Visual-Based Real-Time Detection Using Neural Networks and Micro-UAVs for Military Operations

This article presents a vision-based detection system for a micro-UAV, which has been implemented in parallel to an autonomous GPS-based mission. The research seeks to determine a value objective for decision-making within military reconnaissance operations. YOLO-based algorithms have been used in real-time, providing detection of people and vehicles while fulfilling an automated navigation mission. The project was implemented in the CICTE Military Applications Research Center, as part of an automatic takeoff, navigation, detection, and landing system. The detection based on YOLO V3 offers efficient results from the analysis of sensitivity and specificity in the detection in real-time, in external environments during autonomous navigation and while the recognition of the objective is carried out keeping the UAV in stationary mode, with different angles of the camera.