https://github.com/aman-095/principal-orthogonal-decomposition-on-schlieren-images

Using SVD and PCA to observe the top and most influential modes present in a shock wave generated from a super-sonic wind tunnel.
https://github.com/aman-095/principal-orthogonal-decomposition-on-schlieren-images

canny-edge-detection pca-analysis super-resolution svd

Last synced: 26 days ago
JSON representation

Using SVD and PCA to observe the top and most influential modes present in a shock wave generated from a super-sonic wind tunnel.

• Host: GitHub
• URL: https://github.com/aman-095/principal-orthogonal-decomposition-on-schlieren-images
• Owner: aman-095
• License: mit
• Created: 2022-05-31T10:39:48.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2022-10-26T18:21:57.000Z (about 2 years ago)
• Last Synced: 2023-08-09T15:28:51.025Z (about 1 year ago)
• Topics: canny-edge-detection, pca-analysis, super-resolution, svd
• Language: Jupyter Notebook
• Homepage:
• Size: 1.03 MB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        

# Principal Orthogonal Decomposition on Schlieren Images

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.

## Tools Required

- **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

## Roadmap

1. File Description

- [SVD(POD).ipynb](https://github.com/aman-095/Principal-Orthogonal-Decomposition-on-Schlieren-Images/blob/main/SVD(POD).ipynb)

    This contains the SVD model generated using the Schlieren Images.

- [Fluid_wave_angle.ipynb](https://github.com/aman-095/Principal-Orthogonal-Decomposition-on-Schlieren-Images/blob/main/Fluid_wave_angle%20(1).ipynb) This contains the image tranformation to binary pixel image and application of the Canny Edge Detection and Hough Transform.

- [POD.pdf](https://github.com/aman-095/Principal-Orthogonal-Decomposition-on-Schlieren-Images/blob/main/POD.pdf) This contains the significance of how SVD system works and the matrices associated to it with the physical significance as well.  

    

2. 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](https://github.com/aman-095/Principal-Orthogonal-Decomposition-on-Schlieren-Images/blob/main/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](https://github.com/aman-095/Principal-Orthogonal-Decomposition-on-Schlieren-Images/blob/main/Fluid_wave_angle%20(1).ipynb) which would lead to edge generation and further the properties associated with the fully developed flow.

## LR Top Node

![Sentiment](https://i.postimg.cc/Y2gv1W6h/pod-cov-1.jpg)

## HR Top Node

![Sentiment](https://i.postimg.cc/P5FfGXWg/pod-cov-1-out.jpg)

## Generated Edge

![Sentiment](https://i.postimg.cc/Y9XBF8dJ/POD.png)

## Report

[Report](https://github.com/aman-095/Principal-Orthogonal-Decomposition-on-Schlieren-Images/blob/main/POD.pdf)

## Feedback

If you have any feedback, please reach out to us at [email protected]

## 🔗 Links

[![portfolio](https://img.shields.io/badge/my_portfolio-000?style=for-the-badge&logo=ko-fi&logoColor=white)](https://aman-095.github.io/)

[![linkedin](https://img.shields.io/badge/linkedin-0A66C2?style=for-the-badge&logo=linkedin&logoColor=white)](https://www.linkedin.com/in/aman-bhansali-b4aa26228/)