposed model outperformed the compared approaches across all evaluated aspects. Elsevier published the open-access study online on July 8, 2026, in Volume 283 of Procedia Computer Science.
SOLARbytes brings you the latest news in solar world in bite size; essentially a gist of important solar news in few sentences.
Subscribe to our Newsletter!
Quick Links
© 2026 SOLARbytes | SOLARbytes is a publication of Wat(t) Bytes Media Pvt. Ltd. All rights reserved.
