sreerevanth/sentinelcv

SentinelCV is a computer vision tool that turns ordinary cameras such as CCTV, IP cameras, webcams, or even recorded video files into an AI based safety monitor. Its first and currently only working feature is real time fall detection, meaning it watches a video feed and tries to notice when a person falls down. Under the hood, it captures frames from whatever camera source is connected, runs them through a YOLOv8 object detection model to spot people and analyze their posture, and then decides if a fall has happened. When it detects one, it immediately sends an alert through Telegram, so someone gets notified right away rather than needing to watch the video feed themselves. The README describes fall detection as just the first module of a larger planned platform, with a roadmap that lists many future features such as fire and smoke detection, detecting protective equipment, violence and weapon detection, and a web dashboard, though none of these are built yet. A trained model, a requirements file, and a cleaner project layout are also listed as still in progress. Getting started involves cloning the repository, installing a handful of Python libraries by hand including OpenCV and Ultralytics, setting up a Telegram bot for alerts, and running one of the included Python scripts directly. The project is offered under the MIT license, which allows free use, including commercial use, as long as the license is kept with the code. This project fits people interested in safety monitoring for places like elderly care homes, hospitals, warehouses, or construction sites, and developers comfortable experimenting with an early stage, still evolving codebase rather than a polished product.