Character Recognition Of Plates is a program desgined to detect and recognize character from images. The project can be broken down to 3 independent modules
- License Plate Detection : This part is associated with detection of license plate from the image. The coordinates of license plates are used to crop the plate and pass it onto the next step.
- Character Detection : This step detect the character from the croped region. The coordinates of individual characters are then used to crop them. The sequence of characters are also sorted based on their x coordinate position to make them appear in the same order as in the license plate. This cropped characters are then passed on the final step for recognition.
- Character Recognition : This is the final step that returns a list of possible candidates. Since 8 can sometimes appear as B we need to return a fixed number of possibilities. All possible combinates are generated and are returned after sorting them on the basis of their score.
- License Plate Detection (custom anotation required)
- License Plate Character Detection
- Character Recognition here & here
- Using Python 3.8.10
pip install -r requirements.txt
for the rest of the packages
-
- source : path of image (default is sample_car.png)
- count : the maximum number of possible plates to predict (default is 10)
-
- run
python crop.py
- run
-
- source : path of image (default is sample_car.png)
- count : the maximum number of possible plates to predict (default is 10)
-
- run
python flaskapp.py
- run