English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In this work we pretrain a CLIP/ViT based model using three differentmodalities of satellite imagery across five AOIs covering over ~10\% of Earth'stotal landmass, namely Sentinel 2 RGB optical imagery, Sentinel 1 SAR radaramplitude and interferometric coherence. This model uses $\sim 250$ Mparameters. Then, we use the embeddings produced for each modality with aclassical machine learning method to attempt different downstream tasks forearth observation related to vegetation, built up surface, croplands andpermanent water. We consistently show how we reduce the need for labeled databy 99\%, so that with ~200-500 randomly selected labeled examples (around4K-10K km$^2$) we reach performance levels analogous to those achieved with thefull labeled datasets (about 150K image chips or 3M km$^2$ in each area ofinterest - AOI) on all modalities, AOIs and downstream tasks. This leads us tothink that the model has captured significant earth features useful in a widevariety of scenarios. To enhance our model's usability in practice, itsarchitecture allows inference in contexts with missing modalities and evenmissing channels within each modality. Additionally, we visually show that thisembedding space, obtained with no labels, is sensible to the different earthfeatures represented by the labelled datasets we selected.

Fewshot learning on global multimodal embeddings for earth observation tasks