This solution shows how to process satellite imagery using AWS SageMaker and PlanesNet to build an AI Model to predict aircraft. This readme updates an article "Detect airplanes in Planet imagery using machine learning" by Bob Hammell referenced below and provides a more basic step by step process.
We'll start a Jupyter Notebook using AWS SageMaker. We can then use the Jupyter Notebook to process all the steps required to predict aircraft in the satellite imagery.
Use the AWS Console to configure a SageMaker Instance for processing satellite data. This is a step by step process.
Click on "Notebook instances"
Click on "Create notebook instance"
Notebook instance name: planesnet
Notebook instance type: ml.t2.medium
IAM Role: Create a new role
S3 buckets you specify - optional:
Select "None"
Click on "Create role"
Click on "Create notebook instance"
Notebook/Notebook instances
Name: planesnet
Actions: Open # it will show pending until it's ready to open
This will open the Jupyter Notebook in a new tab in your browser.
Click on "Upload" and Select "aws-sagemaker-planesnet.ipynb" from project jupyter-notebook directory
Once the notebook is uploaded, click on "aws-sagemaker-planesnet.ipynb" to open it.
Run each cell Step by Step
Detect airplanes in Planet imagery using machine learning
https://github.com/rhammell/planesnet-detector