English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

This paper presents deep unfolding neural networks to handle inverse problemsin photothermal radiometry enabling super resolution (SR) imaging. Photothermalimaging is a well-known technique in active thermography for nondestructiveinspection of defects in materials such as metals or composites. A grandchallenge of active thermography is to overcome the spatial resolutionlimitation imposed by heat diffusion in order to accurately resolve eachdefect. The photothermal SR approach enables to extract high-frequency spatialcomponents based on the deconvolution with the thermal point spread function.However, stable deconvolution can only be achieved by using the sparsestructure of defect patterns, which often requires tedious, hand-crafted tuningof hyperparameters and results in computationally intensive algorithms. On thisaccount, Photothermal-SR-Net is proposed in this paper, which performsdeconvolution by deep unfolding considering the underlying physics. Thisenables to super resolve 2D thermal images for nondestructive testing with asubstantially improved convergence rate. Since defects appear sparsely inmaterials, Photothermal-SR-Net applies trained block-sparsity thresholding tothe acquired thermal images in each convolutional layer. The performance of theproposed approach is evaluated and discussed using various deep unfolding andthresholding approaches applied to 2D thermal images. Subsequently, studies areconducted on how to increase the reconstruction quality and the computationalperformance of Photothermal-SR-Net is evaluated. Thereby, it was found that thecomputing time for creating high-resolution images could be significantlyreduced without decreasing the reconstruction quality by using pixel binning asa preprocessing step.

Photothermal-SR-Net: A Customized Deep Unfolding Neural Network for Photothermal Super Resolution Imaging