Forum

Author Topic: Camera altitude vs. GCP Elevation  (Read 11824 times)

RCGeo

  • Newbie
  • *
  • Posts: 3
    • View Profile
Camera altitude vs. GCP Elevation
« on: August 12, 2015, 12:43:40 AM »
Hi all,

As a brief background. I recently surveyed a site with my DJI Inspire 1 at 10m altitude over an area of about 300m X 300m. I created a dense point cloud relying on the camera's known position and selected "reference" before creating the point cloud, so the camera location was used (lat, long, and altitude above ground surface). The results are excellent and now it is time to go back and create a georeferenced model. This is where I have some questions.

Based on other forum topics it appears that the GCPs should be assigned after the initial photo alignment is done. Again, it would be nice to use the camera's known location in this step to speed up processing.  I have dozens of very well located GCPs with elevation values. There is clearly a discrepancy between my GCP elevations (~3000m) and my camera altitudes (~10m above ground surface). When I bring the GCPs in and locate the markers, there is a large error do to this difference after placing the GCP points and "Optimizing Camera."

In a test model I went ahead and ignored this error and proceeded to create a dense point cloud and mesh. I exported the mesh as a TIFF and upon opening the model in ArcGIS it appears it used the marker elevations. Is there a cleaner way to do this? The thing is, I would like to use the camera's known position to help on processing time, but wondering if I need to bring the photos in without spatial data to make sure I am not introducing some error into the model.

Any help/discussion is greatly appreciated.

sakaic

  • Newbie
  • *
  • Posts: 14
    • View Profile
Re: Camera altitude vs. GCP Elevation
« Reply #1 on: August 12, 2015, 08:01:27 AM »
From my experience with my Phantom 2 and S900 the typical altitude values stored with th e internal camera is relative. Recently been messing around with the older photos and I find that if you are processing with GCPs and they are reliable then it produces better results if I remove all location data from the photos.

Keeping the location data tends to generate weird results like sunken or floating point cloud sections.

JMR

  • Hero Member
  • *****
  • Posts: 504
    • View Profile
Re: Camera altitude vs. GCP Elevation
« Reply #2 on: August 13, 2015, 09:24:33 AM »
Camera coordinates will speed up alignment in "reference" mode. You are right, but you can't mix these coordinates with GCP coordinates unless they are in the same system.
Use in-photo geotags for the align stage and after you're done, add your good GCP and disable camera coordinates by unchek all photos in the referece pane. Change refference system settings to match that of your GCP survey and then update  transformation (optimize) and all your photos will be translated to the ground control points coordinates.
« Last Edit: August 13, 2015, 09:26:31 AM by JMR »

RCGeo

  • Newbie
  • *
  • Posts: 3
    • View Profile
Re: Camera altitude vs. GCP Elevation
« Reply #3 on: August 13, 2015, 08:46:20 PM »
Camera coordinates will speed up alignment in "reference" mode. You are right, but you can't mix these coordinates with GCP coordinates unless they are in the same system.
Use in-photo geotags for the align stage and after you're done, add your good GCP and disable camera coordinates by unchek all photos in the referece pane. Change refference system settings to match that of your GCP survey and then update  transformation (optimize) and all your photos will be translated to the ground control points coordinates.

This is very helpful...thank you. I didn't realize you can uncheck the photos. To clarify, here is my workflow with some questions:
1)Align photos using geotags
2)Add GCPs and disable photos in reference pane (uncheck)
3) Optimize the sparse point cloud. Is this the "Optimize cameras" tool?
4) Build Dense point cloud with new optimized sparse point cloud.
5) My goal is to then build a DEM from this dense point cloud.

I appreciate it if you could confirm/clarify this workflow.

Thanks so much,
RC

stihl

  • Sr. Member
  • ****
  • Posts: 410
    • View Profile
Re: Camera altitude vs. GCP Elevation
« Reply #4 on: August 14, 2015, 03:03:00 PM »
Your workflow is spot on!

Optimize camera's is indeed the optimisation tool, leave all checked except K4 which is as far as I know for distortions with a fish eye lens..

nadar

  • Full Member
  • ***
  • Posts: 130
    • View Profile
Re: Camera altitude vs. GCP Elevation
« Reply #5 on: August 18, 2015, 03:57:27 PM »
standard GPS positional accuracy is about 5 m in X&Y and 10m in 5.
If your flight altitude is about 10m above ground, don't expect too much from these data. I would rather orient photos in generic or disabled mode and rely on high-precision GCPs only