Relative vs. Absolute accuracies

[airmap3d]

Example:
Using a UAV for aerial mapping/volume calculations with or without GCP's over small stockpiles (200m x 200m)

With GCP's obviously the absolute accuracies will increase significantly but will the relative accuracies change/improve?

I have a client who just wants volumes of their stockpiles and I am wondering if I would get satisfactory results without the use of GCP's, as they do not need the data to be accurately geo-referenced.

I am hoping that I can get away without the use of GCP's as it would save me a lot of time and my client a lot of money??

[piste]

Same problem!!
Is it possible place markers on the ground without surveying them (GPS, TS or whatelse...)
and then optimize point cloud with no coordinates but only recognizing the exact place of the marker in every image?
Thanks all

[airmap3d]

It seems I have a question that no one knows the answer too?  Surely not??

[Geo]

I have not tried this, but may try it on a project soon. 

I think it is important to have a good methodology that provides quality control and redundant checks especially for stockpile surveys.  You may want to lay out two scale lines along fairly even ground and at appropriate lengths to provide scale for the project area.  I would simply measure forward/reverse along the line with a hand held 30 metre reel tape and check that the results are within measurement tolerance.  Repeat for the second line.  Also take averaged GNSS autonomous positions with a handheld GPS.  Fly the stockpile at two different heights for two independent image sets.

I would seed the GNSS coordinates with about a ten metre accuracy.  I would process the model and scale it using one of the baselines.  I would then determine the stockpile volume.  I would repeat with the second baseline.  I would then use the second set of imagery and repeat the process.  You will have four values of the stockpile volume.  You could also create an ortho with the seeded autonomous GPS coordinates.  The scale lines also give you a rough check on the inverse distance between the two GPS values of the end points.

If you try this, please let us know how it works.

Jim

[airmap3d]

Hi Jim,

Thanks for the input.  Although if I was going to go to those sort of lengths I probably may as well just set up my GPS, log some static and RTK in some GCP's.  You are right though, I would have no redundancies to check against but for some of my 'less important' jobs this is not really an issue.

I have just done some testing of my own with the same set of data (30ha quarry site) both with ground control and completely without.  The results are basically no different... when it comes to volume calculations anyway.  Obviously the 2 data sets are a few metres apart but the relative accuracies look pretty good.

The table attached shows the differences between the 2 tests (note: the base models for the calcs hold a certain degree of estimation so with this in mind you could argue that the results are all but the same??)...

If anyone else has done any testing on this I would be keen to know your results??

Cheers!

[Geo]

I finally tried a volume calc on a small 50 metre long stockpile for a friends data.  He placed two scale lines (15m each in the scene - nice clean targets).  I followed the Volume tutorial and I used one set of images to determine the first volume and then I used a second set of images and the difference is about 30%.  I did a scale check for the length of the stockpile for each model and it looks good.  I imported the model into another program and see similar volume results.  This project needs to be delivered to my friends client so I am a little desperate and at wits end as to why such a large discrepancy.  I would have preferred longer scale lines, but is definitely not that.

Any help would be greatly appreciated.

Jim

[Alexey Pasumansky]

Hello Jim,

We can try to help you with this project if it's possible to provide the projects with the corresponding images.

[Geo]

Hello Alexey,

Thank you for such a quick and kind response.  Do you have an FTP site that I can transfer the images?

Jim