journal article·2024-10

Deep learning-based perspective distortion correction for outdoor photovoltaic module images

Yun Li, Brendan Wright, Ziv Hameiri

Publisher page

Abstract

To address the growing demand for photovoltaic inspection, an increasing volume of photoluminescence, electroluminescence, and infrared image datasets is being produced for research and analysis. Most of these images exhibit perspective distortions, which introduce challenges in subsequent analysis, such as defect detection and classification. Hence, distortion correction is required during the preprocessing stage. However, manually correcting each dataset to eliminate distortions is labour-intensive. This paper addresses this issue by developing a convolutional neural network model to estimate the camera parameters of yaw, pitch, and roll. A heuristic method was referenced as a comparative benchmark, demonstrating that the developed deep learning-based approach is equally accurate and significantly faster. The findings underscore the potential of deep learning techniques in automating and refining image analysis in photovoltaic diagnostics, offering substantial improvements over traditional methods in terms of speed and scalability.
Source: expertdata · Quality: candidate unverified · Source record ↗

Research context

OpenAlex topics: Image Processing Techniques and Applications · Industrial Vision Systems and Defect Detection · Infrared Target Detection Methodologies

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

References (40)

Show source-provided references
1. A collaborative approach to operate high powered devices on small-scale PV systems10.1016/j.egypro.2017.03.252 ↗
2. Development of photovoltaic technologies for global impact10.1016/j.renene.2019.02.030 ↗
3. Review on fault characterization and diagnosis technique in photovoltaic systems
4. A techno-economic review of silicon photovoltaic module recycling10.1016/j.rser.2019.04.020 ↗
5. Luminescence nanothermometry10.1039/c2nr30764b ↗
6. Deep Learning model to denoise luminescence images of silicon solar cells10.1002/advs.202300206 ↗
7. Electroluminescence as a tool to study the polarization characteristics and generation mechanism in silicon PV panels10.3390/app13031591 ↗
8. Outdoor luminescence imaging of field-deployed PV modules10.1088/2516-1083/ac9a33 ↗
9. Development of outdoor luminescence imaging for drone-based PV array inspection
10. Machine learning in solar plants inspection automation10.3390/en15165966 ↗
11. Image preprocessing for outdoor luminescence inspection of large photovoltaic parks10.3390/en14092508 ↗
12. Correcting lens distortions in digital photographs
13. Automatic correction of perspective and optical distortions10.1016/j.cviu.2017.05.016 ↗
14. Hanssens · 2023
15. Virtually secure: a taxonomic assessment of cybersecurity challenges in virtual reality environments10.1016/j.cose.2022.102951 ↗
16. Fault classification for photovoltaic modules using thermography and image processing
17. Decision tree-based fault detection and classification in solar photovoltaic arrays
18. Method for estimation and correction of perspective distortion of electroluminescence images of photovoltaic panels10.1109/JPHOTOV.2020.3019949 ↗
19. A flexible new technique for camera calibration10.1109/34.888718 ↗
20. Correcting for perspective distortion in electroluminescence images of photovoltaic panels
21. Image preprocessing for outdoor luminescence inspection of large photovoltaic parks10.3390/en14092508 ↗
22. Use of the Hough transformation to detect lines and curves in pictures10.1145/361237.361242 ↗
23. Lens distortion correction based on one chessboard pattern image10.1007/s12200-015-0453-7 ↗
24. A flexible new technique for camera calibration10.1109/34.888718 ↗
25. A four-step camera calibration procedure with implicit image correction
26. Learning perspective undistortion of portraits
27. Blind first-order perspective distortion correction using parallel convolutional neural networks10.3390/s20174898 ↗
28. A practical distortion correcting method from fisheye image to perspective projection image
29. A self-adaptive correction method for perspective distortions of image10.1007/s11704-018-7269-8 ↗
30. Effect of roll, pitch, and yaw motions on ISAR imaging
31. Targ · 2016
32. Rotation vector and directional cosine matrix in problems of satellite attitude control
33. Deep learning for anomaly detection: a review10.1145/3439950 ↗
34. Improved adam optimizer for deep neural networks
35. Analysis of the mean absolute error (MAE) and the root mean square error (RMSE) in assessing rounding model
36. Photovoltaic module array luminescence image preprocessing: Heuristic algorithms for perspective correction and cell segmentation
37. Deans · 2007
38. Morphological image processing
39. Bergstrom · 2023
40. Characteristic analysis of Otsu threshold and its applications10.1016/j.patrec.2011.01.021 ↗