Solar Panel Defect Detection Using Geospatially-Aware Deep Learning framework

Notice

This is an unedited manuscript accepted for publication and provided as an Article in Press for early access at the author’s request. The article will undergo copyediting, typesetting, and galley proof review before final publication. Please be aware that errors may be identified during production that could affect the content. All legal disclaimers of the journal apply.

Year : 2026 | Volume : 17 | 02 | Page :
By

Harsha Patil,

Kanchan Borade,

Prajakta Nagmoti,

  1. Associate Professor, Computer Application Department, Ashoka Center for Business and Computer Studies, Nashik, Maharashtra, India
  2. Head of AI, AI Department, Dhaninfo Pvt. Ltd., Maharashtra, India
  3. AI/ML Computational science Specialist, Accenture Solutions Private Limited, Maharashtra, India

Abstract

Large-scale photovoltaic (PV) systems demand reliable inspection techniques to maintain efficiency, as manual methods remain labor-intensive and inconsistent. This study introduces a geospatially informed deep learning framework for defect detection and localization in PV panels from drone and satellite imagery. The framework incorporates an adaptive tiling mechanism that adjusts tile boundaries according to object size, reducing information loss and enhancing detection performance. In addition, coordinate transformation between image pixels and real-world latitude–longitude values support accurate mapping of detected defects. The suggested system also incorporates automatic picture preprocessing, such as noise reduction and contrast improvement, to enhance the visibility of minute flaws under a variety of lighting and environmental circumstances. The method is appropriate for large- scale deployments as it uses lightweight deep learning architecture to accomplish effective inference while preserving high detection accuracy. Additionally, the creation of accurate defect location maps that support maintenance planning, resource allocation, and quick field verification is made possible by the incorporation of geospatial data. When compared to traditional fixed-grid tiling techniques, a thorough experimental evaluation on representative aerial datasets shows enhanced resilience against scale variation, background complexity, and picture distortions. Additionally, the framework’s adaptive picture partitioning reduces computing cost, enabling quicker processing of large amounts of solar farm footage without sacrificing localization accuracy. Experimental validation highlights improved detection precision and localization reliability over uniform tiling approaches, confirming the effectiveness of the proposed methodology for scalable solar farm inspection.

Keywords: Adaptive tiling, coordinate mapping, deep learning, defect detection, GeoTIFF, geospatial analysis, photovoltaic systems, remote sensing, solar panel inspection, UAV imagery

How to cite this article:
Harsha Patil, Kanchan Borade, Prajakta Nagmoti. Solar Panel Defect Detection Using Geospatially-Aware Deep Learning framework. Journal of Remote Sensing & GIS. 2026; 17(02):-.
How to cite this URL:
Harsha Patil, Kanchan Borade, Prajakta Nagmoti. Solar Panel Defect Detection Using Geospatially-Aware Deep Learning framework. Journal of Remote Sensing & GIS. 2026; 17(02):-. Available from: https://journals.stmjournals.com/jorsg/article=2026/view=250443

References

  1. Di Tommaso A, Betti A, Fontanelli G, Michelozzi B. A multi-stage model based on YOLOv3 for defect detection in PV panels based on IR and visible imaging by unmanned aerial vehicle. Renewable energy. 2022 Jun 1;193:941-62.
  2. Bommes L, Pickel T, Buerhop‐Lutz C, Hauch J, Brabec C, Peters IM. Computer vision tool for detection, mapping, and fault classification of photovoltaics modules in aerial IR videos. Progress in Photovoltaics: Research and Applications. 2021 Dec;29(12):1236-51.
  3. Shaik A, Balasundaram A, Kakarla LS, Murugan N. Deep learning-based detection and segmentation of damage in solar panels. Automation. 2024 May 29;5(2):128-50.
  4. Tang C, Ren H, Xia J, Wang F, Lu J. Automatic defect identification of PV panels with IR images through unmanned aircraft. IET Renewable Power Generation. 2023 Sep;17(12):3108-19.
  5. Wen D, Huang X, Bovolo F, Li J, Ke X, Zhang A, Benediktsson JA. Change detection from very-high-spatial-resolution optical remote sensing images: Methods, applications, and future directions. IEEE Geoscience and Remote Sensing Magazine. 2021 Apr 5;9(4):68-101.
  6. Vokhmintcev AV, Abbazov VR, Khater MM, Romanov MA, Vokhmintceva TS. Remote Sensing Detection an Archaeological Sites of Bronze Age Using Digital Terrain Models and YOLO Models Based on a Transformer Architecture. InInternational Conference on Intelligent Information Technologies for Industry 2025 Nov 5 (pp. 280-291). Cham: Springer Nature Switzerland.
  7. Bartulović A, Šerić L, Bejo J, Damjanović I. Geolocalizing Pixels on PTZ Camera Imagery in Natural Environments using Deep Neural Network. In2025 10th International Conference on Smart and Sustainable Technologies (SpliTech) 2025 Jun 16 (pp. 1-6). IEEE.
  8. Wang S, Chen W, Xie SM, Azzari G, Lobell DB. Weakly supervised deep learning for segmentation of remote sensing imagery. Remote Sensing. 2020 Jan 7;12(2):207.
  9. Abro GE, Ali A, Memon SA, Memon TD, Khan F. Strategies and challenges for unmanned aerial vehicle-based continuous inspection and predictive maintenance of solar modules. IEEE Access. 2024 Nov 25;12:176615-29.
  10. Shen Y, Liu D, Chen J, Wang Z, Wang Z, Zhang Q. On-board multi-class geospatial object detection based on convolutional neural network for High Resolution Remote Sensing Images. Remote Sensing. 2023 Aug 10;15(16):3963.
Ahead of Print Subscription Review Article
Volume 17
02
Received 10/06/2026
Accepted 10/07/2026
Published 22/07/2026
Publication Time 42 Days

Login

My IP

PlumX Metrics