fbpx

Build a ML model with Amazon SageMaker

When we talk about IIoT, it is quite challenging to gather, identify, process and analyse any type of product, hood images manually at the EDGE location, near the Assembly Line, Tandem Press Line and Core Mining. But with the help of Amazon SageMaker it is quite east, cost effective and east to quick start, to train and deploy ML models on the top of IoT EDGE Devices. You can use RaspberryPi as an EDGE device to deploy your ML model, Amazon SageMaker to train and built your model and finally Greengrass to give your EDGE Devices most of the cloud capabilities to process your data at the EDGE location.

This solution is based on a ML model for image classification based on the CIFAR-10 dataset, which will be trained with Amazon SageMaker and then downloaded the trained model at AWS Greengrass to perform image classification at the EDGE device, which is connected with a camera to get images of products, car and hood. An EDGE device, here it is RaspberryPi running the Greengrass software to simply copy images to a directory which is regularly scanned and when an image is found it will be classified. Here Lambda function helps to load the model that has been deployed to Greengrass.

Components Involved

IoT Sagemaker

Amazon SageMaker

Amazon SageMaker is a fully managed Machine Learning service, which helps data scientists and developers can quickly and easily build and train machine learning models, and then directly deploy them into a production-ready hosted environment. It provides an integrated Jupyter Notebook instance for easy access to your data sources for exploration and analysis, so you don’t have to manage servers. It also provides common machine learning algorithms that are optimised to run efficiently against extremely large data in a distributed environment.

AWS IoT Greengrass

AWS IoT Greengrass is a software that extends AWS Cloud capabilities to local devices, making it possible for those devices to collect and analyse data closer to the EDGE IoT Devices, while also securely communicating with each other on local networks. More specifically, developers who use AWS Greengrass can author serverless code (AWS Lambda functions) in the cloud and can easily deploy it to devices for local execution of applications at EDGE.

How this Works

Amazon SageMaker
AWS Greengrass

Architecture

Amazon Sagemaker Architecture

JSON Response From the ML Model

JSON Response Model

Conclusion

With the help Amazon SageMaker you can quickly create ML models and deploy it on the top of IoT EDGE Devices to classify Images based on classes, also you can identify dents from the hoods, anomaly detection near Tandem Press Line. This ML Model helps in Object Detection in areas where you have very poor network bandwidth like core mining.

Leave a Reply

Your email address will not be published. Required fields are marked *

Fill out this field
Fill out this field
Please enter a valid email address.

Menu