README.md

Smart Parking Application

The smart parking application is an AI-based solution to monitor individual car parking spaces and their occupancy statuses in a car park. The model vehicle-detection-0201 from Open Model Zoo is used for car detection in this application.

Requirements

Validated hardware

• CPU: 13th generations of Intel Core processors and above
• GPU: Intel® Arc™ graphics
• RAM: 32GB
• DISK: 128GB

Setup

Prequisite

• OS: Ubuntu 24.04 LTS

• Docker and Docker Compose: Ensure Docker and Docker Compose are installed. Refer to Docker installation guide.

• Intel GPU Drivers: Refer to here to install Intel GPU Drivers

Generate SSL certs

bash setup/generate-certs.sh

Download sample video

wget https://videos.pexels.com/video-files/30937634/13228649_1920_1080_30fps.mp4 -O server/resources/carpark_video_1.mp4

Build Image

docker compose build

Start containers

docker compose up -d

Stop containers

docker compose down

Access streams

• View all camera streams: https://localhost/camera/all

  

• View specific camera stream: https://localhost/camera/<no_of_camera_stream>

  

Grafana dashboard

Recommended web browser: Firefox

1. Login

   • https://localhost/grafana/login
   • Username: admin
   • Password: admin

2. Add data source

   • Go to Settings -> Data Sources -> Add new data source
   • Choose InfluxDB as the data source
   • Fill in URL
     • http://influxdb:8086
   • Turn on Auth and fill in the basic auth details
     • Username: admin
     • Password: admin
   • Fill in InfluxDB details
     • Database: carpark_data
     • HTTP Method: POST

   

   • Save and test

3. Import template from grafana folder

   • Import template Lot.json and Summary.json from grafana folder

   

Notes

Model Configuration

You can change the model to pretrained YOLOv8-n by replacing line 37 and 38 in server/utils/video_stream.py with code below:

self.detector_yolo = CarDetectorYolo(f'./model/{model_name}_int8_openvino_model/{model_name}.xml', device="CPU")
self.detector_omz = None