CarPair seeks to establish trust and streamline the car sales process, by establishing a centralized platform that makes use of a webscraper that scours the web for car listings. The gathered data is then further complimented by using Automatic Numberplate Recognition(ANPR) software to extract number plates from car listings, which further enriches the gathered data, by using external web scraping of external data sources. This results in a decentralized platform with car listings that have universally enriched data, whose parameters can then be further searched and filtered through.
- Make sure to install NPM and Node on your machine.
- Clone the repository and CD into the folder \carpair\carpair.
- Run
npm install
to install dependencies. - To start the application, run
node app.js
in the folder. - Visit http://localhost:8080
Please make sure that you have the correct libraries installed in order for the application to properly function.
go to carpair/scripts and open the command prompt and run
python dba.py
go to carpair/scripts/imgscraper.py and change the path in the method DownloadImgFromUrl(), so that it points to the carpair/anpr/Plate_examples folder in the carpair directory.
save_path = 'C:/Users/user/Desktop/carpair/anpr/data/images
run the imgscraper.py script located in carpair/scripts to get all images to run anpr on
python imgscraper.py
Run the following command in the carpair/anpr folder to run the ANPR which ultimately outputs the correctly identified numberplates to the database
python anpr.py
Run the following script with the commandopromot in carpair/scripts
python mvaScrape.py
Your application should now contain the cars, where a number plate has correctly been identified, with additional data, than its original listing had.
_ At the given time the ANPR does not insert number plates into the DB, for lack of interference when presenting the application_