Welcome, Commercial Drone Pilots!
Join our growing community today!
Sign up

Volumes - No GCP Vs GCP

Vantage

New Member
Joined
May 1, 2018
Messages
4
Reaction score
3
Location
UK
Website
vantagedrones.co.uk
Does anyone have any solid comparisons of results (stockpiles for example), from having a typical fly over with no gcp vs RTK/PPK gcps?

I've seen a lot of discussions but cannot seem to find any side by side results to show the level of difference between.

Obviously we know sites like DD say no GCPs are necessary for relative accuracy, and in my experience they are right to a degree. I've had ~2cm accuracy on 2D objects I've measured in real life vs using DD or P4D (the latter being more accurate). So I'm curious what kind of level of accuracy you can get from 3D volumes without GCPs?
 
The answer is read within Pix4D guides. The GCPs provide elevation accuracy. When the interlacing is accomplished and point clouds generated, the GCPs are then used to calculate relatve elevations to the final product(s).
 
any solid comparisons of results
I've got a spreadsheet of compiled data that compares a 3,500cy pile with and without GCPs, and flown at 2 different altitudes. I drew the piles 3 times using the TIN model and 3 times using a ground plane for each test, resulting in 24 different data points. Everything was within 5% of each other and 20 were within 2.5% of each other. For small piles, I don't think you're going to see much of a change, but I could see more error being introduced as you get larger piles, especially depending on the method you're using to draw the bottom of the pile. I can post the information, if you're interested in greater detail.
 
  • Like
Reactions: ArrUnTuS
I just recently did a stockpile with out ground control points. Processed the flight in 3 different progams and the volume from all 3 came out within a few cubic meters. But the comparison to what the contractor tells me is in the pile is not even close. Unfortunately the base of the pile is not very level around the perimeter so this maybe affecting the outcome, also as they are removing the material it looks as though the center of the pile is also dished out. Hoping to get another stockpile and make sure of what the base really is before the pile.
 
what the contractor tells me
How much of a difference? 10% 50% 3,000%? Were they calculating using in-situ volumes prior to moving dirt? There are a ton of different variables that can go into throwing a comparison off, but at least you have three consistent readings.
 
Yes I realize there are many factors involved. I think probably the difference is the weight used by them to calculate their volume. i think they are using a lighter weight per m3 than normally used in our area of the country.
 
How much of a difference? 10% 50% 3,000%? Were they calculating using in-situ volumes prior to moving dirt? There are a ton of different variables that can go into throwing a comparison off, but at least you have three consistent readings.
A lot of volume measurements are based upon weight. The rub is, there is hydration in the material when it is first removed and loaded, and then that water evaporates as it sits in the sun and is exposed to the atmosphere. Water evaporates and the added weight dissipates. So you pay for more material than you actually get. Just my observation.
 
We've done many tests on stockpiles and results between three different processing softwares with and without gcp's and all were within 5% as someone stated above. I also did comparisons against a Robotic Total Station survey and a GNSS receiver. Backup was from trucking tickets and the non-gcp measurements were more accurate than either the robot or gnss. GCP's aren't needed on normal stockpiles, but as the horizontal area gets larger the less that becomes the case. Sometimes you will see stockpiles that are kept low, but cover a large surface area. Once those get to about an acre then we need to start thinking about PPK or 4 GCP's.

Speaking to what someone said about the contractor doubting their results it's most likely caused by the type of base plane that was used in the calculation. We have choices of linear plane, triangulated and lowest point. They all have they're different uses. that said I would be interested to hear more about the discrepancy the contractor saw and how they verified their claim. I work for a contractor and have had no issues with quantities being reported hi there on flights that I have done for us or external clients.
 
We've done many tests on stockpiles and results between three different processing softwares with and without gcp's and all were within 5% as someone stated above. I also did comparisons against a Robotic Total Station survey and a GNSS receiver. Backup was from trucking tickets and the non-gcp measurements were more accurate than either the robot or gnss. GCP's aren't needed on normal stockpiles, but as the horizontal area gets larger the less that becomes the case. Sometimes you will see stockpiles that are kept low, but cover a large surface area. Once those get to about an acre then we need to start thinking about PPK or 4 GCP's.

Speaking to what someone said about the contractor doubting their results it's most likely caused by the type of base plane that was used in the calculation. We have choices of linear plane, triangulated and lowest point. They all have they're different uses. that said I would be interested to hear more about the discrepancy the contractor saw and how they verified their claim. I work for a contractor and have had no issues with quantities being reported hi there on flights that I have done for us or external clients.
Hi Chasco,

Overall, would you recommend GCPs for stockpile analysis with moderate accuracy?
 
Does anyone have any solid comparisons of results (stockpiles for example), from having a typical fly over with no gcp vs RTK/PPK gcps?

I've seen a lot of discussions but cannot seem to find any side by side results to show the level of difference between.

Obviously we know sites like DD say no GCPs are necessary for relative accuracy, and in my experience they are right to a degree. I've had ~2cm accuracy on 2D objects I've measured in real life vs using DD or P4D (the latter being more accurate). So I'm curious what kind of level of accuracy you can get from 3D volumes without GCPs?
It's going to be hard to compare because the data won't align unless you have an external CAD/GIS program.
 
Hi Chasco,

Overall, would you recommend GCPs for stockpile analysis with moderate accuracy?
It depends on the scene. Less than 250ft horizontally or 30-40 ft tall then no. Once the get past that height then you start to get high variance in your GSD and it might not stitch so well. Larger stockpiles like at a quarry tend to get spread out when stored so the material is easy to load into trucks. This causes most of the error to occur horizontally and GCP's will help. Another reason that is more important in my opinion in mining is that the data needs to have a very tight repeatable area if you want to do things like automated reporting.
 

Members online

Forum statistics

Threads
4,277
Messages
37,605
Members
5,969
Latest member
KC5JIM