Deep learning model to land cover classification

Introduction

Classifying pixels is an image processing technique that segments an image by assigning each pixel to a class based on its spectral and spatial characteristics Pixels can be classified individually or in groups of neighboring pixels that form segments Classifying pixels can be used for extracting features from imagery, such as land cover, roads, or buildings.

Supported imagery

The recommended imagery configuration is as follows:

• Resolution—High resolution (80–120 centimeters)
• Dynamic Range—8-Bit Unsigned
• Bands—Three bands (for example, red, green, and blue)
• Imagery—Orthorectified imagery (both on-the-fly and persisted ortho products)

After you have installed all the deep learning libraries to run the deep learning tools in Arc GIS Pro.

• Download the model from ArcGIS Living Atlas of the World.
• Browse to Tools on the Analysis tab.

• Click the Toolboxes tab in the Geoprocessing pane, select Image Analyst Tools, and browse to the Classify Pixels Using Deep Learning tool under Deep Learning.

Set the variables on the Parameters tab as follows:

1. Input Raster—Select the imagery.
2. Output Classified Raster—Set the output feature class that will contain the classification results.
3. Model Definition—Select the pretrained or fine-tuned model .dlpk file.
4. Arguments (optional)—Change the values of the arguments if required

Set the variables on the Environments tab as follows:

1. Processing Extent—Select Current Display Extent or any other option from the drop-down menu.
2. Cell Size (required)—Set the value to 1 (in meters).
3. Processor Type—Select CPU or GPU.

Click Run.

The output layer is added to the map.

Conclusion

Pixel classification with deep learning underpins numerous real-world applications, including:

• Remote Sensing: Classifying land cover (forests, water bodies, urban areas) from satellite imagery for environmental monitoring.
• Medical Imaging: Identifying tumors, organs, and other structures in X-rays, CT scans, and MRIs to aid in diagnosis and treatment planning.
• Autonomous Vehicles: Segmenting objects like pedestrians, vehicles, and lanes in real-time for safe navigation.
• Content-Based Image Retrieval: Searching image databases based on content (e.g., finding images with a specific object or scene).