There are 525 jpg nadir images in a 7-zip file placed the Dropbox link below along with 3 different reconstructed orthos and a copy of this post. The site is an undeveloped construction site that was requested to be documented before groundbreaking this coming week. The jpg’s were processed in WebODM 1.9.12 Build 55 with the following three different outcomes:
(1) Using default conditions, the orthomosaic had a “hole” in it that I initially thought was an absence of images from a problem I had with a battery swap. After reshooting, I decided that the problem was in the reconstruction and not an absence of images.
(2) So I ran the original file a 2nd time enabling PC-Geometric, and again got a hole, but a different size and shape in a similar-but-not-identical location. I concluded that it couldn’t be an absence of images if a different processing set could recreate results in areas that were previously a hole.
(3) I ran the same set a 3rd time but with custom settings that had been found by another forum member to improve edge detail in roofing in an earlier ODM forum thread. This time the reconstruction produced a complete orthomosaic with no holes!
The settings for this 3rd run were: auto-boundary:true, dem-resolution: 2, feature-quality:ultra, gps-accuracy: 8, matcher-neighbors: 10, mesh-size: 250000, orthophoto-resolution 1, pc-geometric: true, pc-quality:high, pc-rectify:true
I figured this might offer some learning opportunities, so I’ve placed the nadir jpgs in Dropbox as wells as TIF’s of the three reconstructions for anyone who wants to look at them.
This isn’t a help request or gripe. Just sharing what appears to be a data set with novel reconstruction outcomes perhaps offering potential for improving the program. Look forward to seeing if anyone finds this useful or can identify the particular software switch that helped fill in the holes.
Thought I’d have a play with these images, but unfortunately another ‘Failed: Network error’ fail to download.
It stopped at 2GB, but I tried pausing and resuming a few times, and the download restarted, but then I went out for a ride for couple of hours, only to find it had failed at some time in that period.
Trying ‘resume’ gives: ‘Failed - Server problem’
I have no idea why, but it seems any zip file much over 2GB fails.
I hate to be a PITA, but that’s not really ideal, if I select a few images and click download, it wants to zip the entire set into the 4.3GB file that previously failed.
Alternatively I have to individually click on each file and download, which is going to take hours of clicking. There doesn’t appear to be a way to download in say 3 separate files in Dropbox, unless you can suggest a method?
Gordon – We’re talking about 2 different zip files here. I uploaded the 7-zip file and Dropbox makes it’s own zip file for downloading large requests. The fact that my 7-zip is presenting troubles downloading doesn’t mean that the Dropbox zip file will fail. If it does, there may be a problem on your machine.
I’m trying it now.
I’m not sure what sort of problem it might be, as I’ve uploaded hundreds of GB of images for a big job from this computer with no problems, although I did upload as individual images, to Google One. It does allow selected groups of files to be zipped there, apparently Dropbox doesn’t.
Failed at 1.6GB. I tried resuming, but it was starting from the beginning again, and I really don’t see any point in waiting for another fail, so I’m giving up on it.
I have downloaded other zipped files of around 2GB with no problems in recent weeks.