---
license: mit
---
# Model Card for UNet-5depth-shuffle: venkatesh-thiru/s2l8h-UNet-5depth-shuffle
## Model Description
The UNet-5depth-shuffle model aims to harmonize Landsat-8 and Sentinel-2 imagery by enhancing the spatial resolution of Landsat-8 images. The model processes Landsat-8 multispectral and pan-chromatic images to output images with Sentinel-2-like spectral and spatial characteristics.

## Model Architecture
This UNet architecture includes 5 depth levels and uses a shuffling mechanism to improve image resolution and spectral alignment. The combination of depth levels and shuffling is optimized to produce high-quality images that closely match Sentinel-2 data.

## Usage

```
from transformers import AutoModel

# Load the UNet-5depth-shuffle model
model = AutoModel.from_pretrained("venkatesh-thiru/s2l8h-UNet-5depth-shuffle", trust_remote_code=True)

# Harmonize Landsat-8 images
l8up = model(l8MS, l8pan)
```
Where:

`l8MS` - Landsat Multispectral images (L2 Reflectances)
`l8pan` - Landsat Pan-Chromatic images (L1 Reflectances)

## Potential Applications
Water quality assessment
Urban planning
Climate monitoring
Disaster response
Infrastructure oversight
Agricultural surveillance


## Limitations
The model shows minor limitations in areas with unique spectral characteristics or under extreme environmental conditions.

## Reference
For more details, refer to the publication: 10.1016/j.isprsjprs.2024.04.026