Important
For the latest version of this application, go to https://github.com/Avnet/SPARK/
Welcome to SPARK, the edge-AI solution for smart EV charging and parking lot management. Bootstrapping your Edge AI with RZBoard and the RZV AI SDK.
SPARK is an effective EV charging prototype that integrates a real-time capable convolutional neural network (CNN) for occupancy detection with the RZBoard V2L. The result: a smart parking lot occupancy/EV charging experience for low cost. This system is designed to automate parking lot analysis, streaming edge compute data to your desired target: an information hub like IoT Connect, or an HDMI output. The result: improved user-experience for EV charging customers and parking lot managers.
Looking for the dataset used to train? Check it out on kaggle.
- Real-Time Occupancy Detection: Utilizing an efficient CNN, SPARK detects vehicle occupancy in real-time, ensuring accurate information on parking availability.
- User-Friendly Interface: Easy-to-use interface demo interface: output data to HDMI for simple viewing. Or, integrate with IoT connect for bootstrapped dashboards.
- Sustainable Solution: Promotes efficient energy use and supports the growing need for EV infrastructure.
- Scalable Architecture: Designed to scale: one image could host up to 200 parking spots. Try adding more parking slots and/or images until your FPS requirement isn't met. Then, scale horizontally with more RZBoards connected to your IoT Connect style HUB!
Looking to demo the software? Simply download and install from the Release instructions.
Want to build and modify the demo yourself? You'll have to run the build script and/or deploy script from within the RZV AI SDK
./build.sh && ./deploy.sh [board_ip]SPARK operation is as simple as clicking a button and dragging some boxes.
operation_demo.mp4
This project is licensed under the MIT License
- This project wouldn't be possible without OpenCV and RZV AI SDK!
© 2024 Lucas Keller. All Rights Reserved.