Frequently Asked Questions

Do I need to code or know how to build AI models to use GeoAI?

No. GeoAI is a no-code platform: you annotate examples on your imagery, and GeoAI trains and applies the model for you. If you do want to automate things, there is a Public API.

Does GeoAI have ready-made AI models?

Yes. GeoAI provides a catalogue of AI models for common tasks, such as detecting structures and objects or classifying land cover, and you can also use models shared by the community.

What is the difference between object detection and segmentation?

Object detection finds objects and draws a box around each one, for example vehicles or buildings. Segmentation goes further and outlines the exact area each object or class covers, down to the pixel. For example, it can separate individual trees, or classify every pixel as road, building, vegetation or water. GeoAI supports both; you choose the model type when you train. See Training Guidelines.

How does GeoAI ensure accuracy?

Your annotations are the ground truth: models learn from them, and their results are checked against them. Annotating training regions completely, and adding negative regions, improves results the most (see the training guidelines).

How long does training a model take?

From minutes to hours, depending on the number of annotations, the size of the training regions and the model type.

How long does applying a model take?

Usually seconds to minutes, depending on the size of the imagery.

What are credits? Will I be charged?

Credits are GeoAI’s usage unit, included in your DestinE Platform subscription. They are not money and you will not be charged. See GeoAI Service Quota.

What imagery can I use?

You can upload GeoTIFF files, connect external imagery such as WMTS services or cloud-optimised GeoTIFFs, or acquire imagery through the Data Market. See Import External Imagery.

How do I get help?

Use the in-app chatbot help, the DestinE Platform support, or write to contact.platform@destine.eu.