Livestock Counting at the Edge
Publication date: July 12, 2022 (Diagram history)
This architecture enables you to build a near real-time, automated counting application for livestock. Use Amazon SageMaker AI and AWS IoT Greengrass to build and deploy a livestock counting application at the edge.
Livestock Counting at the Edge diagram
-
Upload videos and images to Amazon Simple Storage Service (Amazon S3) to train the livestock detection model.
-
Use Amazon SageMaker AI Notebooks to process these videos and create a labeling job using Amazon SageMaker AI Ground Truth.
-
Split the annotated dataset into training and validation sets, and use Amazon SageMaker AI distributed training for livestock detection.
-
Use Amazon SageMaker AI Neo to optimize the livestock detection model for specific target devices like NVIDIA Jetson Nano, TX2, Xavier, AWS DeepLens , or Raspberry Pi.
-
Deploy the machine learning model and counting application AWS Lambda function to the edge device using AWS IoT Greengrass.
-
Consume live video streams from a camera at the farm using real-time streaming protocol (RTSP) through camera serial interface (CSI) or through USB connected to the edge hardware.
-
Run ML Inference on the video frames from step 6 and pass the bounding box outputs to the counting application Lambda function.
-
Connect to the web server running on edge devices and control when to start/stop counting through a mobile application.
-
Submit near real-time counts to an inventory management system.
Further reading
For additional information, refer to
