Pix4UAV showcases

We present 5 datasets captured by different UAVs and processed fully automatically by Pix4UAV.

Showcase 1: Urban area captured by a SwingletCam of senseFly



This dataset was captured by a SwingletCam by senseFly and consists of 73 images, covering 0.13 km2 with an average ground sampling distance of 4cm / pixel. Twelve ground control points have been marked to increase the geolocation.

Interactive map (fullscreen)

Download

Showcase 2: Easter Island captured by the X100 of Gatewing



This dataset was captured by the X100 by Gatewing and consists of 140 images, covering 3.7 km2 with an average ground sampling distance of 30cm / pixel.

Interactive map (fullscreen)

Downloads

Showcase 3: Field captured by Mavinci



This dataset was captured by the Sirius by Mavinci and consists of 525 images, covering 0.7 km2 with an average ground sampling distance of 3.3cm / pixel.

Interactive map (fullscreen)

Downloads

Showcase 4: Mining map captured by the X100 of Gatewing



This dataset was captured by the X100 by Gatewing and consists of 602 images, covering 1.18 km2 with an average ground sampling distance of 4.6cm / pixel.

Interactive map (fullscreen)

Downloads

Showcase 5: Pteryx



This dataset was captured by Pteryx and consists of 741 images, covering 4.58 km2 with an average ground sampling distance of 11.1cm / pixel.

Interactive map (fullscreen)

Downloads

Summary

Images Area [km2] Resolution [cm/pixel] Cloud Time [min] GCP accuracy [m]
Showcase 1 73 0.13 4 40 0.028
Showcase 2 140 3.7 30 57 N/A
Showcase 3 525 0.7 3.3 302 N/A
Showcase 4 602 1.18 4.6 256 0.029
Showcase 5 741 4.58 11.1 612 N/A

Technical Demo Videos

These videos show the power of our technology and its workflow.

Workflow of Pix4D reconstruction pipeline on the EPFL site, using a SwingletCam from senseFly.

The city of Lausanne, Switzerland, 400 images captured by a SwingletCam from senseFly. The 3D model was built from the Digital Elevation Model, using the orthomosaic as texture.

Applying Pix4D technology to ground imagery: 2000 images from a Canon D5.