I cant seem to figure out the right combinations of settings to get my model georeferenced even after making sure the exif data contains the gps data and project coordinate system is set to wgs84. When I click the component, under alignment report it says Geo-referenced false. Do I have to set ground control points? It is my understanding that the Exif data should suffice and just be less accurate.
It seems when i go from raw DNG files to Tiffs through lightroom RC cant read the exif data anymore, even though exiftool shows it to be the same. Don’t know if RC can’t read Tiff Exif data?
Hi Steven,
how do you figure that RC cannot read EXIF data from TIFs?
Did you try it with the RAWs or camera JPGs?
Was it geo-referenced then?
I don’t have any experience with GPS geo-referencing, but I would be surprised if the issue was caused by the TIF format…
After clearing the cache it seems to read it just fine now. Lightroom didn’t include GPS data by default and when I turned it on and replaced the export RC did not read the changes to the gps data.
But my question is, with GPS tagged images, does will RC automatically georeference my model and make things to scale without ground control points?
I keep getting “ill-conditioned constraint” when trying to make a measurement.
What do I need to do to make it so I can make semi-accurate measurements. I need to buy conduit to run under ground from the church to the road but also need to avoid the trees, gas pipe and fiber line in the ground.
Again, I’m no expert but I know that GPS data needs to be taken with a pinch of salt. It can work but it can also cause problems. You have to bear in mind that the GPS module of consumer drones are not overly accurate. Even for “semi-accuracte” measurements you need some kind of ground control points and then you can just as well measure the distance by hand with a tape measure. That would probably also be a lot quicker! 
If you really want or need to use the RC model, you can also scale it by defining a distance between two points of which you know the a-bit-more-than-semi-accurate distance, for example the width of the church…
Yeah I have added defined distances for indoor models. And I have already just measured the run of pipe because I couldn’t get RC to give me the measurements. This is now mostly learning and experimenting and playing with my new drone. How accurate are the results just using the drones GPS? Is it possible to use only the drones GPS? If a model is Geo-referenced when I import it on to Google Earth will automatically aligned to the map? Or is my idea of geo-referencing incorrect all together? I do this mostly do this as hobby and I’m not going to go buy an expensive GPS receiver to get accurate ground control points. I just want to learn what the software can do, the proper work flow, how much accuracy each method will provide so that I can decide when I can get away with cutting corners and when I need to be more precise.
Maybe somebody with experience in that area can weigh in and shed some light on it.
In theory, when the GPS data is very precise, it should work well enough for your purposes. But I think the problem is when the errors get too big, RC might not be able to make any sense of it.
Dear Steven,
you should expect that any drone with a consumer grade GNSS (GPS, GLONASS etc.) with max cca 66 channels L1 receiver will produce georeferenced outputs that have around 2m vertical and 5-10m horizontal inaccuracy in terms of absolute positioning in real world coordinate system. It depends on many factors so these inaccuracies may vary a lot on each project. Last time I run a georeferenced project in RC with geotagged drone imagery from such consumer grade drone and GCPs measured with RTK GNSS with 2cm absolute inaccuracy, it came out as 30m absolute horizontal inaccuracy of this consumer grade drone imagery.
In terms of relative model inaccuracy, such georeferenced model will typically provide cca 1m horizontal inaccuracy on every 10m tall object in your model. So that means 10m ±1m on Z axis. XY axis inaccuracies might be lower so expect cca 30m ±1m.
Generally if you want to to measure anything with some degree of confidence, use GCPs measured with RTK GNSS (which you can borrow for 100USD/day usually) or use tape or any cheap laser disto measurement unit. Just make sure you have correct GCP targets (chessboard pattern 50x50cm is usually best) else you may end up with difficulties finding the GCP middle point in your drone imagery.
After the georeferenced RC outputs (Orthomosaic, DSM) are exported, you can import them in any GIS (use TIF) or Google Earth (use KML/KMZ) without a problem.
For measuring distances in RC you should create Ortho projection first and than measure such distances in this tool. Measuring distances between 2 CPs in 3D view is not primarily used for measurement but for setting a model scale.
Best Regards,
Michal