Towards Selfie Drone: Spatial Localization and Navigation of drone Using Human Pose Estimation | IEEE Conference Publication | IEEE Xplore
Scheduled Maintenance: On Monday, 30 June, IEEE Xplore will undergo scheduled maintenance from 1:00-2:00 PM ET (1800-1900 UTC).
On Tuesday, 1 July, IEEE Xplore will undergo scheduled maintenance from 1:00-5:00 PM ET (1800-2200 UTC).
During these times, there may be intermittent impact on performance. We apologize for any inconvenience.

Towards Selfie Drone: Spatial Localization and Navigation of drone Using Human Pose Estimation


Abstract:

Remotely piloted drones are popular for photography and selfies are the most popular type of pictures. These drones require piloting skills and expertise in controlling m...Show More

Abstract:

Remotely piloted drones are popular for photography and selfies are the most popular type of pictures. These drones require piloting skills and expertise in controlling maneuvers for taking pictures. We create a selfie drone software solution that eliminates the requirement for manual drone maneuvering to the desired location. The RC transmitter is linked to an Android smartphone that runs the software programme. The user selects or provides a template for a selfie and the drone takes off automatically. The software locks the drone’s home position and uses human pose estimation to find the user’s pose. It computes the position vector of the drone camera for the desired selfie, flies the drone to the desired position vector, and captures the selfie. After that, the drone hovers in the same spot, either waiting for the next template or returning to its home position. After careful review of the existing options, we decided to use DJI Phantom 4 as a selfie drone that is controlled through an RC transmitter by default. We control the drone with DJI SDK and use support vector regression to generate the position vector for capturing selfie images similar to the template provided by the user. The algorithm’s performance on the mobile platform is assessed qualitatively: the regressor is tested on real-world settings, yielding an accuracy of 80%, and the images acquired by the drone resemble the desired template images.
Date of Conference: 26-27 October 2021
Date Added to IEEE Xplore: 28 December 2021
ISBN Information:
Conference Location: Rawalpindi, Pakistan

Funding Agency:


Contact IEEE to Subscribe

References

References is not available for this document.