Poor results with one particular mission (DJI RTK)

This project was flown with a DJI Phantom 4 RTK with the D-RTK 2 base station and a surveyor provided us with the GCPs.

I have flown 12 different missions to cover a large project area (735 ac.) and am having issues with one particular area. I’ve flown the mission 3 different times with the latest one using maximum overlap and sidelap (first two missions generated about 1150 photos and the last one was 3377). I also increased the GCPs from 13 to 24 in order to improve my results.

Here are the results from the first flight:
PCR Mission 4_report.pdf (319.8 KB)
Here is the second flight:
Mission 4 Reflight_report.pdf (389.4 KB)
And this is the latest flight:
Mission 4 V3_report.pdf (424.9 KB)

Here is an example of the DSM from this last flight:

…compared to a neighboring mission:

I don’t understand why this particular area keeps giving me the same or similar results as opposed to the other areas of the project that were fine.

I appreciate any help.

Thank you.
-Andrew

Hi,

Which one of the quality reports is the noisy DSM associated with?

Cheers,

Teodora

Hello,

That would be Mission 4 V3_report.pdf

Thank you,

Andrew Poth

Balcones Field Services, LLC

apoth@bfsusa.com

512.913.6683

Confidentiality Notice: This e-mail and any attachment(s) may contain confidential and privileged material intended for the addressee(s) or an individual authorized to receive for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment(s) may be disclosed, copied or distributed, and that any other action related to this e-mail or attachment is strictly prohibited, and may be unlawful. If you have received this e-mail by error, please notify the sender immediately by return e-mail, and delete this message. Balcones Field Services, its subsidiaries and/or its employees shall not be liable for the incorrect or incomplete transmission of this e-mail or any attachment(s), nor responsible for any delay in receipt.

Tweaking the parameters from all the processing steps can improve the DSM quality significantly.

I have obtained better DSM and orthomosaic results after modifying the parameters as follows:
step 1:

  • Keypoints Image Scale set to Full
  • Matching Image Pairs set to Free Flight or Terrestrial
  • Calibration with all the standard parameters (I did not modify anything)

step 2:

  • Image Scale set to 1/2
  • checked option Multiscale
  • Point Density set to Optimal
  • Minimum Number for Matches: 5
  • check option Generate 3D Textured Mesh
  • Settings: Custom (Maximum Octree Depth is 12 , Texture Size 8192x8192 , Decimation Criteria set to Quantitative , with Maximum Number of Triangles of 1200000 , check option Use Color Balancing for Texture )

step 3:

  • Resolution set to Automatic
  • DSM Filters: Use Noise Filtering and Use Surface Smoothing set as Sharp

Hopefully this approach can also improve your DSM results. Of course, there is always room for improvement by tweaking more parameters and then visualizing the results after every change, in order to find the parameters that help achieve an optimal 3D reconstruction.

Cheers,
Teodora

Excellent! I will try these parameters!

Thank you,

Andrew Poth

Balcones Field Services, LLC

apoth@bfsusa.com

512.913.6683

Confidentiality Notice: This e-mail and any attachment(s) may contain confidential and privileged material intended for the addressee(s) or an individual authorized to receive for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment(s) may be disclosed, copied or distributed, and that any other action related to this e-mail or attachment is strictly prohibited, and may be unlawful. If you have received this e-mail by error, please notify the sender immediately by return e-mail, and delete this message. Balcones Field Services, its subsidiaries and/or its employees shall not be liable for the incorrect or incomplete transmission of this e-mail or any attachment(s), nor responsible for any delay in receipt.

I tried those parameters and am having the same results. Could it be that I have too many GCPs for this area?

-Andrew

Hi,

The noise can come from many different reasons:

  • Problems after Step 1 (lack of connections, bad orientations…)
  • Noise in the point cloud itself.

There is one way to get the noise out of the point cloud. Using an Image Scale of 1/4 for example gives a much smoother point cloud, thus decreasing the amount of noise. The improved quality of the point cloud will propagate on the DSM.
For the point cloud, you could also tweak the parameter Minimum Number of Matches by using the values 3 or 4 (should work better than 6).

Cheers,
Teodora