Use the model

You can use the Global location encoder (Sentinel-2) model in the Generate Embeddings Using AI Models tool available in the GeoAI toolbox in ArcGIS Pro.

Complete the following steps to use the Global location encoder (Sentinel-2) pretrained model:

1. Download the model and add the feature class layer for which embeddings will be generated to an ArcGIS Pro project.

   

2. Either zoom to an area of interest or use the entire extent.
3. Click the Analysis tab and browse to Tools.

   

4. In the Geoprocessing pane, click Toolboxes, expand GeoAI Tools, and browse to Generate Embeddings Using AI Models tool in the Embeddings Based Analysis toolset.

   

5. On the Parameters tab, set the variables as follows:
   1. For Input Feature Class Or Raster, use a location feature layer for which embeddings will be generated.
   2. For Output Embeddings Feature Class, set the feature class that will store generated embeddings in a compatible embedding format.
   3. For Input Model Definition File, set the location encoder models that will be used to compute embeddings. Select the pretrained model .dlpk file.

      The model determines the semantic representation of raw data.

   4. Optionally, under Arguments, for Batch Size, set the number of features processed in each step of the model inference.

      This depends on the memory of your graphics card.

   
6. On the Environments tab, set the variables as follows:
   1. For Processing Extent, select Default or any other option from the drop-down menu.
   2. For Processor Type, select CPU or GPU as needed.

      If GPU is available, it is recommended that you select GPU and set GPU ID to the GPU to be used.

   
7. Click Run.

   Once processing is complete, the output layer is added to the map.

   

Explore use cases

1. Find Similar Features Using Embeddings

1. The map below shows the previously generated embeddings created using the Generate Embeddings Using AI Models tool geoprocessing tool. For this analysis, india_desert was added as the feature of interest to identify regions with similar characteristics. This can be further used for finding similar features.

   
2. Click the Toolboxes tab in the Geoprocessing pane, expand GeoAI Tools, and browse to the Find Similar Features Using Embeddings tool under Embeddings Based Analysis.

   

3. Set the variables on the Parameters tab as follows:
   1. For Embedding Features, set the feature class containing embeddings to find features most similar to the query features.
   2. For Query Features, set the feature layer containing one or more features representing the objects of interest.
   3. For Output Embeddings Feature Class, use the feature class that stores the results of similarity search.
   4. For Threshold, specify the minimum similarity scores an embedding feature must have when compared with the query features to be included in the output.

   

   An output feature class containing similar embedding, a query showing similar desert features, will be added to the map.

   

.

2. AutoML using location embeddings

This model is used for generation location embedding in the AutoML workflow, as described in section 1.f of the Train Using AutoML workflow. For more information, see How AutoML works.