Project Specification
- Project Type: Academic
- Project Mode: Independent
- Project Status: Complete
- Project Duration: 3 Months
- Project Involvement
- Extended all scripts
- Project management
- Produced all documentation
- Developed independent script
- Engineered model modifications
- Software Applications
- QGIS
- Anaconda
- Notepad++
- Adobe Photoshop
- Google Colaboratory
- Microsoft Visual Studio
- Supported Platforms
- Windows
- Download: Google Drive
- Documentation: .pdf .docx
Synopsis
A modified state-of-the-art super-resolution neural network model purposed to furthering the perceptual quality of super-resolved
satellite imaging. The base model, namely DSen2,
is designed to super-resolve the low-resolution spectral band images of Sentinel-2 satellite imaging.
The super-resolution neural network model employs two state-of-the-art data augmentation techniques, in the testing phase, to reduce
the presence of artifacts and noise articulation in super-resolved, low-resolution Sentinel-2 satellite imaging.