Ready-to-Use Geospatial Deep Learning Models

[CanserinaKurnia]

With the vast availability of imagery that is streaming down daily from a variety of sensors, the need for artificial intelligence (AI) to automate feature extraction is increasing.  Many use cases are now available, showing the benefit of using this cutting edge of technology.  

Universities have started incorporating deep learning/AI into their teaching and research. To support, Esri is providing ready-to-use geospatial AI models in the ArcGIS Living Atlas of the World.   There are currently over 60 pretrained deep learning models have been made available.   The ability to train deep learning models on geospatial datasets and derive information products is available using the ArcGIS Online raster analysis tools, ArcGIS Pro or ArcGIS API for Pytho. Users can also scale up processing using ArcGIS Image Server.

Following are a list of resources to get familiar with these ready-to-use deep learning models and how to get started. 

Ready-to-use Deep Learning Models

• Ready-to-Use Geospatial Deep Learning Models (blog)
• Introducing pre-trained geospatial deep learning models (blog)
• List of ready-to-use deep learning model in ArcGIS Living Atlas of the world
• Deep learning models in arcgis.learn
• Building footprint extraction sample (StoryMap)

DL Libraries Installer

• Esri GitHub repo on Deep Learning

GeoAI

• Access to GeoAI resources and examples
• GeoAI Medium (technical blogs)

Notebooks

• Sample Notebooks

Community

• Imagery and Remote Sensing community page on GeoNet
• Official Imagery and Remote Sensing in Higher Education

Please feel free to contact me at ckurnia@esri.com if you are interested in adopting these deep learning model for your classroom, research, or have any additional questions.