🔫 Weapon Detection System using YOLOv8

An AI-powered real-time weapon detection system that identifies Guns and Knives from images, videos, and live CCTV feeds using YOLOv8.

---

📌 Overview

Manual CCTV monitoring is error-prone and slow. This system automates the detection of weapons to improve surveillance response time and public safety using deep learning.

---

✨ Features

• Detects Gun and Knife with bounding boxes & confidence scores
• Real-time webcam / CCTV feed detection
• YOLOv8 lightweight architecture for faster inference
• Scalable for additional weapon classes

---

👥 Team Details

Team Name: AI Defenders

Name
Vineet Unde
Shraddha Bhadane

---

🚨 Problem Statement

Traditional surveillance depends on human operators who may miss critical events due to fatigue or distractions, causing delayed reactions to violent incidents.

---

💡 Proposed Solution

A YOLOv8-based automated detection system trained on thousands of weapon images to identify guns and knives instantly in real-time.

---

🏗️ System Architecture

Input (Image / Video / CCTV) │ ▼ Preprocessing (Resize, Normalize, Augmentation) │ ▼ YOLOv8 Model │ ▼ Bounding Boxes + Confidence Scores │ ▼ Alert System (Future Scope)

---

🛠️ Tech Stack

Component	Tools
Language	Python 3.12
Framework	PyTorch
Model	YOLOv8 (Ultralytics)
Libraries	OpenCV, NumPy, Pandas, Matplotlib
Tools	Jupyter Notebook, VS Code, GitHub

---

📂 Dataset

Category	Details
Total Images	4000–5000
Classes	Gun, Knife
Split	80% Train / 20% Validation
Format	YOLO TXT

---

📁 Dataset Structure

Weapon_Detection_System/ └── dataset/ ├── train/ │ ├── images/ │ └── labels/ ├── val/ │ ├── images/ │ └── labels/ └── test/

📑 data.yaml

path: dataset
train: train/images
val: val/images
test: test/images
nc: 2
names: ["Gun", "Knife"]
⚙️ Installation
```bash
Copy code
git clone https://github.com/VineetC137/Weapon-Detection_System.git
cd Weapon-Detection_System
python -m venv yolov8env
yolov8env\Scripts\activate
pip install ultralytics opencv-python matplotlib numpy pandas
🏋️ Training
```bash
python
Copy code
from ultralytics import YOLO

model = YOLO("yolov8n.pt")

model.train(
    data="dataset/data.yaml",
    epochs=30,
    imgsz=640,
    batch=16,
    name="weapon_detection"
)
-----

🔍 Inference
```bash
Run on Image
python
Copy code
model.predict("test.jpg", show=True)
Run on Webcam
python
Copy code
model.predict(source=0, show=True)
-------
📊 Results
Metric	Value
mAP50	0.75 – 0.85
Precision	High
Limitation	False positives on cluttered backgrounds
-------
🚀 Future Enhancements
Add detection for more weapon classes

Integrate real-time alert notifications

Deploy as a Flask/Django web application

Optimize model for edge devices
-------
📜 License
This project is for educational and hackathon use only.
--------
🙌 Acknowledgements
Ultralytics YOLOv8

OpenCV Community

Hackathon Organizers