GeoPl@ntNet was designed with multiple issues in mind – First the lack of comprehensive habitat maps, secondly the lack of a tool to produce lists of species on a given spatial zone, thirdly the lack of geographical context to improve the accuracy of the Pl@ntNet species identification, fourthly the lack of a tool to access mapped biodiversity indicators and lastly a tool that predicts which species and habitats are in a given zone. We want to improve GeoPl@ntNet by allowing a temporal exploration of the predicted indicators, as well as extending it globally.

The platform covers all continental European countries, and ongoing development are planned to (i) replicate the workflow based on a new deep learning model, (ii) open the service to a growing number of use cases in order to gather user feedbacks and improve the interface.

We aim to ensure its robustness and accuracy while continuing to expand it a larger spatial scale.

The platform allows a large number of end-users to download contextualised biodiversity reports. We plan to scale the product in order to extend it to a larger geographical scale, and a higher resolution, which can contribute to respond to new needs.

The technology developed can be adapted to new geographical areas, at different geographical resolutions.