PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.
While deep learning methods have proven their superiority over conventional image processing techniques in many domains, their use in airborne heliostat monitoring remains limited. Our aim is to bridge this gap by developing models to improve and extend existing image-based measurement methods in this field. We use Blender and BlenderProc to generate synthetic image data, which grants us access to vast amounts of training data essential for developing effective deep learning models. The exemplary model we train can potentially solve the following tasks related to airborne heliostat field monitoring: detection of heliostats and detection of mirror facet corners. Our promising preliminary results demonstrate the applicability of our approach to use synthetic training data for the development of the intended deep learning models.
Conference Presentation
(2023) Published by SPIE. Downloading of the abstract is permitted for personal use only.
Rafal Broda,Alexander Schnerring,Julian J. Krauth,Marc Röger, andRobert Pitz-Paal
"Towards deep learning based airborne monitoring methods for heliostats in solar tower power plants", Proc. SPIE 12671, Advances in Solar Energy: Heliostat Systems Design, Implementation, and Operation, 1267108 (4 October 2023); https://doi.org/10.1117/12.2676821
ACCESS THE FULL ARTICLE
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.
The alert did not successfully save. Please try again later.
Rafal Broda, Alexander Schnerring, Julian J. Krauth, Marc Röger, Robert Pitz-Paal, "Towards deep learning based airborne monitoring methods for heliostats in solar tower power plants," Proc. SPIE 12671, Advances in Solar Energy: Heliostat Systems Design, Implementation, and Operation, 1267108 (4 October 2023); https://doi.org/10.1117/12.2676821