merge nadir image with 45° image.

Hi, I want to run a 3D mapping of a small brick bridge located above a stream of water , using my UAV (APM 2.6 and grounstation Mission Planner)
For nadir image I’ll use Canon S100+CHDK (altitude for example 50 mt) and, to detect the best profile of the bridge and improve more accuracy reconstruction , I would like to take pictures to 45° and horizontal also (always with UAV)
My question:
1°: - Can I use, for 45° and horizontal images, different camera on my UAV (for example GoPro hero4 with Gimbal) or for best accuracy reconstruction should I use same “nadir” camera (Camera S100)
2°: - 45° and horizontal images from UAV, must be georeferenced also (like nadir image) ?

Best Regards - Giuseppe - Italy

1 Like

Hello Giuseppe,

It is possible to create projects using different datasets ( aerial nadir / aerial Oblique / terrestrial) using the same or different cameras.

In our Chillon Castle project, we used several drones, cameras, and used aerial nadir, aerial oblique and terrestrial (indoor and outdoor) images to create one single project: https://support.pix4d.com/hc/en-us/articles/202557199

It is suggested to:

Take the nadir dataset with 85% frontal overlap and at least 70% side overlap:
https://support.pix4d.com/hc/en-us/articles/202557459#label1

Take the oblique dataset with at least 85 % overlap and if possible make two rounds around the object: https://support.pix4d.com/hc/en-us/articles/202557459#label4

Fly with same lightening conditions in all the cases.

Fly when there is a little bit of clouds if possible, so the lightening is homogeneous and one facade will not be very bright and another dark.

Use GCPS if possible and in any case, with or without GCPs, use manual tie points:
https://support.pix4d.com/hc/en-us/articles/202560239
https://support.pix4d.com/hc/en-us/articles/202560619

So the ideal is create 2 sub projects, one with the aerial images, another with the oblique images, add GCPS and add manual tie points (using the same manual tie points):

  1. Create sub project A (Aerial images)
  2. Create 10 -15 Manual tie points, for example, pp1, pp2, pp3, pp4, pp5…
  3. Mark each Manual tie point in at least 3 images.
  4. Process step 1.
  5. Create sub project B (oblique images)
  6. Create the same 10 -15 Manual tie points created in the sub project A and give them the same names than in the sub project A (pp1, pp2, pp3, pp4, pp5…)
  7. Mark each Manual tie point in at least 3 images.
  8. Process step 1.
  9. Merge the projects by creating a new project: the software will ask if the Manual tie points with the same names are the same tie points, click YES and once is processed.
  10. Add the GCPs, mark them and reoptimize.
  11. Process the steps 2 and 3 if desired.

For more detailed step by step instructions (it is the same procedure than when processing large datasets):
https://support.pix4d.com/hc/en-us/articles/entries/27572236

Regards,

Thanks, very good explanation, I’m studying
all your webinars, and I’m beginning to see the first good results.

Best Regards - Giuseppe - Italy

use good overlap and gcps, add manual tie points and all should be fine :slight_smile:

Might this workflow also apply to vegetated or forested areas? I have been having difficulty processing a 250 acre property site that is approx 50% covered by waist high grass and jungle canopy. Been flying 300’ in nadir with Phantom Pro 4 v.2.0 but approx 15% of images wont calibrate and are lacking automatic tie points. Thanks for your help, Marc.

Yes, you can apply to vegetated or forested areas as well. I also recommend checking the link below to improve the results.

Kind regards,