Method and system for tracking and landing of unmanned aerial vehicles

Abstract

The present invention proposes a UAV tracking and landing method and system, including a long-distance tracking method and a short-distance tracking method; yolov4-tiny target recognition is carried out on the captured video frames, and tracking is performed from a long distance to the landing platform through an ECO correlation filter , detect the Apriltag code, and perform a close-range tracking method, the Apriltag code is the nested code of the Apriltag center, the nested code of the Apriltag center is detected globally and locally, the nested code center of the Apriltag center is calculated, and the unmanned aerial vehicle and the target landing platform are estimated The position difference dx, dy and dz; the height H of the UAV and the target landing platform is measured by laser, and the ratio of dx and dy is corrected based on H, and the corrected data is used to control the motion of the UAV. The present invention Utilize the yolov4-tiny target recognition algorithm and the mutual switching of the RCO correlation filter tracking to realize the relocation of the target, improve the efficiency of the tracking detection algorithm, and accurately locate the target; the Apriltag center nesting code improves the utilization of spatial coding rate, which is conducive to the identification of UAVs.

Description

technical field [0001] The invention relates to the technical field of UAV landing, in particular to a UAV tracking and landing method and system. Background technique [0002] Current tracking algorithms are mainly based on deep neural networks or correlation filtering, or a combination of both. The tracking algorithm based on the deep neural network has high precision, but the running speed is slow; while the tracking algorithm based on the correlation filter is fast, but the precision is slightly lower, and when the target leaves the field of view or is blocked, the tracking target is prone to loss. As for the tracking algorithm that combines the two, it generally uses convolutional neural networks to extract features to replace manually selected features, and then perform correlation calculations to achieve tracking. Although there is a better balance between speed and accuracy, due to The calculation time of convolutional neural network is long, and it is still difficu...

AI Technical Summary

Problems solved by technology

[0005] In view of the above defects, the purpose of the present invention is to propose a UAV tracking and landing method and system to solve the problems of poor control effect of UAV landing, lack of recalibration of tracking, and poor accuracy

Images