Jetson Nano Machine Learning Projects You Need To Try!

By Gigi 3 years ago

NVIDIA’s Jetson Nano has excellent GPU capabilities, making it a popular choice for Machine Learning (ML) and often used for gaming and CUDA-based computations. The NVIDIA Jetson Nano Developer Kit is currently out of stock. Check out Seeed’s reComputer J1010 and J1020 built with the Jetson Nano production model, rich onboard interfaces, heatsink, and enclosure! 

reComputer with Jetson Production Module: Real-World AI at the Edge, starting from $199

Built with Jetson Nano 4GB

Buy Now
  • Edge AI box fits into anywhere
  • Embedded Jetson Nano Module
  • Pre-installed Jetpack for easy deployment
  • Nearly the same form factor as Jetson Developer Kits, with a rich set of I/Os
  • Stackable and expandable

Without further ado, let us look at some ML projects you can create yourself powered by Jetson Nano!

The projects are categorized into 2 of the most popular frameworks: TensorFlow and PyTorch. Also, scroll to the bottom of the page under resources if you’re unfamiliar with ML frameworks to learn how to get started with Jetson Nano!

TensorFlow Projects

Autonomous Tank

Ref: Hackster

This project will teach you how to use Jetson Nano to connect to TensorFlow and make a tracked vehicle made with LEGO parts.

What you‘ll need

  • Jetson Nano
  • Pixy2 Camera
  • LEGO Mindstorms EV3 Programming Brick / Kit
  • YDLIDAR X4, 2D Laser Scanner
  • microSDHC Card (EV3 only supports MicroSDHC cards)
  • USB-A to Micro-USB Cable
  • USB-A to Mini-USB Cable
  • USB Power Bank
  • Wi-Fi USB Adapter
  • Generic USB Sound Card Adapter
  • Small 3.5mm Microphone
  • TensorFlow

AI-Powered Irrigation Robot

Ref: Hackster

Constantly forgetting to water your plants on time? You can build an autonomous, sustainable robot to help you care for your indoor and outdoor plants. It is also able to dispense fertilizers and pesticides, detect the water levels in the soil, and detect sick plants.

What you’ll need

Clean Water AI

Ref: Hackster

Ever thought of detecting water contamination without the use of chemicals? By building Clean Water AI and utilizing TensorFlow Object Detection Method and WebRTC, you’ll be able to check your water source!

What you’ll need:

Fruits Classification

Ref: Hackster

Just completed the Hello World tutorial and not sure which project to try out next? This project shows you how to make your dataset and classify the different fruits using TensorFlow example files and the camera-capture utility.

What you’ll need:

Deploying TensorFlow Models at the Edge

If you’re into exploring TensorFlow running at the edge with Kubernetes, then this tutorial will show you how you can do so! By running TensorFlow models as microservices at the edge on K3s distribution, you’ll be able to combine AI with IoT to Kubernetes infrastructure.

What you’ll need:

Custom Object Detection

Do you already know how to utilize the Jetson Nano in your projects and want to create your own custom object detection neural network? This project will show you everything you’ll need to do so!

What you’ll need:

  • Jetson Nano
  • TensorFlow
  • OpenCV
  • Jupyter Notebook
  • RPI Camera
  • Python

Mommybot

Ref: GitHub

You might be wondering what is a Mommybot? It is a personal sleep assistant that helps you manage your sleeping hours. If you’re constantly sleep-deprived, this project might be your solution to that!

What you’ll need:

Reading Eye For The Blind

Ref: Hackster

Ever wanted to build something to help the visually impaired? This project aims to aid those with poor vision or reading disabilities and hear printed and handwritten text by converting recognized sentences into synthesized speech.

What you’ll need:

  • Jetson Nano
  • Raspberry Pi Camera Module V2
  • Flex Cable for Raspberry Pi Camera
  • USB Keyboard and Mouse
  • 5V 2.5A Power Supply With Micro USB Cable
  • 32GB Micro SD Card
  • Computer Display
  • Arducam 8MP Wide Angle Drop-in Replacement
  • Power Bank
  • Edimax Wi-Fi USB Adapter
  • USB to 3.5mm Jack Audio Adapter
  • 2.0 A-Male to Micro B USB Cable
  • Mini LED Arcade Button
  • Illuminated Toggle Switch with Cover
  • Noctua NF-A4x20 5V Fan
  • TensorFlow
  • OpenCV
  • Google Cloud Vision & Text-to-Speech API
  • 3D Printer

MixPose

Ref: Hackster

Want to attend yoga or fitness classes in the comfort of your home? With this project, your yoga or fitness instructors could guide you through without the need to meet face to face, which could be very useful during times of pandemic!

