I have a flight with an elevation difference of 30 meters from top to bottom. I’m flying a DJI P1 with the intent of making a DEM. I’m wondering at what elevation difference terrain follow really becomes necessary (in terms of gsd, not crashing into things). Presumably for differences of 100 meters you would need to do it, but 30 seems like it might be ok not to?
Obviously the least complicated you can make the processing the more accurate and consistent your results are going to be so it’s always been interesting to me that not all flight softwares have the ability but also that it took as long as it did just to get a few of them decent at it. Personally I wish following a TA was just part of the process and as we have discussed before work towards the ability to do a quick 400ft flight to generate a sparse DSM to follow almost immediately.
I think as long as all of your imagery is 2-3cm/px (1in) or better you are going to turn out good data but even moreso it depends on what kind of work you are doing and what the requirement is. I can fly near 300ft on a large quarry and still provide enough detail to provide good stockpile volumes but if I am doing a construction subgrade verification I need to get down to 0.5in/px in order to check very small percentage slopes and do flatness tests. You may see some slight variations across the map but that is just visual and there should still be plenty of points in the lower lying areas to process good data. The most important thing is to get just enough detail in the higher elevation areas. You’ll want at least 8 images per point which is usually around 70/65 overlaps. We always start at 80/65 and adjust by percentage points until the flight plan is doing exactly what we want it to do.
More specifically to your question I think 30m is way too much. Personally I think more around 10m but even that seems too much for allot of what we do.