Meshroom – Open-source photogrammetry software
51–60 of 77 posts
Re: Meshroom – Open-source photogrammetry software
#52Earlier quoted context omitted.
They have binaries? I must have missed that, the only install instructions I saw were for building from source.
Yep and it's using the GPU i just noticed. Binaries are linked here: https://alicevision.org/#meshroom It's really nice! I read more about the project and it's apparently always been open source. It looks like a high quality project run by some researchers with university and EU backing. I've been trying to reconstruct large outdoor environments for robotics simulation and this might be just what I need!
Re: Meshroom – Open-source photogrammetry software
#53Earlier quoted context omitted.
Yep and it's using the GPU i just noticed. Binaries are linked here: https://alicevision.org/#meshroom It's really nice! I read more about the project and it's apparently always been open source. It looks like a high quality project run by some researchers with university and EU backing. I've been trying to reconstruct large outdoor environments for robotics simulation and this might be just what I need!
Nice, I'm working on something similar but mostly indoor. Are you using SLAM for the robot itself?
If you can share info on your project I'd love to see it!
Re: Meshroom – Open-source photogrammetry software
#54Re: Meshroom – Open-source photogrammetry software
#55I'd be very interested in a comparison Meshroom vs. RealityCapture (which seems to become the standard in non-free photogrammetry).
Agisoft Metashape is pretty good as well. Results are comparable to RealityCapture (depending a bit on what you want) but personal license is only 179$. I tried all sorts of free photogrammetry softwares way back but after I tried Metashape I decided to buy a license. With 32 GB of memory it's great. I'm taking photogrammetry models as a hobby - here's a recent one made with Metashape https://skfb.ly/6S6NR
Re: Meshroom – Open-source photogrammetry software
#56Earlier quoted context omitted.
Nice, I'm working on something similar but mostly indoor. Are you using SLAM for the robot itself?
Nice. I don't have a clear plan for localization yet. Right now I am working on deep net based trail following. [1] Eventually I will probably do some kind of vision based localization. [2] If you can share info on your project I'd love to see it! [1] https://reboot.love/t/new-cameras-on-rover/277/4 [2] https://github.com/ethz-asl/hfnet
Re: Meshroom – Open-source photogrammetry software
#57Earlier quoted context omitted.
Yes, we're getting there. Of course a > $50k laser scanner will still outperform it in terms of accuracy but the new generation of depth sensors (e.g. Intel RealSense L515) is really quite good. We're going to announce support for this very soon and will also put some datasets online for comparison.
Hey, that sounds potentially relevant to my employer. I'm a small fry in our org but would like to get some of our heavy hitters to take a closer look at what you guys have. Who should we contact (if folks find this relevant)? I prefer not to link my private online persona to my professional role but the page https://www.dotproduct3d.com/workflows.html has several references to our products already so you likely are…
Re: Meshroom – Open-source photogrammetry software
#58Has anyone tried it? I've been trying for a while to find a way to 3D scans stuff to 3D print but the quality of the final file has always been so crappy that I didn't even want to waste time and filament printing it
I've used Scandy Pro[1] a few times and was surprised by it's resolution and ease-of-use. Only downsides are it's iPhone only and you get 1 free scan/24 hrs.
Re: Meshroom – Open-source photogrammetry software
#59Has anyone tried it? I've been trying for a while to find a way to 3D scans stuff to 3D print but the quality of the final file has always been so crappy that I didn't even want to waste time and filament printing it
From memory, 3DF Zephyr had the best quality, at least in my limited experience. It was also very easy to use to clean up unnecessary points, which is an important step for high quality output.
Meshroom to me feels like a research testbed. It's not a "product" that's ready to consume. The user interface especially is an unmitigated disaster.
For example: What are the little connected boxes at the bottom, and why do I care? If I need to care about what they are, why many of them have text that is cut off or contracted with ellipses? E.g.: "dept...lder".
Next to this is a node property editor where the labels line wrap and can't be resized to fit (despite there being plenty of room). The text next to it aligns to the right, so I see only the random suffixes of the files, not the actual path prefix. It's literally a window showing me a bunch of random numbers I don't care about.
I could go on and on about how unusable it is, but it feels like kicking a puppy.
The real issue is that overall it's quite slow. I suspect because despite requiring a GPU, they're not very efficient with it. The concurrent CPU usage is high, yet progress doesn't seem to go forward very much. I rarely see the GPU utilisation exceed 20% or thereabouts. Other tools breeze through the same data sets in far less time.
Re: Meshroom – Open-source photogrammetry software
#60They ever get around to HIPifying the CUDA bits?
I very much second this, as I hate relying on proprietary software of any sort for stuff that's essential, has a steep learning curve, or that I am going to use on a regular basis (plus other criteria). GPU drivers match the first and last point.