Skip to content

An OpenCV application that uses YOLOv3 and YOLOv3-Tiny object detection and weights trained on a custom dataset to detect firearms in a given image, video and in real-time.

License

Notifications You must be signed in to change notification settings

parikshitkumar1/Optimized-Firearm-Detection

Repository files navigation


Firearm_Detector_v2

Motivation

To detect firearms and enhance capabilities of cctv based security systems, in images, videos, and in real time.

Requirements

Python 3.8 or above with all requirements dependencies installed. To install run:

$ pip3 install -r requirements.txt

Downloads weights from here and put them in the root of this directory:

rename the weights to "yolov3.weights"

To run for images

$ python3 images.py

To run for videos

$ python3 video.py

To run in real-time

$ python3 realtime.py

Image result:

Video result:

Real-time result:

(used phone as a webcam and made it detect firearms in random images on the internet and youtube videos)

Might Do

  • Try out Yolo v4 and v5
  • Add training notebook to the repository

About

An OpenCV application that uses YOLOv3 and YOLOv3-Tiny object detection and weights trained on a custom dataset to detect firearms in a given image, video and in real-time.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages