Abstract
This research shows how Hybrid Machine Learning (HML) techniques may
be used in real-time to identify an Army’s personal fighting zone or any other specified
location in order to reduce safety risks via the detection of an invasion or enemies.
Deep Learning (DL) techniques, such as Faster R-CNN, YOLO, and DenseNet, were
used to find employees, categorize objects, and detect subtle characteristics in a variety
of datasets. Testing showed that a 95% recall rate and a 90% precision rate were
possible. This indicates high detection. A cleanness of 85 percent and a correctness of
80 percent were achieved in a real-world construction site application. To some things
up: The recommended approach may enhance current safety management methods in
conflict zones, borders, and beyond.