Well what’s weird is I have been using the DJI pilot app, and the geotags on those photos compared to the geotags on the photos with this other app look the same. But when I put the DJI app photos into Pix4D it works fine just like I’m doing it.
Ok correction. I don’t guess I have done a project since I updated my Pix4D software. Its doing this with photos taken with the DJI pilot app as well. Figured I would test it to see and sure enough its doing the same thing. So it must be some setting or something that has changed on version 2.0.83.
Yes, it was changed. I forget the exact version number, but there was a bug that was fixed in a later version that was not applying the geoid separation correctly. The behavior you describe is consistent with my testing using our DJI units, The P4D file corrects this problem. Most of the higher end mapping systems report LL and ellipsoid height in the EXIF data then when you apply the geoid separation in P4D you get correct MSL heights. For assumed or local coordinate systems I just use the Basic editor and run the whole project.
So is there another version since the 2.0.83 that fixes this problem?
I may try to go back to the 1.4 version and see if it works.
The problem is fixed and the current version works correctly. The bug was in a previous 2.0 version that was out earlier. I don’t remember the version. For correct MSL elevation the height of the droe should be reported with the ellipsoid elevation not the height above the ground. Then when you use one of the supplied geoid models or the geoid separation you will get the correct elevation at MSL within the tolerance of the GPS on-board the aircraft.
So how should I go about processing when using a arbitrary coordinate system? When you start a new project and you import your photos, and choose 3D map, then it gives the option to choose a coordinate system. Should you choose anything at this point or leave it alone?
You read in you arbitrary coordinate system and then pick the photos using the basic editor. The gotcha is tha you have to know where the control points are located. The best way is to paint a number by the GCPs. After you pick the first point, you can sort the images by nearest which helps in picking. After all your GCPs are picked you can run all three of the processes. I haven’t tried it the way you explained in your earlier post. The above is the way explained in the P4D knowledge base and it works well. Definitely not as easy as the Ray cloud editor in my opinion, but it does work.
https://support.pix4d.com/hc/en-us/articles/202560239-How-to-include-GCPs-in-the-project#label3
I think I may have figured out what the problem has been. Does the image dimensions play a big role in it? I think that may be what happened. It was changed somehow.
The Map Pilot app is taking photos in a 4:3 ratio, where as the DJI Pilot app you can take them in a 16:9 ratio, which works in Pix4D. So unless there is a way to change a setting in Pix4D to be able to use images taken with a 16:9 ration I will have to go back to doing manual flight paths and photos. Or if there is a way to change this setting inside the Map Pilot, but I don’t see any way to change it.