2026-04-21T03:00:36Z
https://rodare.hzdr.de/oai2d
oai:rodare.hzdr.de:2704
2024-05-15T08:30:41Z
openaire_data
user-hzdr
user-rodare
true
3.1
HZDR.RODARE
10.14278/rodare.2704
Arbash, Elias
Helmholtz Institute Freiberg for Resource Technology
Fuchs, Margret
0000-0001-7210-1132
Helmholtz Institute Freiberg for Resource Technology
Rasti, Behnood
Lorenz, Sandra
0000-0001-8464-2331
Helmholtz Institute Freiberg for Resource Technology
Ghamisi, Pedram
0000-0003-1203-741X
Helmholtz Institute Freiberg for Resource Technology
Gloaguen, Richard
0000-0002-4383-473X
Helmholtz Institute Freiberg for Resource Technology
PCB-Vision: A Multiscene RGB-Hyperspectral Benchmark Dataset of Printed Circuit Boards
Rodare
2024
circular economy
automated data processing
optical sensors
recycling
e-waste
printed circuit board
hyperspectral
dataset
RGB
conveyor belt
sensors
machine learning
deep learning
PCBVision
open-source data
digitalization
2024-01-29
https://rodare.hzdr.de/record/2704
10.48550/arXiv.2401.06528
https://www.hzdr.de/publications/Publ-38529
https://www.hzdr.de/publications/Publ-38684
https://www.hzdr.de/publications/Publ-38529
10.14278/rodare.2703
https://rodare.hzdr.de/communities/hzdr
https://rodare.hzdr.de/communities/rodare
1
Creative Commons Attribution 4.0 International
Open Access
<p><strong>PCB-Vision Dataset</strong></p>
<p>Description:</p>
<p>The PCB-Vision dataset is a multiscene RGB-Hyperspectral benchmark dataset comprising 53 Printed Circuit Boards (PCBs). The RGB images are collected using a Teledyne Dalsa C4020 camera on a conveyor belt, while hyperspectral images (HSI) are acquired with a Specim FX10 spectrometer. The HSI data contains 224 bands in the VNIR range [400 - 1000]nm.</p>
<p><strong>Data Format</strong></p>
<ul>
<li>RGB Images: .png files</li>
<li>PCB Masks: .jpg files</li>
<li>HSI Data: Each hyperspectral data cube is accompanied by a data file and a .hdr file.</li>
</ul>
<p><strong>Folder Organization</strong></p>
<ul>
<li>PCBVision
<ul>
<li>HSI/
<ul>
<li>53 subfolders (one for each PCB)</li>
<li>'General_masks' folder for 'General' segmentation ground truth</li>
<li>'Monoseg_masks' folder for 'Monoseg' segmentation ground truth</li>
<li>'PCB_Masks' folder for masks of the 53 PCBs in the hyperspectral cube</li>
</ul>
</li>
<li>RGB/
<ul>
<li>53 .jpg images</li>
<li>'General' folder for RGB images 'General' segmentation ground truth</li>
<li>'Monoseg_masks' folder for RGB images 'Monoseg' segmentation ground truth</li>
</ul>
</li>
</ul>
</li>
</ul>
<p><strong>Data Classes in Masks</strong></p>
<ul>
<li>Masks (both 'General' and 'Monoseg') contain 1 to 4 segmentation classes:
<ul>
<li>0: "Others"</li>
<li>1: "IC"</li>
<li>2: "Capacitors"</li>
<li>3: "Connectors"</li>
</ul>
</li>
</ul>
<p><strong>Code Repository</strong></p>
<p>To facilitate reading and working with the data, Python codes are available on the GitHub repository:</p>
<p>https://github.com/hifexplo/PCBVision</p>
<p><strong>Citation</strong></p>
<p>If you use this dataset, please cite the following article:</p>
<p><strong>Word</strong>:</p>
<p>Arbash, Elias, Fuchs, Margret, Rasti, Behnood, Lorenz, Sandra, Ghamisi, Pedram, & Gloaguen, Richard. (2024). PCB-Vision: A Multiscene RGB-Hyperspectral Benchmark Dataset of Printed Circuit Boards (Version 1) [Data set]. Rodare. <a href="http://doi.org/10.14278/rodare.2704">http://doi.org/10.14278/rodare.2704</a></p>
<p><strong>Latex:</strong></p>
<p>@article{arbash2024pcb, title={PCB-Vision: A Multiscene RGB-Hyperspectral Benchmark Dataset of Printed Circuit Boards}, author={Arbash, Elias and Fuchs, Margret and Rasti, Behnood and Lorenz, Sandra and Ghamisi, Pedram and Gloaguen, Richard}, journal={arXiv preprint arXiv:2401.06528}, year={2024} }</p>
<p><strong>Contact</strong></p>
<p>For further information or inquiries, please visit our website:</p>
<p>https://www.iexplo.space/</p>
<p>Contact Email: e.arbash@hzdr.de</p>