Hot Most Recent
This video is adapted from 10.3390/electronics10161931
The video contains the three experimental conditions of hovering, tracking, and traversal of the UAV mentioned in the video, and provides a visual display of data processing from the first perspective of the aircraft in the form of picture-in-picture.
For a small quadrotor UAV equipped with a lightweight monocular sensor, a Single-frame Parallel-features Positioning Method (SPPM) is proposed and verified for a real-time dynamic target tracking problem. The solution is featured with systematic modeling of the geometric characteristics of moving targets, and the introduction of numeric iteration algorithms to estimate the geometric center of moving targets. Experiments show that the root mean square error percentage of static target tracking is less than 1.03% and the root mean square error of dynamic target tracking is less than 7.92cm.
The main technical contributions of this video are summarized as follows:
1) Aiming at the problem of tracking dynamic targets, a Single-frame Parallel-features Positioning Method (SPPM) is proposed. Compared with a standard solution to the perspective 3 points problem of moving targets, authors method extracts parallel constraint relations between target feature points to construct a high-order nonlinear over-determined equation with unknown depth values.
2) Based on the SPPM, we designed a 2D feature recognition algorithm for parallel feature extraction. Then we introduced a monocular SLAM algorithm based on PTAM for navigation. Finally, we constructed an indoor UAV visual positioning and tracking framework that integrates target feature recognition, monocular SLAM positioning, and dynamic target tracking.
3) To verify the effectiveness and robustness of the framework, we have carried out several indoor flight test experiments for a small UAV AR. DRONE2.0 to track, and the UAV is equipped with a lightweight monocular camera as the visual front-end. Their method is systematically evaluated by considering the computation amount, the convergence speed of the depth value, and the tracking accuracy.