Loughborough University
Leicestershire, UK
LE11 3TU
+44 (0)1509 263171
Loughborough University

Loughborough University Institutional Repository

Please use this identifier to cite or link to this item: https://dspace.lboro.ac.uk/2134/35603

Title: 1st International round robin on EL imaging: automated camera calibration and image normalisation
Authors: Bedrich, Karl G.
Chai, Jing
Wang, Yan
Aderle, Armin G.
Bliss, Martin
Bokalic, Matevz
Doll, Bernd
Kontges, Marc
Huss, Alexandra
Lopez-Garcia, Juan
Sheng Khoo, Yong
Keywords: Electroluminescence
PV module
Issue Date: 2018
Publisher: WIP
Citation: BEDRICH, K.G. ... et al., 2018. 1st International round robin on EL imaging: automated camera calibration and image normalisation. Proceedings of the 35th European Photovoltaic Solar Energy Conference and Exhibition (EU PVSEC 35), Brussels, Belgium, 24 - 28 Sept, pp.1049-1056.
Abstract: Results from the first international Round Robin on electroluminescence (EL) imaging of PV devices are presented. 17 Laboratories across Europe, Asia and the US measured EL images of ten commercially available modules and five single-cell modules. This work presents a novel automated camera calibration and image scaling routine. Its performance is quantified through comparing intensity deviation of corrected images and their cell average. While manual calibration includes additional measurement of lens distortion and flat field, the automated calibration extracts camera calibration parameters (here: lens distortion, and vignetting) exclusively from EL images. Although it is shown that the presented automated calibration outperforms the manual one, the method proposed in this work uses both manual and automated calibration. 501 images from 24 cameras are corrected. Intensity deviation of cell averages of every measured device decreased from 10.3 % (results submitted by contributing labs) to 2.8 % (proposed method), For three images the image correction produced insufficient results and vignetting correction failed for one camera, known of having a non-linear camera sensor. Surprisingly, largest image quality improvements are achieved by spatially precise image alignment of the same device and not by correcting for vignetting and lens distortion. This is due to overall small lens distortion and the circumstance that, although vignetting caused intensity reduction of more than 50%, PV devices are generally positioned in the image centre in which vignetting distortion is lowest.
Description: This is a conference paper.
Version: Accepted for publication
DOI: 10.4229/35thEUPVSEC20182018-5BO.11.1
URI: https://dspace.lboro.ac.uk/2134/35603
Publisher Link: https://doi.org/10.4229/35thEUPVSEC20182018-5BO.11.1
ISBN: 3936338507
ISSN: 2196-100X
Appears in Collections:Conference Papers and Presentations (Mechanical, Electrical and Manufacturing Engineering)

Files associated with this item:

File Description SizeFormat
Karl EU-PVSEC 35 - paper15.pdfAccepted version21.06 MBAdobe PDFView/Open


SFX Query

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.