WebODM 3 appears to be locking up, on photo arrays that 1.9.13(?), computed
FIRST 1200 photos(prev used on ODM), appeared to lock up after about 10 or so hours, the manually stopped by me after another 12-13 hours of seemingly no activity
SECOND 700 photos appeared to have locked up up after 3 or 4 hours and has been stopped(?), for the past 11 hours. I am about to kill the project and use a less than 30 or so photo project
I use a Hubsan Zino Mini SE
My methodology is somewhat questionable, since I have no auto control of pictures … I can control location, speed, direction and distance, but have no worthwhile control over photos. I can take photos every x seconds, but not every x feet(meters).
North is usually the same direction throughout(plus or minus a degree or two), speed is generally about 3 meters/sec and there is "generally ample overlap … sometimes more or less but almost always(if not always), about 80-85+ percent overlap on line and about the same laterally.
I somehow wonder if 1.9.12(?), was more forgiving of my shortcomings. In any case I can no longer run 1.9.12(?), and have lost ready access to those projects
just ran 167 photos … all 80% or better overlaps, along a road(newly paved & marked/sharp), and back & forth on 3 lines
Bounds Not Valid & Cannot process data set …
Maybe it’s time to try Lightning, but I installed Lightning, a month ago and it doesn’t seem to work …
all the processing settings are default, except resolutions was set to default(after the high resolution didn’t work)
Non-Cuda, Intel(R) Core™ i5-4570T CPU @ 2.90GHz 2.90 GHz, Windows 10pro, 16GB memory 2TB external HD …
I have the data sets … some up to 2251 photos(worked great on previous WebODM, appears to not work on latest version, as 2 smaller dataset also did not complete. I have a very short 30 photo dataset that looks a little “goofy”, along a roadway centerline. I also have the report.txt, on the projects
A quick look through my recent tasks and with high quality feature extraction and point cloud, typically my i7 3.8GHz CPU with 96GB RAM takes 30-40 hours per thousand, so I’m not sure why you can do it so quickly with just 16GB and 2.8GHz processor!
There are a number of presets, which one are you using? Fast orthophoto will be significantly faster, but it doesn’t do as much as most of the others.
I reran a series, which I ran about 2 weeks ago, with 2251 photos, and had great results on WebODM(early 2022). default settings. Levono ThinkCenter M93, 16GB, on a 250 GB HD(about 70GB available) … the processing took 17 hours
I am now running the same 2226 photos from the same file, in a new project on WebODM3. I removed extra oblique photos of tree trunks, and using default settings(default quality).
Same computer, but now running from a 2TB HD(about 1.5TB available), and I am in the 23rd hour of processing. Reported processing was mismatch by 2 hrs, a few hours after beginning, and is presently mismatched by 10 hours(WebODM is reporting 33-1/2 hours).
The marker on the processing indicator has not moved for almost a full day(24hrs). Task Manager indicates WebODM active and using about 3%, and DensifyPointCloud active and using about 50-80% resources.
The gauge indicator is showing about 22 percent completed.
What about the difference between clear/open land, mature Treed land, and first or secondary growth(which is often “scrubby”? Would WebODM3 have numerous problems with “willowy”, always blowing immature wet area growth?
Along with forcing pc-geometrics, would clearly marked ground control “markers”, or even a substantial number of undetermined/randomly placed “markers”, help produce final products in shorter time?
My thought being flight 1, very high(300’-400’), with lots of well defined markers … some with very accurate control coordinates, but maybe most with NO control coordinates(just a group of common points on a series of photos).
Then run flights of the specific areas of needed data(at 100’-200’), with all the same “markers”. This(maybe), would allow me to “skip”, that huge 50% area(of wet scrubby constantly moving, willowy limbed immature trees. which offers virtually nothing worthwhile that a higher flight would not have gathered.
This will be my next project & it’s bright cloudy right now …