This tool, built using Darknet and YOLO, aims at helping Policemen (can be extended to CCTVs too), to identify a potentially dangerous situation like a person holding a deadly weapon and is trained especially for detection of the presence of GUNS in an image. All the files required for this tool to work are present here, in this repository, except the .weights file, which got generated after rigorous training of the model and can be downloaded from here.
This tool has been trained well to work with low-resolution images, therefore images from CCTV cameras won't be an issue. The tool is being made Open-Source and hence, can be extended to work as a real-time Weapon Classifier, after making a few modifications to the source code.
- Clone the repository. Do not forget to download the .weights file.
- Open the Terminal/CMD in the same directory as the main.py file. (python main.py)
- Follow the instructions given in the program.
For the project to be fully functional, make sure that you have latest version of Python installed on your system.
Kindly note that all the paths to be provided to the program are absolute paths. DO NOT provide relative paths.
This is how we built and trained the model.
<iframe width="560" height="315" src="https://www.youtube.com/embed/KkErB0YqrEA" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe><iframe width="560" height="315" src="https://www.youtube.com/embed/RPXL25qcutw" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>
First clone this repo to your machine and then watch this video on how to make use of this tool.
Thanks and Regards
Samhita Joshi and The Team