Gravel pile over complex surface. Wildly different results depending on method

I am making some tests trying to discover the best software to calculate volume of the piles on our quarry.

Our quarry is on the side of a big hill, so the idealistic “pile over flat surface” is not an option.

Particularly problematic is one of our biggest piles, which is located over a 45º declivity, spilling on roughly flat areas on the top and bottom of that declivity.

Selecting the entire contour I get only 20k cubic meters. We know that result is not accurate. It should be larger.

I wonder if DroneDeploy somehow tries to create a base surface with those 2 flat areas connected by a 45º declivity? Or if it creates a SINGLE base plane, inclined, connecting the top to the bottom vertexes?

It would be REALLY HELPFUL if DroneDeploy showed the base surface it created, in 3D, so we could better understand how it is calculating that base surface, so we could better located the vertex points of even adjust the base surface.

Anyway, as the single polygon selection resulted in a volume quite smaller than expected, I tried to divide the same pile in 3 parts.

That is… a polygon for the bottom part, a polygon for the upper part and a polygon for the middle part, over the 45º incline.

Result was about 30k cubic meters. Close to our expectations But does it REALLY works separating a pile in 3 parts to calculate volume, or was it a fluke?

Additionally, we have some piles that are located under the gravel conveyor belts. Any tip on how to calculate those piles, considering the conveyor belts will be shown and calculated as solid objects that are quite taller than the pile?

The base surface is generated by triangulation of the elevations of the points that you place. If you have a pile one the side of the hill that continues onto a flat area make sure you have points right at the grade break.

I occasionally have machines sitting on top of the piles so I take-off the machine itself and subtract that from the file total. I usually include this as a note in the annotation.

It would be really nice if we could SEE the base surface, in the 3D view, just to be sure that it did really triangulated correctly at the grade breaks.

We have two licenses of Autodesk Civil 3D at my company. Although much more complicated, I guess we can or should import the point cloud generated by DroneDeploy, import at Civil 3D and verify the volume values. It’s much more complicated and time consuming to do it in Civil 3D, but the purpose would be to verify values before using only the simpler and faster DroneDeploy.

ps: is the grade break in your photo really in that place? Seems to be left and right tire marks of the same truck on each side of the grade break.

Regarding “seeing” the triangulated base mesh…

How do you know that DroneDeploy is triangulating in the right direction, from vertex to vertex somewhat parallel to the grade break? What if it triangulates perpendicular to the grade break, thus creating a straight line that goes across the grade break without considering it?

What if the polygon shape is much more complex than your roughly oval/rectangular shape? Maybe DroneDeploy does the base mesh correctly, and is able to follow the grade break elevations. What if it doesn´t for a few of your maps? You won’t know!

The image I provided is merely an example of where to put the triangulation points. The is a very small grade break on this one, but you can see from the pile to the north that they drive perpendicular to the pile and then drive up it.

Seeing the bottom mesh and even having an option for averaged surface would be great. Right now we just have actual and low points. The triangulation is like most other software where the closest points are used.

We use Carlson and I agree that it can be time consuming to recreate this workflow. We have started using Carlson Precision 3D which is specifically for point clouds and is much easier than Civil, but still way more time consuming than DroneDeploy. Piles within 2-5% of actual volume works for us.

For those who may not quite understand triangulation… Not Rogerio of course. :wink:

I guess the triangulation algorithm may differ a bit from software to software, hope DroneDeploy’s one is like your example (and again, seeing that base mesh triangulation to be sure it conforms to our terrain would be great)

Nice example you created, as it’s quite similar to the pile and terrain it’s located over.

Here’s a quick comparison between DroneDeploy and Precision 3D. Within 2.2% and that’s probably not being able to get the exact toe points. Thanks for the topic!

Is it possible to send you my pile, so you can compare with PrecisionPoint to check if DroneDeploy is being precise for it too?

If yes, how do I do it?

After some good offline discussion about his project in particular, I wanted to pull this back into DroneDeploy because I think it will be good for many to see from a civil estimating standpoint. Please visit the new topic so we can have a clear and concise discussion about civil estimating and quantities take-offs.

Civil Estimating and Quantities Take-offs