Measure elevation data without GCPs

[miloco2016]

Hello, im working on several projects, and some of them don't need to be very accurate. So im trying to find the way to not use GPCs cause of the time and cost it has.

Is there any way that i can use the takeoff point of the drone and mark it as a reference for the other points in the project to get approximate elevation data?

I know it won't be related to a particular coordinate system, but in some of the projets i just need to generate data like level curves, and without any reference im getting some random data.

You can see in this article that drone deploy already does this.

https://blog.dronedeploy.com/new-elevation-and-volume-tools-for-mining-and-construction-89e798b7e1a8#.xodj461e9

HELP PLEASE!!

[stihl]

Hi,

On your link it says the following:
"Note: Unless corrected with ground control points (GCPs, available on our Premier plan), elevation data reflects height in feet or meters relative to point at which the drone took off."

In short this means that it is implied that the UAV that's being used has some logging system of its altitude.
Without this, drone deploy wouldn't know your starting altitude.

So if the UAV you use already logs it's position, you can use this information as camera positions in WGS84 coordinate system to calculate relatively accurate stockpiles within photoscan.

A terrain without GCP's or flight positions has no scale and it is therefor impossible to extract and meaningful volume as photoscan has no idea how large it is in the first place.

If you want meaningful volumes for your client you'll either have to use flight positions (with an absolute accuracy of 5-10m) or GCP'S (with an absolute accuracy of 2-3cm). With both methods you can extract accurate *relative* stockpile volumes. However if you need to have volumes compared to a reference height you will *have* to use GCP's.

[miloco2016]

Ok, so is there any way to use a total station for example to create markers and reduce the error to cm? or the only way is to use a hi precision GPS?

Im trying to reduce the costs of the works i do so that's why im asking this things, and im new on this field also

So thanks for your help and patience!

[stihl]

A total station will do that fine. Would it be referenced or is it meant for a local coordinate system?

[miloco2016]

Ok, actually what's important in this particular project is to generate level curves in photoscan and we need them to be precise! We are getting 2 meters errors now cause we are not using gcps. We are new in this field as i said before, soi dont really know how total station works. We rented for the last project a GPS opereted by the owner and mark some points and he gave us X, Y and Z coordinates that we put into photoscan and get errors only of 2cm wich is grate. But the rent of the GPS is expensive.

I google the total station thing and i cant understand how it gives this coordinates to reduce the problem. so im sorry but i don't understand your question, we want the final product to be georeferenced, so if we generate a KML file we can open it on google earth for example, we dont need the location to be precisely at cm, but we do need to have precise altitud data.

In resume, we want to create level curves that have to be very precise. But not necessarily need to be geotagged

[stihl]

If you want precise contour lines (the level curves you speak of) you have to, by definiton, geo-refecence or scale your project. Otherwise it's useless as Photoscan has no idea how large the object or terrain is that you've photographed.

A total station is a device used to very precisely measure distances and angles. With it you can measure 3 dimensional points (XYZ coordinates) in a coordinate system, but *only* if you know the exact location of the Total station in the same coordinate system. The exact location can be interpolated by measuring at least 3 base points with RTK GPS (The GPS system that has a 1-2cm accuracy) and having the total station calculate it's position based on those measurements..... in other words this is far too much work for what you want to achieve.

From what I understand from your post you'll need a RTK GPS (1-2 cm accuracy) to measure in GCP's (ground control points) on your terrain.
More accurate measured GCP's means a more accurate model and thus lower errors. A total station far succeeds the accuracy of a RTK GPS and thus means a better model. However for nearly all aerial surveys, RTK GPS accuracy is more than enough.

Is this an assignment for a client? If so you can just add the cost of renting a RTK GPS to the client.
If it's not for an assignment but for study applications you can try using your phone GPS. It's far from accurate, but accuracy comes at a $$$Cost$$$.

[miloco2016]

Thank you so much for your patience and time.

Ok, so i unerstand mow that the only way to have low error is to meassure gcps with the rtk gps.
It is for a client so im planning on adding the cost of the renting.

What i dont see clear now is the total station option, you can use a total station to mark points but still need to locate the location of some points of the total station in the project to make it work?

And what about the iPhone GPS, how accurate is it? Can i get X, Y and Z coordinates with it? Or is it less acurate than the drone GPS tags? Im using a phantom 4. (Just asking this for some personal projects)

[stihl]

Your Phone GPS might be better than your Drone GPS since it probably uses something called "DGPS" which uses information from the internet to calculate your position to within 2-5 meters accuracy whereas your drone probably has an accuracy of 5-10 meters.
There are several apps available on any mobile system that allows you to log your XYZ position. Just position it as lowest to the ground possible before logging the position.

Most total stations do not have GPS onboard (besides additional modules) which means that they have no idea where they are located. If you turn on a total station and just start measuring, you're measuring in a local coordinate system, i.e. an arbitrary system.
A total station that's measuring in a coordinate system has it's position made known by measuring other known points (which are for example measured by GPS) and it's location relative to those points. If you have 3 or more known points, the total station can triangulate it's position relative to those points.
With additional post processing you can then refine the total stations' location even further to get within the 1 cm range. From what I can tell this is mostly irrelevant to your needs.

[miloco2016]

Ok! I understand now!...

I have another question now, the last time we used the RTK GPS, we just use natural points to get the points, now we now that we can print makers directly from the software and place them in the terrain before starting the photos and use those markers coordinates, and the software will recognize them automatically and then i just need to put in the X,Y and Z coordinates.

What im worried about is the size of the printed markers, im planning to fly up to 80 meters so i will have like 3.5cm/px.
How big do i have to print them? i don't have the chance of printing them bigger than a A4 paper, but im asking this cause im scared of printing them that big as i can and measure those points and then can't be able tu use them correctly

[stihl]

A4 size probably will not do with that GSD. I wouldn't count on it.

We've made our own GCP's and purchased some aluminum discs too (expensive) which work well. The smallest ones are at least 30 cm in diameter but we usually use 60 cm.

You could try making your own by using those window blinds and using two to form a cross which you can measure the center of.