3. PROBLEM AND SOLUTION
Systematic Bridge Inspection with Drones, EuRoC Challenge 3
I4MS 2016: FOSTERING DIGITAL INDUSTRIAL INNOVATION IN EUROPE · 23rd and 24th of June – Amsterdam
Main Partners:
Computer Vision Group, Technical University of Munich, Germany
Vision-Based Localization and Mapping
Vision-Based Drone Position Control
Schällibaum AG, Switzerland
Family business since 1964
Civil Engineering, Architecture and Geomatics
80 employees
4. PROBLEM AND SOLUTION
Systematic Bridge Inspection with Drones
I4MS 2016: FOSTERING DIGITAL INDUSTRIAL INNOVATION IN EUROPE · 23rd and 24th of June – Amsterdam
Current Status
Visual Inspection
Climbing Teams / Access
Vehicles
Dangerous Work
Lack of Accessibility
Time and Cost Intensive
Referencing of a damage is not
given
5. PROBLEM AND SOLUTION
Systematic Bridge Inspection with Drones
I4MS 2016: FOSTERING DIGITAL INDUSTRIAL INNOVATION IN EUROPE · 23rd and 24th of June – Amsterdam
Current Drone Technology
No secure positioning without GPS
No collision avoidance
Difficulties for the pilots to estimate the
position of the UAV
External conditions can be harsh
(wind, temperature, dark areas,
narrow spaces, vegetation as
obstruction)
Referencing of a damage is not given
Data Analysis is not state of the art
6. PROBLEM AND SOLUTION
Systematic Bridge Inspection with Drones
I4MS 2016: FOSTERING DIGITAL INDUSTRIAL INNOVATION IN EUROPE · 23rd and 24th of June – Amsterdam
Goals of the project:
To plan the inspection flight
To fly systematically around the entire
structure
To support the ground staff with real-
time data
3D acquisition of the structure or the
structure damage
To store the collected data in an
information system
To evaluate systematically the
collected data
To increase the repeatability of the
inspection
7. EuRoC Challenge 3 Simulation Stage
(Completed):
Publically available datasets to evaluate
and develop core components of the
system (Vision based position tracking,
3D reconstruction)
Drone simulator for faster development
EuRoC Challenge 3 Lab Experiment Stage
(Ongoing):
Hardware Platform
Tools for Performance evaluation
HOW I4MS HELPED
Systematic Bridge Inspection with Drones
I4MS 2016: FOSTERING DIGITAL INDUSTRIAL INNOVATION IN EUROPE · 23rd and 24th of June – Amsterdam
FP7-2013-NMP-ICT-FoF Call for Challengers
T_firefly: Pose of the helicopter, i.e. origin of the helicopter frame, expre
coordinates. In ROS tf terminology, “world” is the parent frame and “firefly
frame. All given waypoints are expressed for this pose.
T_sensor1: Pose of sensor1 expressed in the world frame. This is what sensor1
low rate and is corrupted by noise and delay. In ROS tf terminology, world
frame and sensor1 is the child frame.
T_sensor2: Like sensor1, but mounted at a different location / orientation (not
image).
3.5.3 ROS Interface:
Published topics:
/firefly/imu (sensor_msgs/Imu)
Measurements from the IMU, located at the base_link (approx. center of g
helicopter. The measurements are corrupted by additional noise, constant
slowly drifting bias, but not delayed. The attitude in the IMU message is only es
8. Schällibaum AG:
Expected Reduction of Costs and Time:
From 4-5 days now to 2 days
From €14k to €7-9k
Improvements to safety of bridge inspection
Overall:
Hundreds of medium to large bridges in
each country must be inspected
Considerable market potential (15000
bridges only in Germany and
Switzerland require inspection)
This technology can also be applied to
inspecting other types of objects
IMPACT
Impact
I4MS 2016: FOSTERING DIGITAL INDUSTRIAL INNOVATION IN EUROPE · 23rd and 24th of June – Amsterdam