We’re getting closer. I am starting to get some pretty good results with the 3D Mode, but… using Windows Explorer I arrange the tiles so that I could easily take out half of the images. I think that if I ran two flights, one of them much closer to the building I might get an even better result. The pattern on the CMU is almost impressive.
What is everyone else’s experience with trying to create better 3D models? I have stated this before, but this is just something for the client to look at who knows nothing about what we are processing. If I wanted to show them an actual 3D model I would use the BIM, but who has the time and computing power to render all of this…
Sculpting good 3D models with a drone is part art and part science at this time. Gaining experience in how best to apply the techniques to each unique subject matter to optimize the models is the best approach.
Along with the Nadir grid photos, getting good sequential overlapping oblique images without the sky in them is key. Many people use the POI feature in the DJI Go app or fly manually to get good overlapping oblique images.
Sometimes uploading just the obliques can come out better than using the nadirs and obliques.
These documents give good insight on how to go about it.
Great, thanks Gary. The example above was shot with no Nadir and 2 orbits in Litchi.