Jacob Prince-Bieker

Train Model on Optimal Cloud Analysis product

4

DARDAR has some very detailed maps from CALIPSO and CloudSat satellites. Unfortunately, the data ends around 2016-2017. There could be some use in still training a model on historical data,...

Give network optical flow future frames as input

1

Give the network future frames from optical flow as well as the GT past frames, have it try to predict the future frames, or corrections to the optical flow frames...

Try as classification problem, instead of regression

1

Classify each pixel as the output for if there is a cloud, or type of cloud with the #3 instead of regression. Attempting to predict all 12 spectral channels as...

Pretraining without labels, using contrastive learning for example, before then training on the actual predictions could help.

Use Pixel Shuffle and other ones instead of the standard transpose convolutions, self-attention in the lowest downsample layer, add upsample with blur

Try superresolution images

1

While having a network predict super-resolution images from just the geo satellite images is probably not helpful, as it would be making up data that does not exist there, one...

## Detailed Description https://eumetsatspace.atlassian.net/wiki/spaces/EUM/blog/2022/10/24/1908965377/Data+Store+APIs+1.0.0+released ## Context Satip needs to be up to date to keep working with the API. And there might be some changes to how to delete customizations...

Add pipeline for pulling Indian Ocean Satellite Imagery

1

## Detailed Description Use stuff built for #133 to also start collecting Indian Ocean 15 minutely data. Would be really great if we could try pushing it onto HF or...

Add automated pipeline for pulling, processing, and saving to HF/GCP/Wherever

1

## Detailed Description Instead of relying on someone to manually run the data pipeline to get more satellite data, we should try automating it, so the dataset just grows on...