Workshop Organizer

Ronny Hänsch
Technische Universität Berlin, Germany
r.haensch@tu-berlin.de
www.rhaensch.de

teaser

 

Why Participate?

The full-day workshop UAV-3D² is meant as a complementary forum to the UAV-g 2017 main conference where participants can learn more about the hottest topics in UAV-based photogrammetry as well as related fields of research. It aims at providing a forum for collaboration between the computer vision, photogrammetry, and geomatics communities to discuss modern challenges and ideas, propose new and contemporary benchmarks, and elaborate on the overlap to machine learning, mathematics, and other related areas.

UAV-3D² 2017 welcomes presentations of original research in computer vision and photogrammetry in the context of UAVs and invites submissions of high-quality research results as full papers. Topics include, but are not limited to:

  • Feature extraction, matching, and sensor orientation
  • Image-based navigation and position/orientation determination in outdoor and indoor environments
  • Vision-aided computation of velocity and attitude components of a moving platform in the 3D space
  • Vision-based environmental perception and obstacle avoidance
  • Structure from motion and SLAM
  • Multi-view stereo reconstruction
  • 3D point cloud processing, segmentation, and classification

Workshop Description

By definition, UAVs and the 3D world are strongly connected since the UAV needs to navigate through the three-dimensional space. Already this very basic fact states unique challenges in 3D mapping, path planning and estimation, as well as obstacle avoidance. Furthermore, since UAVs potentially observe the complete 3D scene, they are able to capture images of the object of interest from very different perspectives and offer unique advantages over earth-bound vehicles which perceive the object only from the ground. Consequently, the usage of UAVs as data acquisition platform has become very attractive for many applications.

The acquired image data are often used for the automatic generation of accurate three-dimensional models of the scene that have a wide range of applications including navigation, documentation, computer-graphics, virtual reality, and visual inspection. Especially the three-dimensional reconstruction of buildings is important due to its potential for 3D city models in city planning, damage assessment, monument conservation, architecture, and digital tourism.

While there are several ways to obtain 3D information of the scene and/or for and about the UAV, this workshop focuses on image-based solutions and in particular on the relationship of UAVs and photogrammetry. On the one hand, these include image-based navigation and orientation using external cameras observing the UAV as well as cameras attached to the UAV in outdoor and indoor environments, computation of velocity and attitude components of a moving platform in the 3D space, vision-aided navigation and control, vision-based environmental perception and obstacle avoidance. On the other hand methods are of interest, that use the acquired image data for a subsequent and usually offline image-based 3D reconstruction of the scene. Applications include mapping, damage assessment, cultural heritage, and digital tourism.

Important Dates

24.07. – Full Paper Submission
07.08. – Author Notification
21.08. – Camera Ready Submission
04.09. – UAV -3D² Workshop

Paper Submission

Paper submission opens soon.

Preliminary Program

09:15 – 09:30 Welcome
09:30 – 10:30 Keynote
Morning coffee break
11:00 – 12:30 Orals Session 1
Lunch break
14:00 – 15:30 Orals Session 2
Afternoon coffee break
16:00 – 17:30 Orals Session 3
17:30 – 18:00 Closing Remarks