[17a-K506-1~11] 3.6 Laser processing

11:00 AM - 11:15 AM JST(2:00 AM - 2:15 AM UTC)

[17a-K506-8] Improvement of Spatial Resolution in Supersonic Flow Using Wavelet-based BOS Method with Deep Learning Super-Resolution

〇(B)Kazuki Maeda¹, Atsushi Suda¹, Kazuki Hiro¹, Shigeto Nakamura¹, Hiroshi Fukuoka¹ (1.NIT. Nara College)

Keywords:

Background Oriented Schlieren,Super-Resolution,Supersonic Flow

The Wavelet-based BOS (W-BOS) method calculates quantitative density gradients using a continuous wavelet transform on a sinusoidal background image. In high-speed imaging, there is a trade-off between frame rate and resolution. Therefore, this study aims to improve the spatial resolution in supersonic flows by applying a super-resolution technique based on deep learning to the W-BOS method. In this report, we clarify the effect of the type of dataset trained on the super-resolution model on the accuracy of restoring high-resolution images.