YOLOv5 Real-Time Object Detection for CCTV

Introduction

The YOLO (You Only Look Once) series of object detection models are known for their real-time performance and accuracy. In this project, we will be using YOLOv5 to detect guns and knives in CCTV footage in real-time. This project is a proof-of-concept for using YOLOv5 in a security setting to improve public safety.

Requirements

NVIDIA GPU
CUDA and cuDNN (version 10.2)
Python 3.8 or later
OpenCV
PyTorch
YOLOv5
A dataset of CCTV footage containing guns and knives
Installation
Install CUDA and cuDNN. You can find the installation instructions on the NVIDIA website.

Clone this repository.

Copy code

 > git clone https://github.com/AlexeyAB/darknet

Install the required Python packages by running the following command in the project directory:

Copy code

 > pip install -r requirements.txt

Download the pre-trained YOLOv5 weights from the official YOLO website.

Copy code

 > https://github.com/AlexeyAB/darknet

Build the darknet library.

Copy code

 > cd darknet
 > make

Usage

• Collect CCTV footage containing guns and knives.

• Run the object detection script on the CCTV footage.

Copy code

 > python detect.py --input path/to/video.mp4 --output path/to/output.avi --weights path/to/weights.pt

The script will detect guns and knives in the footage and save the output to the specified location.

Conclusion

In this project, we have shown how YOLOv5 can be used for real-time object detection in CCTV footage to detect guns and knives. This proof-of-concept demonstrates the potential for using YOLOv5 in a security setting to improve public safety. However, it is important to note that this project is for educational purposes only and should not be used in a real-world setting without proper testing and evaluation.