Assess wildfire damage with Amazon SageMaker geospatial capabilities

This notebook's CI test result for us-west-2 is as follows. CI test results in other regions can be found at the end of the notebook.

This notebook demonstrates how to use Amazon SageMaker geospatial capabilities to assess wildfire damages using multi-temporal Sentinel-2 satellite data.

The area of interest for this example is located in Northern California, from a region which was affected by the Dixie Wildfire in 2021.

The workflow is as follows:

• Step 1: Import SageMaker geospatial capabilities SDK
• Step 2: Inspect the area of interest
• Step 3: Create an Earth Observation Job (EOJ) to perform landcover segmentation
• Step 4: Visualize EOJ results in Amazon SageMaker geospatial Map SDK
• Step 5: Export EOJ output to S3
• Step 6: Quantify loss of vegetation and wildfire impact area

Prerequisites

This notebook runs with Kernel Geospatial 1.0. Note that the following policies need to be attached to the execution role that you used to run this notebook:

• AmazonSageMakerFullAccess
• AmazonSageMakerGeospatialFullAccess

You can see the policies attached to the role in the IAM console under the permissions tab. If required, add the roles using the 'Add Permissions' button.

In addition to these policies, ensure that the execution role's trust policy allows the SageMaker-GeoSpatial service to assume the role. This can be done by adding the following trust policy using the 'Trust relationships' tab:

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Principal": {
                "Service": [
                    "sagemaker.amazonaws.com",
                    "sagemaker-geospatial.amazonaws.com"
                ]
            },
            "Action": "sts:AssumeRole"
        }
    ]
}

Import SageMaker geospatial capabilities SDK

Inspect the area of interest

The raster data collection is queried with coordinates of the area impacted by the wildfire, and returns a list of satellite imagery matching the selected filters.

The data in cloud optimized GeoTIFF (COG) format allows a visual inspection of the impacted area before and after the wildfire.

Create an Earth Observation Job to perform landcover segmentation

The following cell shows how to launch an Earth Observation Job (EOJ). In this example, a pre-trained machine learning model for land cover segmentation is used. Depending on your use case, you can choose from a variety of operations and models when running an EOJ.

In addition to the type of operation, you can also select the area of interest, choose the data providers, and set time-range based and cloud coverage percentage filters.

Visualize EOJ results in Amazon SageMaker geospatial Map SDK

The following cells show how to create a embedded map instance with the geospatial Map SDK and visualize input and output of the Earth Observation Job in the map.

Land Cover Segmentation Visualization Legend

Export EOJ output to S3

Quantify loss of vegetation and wildfire impact area

The following cells show how the exported EOJ data can be processed further to quantify the vegetation loss caused by the wildfire and visualize the area which has been impacted.

Visualize difference in vegetation before and after the wildfire

Quantify of loss in vegetation caused by wildfire

Notebook CI Test Results

This notebook was tested in multiple regions. The test results are as follows, except for us-west-2 which is shown at the top of the notebook.