Unmanned aerial vehicle visual guidance air refueling method based on deep learning

Abstract

The invention discloses an unmanned aerial vehicle visual guidance air refueling method based on deep learning, and belongs to the technical field of navigation positioning and control. The method comprises the following steps: 1, identifying and tracking a moving target according to the relative background movement of the target by adopting an inter-frame difference method; 2, adopting a deep learning target detection Faster RCNN method, and identifying and tracking a moving target according to the relative background movement of the target; And 3, driving the holder to track the target so that the target is always in the center of the image. According to the method, the unmanned aerial vehicle is tracked in a mode of combining the inter-frame difference and the improved target detectionalgorithm Faster RCNN, an original algorithm framework is improved, the detection precision and the test speed are improved, and rapid detection and tracking of an oiling machine and a taper sleeve inthe air refueling process are achieved.

Description

technical field [0001] The invention discloses an aerial refueling method for unmanned aerial vehicles based on deep learning and vision guidance, and belongs to the technical field of navigation, positioning and control. Background technique [0002] Aerial refueling technology is an important means to increase aircraft combat radius, increase ammunition capacity, and solve the contradiction between take-off weight and flight performance, and has always been highly valued by aviation developed countries. Since aerial refueling was first proposed in the 1950s, the artificially operated aerial refueling technology has been relatively mature, but the artificially operated aerial refueling technology has low efficiency, high requirements on the pilot's driving skills, and is easily affected by the driver's psychology, physiology and Influenced by the technical and tactical status, it is urgent to carry out research on automatic aerial refueling technology. [0003] Aerial refu...

AI Technical Summary

Problems solved by technology

Sensors such as laser, radar, and GPS have been used in autonomous aerial refueling systems, but there are more or less defects, and a single sensor is difficult to meet the requirements of the aerial refueling docking section

For example, the GPS signal cannot cover all corners of the world, and when the tanker and the tanker dock, the relative position of the drogue and the tanker is not fixed, the droop height of the drogue is related to the flight speed of the tanker, and the refueling drogue Both the refueling hose and the refueling hose are disturbed by the airflow, and the GPS antenna cannot be installed on the real refueling drogue. In addition, the bandwidth of the GPS system may not meet the frequency requirements under high-speed docking; laser and radar are easily interfered by the external environment, and the obtained Insufficient signal reliability

Images