Drogue detection and tracking method for monocular-vision-based autonomous aerial refueling

Abstract

Drogue detection for vision-based probe-drogue autonomous aerial refueling is a challenging task due to disorderly motion of drogue caused by both the tanker wake vortex and atmospheric turbulence. In this paper, a drogue detection method based on drogue inner part's shape features is proposed. Drogue detection is divided into two processes: rough detection and fine location of drogue. The drogue inner part's circular feature is introduced into the drogue rough detection model. Row-column scanning method is proposed to the fine location of drogue. The ellipse least-square fitting is used to obtain the drogue contours' center. Based on drogue image size and drogue image position, we can obtain the tracking region of interest (ROI) area. We proposed a switch rule between drogue detection and drogue tracking. We have experiments on three real autonomous aerial refueling videos. The results show that the proposed method is effective.