DronePose: Photorealistic UAV-Assistant Dataset Synthesis for 3D Pose Estimation via a Smooth Silhouette Loss

Albanis, Georgios; Zioulis, Nikolaos; Dimou, Anastasios; Zarpalas, Dimitrios; Daras, Petros

doi:10.1007/978-3-030-66096-3_44

Cited by 5 publications

(5 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Computer Vision approaches have been used in AR applications for object detection and tracking [18,19], object pose estimation [20], localization [21], and even gaze-based UAV navigation [22]. In our application, we utilize computer vision techniques for visual pose drone detection [23].…”

Section: Computer Vision For Armentioning

confidence: 99%

“…To that end, we propose a method that exploits the latest advances in deep learning to automatically retrieve the 6DoF pose of a drone from a single image. An early version of this architecture, described in more detail in [23], uses a CNN encoder backbone followed by three fully connected layers that output two predictions, one for the translation components and one for the rotation. The translation prediction uses an L 2 loss, while the rotation prediction aims to minimize the angular difference L R .…”

Section: Visual Drone Pose Estimationmentioning

confidence: 99%

“…For training our models and evaluating their performance, we have compiled and made publicly available the UAVA dataset (also in [23]), an extensive set of synthetic, photorealistic images of drones in indoor and outdoor environments, which provides ground truth annotations of 2D keypoints, pose, depth maps, normal maps, and more.…”

Section: Visual Drone Pose Estimationmentioning

confidence: 99%

See 2 more Smart Citations

Drone Control in AR: An Intuitive System for Single-Handed Gesture Control, Drone Tracking, and Contextualized Camera Feed Visualization in Augmented Reality

Konstantoudakis

Christaki

Tsiakmakis

et al. 2022

Drones

Self Cite

View full text Add to dashboard Cite

Traditional drone handheld remote controllers, although well-established and widely used, are not a particularly intuitive control method. At the same time, drone pilots normally watch the drone video feed on a smartphone or another small screen attached to the remote. This forces them to constantly shift their visual focus from the drone to the screen and vice-versa. This can be an eye-and-mind-tiring and stressful experience, as the eyes constantly change focus and the mind struggles to merge two different points of view. This paper presents a solution based on Microsoft’s HoloLens 2 headset that leverages augmented reality and gesture recognition to make drone piloting easier, more comfortable, and more intuitive. It describes a system for single-handed gesture control that can achieve all maneuvers possible with a traditional remote, including complex motions; a method for tracking a real drone in AR to improve flying beyond line of sight or at distances where the physical drone is hard to see; and the option to display the drone’s live video feed in AR, either in first-person-view mode or in context with the environment.

show abstract

Section: Computer Vision For Armentioning

confidence: 99%

Section: Visual Drone Pose Estimationmentioning

confidence: 99%

Section: Visual Drone Pose Estimationmentioning

confidence: 99%

See 1 more Smart Citation

Drone Control in AR: An Intuitive System for Single-Handed Gesture Control, Drone Tracking, and Contextualized Camera Feed Visualization in Augmented Reality

Konstantoudakis

Christaki

Tsiakmakis

et al. 2022

Drones

Self Cite

View full text Add to dashboard Cite

show abstract

“…Pedro et al [18] use the Unreal Engine 4 simulator to generate labelled images of spacecraft orbiting the Earth and provide 3D pose information that can be used to train and evaluate neural networks for the pose estimation task. Georgios et al [15] designed a data synthesis pipeline to create a multimodal dataset that provides vehicle pose labels for quadrotor UAV-assisted localisation studies. However, none of these datasets involve fixed-wing vehicles.…”

Section: Related Studiesmentioning

confidence: 99%

“…Georgios et al. [15] designed a data synthesis pipeline to create a multi‐modal dataset that provides vehicle pose labels for quadrotor UAV‐assisted localisation studies. However, none of these datasets involve fixed‐wing vehicles.…”

Section: Related Studiesmentioning

confidence: 99%

Air2Land: A deep learning dataset for unmanned aerial vehicle autolanding from air to land

Zheng

2022

IET Cyber-Syst and Robotics

View full text Add to dashboard Cite

In this paper, a novel deep learning dataset, called Air2Land, is presented for advancing the state-of-the-art object detection and pose estimation in the context of one fixed-wing unmanned aerial vehicle autolanding scenarios. It bridges vision and control for groundbased vision guidance systems having the multi-modal data obtained by diverse sensors and pushes forward the development of computer vision and autopilot algorithms targeted at visually assisted landing of one fixed-wing vehicle. The dataset is composed of sequential stereo images and synchronised sensor data, in terms of the flying vehicle pose and Pan-Tilt Unit angles, simulated in various climate conditions and landing scenarios. Since real-world automated landing data is very limited, the proposed dataset provides the necessary foundation for vision-based tasks such as flying vehicle detection, key point localisation, pose estimation etc. Hereafter, in addition to providing plentiful and scenerich data, the developed dataset covers high-risk scenarios that are hardly accessible in reality. The dataset is also open and available at https://github.com/micros-uav/micros_ air2land as well. The cover image is based on the Research Article

show abstract

A Survey of Cooperative Path Planning for Multiple UAVs

Sun¹,

Hao²

2022

Lecture Notes in Electrical Engineering

View full text Add to dashboard Cite

DronePose: Photorealistic UAV-Assistant Dataset Synthesis for 3D Pose Estimation via a Smooth Silhouette Loss

Cited by 5 publications

References 55 publications

Drone Control in AR: An Intuitive System for Single-Handed Gesture Control, Drone Tracking, and Contextualized Camera Feed Visualization in Augmented Reality

Drone Control in AR: An Intuitive System for Single-Handed Gesture Control, Drone Tracking, and Contextualized Camera Feed Visualization in Augmented Reality

Air2Land: A deep learning dataset for unmanned aerial vehicle autolanding from air to land

A Survey of Cooperative Path Planning for Multiple UAVs

Contact Info

Product

Resources

About