mvp-anpr-system

Automatic Number Plate Recognition (ANPR) System

This project implements an Automatic Number Plate Recognition (ANPR) system using YOLO for object detection and Tesseract for Optical Character Recognition (OCR). The system processes video frames in real-time, detects number plates, and performs OCR to extract the text from the detected plates.

Colab Notebook

Features

Real-time object detection using YOLO.
Optical Character Recognition (OCR) using Tesseract.
Bounding box drawing for detected objects.
Integration with camera for real-time processing.

Requirements

Python 3.x
OpenCV
PyTorch
Ultralytics YOLO
Pytesseract
PIL (Python Imaging Library)
NumPy

Installation

Clone the repository:

git clone https://github.com/megamindco/mvp-anpr-system.git
cd anpr-system

Install the required dependencies:

 python -m venv anpr
 venv\Scripts\activate
 pip install -r deps.txt

Download the YOLO model weights (e.g., yolov5s.pt) and place them in the project directory.
Download Tessrect-ocr for your platform

Usage

Run the main script to start the real-time ANPR system:

python main.py

The system will open the default cam and start processing video frames. Press q to exit the application.

Code Overview

Initialize Model

def initialize_model(model_path="yolov5s.pt"):
  model = YOLO(model_path)
  return model

Preprocess Frame

def preprocess_frame(frame):
  return cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)

Perform Detection

def perform_detection(model, frame):
  results = model(frame)
  return results[0].boxes.data.cpu().numpy()

Draw Bounding Boxes

def draw_bounding_boxes(frame, detections, threshold=0.25):
  for detection in detections:
      x0, y0, x1, y1, confidence, class_id = detection
      if confidence > threshold:
          x0, y0, x1, y1 = map(int, [x0, y0, x1, y1])
          cv2.rectangle(frame, (x0, y0), (x1, y1), (0, 255, 0), 2)
          cv2.putText(frame, f"Conf: {confidence:.2f}", (x0, y0 - 10),
                      cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 1)
  return frame

Perform OCR on Box

def perform_ocr_on_box(frame, box):
  x0, y0, x1, y1 = map(int, box)
  cropped_image = frame[y0:y1, x0:x1]
  try:
      gray = cv2.cvtColor(cropped_image, cv2.COLOR_BGR2GRAY)
      gray = cv2.adaptiveThreshold(gray, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C,
                                    cv2.THRESH_BINARY, 11, 2)
      config = r'--oem 3 --psm 7 -c tessedit_char_whitelist=ABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789'
      text = pytesseract.image_to_string(gray, config=config).strip()
  except Exception as e:
      text = f"Error: {e}"
  return text

Main Real-time Processing


def main_realtime(model_path="yolov5s.pt"):
    model = initialize_model(model_path)
    cap = cv2.VideoCapture(0)  # Open default webcam

    if not cap.isOpened():
        print("Error: Unable to access the camera.")
        return

    while True:
        ret, frame = cap.read()
        if not ret:
            print("Error: Unable to read frame from camera.")
            break

        rgb_frame = preprocess_frame(frame)
        detections = perform_detection(model, rgb_frame)
        frame_with_boxes = draw_bounding_boxes(frame, detections)

        for detection in detections:
            x0, y0, x1, y1, confidence, class_id = detection
            if confidence > 0.25:  # Ensure only confident detections are processed
                number_plate_text = perform_ocr_on_box(frame, (x0, y0, x1, y1))
                cv2.putText(frame_with_boxes, f"Plate: {number_plate_text}", (int(x0), int(y0) - 30),
                            cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255), 1)

        cv2.imshow('ANPR Realtime', frame_with_boxes)

        if cv2.waitKey(1) & 0xFF == ord('q'):
            break

    cap.release()
    cv2.destroyAllWindows()

if __name__ == "__main__":
    main_realtime()

Contributing

Contributions are welcome! Please open an issue or submit a pull request.