even though there are quite a couple of postings about incomplete output or input images being ignored, this thing seems to be a bit different, at least I haven’t found any duplicate:
First, let me share with you the data I’m talking about:
Input set of images: https://drive.google.com/drive/folders/1np1w3S9B-gCciH7F_H1NI9vCPY5Pnh3j?usp=sharing
Task output: https://drive.google.com/file/d/12NBpYoi-nfKVjNmHxkzIKysbqFIEcgSd/view?usp=sharing
I’m using ODM through WebODM installed from git and updated (webodm.sh update) just recently. My server OS is Linux 5.7.0 AMD64, Debian Bullseye (11). I installed WebODM according to this guide: https://docs.opendronemap.org/installation.html#linux
My (client) browser is firefox 77.0.1 amd64.
I have tried with original as well as downsized images (2048px), and several different parameters (Default, Forest, Fast Orthophoto), the result is always the same:
The dataset comprises 301 5k images captured at constant altitude (84.5m) and angle (-90°) using a DJI P4P.
The capture area has a “neck” to it, i.e. it’s two bigger areas connected by a narrow bridge.
The output however only consists of the area behind the neck. It seems all images are somehow processed, but ODM seems to yield two separate chunks and gives me only one of them (just a guess).
What’s going wrong here and is there a way to fix it using the given image set?
Thanks a million and best regards!