ParkLens is a mobile application that utilizes computer vision techniques to recognize vacant parking spots by analyzing video footage of a parking lot.
Note: This repository goes along with the computer vision model, the repository of which can be found here
This guide is for the app development only. Please refer to other repo for Python environment setup guidelines.
- Parking Spot Detection: Accurately identifies empty and taken parking spots in real-time.
- Live Streaming: Streams processed video data to the iOS application via HLS.
- RTMP Support: Uses an RTMP protocol to capture and analyze live video from a real security IP camera.
- Easy-to-use UI: User-friendly UI that guides user to open parking spots seamlessly.
Note: If you are a contributor, I have made a Figma UI Prototype of the app that you can view here
ParkLens operates through this pipeline:
-
A live IP camera records video of the parking lot.
-
The video feed is streamed via RTMP to a processing server (MacBook running OpenCV and YOLO for object detection).
-
The processed footage, with detected parking spots, is sent via RTMP to an Nginx server.
-
The Nginx server converts the stream to HLS format for viewing on the ParkLens iOS app.
- iOS(17.2 or higher)
- Stable Wifi connection
tbd...
tbd...