Vision based ID Reader

The objective of the project is to read and store data from an ID Card using a camera using OCR. The OCR package used here is pyocr.

The front of the ID card of VIT University contains Registration Number and Name of the student which is what we aim to extract. The OCR dumps all kinds of data, which we clean by means of Regex and stored in a file named data.txt We clean that data by the logic, 'Most occured entry is the right entry'. Name and RegNo is now stored. Same is done with the back side for phone number.

Packages and Dependencies

pip install -r requirements.txt on your virtual environment.
If by any means you recieve the error "Index out of range", follow this link: https://stackoverflow.com/questions/31892413/no-tools-available-from-pyocr

Procedure for Computer

1. Install dependencies on your virtual environment.
2. Show the front of your ID card to the camera for say 3 seconds followed by the back for the same duration.
3. Press 'q' to close the window.
4. A prompt shows the details it has read, if right press "y" and it will get updated in the database

Procedure for Raspberry Pi

1. Install dependencies.
2. Show the front of your ID card to the camera for say 10 seconds at the focus for better results.
3. Hold the Button until the LED flashes.
4. A prompt shows the details it has read. If right,hold the button and it will get updated in the databas. If not, leave it be and it is regarded as a false entry.