This is an ongoing project dedicated to creating a comprehensive dataset of thermal images acquired from the DJI Mavic 2 Enterprise Dual drone with a thermal resolution of 160 x 120 px.
Please note that the PV module faults that would be presented below are not annotated at this phase, while also no supplementary in-situ/on field measurements have been gathered to verify the human identified PV module faults.
This work, aims to support the Open Science Practices, from this initial stage, focusing on:
- Open data sharing;
- Community evaluation and expert-oriented feedback;
The first objective is to establish current and forseen collaborative technology actions assisting knowledge sharing. The second objective is mainly defined for stakeholder engagement in PV module faults verification, while also for the enhancement of the dataset's general perspective and description.
Different photovoltaic faults are identified:
Type 1: Hot spot(s) due to glass breakage - shattered glass.
Type 2: PV module with one inactive (open-circuited) submodule, due to a bypass diode failure.
Type 3: Open-circuited module
Type 4: Hot spot due to vegetation - shading.
Indicative visual examples of the dataset are given:
| 
| 
| 
| 
|
Figure 1. Different captures of the same fault: Type 1.
| 
| 
| 
|
Figure 2. Different captures of the same fault: Type 2.
|  |
Figure 3. Example of fault: Type 3.
| 
|  |
| 
|
Figure 4. Different day captures of fault presented in Figure 3 (Type 3), with also different camera configuration settings.
|  |
|  |
| 
|
Figure 5. Different day captures of fault presented in Figure 3 (Type 4), with also different camera configuration settings.