Used POD to analyze the most influencing modes in the fully developed fluid flow using a schlieren imaging dataset generated from a supersonic wind tunnel. Further implemented ESRGAN model and performed Canny Edge Deatection and Hough Transform to gain the physical entities associated with the fluid i.e Mach No. and Fluid velocity.
-
Pandas - Python data manipulation libraries
-
Open-CV - Working with images
-
Scipy - Performing SVD
-
Matplotlib - Visualizing the images
-
ESRGAN - Enhance the image resolution
-
Canny Edge Detection - Generates boundaries from image
-
Hough Tranform - Detects line and provide angle between two lines
- File Description
- SVD(POD).ipynb This contains the SVD model generated using the Schlieren Images.
- Fluid_wave_angle.ipynb This contains the image tranformation to binary pixel image and application of the Canny Edge Detection and Hough Transform.
- POD.pdf This contains the significance of how SVD system works and the matrices associated to it with the physical significance as well.
- Pipeline
- Installing libraries and dependency
- Schlieren dataset generated is of High Resolution and SVD model generates a covariance matrix which makes computation difficult as the space required is huge in Tbs.
- Reduced the dimensions of the images which also reduced the quality of the images.
- Perform SVD with the code mentioned in SVD(POD).ipynb which would generate the top modes of the fluid flow.
- Apply ESRGAN model to re-enhance the resolution.
- Run the code file Fluid_wave_angle.ipynb which would lead to edge generation and further the properties associated with the fully developed flow.
If you have any feedback, please reach out to us at [email protected]
