Enhance Operational Efficiency with Edge Computer Vision using AWS Panorama

Overview

In today’s data-driven world, organizations increasingly turn to computer vision (CV) to gain insights from their visual data. However, traditional CV approaches often rely on centralized cloud computing, introducing latency and bandwidth constraints that can limit real-time decision-making and operational efficiency.

To address these challenges, AWS Panorama emerges as a powerful solution for edge computer vision. It provides a comprehensive platform for deploying and managing CV applications directly on edge devices, enabling real-time processing and analysis of visual data at the source.

Pioneers in Cloud Consulting & Migration Services

Reduced infrastructural costs
Accelerated application deployment

Get Started

Introduction

AWS Panorama is a machine learning solution consisting of a dedicated appliance and a software development kit (SDK) designed to enable on-premises cameras to perform accurate and low-latency predictions using computer vision. The AWS Panorama Appliance facilitates the automation of tasks typically reliant on human inspection, enhancing the detection of potential issues.

This technology can assess manufacturing quality, pinpoint bottlenecks in industrial processes, and enhance workplace security, especially in settings with restricted or no internet connectivity. Additionally, the SDK empowers camera manufacturers to integrate similar capabilities directly into their IP cameras.

The AWS Panorama Appliance operates locally as an edge device, facilitating real-time video analysis. Applications comprise nodes representing components like cameras and models, bundled into packages and uploaded to Amazon S3. An application manifest configures connections between these nodes.

AWS Panorama

AWS Panorama is compatible with models created using various frameworks, including Apache MXNet, DarkNet, GluonCV, Keras, ONNX, PyTorch, TensorFlow, and TensorFlow Lite. You can construct models using Amazon SageMaker or import them directly from an Amazon Simple Storage Service (Amazon S3) bucket.

Configuration of Appliance

First, we should configure it to the network (cable + DHCP, also static IP configuration) and register it to securely connect to the AWS Account. We must navigate the AWS Management Console and enter network configuration details to generate configuration files and certificates. Later, it is copied to the appliance using the provided USB key.

Subsequently, retrieving a sample application from the AWS Panorama GitHub repository and experimenting with the Test Utility for AWS Panorama is essential. This utility utilizes Jupyter Notebooks to swiftly test sample applications or custom code before actual deployment on the appliance. It also provides a set of commands to deploy applications to the appliance programmatically.

AWS Panorama Command Line

The Panorama command line facilitates creating a project, importing assets, packaging it, and deploying it onto the AWS Panorama Appliance.
Once the application is prepared, it is necessary to build the container utilizing a Linux machine equipped with Docker Engine and Docker CLI, thereby avoiding using Docker Desktop for macOS or Windows.

$ panorama-cli build-container                               \
   --container-asset-name {container_asset_name} \ 
   --package-path packages/{account_id}-{package_name}-1.0

$ panorama-cli build-container \

--container-asset-name {container_asset_name} \

--package-path packages/{account_id}-{package_name}-1.0

2. Adding an ML Model

The AWS Panorama Appliance is compatible with various ML model frameworks. Models can be trained on Amazon SageMaker, and in this case, the ML model is downloaded from S3 and imported into the project.

panorama-cli add-raw-model                                                 \
    --model-asset-name {asset_name}                                        \
    --model-s3-uri s3://{S3_BUCKET}/{project_name}/{ML_MODEL_FNAME}.tar.gz \
    --descriptor-path {descriptor_path}                                    \
    --packages-path {package_path}

panorama-cli add-raw-model \

--model-asset-name {asset_name} \

--model-s3-uri s3://{S3_BUCKET}/{project_name}/{ML_MODEL_FNAME}.tar.gz \

--descriptor-path {descriptor_path} \

--packages-path {package_path}

3. Packaging the Application

Now that the ML model and application code are enclosed in a container, the next task involves packaging the application assets for the AWS Panorama Appliance.

The below command uploads all application assets, along with all manifests, to the AWS cloud account.

panorama-cli package-application

1	panorama-cli package-application

4. Deploying the Application

In the last step, we deploy the application to the AWS Panorama Appliance. During deployment, the application and its configuration details, such as camera stream selection, are transferred from the AWS cloud to the on-premise AWS Panorama Appliance. In this case, the deployment is carried out programmatically using Python code.

client = boto3.client('panorama')
client.create_application_instance(
    Name="AWS News Blog Sample Application",
    Description="An object detection app",
    ManifestPayload={
       'PayloadData': manifest 
    },
    RuntimeRoleArn=role,               
    DefaultRuntimeContextDevice=device  
)

client = boto3.client('panorama')

client.create_application_instance(

Name="AWS News Blog Sample Application",

Description="An object detection app",

ManifestPayload={

'PayloadData': manifest

RuntimeRoleArn=role,

DefaultRuntimeContextDevice=device

)

Conclusion

Edge computer vision, powered by AWS Panorama, offers a transformative approach to unlocking operational efficiency and gaining real-time insights from visual data. By bringing computer vision capabilities closer to the data source, organizations can optimize operations, enhance decision-making, and improve overall business outcomes.

Drop a query if you have any questions regarding AWS Panorama and we will get back to you quickly.

Making IT Networks Enterprise-ready – Cloud Management Services

Accelerated cloud migration
End-to-end view of the cloud environment

Get Started

About CloudThat

CloudThat is an official AWS (Amazon Web Services) Advanced Consulting Partner and Training partner, AWS Migration Partner, AWS Data and Analytics Partner, AWS DevOps Competency Partner, Amazon QuickSight Service Delivery Partner, Amazon EKS Service Delivery Partner, Microsoft Gold Partner, and many more, helping people develop knowledge of the cloud and help their businesses aim for higher goals using best-in-industry cloud computing practices and expertise. We are on a mission to build a robust cloud computing ecosystem by disseminating knowledge on technological intricacies within the cloud space. Our blogs, webinars, case studies, and white papers enable all the stakeholders in the cloud computing sphere.

To get started, go through our Consultancy page and Managed Services Package, CloudThat’s offerings.

FAQs

1. Is AWS Panorama available in all AWS Regions?

ANS: – No, AWS Panorama is not yet available in all AWS Regions. It is currently available only in Europe (Ireland), Asia Pacific (Singapore), Asia Pacific (Sydney), the US East (N. Virginia), Canada (Central), and the US West (Oregon).

2. How is storage for AWS Panorama assets handled, and what are the associated costs?

ANS: – AWS Panorama stores versioned copies of all deployed assets, including ML models and business logic, in the cloud. Users are charged $0.10 per GB per month for this storage.

WRITTEN BY Abhilasha D

Abhilasha D is a Research Associate-DevOps at CloudThat. She is focused on gaining knowledge of Cloud environment and DevOps tools. She has keen interest in learning and researching on emerging technologies.