MMW-YOLO for Robust Personnel Detection in Obscured Environments Using Passive Millimeter-Wave Imaging

Passive millimeter-wave (mmW) imaging offers significant potential benefits across a wide range of applications, such as weather and climate monitoring, search and rescue, surveillance, and contraband detection. The ability of mmW radiation to penetrate materials allows for the detection of personnel in visually obscured environments where conventional imaging techniques are ineffective. This work presents a novel method for personnel detection with the YOLOv8 algorithm, utilizing transfer learning with augmented passive mmW data. The detection model is evaluated under two conditions: scenarios with available visual imagery and those without. Performance analysis of both cases is presented, demonstrating the effectiveness and robustness of the proposed mmW-YOLO model.