mkdir backup
plate_detect.py: This file contains code to detect number plates in an image. It uses OpenCV to perform image processing tasks such as thresholding, blurring, edge detection, and contour detection to locate regions of interest that may contain number plates. It then applies a series of filters to eliminate non-plate regions and extracts the plate region using perspective transformation.
character_recognition.py: This file contains code to recognize characters on number plates. It uses the Tesseract OCR engine to perform character recognition on the extracted plate regions. The recognized characters are then filtered and processed to obtain the final license plate number.
database.py: This file contains code to connect to a MySQL database and perform CRUD operations on it. It includes functions to create a new record, read existing records, update records, and delete records.
criminal_record.py: This file contains code to check if a given license plate number is present in the criminal record database. It connects to the MySQL database using database.py and fetches the relevant records.
main.py: This file is the entry point of the application. It takes an input image and performs license plate detection and character recognition using plate_detect.py and character_recognition.py. It then checks if the license plate number is present in the criminal record database using criminal_record.py.
The website_method() function takes images from the "./data/img/" directory, sends them to an API provided by PlateRecognizer for license plate recognition, and returns the recognized license plate number.
The algo_method() function first generates a list of image paths, which is then passed into the YOLOv3 detector to identify license plate objects in the images. The detected objects are then filtered to include only the ones that contain license plates. The bounding box coordinates of the license plate objects are used to crop the license plate images out of the original images, which are then passed to OCR.Space API for optical character recognition. The recognized characters are then cleaned up, and if it matches the format of a license plate number, it is returned.