Drone Detection by Using Multi-Band Image

MULTIBANDID addresses security challenges posed by unmanned aerial systems (UAS) by integrating visible (RGB), ultraviolet (UV), and long-wave infrared (LWIR) imaging to overcome limitations of single-band methods. Independent YOLO deep learning models process each band simultaneously, and their outputs combine in a decision-level spatial fusion pipeline based on homography-driven coordinate alignment. A PyQt6 GUI provides synchronized visualization. UV models achieved the highest single-band accuracy, while spatial fusion effectively aggregates multimodal detections for reliable counter-UAS performance.