What you’ll need:

Qrio

Ref: towardsdatascience

If you have a toddler who is curious about everything, this might be the perfect project to build for your kid! Using an object detection AI model, game engine, and a few other software, you’ll be able to get your bot which can speak, recognize a toy, and play a relevant video on YouTube for him/her!

What you’ll need:

  • Jetson Nano
  • TV
  • IMX219 Camera
  • TensorFlow
  • Amazon Polly Text-to-Speech
  • Selenium Automation Framework
  • Game Engine with Skeleton Animation System

PyTorch Projects

Fire Detecting Drone

Ref: Hackster

This project seeks to detect wildfires early to prevent casualties. An autonomous drone is deployed to detect wildfires using Computer Vision and Convolutional Neural Networks. To train the ML of the drone, a domain-specific fire dataset of 2,000 images consisting of drone imagery from real-world fire scenarios can be found here.

What you’ll need:

Drowsiness, Blind Spot and Emotions (DBSE) Monitor

This project is an in-car assistance system that alerts drivers if they are drowsy or distracted and notifies them about the objects in the blindspots. 

There are 3 different systems built in: 

  • Drowsiness Detection – detects a person’s drowsiness level and notifies the user about his state and whether he can drive. It also measures the driver’s attention and if he is falling asleep while at the wheel. 
  • Emotion Detection – detects the driver’s state (drowsy or distracted). If it is positive, a powerful alarm will sound off to wake the driver. It also plays music to generate positive responses for the driver using a facial feature analyzer.
  • Driving Monitor – detects the traffic condition and acts as an automobile’s blind spot detector. 

It also has an inbuilt accelerometer to generate a call to emergency services if an accident were to happen.

What you’ll need:

  • Jetson Nano
  • ESP32
  • 0.96″ OLED Display
  • Power Inverter for Car
  • Bluetooth Speaker or Bluetooth Audio Car System
  • USB TP-Link USB Wifi Adapter TL-WN725N
  • UGREEN USB Bluetooth 4.0 Adapter
  • HD Webcam
  • 32 GB MicroSD Card
  • 5V-4A AC/DC Adapter Power Supply Jack Connector
  • VMA204
  • PyTorch
  • JetPack 4.3
  • YOLOv3
  • OpenCV
  • Twilio
  • Arduino IDE
  • Mosquitto MQTT

Auto Car 

Ref: Github

Want to build an autopilot remote car to play with during your free time? With this project, you’ll be able to do so with the help of a few software!

What you’ll need:

  • Jetson Nano
  • IMX219 Camera
  • Servo Driver
  • WiFi Chip
  • Wifi Antenna
  • PS4 Controller
  • 1/10 4wd Car Chassis
  • Brushed Motor
  • Standard Analog Servo
  • 45A Brushed Car ESC
  • 5000mAh 2S 7.4V 60C Hardcase Pack
  • PyTorch

Image Recognition 

Ref: Medium

If you’re a beginner who has just learned to set up your Jetson Nano, this project will teach you how to use pre-trained models to train a model from scratch and do transfer learning!

What you’ll need:

Interested? Do check out this project and learn how to train models!

Tipper

Ref: Hackster

Have a spare nerf gun you’d like to repurpose as a ball launcher? This project shows you how to pitch a ball, and then it’ll tell you whether it is in or out of the strike zone!

What you’ll need:

AIoT – Artificial Intelligence on Thoughts

Ref: Hackster

Do you want to build an artificial intelligence-powered device that allows you to interface your brain with your hardware and software project directly? This project will show you how to build a neural network from scratch!

What you’ll need:

  • Jetson Nano
  • Breadboard
  • EEG device
  • TPLINK TLWN823N Wifi stick
  • 2 x Resistor 10k ohm
  • 2 x Resistor 330 ohm
  • 2 x General Purpose Transistor NPN
  • Male/Female Jumper Wires
  • PyTorch
  • Python 3
  • NVIDIA Jetson Nano Image Software

Summary

And that’s all the Jetson Nano projects for now! We hope you’ll find some of the projects interesting. Should you need further resources on Jetson Nano and Machine Learning in general, do click on the links under resources to find out more!

Resources

About Author

Gigi

See author's posts

Tags: , , , , ,

Calendar

February 2021
M T W T F S S
1234567
891011121314
15161718192021
22232425262728