📌 Overview

This project detects: Helmet usage (Helmet / No Helmet) Vehicle number plates (for OCR recognition)

It uses: YOLOv8 (Ultralytics) for object detection Tesseract OCR for reading number plates OpenCV for image preprocessing and visualization

Team Members and Responsibilities

Member	Role	Main Tasks
Member 1 (Data Engineer)	Dataset Preparation	Gather images, annotate using LabelImg, denoise and preprocess dataset, create YOLO structure
Member 2 (Model Trainer)	Model Training	Train YOLOv8 on the dataset, tune hyperparameters, and save weights
Member 3 (OCR Specialist)	Number Plate Recognition	Implement and test OCR (Tesseract) to extract plate numbers from detected regions
Member 4 (Integration & Testing)	Integration + Final Testing	Integrate YOLO + OCR pipeline, run predictions, and test on real videos/images

Step 1: Install Dependencies (All Members) Each member should have Python 3.10+ installed.

Create a virtual environment python -m venv venv venv\Scripts\activate

Install required libraries pip install ultralytics opencv-python pytesseract numpy matplotlib If you face errors with Tesseract:

Download and install Tesseract from: 👉 https://github.com/UB-Mannheim/tesseract/wiki Then note its path (for example): C:\Program Files\Tesseract-OCR\tesseract.exe

Add this to your Python script later: pytesseract.pytesseract.tesseract_cmd = r"C:\Program Files\Tesseract-OCR\tesseract.exe"

tep 4: Install Required Packages pip install -r requirements.txt

Your requirements.txt should include:

ultralytics==8.3.205 opencv-python pillow numpy pandas torch tqdm pytesseract

Step 2: Member 1 — Dataset Preparation 🔹 1. Folder structure Create folders like this: HelmetNumberplate/ │ ├── data/ │ ├── images/ │ │ ├── train/ │ │ ├── val/ │ └── labels/ │ ├── train/ │ ├── val/ ├── scripts/ │ ├── noise_filter.py │ ├── preprocess.py ├── dataset.yaml └── runs/

Step 3: Member 2 — YOLOv8 Model Training 🔹 1. Create dataset.yaml train: data/images/train val: data/images/val

nc: 3 names: ['Helmet', 'NoHelmet', 'NumberPlate']

🔹 2. Train the model yolo detect train data="dataset.yaml" model="yolov8n.pt" epochs=50 imgsz=640

This will create: runs/detect/train/ └── weights/ ├── best.pt └── last.pt

🔹 3. Validate model yolo detect val model="runs/detect/train/weights/best.pt" data="dataset.yaml"

Step 4: Member 3 — Number Plate OCR 🔹 1. Import Tesseract import pytesseract import cv2

pytesseract.pytesseract.tesseract_cmd = r"C:\Program Files\Tesseract-OCR\tesseract.exe"

🔹 2. Apply OCR on cropped number plate image = cv2.imread('car.jpg') gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY) text = pytesseract.image_to_string(gray) print("Detected Text:", text)

🔹 3. Integrate OCR after YOLO detection

Use YOLO’s bounding box to crop the number plate region and pass it to pytesseract.

🧩 Step 5: Member 4 — Integration & Testing 🔹 1. Run predictions yolo detect predict model="runs/detect/train/weights/best.pt" source="data/images/val"

🔹 2. Combine YOLO + OCR

In one Python file (main.py):

from ultralytics import YOLO import cv2, pytesseract

model = YOLO("runs/detect/train/weights/best.pt") results = model.predict(source="test_images/car1.jpg", save=True)

for r in results: for box in r.boxes: x1, y1, x2, y2 = map(int, box.xyxy[0]) crop = r.orig_img[y1:y2, x1:x2] text = pytesseract.image_to_string(crop) print("Detected Number:", text)