Satellite imagery is images of our Planet collected by imaging satellites all around the world. They have many applications in various domains like agriculture, climate monitoring, food security, the media industry.
While these images are valuable sources of information, they remain challenging for the non-expert to access.
Today these images are not indexed by search engines, so you must already know that they exist and know where to find them. Besides, to select images in the services that host them, you need some knowledge about the satellites' sensors or these satellites' products.
The EarthSearch portal is an innovative service aimed at making satellite imagery accessible to a broader audience. We aim to make these images available via the Qwant search engine when a remote sensing query is detected. When we notice a query related to remote sensing on our search engine, we will display one or several Instant Answers. The Instant Answers will display information about Sentinel-2 images and a link to the original images.
We use recent advances in Natural Language Processing (NLP), and Computer Vision (CV) applied to remote sensing images to achieve this goal. Our artificial intelligence module allows us to understand the image's content without using metadata and matching it with the query to deliver quality search results.
In this example, we show an example of a SnapEarth related Instant Answer after a user sends a remote sensing query. We first have an article from the SnapEarth website and below a Sentinel-2 image related to the query on the right column.