Filtern
Erscheinungsjahr
- 2020 (3) (entfernen)
Dokumenttyp
Sprache
- Englisch (3) (entfernen)
Referierte Publikation
- nein (3) (entfernen)
Schlagworte
- Active thermography (1)
- Deep learning (1)
- Diffusion (1)
- High-power laser (1)
- Iterative shrinkage thresholding algorithm (1)
- Laser scanning (1)
- Laser thermography (1)
- Neural network (1)
- Nondestructive testing (1)
- Photothermal imaging (1)
Organisationseinheit der BAM
Using spatial and temporal shaping of laser-induced diffuse thermal wave fields in thermography
(2020)
The diffuse nature of thermal waves is a fun-damental limitation in thermographic nonde-structive testing. In our studies we investigated different approaches by shaping the thermal wave fields which result from heating. We have used high-power laser sources to heat metallic samples. Using these spatial and temporal shaping techniques leads to a higher detection sensitivity in our measurements with the infra-red camera. In this contribution we show our implementation of shaping laser-induced diffuse thermal wave fields and the effect on the defect reconstruction quality.
In this work we focus on our most recent studies to super resolution (SR) laser thermography. The goal of SR nondestructive testing methods is to facilitate the separation of closely spaced defects. We explain how to combine laser scanning with SR techniques. It can be shown that stepwise as well as continuous scanning techniques are applicable. Finally, we discuss the effect of experimental parameters and im-age processing techniques to find the optimal SR technique which leads to the highest reconstruction quality within laser thermography.
Learned block iterative shrinkage thresholding algorithm for photothermal super resolution imaging
(2020)
Block-sparse regularization is already well-known in active thermal imaging and is used for multiple measurement based inverse problems. The main bottleneck of this method is the choice of regularization parameters which differs for each experiment. To avoid time-consuming manually selected regularization parameters, we propose a learned block-sparse optimization approach using an iterative algorithm unfolded into a deep neural network. More precisely, we show the benefits of using a learned block iterative shrinkage thresholding algorithm that is able to learn the choice of regularization parameters. In addition, this algorithm enables the determination of a suitable weight matrix to solve the underlying inverse problem. Therefore, in this paper we present the algorithm and compare it with state of the art block iterative shrinkage thresholding using synthetically generated test data and experimental test data from active thermography for defect reconstruction. Our results show that the use of the learned block-sparse optimization approach provides smaller normalized mean square errors for a small fixed number of iterations than without learning. Thus, this new approach allows to improve the convergence speed and only needs a few iterations to generate accurate defect reconstruction in photothermal super resolution imaging.