ProductionEvidence: Low40/100

AWS IoT Greengrass and Amazon SageMaker Edge Manager for Livestock Monitoring Using Computer Vision

Multiple farms in the agriculture industry faced difficulties counting and tracking livestock in remote areas without constant cloud connectivity, leading to inconsistent and inefficient manual counting. They used Amazon SageMaker's built-in object detection model, optimized with Amazon SageMaker Neo for edge devices such as NVIDIA Jetson Xavier, to detect livestock in video streams. Deployment and management of edge models and applications were handled via AWS IoT Greengrass V2, which allowed running inference and counting livestock near real-time at the edge from live camera streams. This solution improved operational efficiency with consistent and accurate livestock counts, reduced manual counting errors, and enabled workers to focus on higher-value tasks.

Organization
Multiple farms
Industry
Agriculture
Published
September 2021

Reported outcomes

Strategic outcomes

Better decisions & insightProvided consistent, accurate livestock countsCost efficiencyReduced manual labor and counting errorsEmployee experienceFreed workers for higher-value tasksSpeed & agilityEnabled near real-time edge counting
Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks
Customer
Multiple farms
Provider
AWS
Maturity
Production

This solution improved operational efficiency with consistent and accurate livestock counts, reduced manual counting errors, and enabled workers to focus on higher-value tasks

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 3 of 3

  • 1Edge AI for Computer Vision
  • 2Livestock Monitoring
  • 3Operational Efficiency Enhancement
  • Counting and tracking livestock manually in remote farms was inconsistent and inefficient due to lack of continuous cloud connectivity.
  • Manual counting was error-prone and time-consuming, affecting operational decisions on feed and weight management.
  • Trained an object detection model using Amazon SageMaker built-in algorithms on annotated livestock images.
  • Optimized the model for edge deployment using Amazon SageMaker Neo tailored for NVIDIA Jetson Xavier hardware.
  • Packaged and securely deployed the model and accompanying inference client as AWS IoT Greengrass V2 components to run inference locally at the edge.
  • Edge Manager monitored model health and enabled updates without disrupting the overall application.
  • Inference used live video feeds for real-time livestock counting using computer vision and tracking algorithms.
  • Provided consistent, accurate, near real-time livestock counts for better operational decisions.
  • Reduced manual labor and counting errors, improving worker productivity and focus on higher-value tasks.
  • Enabled technology adoption in agriculture to enhance economic outcomes and animal care quality.
Architecture

Architecture integrating Amazon SageMaker model training with SageMaker Neo optimization for edge devices, deployment via AWS IoT Greengrass V2, and Edge Manager for model monitoring and management on NVIDIA Jetson Xavier edge devices conducting computer vision inference on live camera streams.

Sources & evidence1
Evidence: Low40/100Evidence strength
  • Customer explicitly identified
  • Deployment status explicitly supported
  • Technical implementation details available
Type: Blog PostPublished: Sep 2, 2021Publisher: AWS Machine Learning BlogEvidence: VendorConfidence: Medium

AI-generated summary. Verify important details with the linked sources before relying on this case.

Explore related AI use cases

Was this useful?

Community

Comments

Loading comments...

Similar cases