Application for threat surveillance using Python and Yolo-V7

This paper describes the functionality and the results of the Threat Surveillance mini-project on computer vision. Its main objective is to create and offer the possibility of monitoring and detecting cold and fire weapons, while maintaining public safety. Created not only for military purposes but also for civilian ones, the project proposes a simple graphical user interface (GUI) which is aimed at maximum user convenience. The core of the functionality is however based on advanced commercial software belonging to artificial intelligence (AI), called Yolo - V7. It is an algorithm of real-time object detection. The GUI allows personalizing the application execution by choosing the targeted weapon category along with loading a photograph or video which contains the category to be detected/recognized. The weapon detection part provides possible bounding boxes where the desired weapons might be situated in the chosen media and classifies them according to the object classes previously created by training.