Wondering if anyone can help.
I’m after software that can do photogrammetry, however I dont want to use the automated SIFT type methods. I would like to be able to load multiple photos, and actually mark points or lines on each photo manually.
Does anyone know of any software that is capable of this?
Hi. If you mean a user annotates / marks some frames and the rest are inter/extrapolated, I can’t think of any.
Some recognise fiducial markers placed in the shot scene. You could try annotating (in mspaint / gimp / ps) a frame sequence with some garish / contrasting colors and testing the output (app agnostic). A contrasting pixel per feature point should do.
Photoscan lets you add manual tags.
Hadn’t thought of that - I might give it a try. I don’t need the software to actually create a photo realistic mesh - just a general idea of the proportions that I can use as a template for a model I am building.
Thanks - i’ll take a look, but I probably cant justify the cost of commercial software for what will likely be a one off use case.
I see, in that case techniques like this may be enough:
There’s also fSpy, BLAM, pvr, etc tutorials / addons.
You can also add manual control points in Reality Capture (it’s the best photogrammetry software, period).
If that is all you need then you can try using Blender’s internal tracker.
Machine learning solution…
Pixel2Mesh: Generating 3D Mesh Models from Single RGB Images
Cheers i’ll take a look
I think this addon could really help you : https://www.blendernation.com/2019/08/15/blender-photogrammetry-addon-v1-0/