There are a few user interfaces for process, managing, and serving drone and other photogrammetry data, from DroneTM, WebODM, DroneDB, MICMAC and others[1] on the free and open source side to Pix4D, Agisoft, etc. on the proprietary side.
What are the biggest pain points / opportunities, missing features, cool features you see in photogrammetry / drone image processing software? What do you like / dislike / cannot-live-without in the above software?
I have personally used several programs. Since I use ground control points (GCPs), it is very important to me that the process of identifying and marking them is simple and as automated as possible (e.g., preliminary point alignment / automatic or semi-automatic visual detection) it is possible that modern algorithms can bring a lot of innovation here. Another issue is full process automation, this is welcome so that the entire processing workflow does not require intermediate steps. Another issue is the mapping of buildings, their roofs, walls, and edges, as well as numerous artifacts - this remains a weak point of UAV photogrammetry.
+1 on making GCP handling as painless as possible. Commercial software is way ahead in this area specifically.
In general though I find the question of what works and doesn’t across all platforms is so broad I personally get jammed up trying to answer with any specifics. I appreciate this is a good time to revisit the foundations and assumptions in the ODM project but it might be easier to get feedback if we can break up the question into specific areas and ask over time (e.g., prep, processing, troubleshooting, 2d, 3d, docs, community – dunno the best list; just spitballing).