Adaptive Lightweight YOLO v5-Based Network for Military Target Detection and Classification under Low-Visibility Conditions

The Weapon and Militant Detection Using Convolutional Neural Networks (WMD-CNN), a model designed to enhance automated surveillance through deep learning. The developed WMD-CNN method utilizes two datasets: one with prelabelled images and another manually labelled dataset to improve robustness. To ensure precise and efficient identification of weapons and militant activity, Convolutional Neural Networks (CNNs) are utilized for the extraction and categorization of visual features. The model’s precision and speed are balanced in its design to ensure its applicability in practical settings. Furthermore, WMD-CNN seeks to assess the accurate operation of the detection system and forecast crime using an image dataset. In order to increase dependability in practical applications, the suggested algorithm is made to notify human operators when a weapon or militant is identified in an image while reducing false alarms. Future work focuses on deploying the system in live applications, enhancing detection accuracy, and further reducing crime through advanced real-time monitoring.