A collection of image processing algorithms that aim to extract the license plate from a car’s photograph taken at parking lots/tolls etc.

 

The whole process is consisted by 3 parts:

  1. Locate the plate inside the photograph
  2. Extract each character as a separate part of the photograph
  3. Transform the characters to text using OCR

1. Locating the plate inside the photograph

  • Convert the image to grayscale
  • Apply a sobel mask for later edge detection
  • Convert the image to binary by Thresholding
  • Calculate edges
  • Use the edge to locate the plate

2. Extracting each character as a separate part of the photograph

  • Convert the plate image to binary by Thresholding
  • Calculate edges
  • Use the edges and extract the characters

3. Transform the characters to text using OCR

Not implemented

Developed by Stavros Tsourlidakis, and Philip Dinenis