Weapon Detection using YOLOv3

This Python script is designed to detect weapons in images using a pre-trained YOLO (You Only Look Once) object detection model.

Table of Contents

• How to Use
• How It Works
• Steps
• Requirements
• Installation
  • Prerequisites
  • Steps
• Result

How to Use

Input:
Provide the path to an image containing objects, including possibly a weapon.

Output:
If the script detects a weapon in the image with confidence above a certain threshold, it will draw a bounding box around the detected weapon.

How It Works

• Loading YOLO Model:
  The script loads a pre-trained YOLO model, which has learned to recognize various objects including weapons.

• Setting Confidence Threshold:
  It sets a minimum confidence level for detecting objects. Only objects with confidence above this threshold will be considered valid detections.

• Processing Image:
  Load the specified image using the provided path.

• Detecting Weapons:
  The script analyzes the image using the YOLO model to identify objects, focusing on detecting weapons based on learned patterns.

• Drawing Bounding Box:
  If a weapon is detected with sufficient confidence, the script draws a box around the detected weapon on the image.

Steps

1. Run the script.
2. When prompted, enter the path to the image you want to analyze.
3. The script will process the image and display it with a bounding box around any detected weapons.

Requirements

Make sure to have the necessary YOLO model files (yolov3.weights, yolov3_t.cfg) and class labels (dataset info.txt) in the specified paths for the script to work correctly.

You can download the required YOLOv3 weights file from the following Kaggle dataset:

• Kaggle Weapon Detection Dataset

Installation

Prerequisites

• Python 3.x
• OpenCV
• NumPy

Steps

1. Clone the Repository

   git clone https://github.com/hija-happy/Weapon-Detection-Yolo.git
   cd Weapon-Detection-Yolo

2. Install Required Packages

   pip install opencv-python numpy

3. Download YOLO Weights

   Download the YOLOv3 weights file from Kaggle and place it in the appropriate directory.

Result