Perfect Parking

Perfect Parking aims to create a parking system that will replace outdated systems and help stop the widespread problem that is parking in our cities. The goal of this project is to improve the effectiveness of finding parking spaces and to also relieve the stress of the users looking for parking by implementing new and innovative features. This will be done by building a client-server API and website that will show the user exactly where the parking is. The server will be supported by parking monitor clients powered by OpenCV to detect if a parking spot has been filled or made empty. To read more, please see the thesis.

Requirements

• Python 3.9 - Python is a programming language that lets you work quickly and integrate systems more effectively.

IDEs

• VS Code - Code editing.Redefined. Free. Built on open source. Runs everywhere.¹

Source Code Management

• Git is a free and open-source distributed version control system designed to handle everything from small to very large projects with speed and efficiency.
• GitHub is a web-based hosting service for version control using Git. It offers all of the distributed version control and source code management (SCM) functionality of Git as well as adding its features.
• GitHub Desktop - Simple collaboration from your desktop
• WinMerge is an Open Source differencing and merging tool for Windows.

Mobile Apps for Camera Feeds

• DroidCam - Use your Android or IPhone as a webcam on your computer over WiFi or USB.

Run Application

Clone the repository

git clone https://github.com/rhys-lit/perfect-parking-fyp.git

There are three applications in this project: the website server, a client parking monitor and a mock client parking monitor.

Run Website Server

1. Install dependencies

   • With Pip (To fast execute, run on Windows pip-install.bat)

     pip install -r requirements.txt

   • With conda (To fast execute, run on Windows conda-install.bat)

     conda install --file requirements.txt

2. Run the app Open VS Code and run the app by selecting the Run and Debug (Ctrl + Shift + D) tab and selecting Server - Django from the dropdown menu. Then click the green play button to run the app, or Open a terminal and run the following command:

   python manage.py runserver

3. Open the app in your browser http://localhost:8000

Built-in User Accounts

Built-in user accounts are:

Username	Password	Role	Note
john	John123456	admin	Built in admin account
admin1	letmein	admin	Built in admin account
parkingMonitor	Letmein1$	user	Account for parking monitor apps

Run the Client Parking Monitor

For the client parking monitor to run, the server must be running.

1. Install dependencies

   • With Pip (To fast execute, run on Windows PerfectParkingClient/pip-install.bat)

     pip install -r PerfectParkingClient/requirements.txt

   • With conda (To fast execute, run on Windows PerfectParkingClient/conda-install.bat)

     conda install --file PerfectParkingClient/requirements.txt

2. The server must be running

3. Open VS Code and run the app by selecting the Run and Debug (Ctrl + Shift + D) tab and select one of the client configurations e.g., Client - Config Model Car Park Recording from the dropdown menu. Then click the green play button to run the app, or Open a terminal and run the following command:

   python PerfectParkingClient/main.py '--image' 'PerfectParkingClient/images/live-sample-3.png' '--data' 'PerfectParkingClient/data/coordinates-live-sample.yml' '--video' 'PerfectParkingClient/videos/live-sample-3.mp4' '--start-frame' '400'

4. Follow the instructions in the Window Title Bar.

Sample Live Client Run

The live client configuration is used to get video from a phone app such as DroidCam. Note that the IP address of 192.168.188.106:4747 for Client Live configurations may need to be changed depending on your Network Configuration. This can be changed in the launch.json file. After running the app, the following steps will be shown in the Window Title Bar.

1. Press s to select an image from the video feed to make your parking spaces on. Ideally, this should be an image of the car park with no cars in it.
2. Mark the car parking spaces by clicking on the four corners of each parking space. Try not to include any white lines in the parking space.Press q when finished.
3. Step 3: Watch the video feed and see the parking spaces being detected. Press q to quit when you are finished.

Mock Client Parking Monitor

The mock client parking monitor is used to test the server without the need for a client parking monitor. It loads the configs for five parking monitors and sends randomly generated data to the server at random times.

For the mock client parking monitor to run, the server must be running. The mock client dependencies are the same as the client parking monitor.

To run the mock client parking monitor, Open VS Code and run the app by selecting the Run and Debug (Ctrl + Shift + D) tab and selecting Mock Client - ParkingLot from the dropdown menu. Then click the green play button to run the app, or

Open a terminal and run the following command:

python PerfectParkingClient/mock.py

You can stop the mock client by pressing q at any time.

---

Copyright © 2023, Rhys Quilter

Footnotes

1. The Extensions recommended for this project are in extensions.json ↩