Work Flow for GCP

[LFSantosgeo]

[...] future lies in UAVs that have PPK/RTK capabilities which completely eliminates the need of ground control points and instead uses the highly accurate location of each image ('air control points'). This means that you will then have as many 'air control points' as you have images, which is never the case when using ground control points.

As a reminder, if you plan to deliver your product/research with afered error then even with a drone capable of RTK/PPK surveys you would need some GCP to use as check points to validate the precision of the flight. Unless you present a whole study comparing the accuracies of both survey methods on different survey scenarios (slopes, flat terrain, urban areas etc.). Even so you wouldn't be able to prove that in some particular flights they have low errors and high accuracy out of the RTK/PPK system boarded in the drone.

Am I correct? I mean... despite the promising future of RTK/PPK on drones it still doesn't eliminate the need of GCP for some uses.

[stihl]

You're correct. It's still necessary to verify your deliverables with an independent check. Whether that's by checking the XYZ displacement on on-site measured check points or by comparing the model to another trusted accurate model usually means having to go out in the field with a RTK system.

[SAV]

Hi lfsantosgeo & stihl,

There have been quite a number of studies in which current RTK/PPK technology on UAVs was tested and compared to 'traditional UAV surveys' using GCPs. For example, see here https://pix4d.com/rtk-ppk-drones-gcp-comparison/

The results are very promising, and RTK/PPK technology on UAVs will be getting better and more affordable in the future.

RTK/PPK systems on drones will just be as reliable as a 'traditional' RTK/PPK systems which you would use to survey ground control points. This is also when, IMO, I don't see the need for an independent test anymore.

Let's turn the whole story around, which means one could also say that traditional GCPs need to be tested and independently checked with (e.g., LIDAR) measurements from the air or non-GPS, ground-based techniques (e.g., total station) to be sure that they are correct. They are simply accepted to be correct.

At the moment, I agree, it might be still useful to test/double check using a few GCPs. But that will change within a few years, IMO.

Regards,
SAV

[stihl]

Whilst I agree that technology is getting better and more sophisticated each year, it's easy to forget the most error prone part of a project, the human factor. Because of that there's always a need for an independent check. There's no guarantee that the base station used for your RTK/PPK UAV solution is correctly set up without proper checks. If this part of the chain goes wrong the results are practically useless without checkpoints to verify your results.

We always verify our GCP coordinates compared to other reference data since a simple mistake like a wrong antenna height input before the measurement *can* happen. It doesn't matter if all goes well in 999 out of a 1000 projects. If stuff goes wrong and you have no way to check for errors or verify your findings with an independent method then you really need to consider whether you're a serious surveying company or not.

They are simply accepted to be correct.

Perhaps if your client is satisfied with sub meter accuracy. Most clients will want adequate proof of your blackbox principle height model.

[SAV]

Hi stihl,

I agree, you should always check with independent measurements to 'prove' that your results are correct and any human error can be excluded.

At some stage, however, it becomes a 'philosophical' question, IMO. The Earth is a quite dynamic system that changes over time. For example Australia is moving NE by about 7cm each year (!!!). One should be aware of such 'natural error' that needs to be considered as well.

I think in the end it boils down to the project requirements. For some, the absolute accuracy is not crucial. For example, if I simply want to calculate a volume of an object (e.g, stockpile), I don't really care if the whole survey is off to the E/N/S or W by a few meters. As long as it is properly scaled and internally consistent, it will deliver the correct measurements/results.

Merry X-Mas.

Regards,
SAV

[Walter Volkmann]

Hi Geomaticist,

If you don't have PPK on board yet, no need to buy a new drone. Consider the V-Map system which is completely platform independent and weighs only 120g (about 4oz). If your drone can handle the extra 120g and has a space for the dual frequency antenna, you should be able to mount it on your existing drone. The only requirement is that the camera has an external flash hot shoe connector that delivers a pulse at mid-exposure. The V-Map system comes with free post-processing software that takes you through the whole work flow from RINEX conversion to trajectory computation to event position calculations to geo-tagging. You can use your existing survey grade GNSS receiver, a VRS, a network service or a second V-Map receiver as reference station.Have a look at www.v-map.net. There's a white paper there as well. Feel free to leave an enquiry. We will respond with technical info and pricing.

[SAV]

... or wait until the Phantom 4 Pro RTK is released 

https://dronedj.com/2017/12/24/more-details-emerge-of-dji-phantom-4-rtk/

Regards
SAV

[dpitman]

I'm looking forward to it so long as it is not as lame as the M210 or M600 implementation of "RTK".

[SAV]

Yes, I agree dpitman.

So far 'RTK' on DJI drones was only used for 'navigational' purposes of the UAV itself, but not really for the generation of cm-accuracy maps & models. Maybe its also easier to sell such a UAV to surveyors because it has RTK in the name  A lot of people were quite upset when the realized that they spend a lot of money on something that does not do what it says in the name 

I hope that DJI has listened to their users and allows for proper RTK / PPK capabilities in the future. #fingerscrossed

Regards,
SAV

[rossnixon]

[...edit...]
3. Convert GPS coordinates of your geotagged images (WGS84) to match the coordinate system of your ground control points (GCPs) which will be imported later. Note that the altitude information stored in the EXIF data of imagery acquired by DJI drones is the relative altitude from the point of take off and not the absolute/real world altitude. Here are 3 ways to fix this issue: http://www.agisoft.com/forum/index.php?topic=4986.msg38769#msg38769
[...edit...]

