Volumes not matching

Hey all,
I’m trying the free trial of DD.
I’ve done the volume calculations on some stockpiles I’ve flown…very easy to use (although it would be nice to see the base surface as mentioned in other threads.)
I’m finding that the volumes I’ve calc’d are 11% on average different from the actual volume that I have measured using RTK GPS. I haven’t adjusted anything to GCP’s…I don’t need to match any coordinate system or use it for design work, just need the volume of piles.
Am I doing something wrong? I have calc’d the volumes using another platform and I’m very pleased with the results, so I think my data is OK.

Just a few backend questions before we can even begin to analyze this.

What software did you use for comparison and what were the processing settings?
What did you use for field collection?
What was your spacing for your field topo?
Did you catch all grade breaks?
What plane did you use for the base in DroneDeploy?

Hi Michael, thanks for your reply.

I used Maps made easy for a comparrison with the following settings
22760 x 14908 @ 1.6 cm/Pixel
Average AGL=58.2 Elevation Range= 102m to 122m

I climbed the piles using Leica GPS. Flew the site with P4P V2.0

I picked up all breaklines and the usual. Very tight grid. I’ve been doing volumetric
surveys for quite some time and very confident in my calculations

I used Linear fit. I just tried the other methods on 1 pile with similar results.
The Cad method I use for these piles would be similar to the linear fit, creating a surface from the points around the base of the pile.

Again, thanks, and I appreciate your input.

GCPs should not make a difference as long as you use the same data between the two programs. Gcp will help with precision, but should not make a difference in this case. I’m a little confused by your elevations. Approximately 60 ft AGL or 300 feet? Or is the 300 feet MSL? Given your description of the survey method I trust that your ground survey was good. The base plane is the part that I question. When you do a traditional ground survey and use the toe of the stockpile as a surface you are actually using a triangulation method. This is what I use for pretty much all of my stockpile takeoffs. If you were to run a cross section of the lowest and highest points of the stockpile how much difference do you see between MME and DD?

I flew at 60m AGL (Avg 58.2m ) the The 102m to 120m would be Geodetic elevations of the site
I’m getting an elevation diff of 2 m btwn CAD and DD. Not sure how to get a cross section from MME

Pile 1 Cad E diff 8.2 DD 6.75
Pile 2 Cad E Diff 9.5 DD 9.0
Pile 3 Cad E Diff 6.2 DD 5.8

Piles 1&2 are 165m apart
1&3 are 204m apart
2&3 are 96m apart, all in a triangular position with Pile 2 to the North, Pile 3 at SE and Pile 1 at SW

So by that it looks like your quantities in dronedeploy were quite a bit less? What data format from DroneDeploy did you use in CAD? DXF? I wish they would give us the option to download 6-inch contours, but I guess it is what it is. I always use the point cloud.

I haven’t used any DD data in CAD. Just comparing my survey to DD Volumes. I can’t export any data on the free version, and certainly not willing to pay if this is going to be an issue.
MME Volumes were within 1%-1.5% (surprising really) of my survey.
Now I’m second guessing a couple quarries that I flew and didn’t measure with GPS.

I find that my quantities are within 2%, but since I have already proven it with traditional survey we now tracking by trucking volumes. Believe it or not the volumes that we are getting out of DroneDeploy end up being closer to the trucking volumes then the traditional survey. That is, unless you want to spend three hours topo’ing stockpile. I have found that the key to the DroneDeploy volume takeoffs is the base plane. Where you put your points and what type of plane you are choosing. You might try taking a second look and using the elevation overlay as a guide to establish your base. Additionally you can use the distance measurement to see the cross-section, but honestly I have not had to use this in a while with the addition of transparency to the elevation layer. The part that bothers me is the somewhat major difference you are seeing and elevation of the peaks of the piles.

Ya I can see that there could be a lot of diff with selecting base of pile, especially if there is a bunch of junk along the back of the pile. The Elevation layer is def the way to go.
I have to get some numbers out so I guess its back to climbing piles again!!!

At least we aren’t using the hand-level method anymore, LOL. If you like, feel free to PM me a download for the images so I can process in our account and we can compare.

I appreciate that. I’ll send you a link to google drive in the next few days.
I’m doing the same comparison on another quarry as we speak.
It wouldn’t have anything to do with the trial version? I wouldn’t think

Sounds good, and I would hope not. I mean to analyze in Precsion 3D and Civil 3D as well since the point Cloud will become available.

Could you copy me on the link also? I would like to process your photos thru Metashape and use my Rhino Drone App to analyze the volumes. I have done this on one of Michael’s projects (Crosswinds) which has several piles on site. Michael and I could cross check on his piles and then I could cross check with yours to hopefully get a clearer picture of any discrepancies.


1 Like

Good idea @SolarBarn. I think the more softwares and methods that we can get involved will eventually help us nail down some pretty concrete information as to what is defining the information we are looking at. I can also process in Pix4d and Recap Photo

One thing that we will need to do is define a standard set of base plane polylines so that we can ensure that all of our calculations are coming from the exact same points. Also, I think we should narrow it down to moderate settings to achieve the best time efficient processing. Nobody has time to sit around for 8 hours waiting for a couple of stockpiles to process.

I assume you both can handle a DXF for use as the perimeter?

@SolarBarn, @MichaelL, I’ve sent you both a link to the data.
Yes, DXF should be fine to use Michael.