I need to survey a small area without GCPs (and without RTK)

[PeteM]

As the title says.

Unfortunately I only have a non-RTK P4 and this time I cannot access the ground to measure some GCPs with a total station (as I usually do).

The survey must NOT be georeferenced and must NOT be precise.

Furthermore, I already have a precise survey of the area and I just need to add something that has changed over time to the (precise) map I already have.

I thought about sending the drone into flight at a very low speed (like 2 m/s) and at an altitude of only 50 m to and have a more or less correct photo to use to integrate the map I have.

Advice ?

 

[Miner_jeff]

As the title says.

Unfortunately I only have a non-RTK P4 and this time I cannot access the ground to measure some GCPs with a total station (as I usually do).

The survey must NOT be georeferenced and must NOT be precise.

Furthermore, I already have a precise survey of the area and I just need to add something that has changed over time to the (precise) map I already have.

I thought about sending the drone into flight at a very low speed (like 2 m/s) and at an altitude of only 50 m to and have a more or less correct photo to use to integrate the map I have.

Advice ?

Hi @PeteM

Sounds like you might heading into photogrammetry territory. (Making photo mosaics from smaller overlapping photos.)
Depending on how small your area of interest is, does it fit in one photo? How much you capture in the photos is 100% altitude dependant and how much detail you want to see. If you want to match the existing data you would have to know what the Ground Sample Distance.
Different cameras/drones will fly at the appropriate altitude to achieve a particular level of detail in the captured images.
I suggest you search for YouTuber for Rich Charpentier

and Do You Know Drones (mapping flight apps) https://youtube.com/playlist?list=PL_EEcG1Zv5elxMLrsvgRojLH4N0BJMegq&si=1NGN-U2i5HglGTO6

There is lots to learn on the subject.
Reply with questions (there will be many).
Cheers,
Miner_Jeff

 

[PeteM]

I have been involved in drone photogrammetry for many years and I know how to operate.

The problem is that this time I can't physically measure GCPs on the ground.

However, since I don't need a survey that is precise or georeferenced, I wanted to understand how to obtain an "acceptable" result without GCPs, having a drone without RTK.

I need a survey that planimetrically (x,y) is not too distorted. I don't care about the z coordinate.

 

[Miner_jeff]

Fly the mission and geolocate the newly processed area using existing common points of interest between maps. I believe this could be achieved using QGiS.

 

[PeteM]

Fly the mission and geolocate the newly processed area using existing common points of interest between maps. I believe this could be achieved using QGiS.

Having already had a survey of the area, I was thinking of something like this.

Maybe I can impose a distance between two distant points to scale the image correctly.

 

[Miner_jeff]

Having already had a survey of the area, I was thinking of something like this.

Maybe I can impose a distance between two distant points to scale the image correctly.

If your small map area is processed correctly using the GPS exif data your small area should at most be a few meters off. The few meters can be fixed by translating the new area to correspond to the previous.

 

[jaja6009]

Having already had a survey of the area, I was thinking of something like this.

Maybe I can impose a distance between two distant points to scale the image correctly.

You can take a single image and then rectify it in Arc Pro. You would use control points that you can find in both the image and in the accurate survey you already have.
I did this to look at social distancing during Covid on our beaches. It will meet your goal of not accurate or precise, but after rectification will be georeferenced.

Below are simply images taken with a Mavic 2 Dual that were rectified and georeferenced so that they can now work with Arc GIS. A 6ft x 6ft grid was then overlaid to show how far apart people were staying on beach during the height of Covid.

There are also raster tools available that could let you merge the two rasters, but I have never done it with orthophoto rasters.

 

[PeteM]

I solved the problem by inserting some GCPs not measured directly in the field but known from previous maps (edges of tanks and other unique elements never moved).

I was not able to insert the point with the precision of a colored marker on the ground but with a good approximation I scaled the output coming from the drone.