I'm currently having a problem aligning a set of 500+ UAV photos over a forest.
Does the above step#3 mean than I need to somehow replace the longitude/latitude EXIF date with the coordinate system that my (very accurate) GCPs use? If so, how do I do that?
Also, is the altitude data used by Photoscan? Should I remove those?
Thanks!

[SAV]

Hi rossnixon,

Yes, that's correct. Import your images and then click on the CONVERT icon in the reference pane (the icon shows a spreadsheet). Then choose the coordinate reference system (CRS) that matches the CRS of your accurate GCPs. To find the right CRS it is best to use the unique EPSG number.

Yes, altitude data is also used by Photoscan. In case you acquired the aerial imagery with a UAV from DJI, the altitude value might not be very accurate (has been discussed in detail in this forum). You can then either uncheck all images in the reference pane and only use your highly accurate GCPs before running your photo alignment or change the camera accuracy settings to a large value in order to account for the large (vertical) positioning error.

All the best.

Regards,
SAV

[...edit...]
3. Convert GPS coordinates of your geotagged images (WGS84) to match the coordinate system of your ground control points (GCPs) which will be imported later. Note that the altitude information stored in the EXIF data of imagery acquired by DJI drones is the relative altitude from the point of take off and not the absolute/real world altitude. Here are 3 ways to fix this issue: http://www.agisoft.com/forum/index.php?topic=4986.msg38769#msg38769
[...edit...]

I'm currently having a problem aligning a set of 500+ UAV photos over a forest.
Does the above step#3 mean than I need to somehow replace the longitude/latitude EXIF date with the coordinate system that my (very accurate) GCPs use? If so, how do I do that?
Also, is the altitude data used by Photoscan? Should I remove those?
Thanks!

[JMR]

UAV from DJI, the altitude value might not be very accurate (has been discussed in detail in this forum). You can then either uncheck all images in the reference pane and only use your highly accurate GCPs before running your photo alignment or change the camera accuracy settings to a large value in order to account for the large (vertical) positioning error.

My two cents: If your altitude tags are way wrong, I'd suggest to do the following after alignment and once the project CRS has been converted to that of the GCPs survey.
(optional but highly recommended if you used well defined ground targets) goto preferences/advanced and check refine marker projections based on image content.
1 import CGP coords
2 set model view to top and ortho by pressing 5 and then 7 numeric keys
3 pick circular selection and select a small region around first gcp, then rightclick and filter photos by points so you can pin the marker in at least two instances.
4 repeat in two more GCP forming the biggest triangle in your area, then click on "Update" button in reference panel toolbar

Since from now the sparse point cloud and the markers should be in the same altitude, you can keep pinning all markers on GCP using filter by target. The automated marking works really well, but it is highly recommended to review all instances (blue flags) created by PScan, that can be fooled in some occasions. This review process is done quite rapidly if you first filter by marker and then, with the marker highlighted in the list,  use av-pag re-pag key to browse filtered photos focussing and zoomed on marker.

There is no obligation to clear or uncheck camera coordinates even if their altitude is way off meters, because once that surveyed GCP are in use, their much higher relative weight will rule the adjustment. That said, is correct to say that wrong camera coordinates can contaminate the final adjustment a little bit, but usually meaning nothing significant with respect to the survey real life accuracy.

[Marija]

Dear all,

Can somebody give an advice?
I was trying to combine yours and tutorial (steps in manuals) and whaching videos,but still I didn't found perfect solution, so can somebody check and comment my steps?

1. I import photos (I already have camera locations)
2. I align photos with HIGHEST accuracy (with on generic and reference preselection; in advanced I turn off "Adaptive camera model fitting")
3. I select location of GCP (accuracy 0.001m)
4. After I found location (I am looking that the error is as minimum as I can define them), I press update
5. I optimize camera location
6. I build dense cloud
7. I export dense cloud

But still, when I am controling position in Autocad there is an error around 20cm between "real" and "Agisoft pointcloud".


So I have questions:
1. How can I manage to have better resault?
2. After adding GCP and unchecking camera, should i need again alighn photos or I can go to creating dense cloud?
3. Is it necessary to build mesh or I can add photos on just alighned photos
4.  How and is this good,to delete (or uncheck or disabled) photos which are not so good, or blury? Or this I can see in camera estimate? Or I can skip this step?
5. Can you, please, tell me which additional steps I should do to have better and more accurate point cloud?

Thank you.

[Alexey Pasumansky]

Hello Marija,

Usually for aerial datasets it is not reasonable to use Highest accuracy option, so we suggest to use High instead.

Which calibration parameters you are optimizing on your step 5? (answering to your question, you do not need to run alignment again after placing markers, just use optimization).

Extremely blurred images can be disabled before building the dense cloud and orthomosaic to avoid the utilization of the poor quality data on input.

Question #3 is not clear, please elaborate.

[Marija]

Dear Alexey,

Thank you for answering.

On my step 5., on Optimize Camera Alignment I am using Fitf, Fit cx cy, Fit k1, Fit k2, Fit b1, Fit b2, Fit p1, Fit p2  (only not Fit k3, Fit k4, Fit p3, Fit p4) and I optimize it with unchecked cameras (only selected my GCP). Is this ok or I can select others to have better resaults?

In manuals, when I import photos first i need to Align them. After this, I need to Build Mesh and then import GCP.
But I skip this step, I just import photos and align them and then I am adding GCP (because I know on each photo is each marker). How is this important?

But still the accuracy is not the best.

Thanks.