Accompanying dataset for the FSR 2015 submission:

Philipp Oettershagen, Thomas J. Stastny, Thomas A. Mantel, Amir S. Melzer, Konrad Rudin, Gabriel Agamennoni, Kostas Alexis, Roland Y. Siegwart; Long-Endurance Sensing and Mapping using a Hand-Launchable Solar-Powered UAV

Provided data

• Recorded data in ROS bag file (Version 2.0) for:
  • 1h flight segment
  • Lawnmower pattern
  • Rectangular spiral
• Camera calibration for grayscale camera as text file
• Derived data:
  • Dense point cloud in PLY format
  • Mosaic as GeoTIFF
  • Digital Surface Map
  • Flightpath in KML format

Topics in ROS bag

• Oblique-view grayscale camera images (/cam0)
• Nadir-looking RGB camera images (/cam1)
• Oblique-view (as grayscale) IR camera images ⁽¹⁾ (/ircam)
• Raw GPS measurements (/px4/raw/gps)
• Raw IMU measurements (/px4/raw/imu)
• Raw Magnetometer measurements (/px4/raw/mag)
• Position in world coordinates ⁽²⁾ (/px4/position_world)
• Velocity in world coordinates ⁽²⁾ (/px4/velocity_world)
• UAV orientation (UAV body frame) ⁽²⁾ (/px4/orientation)
• UAV angular rates (UAV body frame) ⁽²⁾ (/px4/angular_rates)

Employed sensors

• Aptina MT9V034 with f=2.8mm lens
• IDS ueyeXS2
• FLIR Tau 2 with f=9mm
• Analog Devices ADIS16448 IMU
• ublox LEA-6H GPS receiver

⁽¹⁾ IR camera images are stored in 14 bit/pixel. ROS tool for false coloring is available on github.com.
⁽²⁾ Output of UAV state-estimator

Flight data

Camera calibration

• Grayscale camera intrinsic calibration

Derived data

• Dense Point Cloud PLY 1 PLY 2 PLY 3
• Mosaic reconstruction and Google tiles ZIP
• Digital Surface Map and tilesZIP

All topics in the namespace px4 are streamed from the UAV autopilot via an UART link and are time stamped on arrival on the embedded computer.

The /px4/orientation topic gives the (estimated) orientation of the UAV in quaternions in ENU coordinates (using the ROS quaternion function tf::createQuaternionFromRPY(roll, -pitch, -yaw)). Hence, the /px4/angular_rates topic stores [roll_rate, -pitch_rate, -yaw_rate].

The (estimated) position in world coordinates (/px4/position_world) is given as a vector of the form [latitude, longitude, altitude]. The same yields for the velocity (/px4/velocity_world), which is the ground speed of the UAV given as a vector of the form [latitudinal speed, longitudinal speed, vertical speed].