Finding parking spots with Custom Vision and IoT
summary
Camera plus a standard USB webcam trains on labeled images of parking spots and runs an exported Custom Vision model on a $40 Raspberry Pi (dockerized) to classify spots every 3–5 seconds; demo shows spot 1 empty, 2 and 3 full. Setup uses Custom Vision upload and export to an on‑device model (Custom Vision export), keeps inference local for privacy, and generalizes to fridge checks, trash, garage door or detecting a cat bringing objects.
AI-written summary · about
chapters (6)
- 0:00Introduction to parking AI
- 1:02Toy parking demo results
- 2:03AI training and hardware
- 3:05Expanding AI to other uses
- 4:40Implementation and training
- 5:43Privacy and conclusion
related talks
29mApplied Custom Vision
26mWhere can I park? Deep Learning is here to help you!
54mSurveillance Detection Scout – Your Lookout on Autopilot
18mRecognition and Localization of Parking Signs using Deep Learning
43mIntersection of Awesome: Building AI enabled IoT solutions powered by the cloud
35mComputer vision: British Clay
26mSelf Driving cars with Opencv
28mVision, Big Data and AI