Don't know anything about this except that they have a free version.
https://community.acute3d.com/
I only had a look and thought it might be interesting for some of you.
Recommended at least 16gig of ram iirc,so not for everyone me included.
Interesting.I saw their site ones ( http://www.acute3d.com/ ) and didn't bother to try their software because
it was very expensive and even for a demo you had to contact them...
This is a new version and new site for the basic edition.
Only for NVidia cards for now but they say other cards will be supported within a few weeks.
I tried this with the old photo set i used for 123D catch.
Not bad for a quick test.
Someone like Ben could say more about this program.
What's obvious here is that you need to light the object from all sides; the shadow of the right front foot is interpreted as the black background.
Yes a more proper photographing procedure is needed.
Not exactly he place I'd expect to keep up with this news but good to know. Very interesting. AFAIK Aerometrex use the grown up version of this for their aerial work. There has been some comparison of Acute3D on the Agisoft forum but I'm going to have to test this for myself. Whilst Acute3D is behind 123D Catch I suspect anything that gives you more control over settings has the potential to provide improved models. I'm currently doing some testing of hardware for this sort of stuff and I'll throw this program into the mix. To think I get paid for this now :)
Quote from: Dune on August 25, 2014, 02:23:20 AM
What's obvious here is that you need to light the object from all sides; the shadow of the right front foot is interpreted as the black background.
Yep. Deep shadows are a killer regardless of the software. If you're converting from camera raw make sure blacks are set to 0 and reduce the contrast (-20 to -40)
Won't be testing at home though...
[attachimg=1]
hmmm... Running the basic benchmark test for Photoscan through it (http://www.agisoft.ru/tutorials/photoscan/02/). 50 x 12 megapixel images.
1. Only aligned 33 images
2. exceeds 500 megapixel limit for exporting obj (both free and basic versions)
3. standard version of Photoscan is ~135 euro compared to 199 for basic version, but has no restriction
My test was very slow compared to Photoscan (~30% longer) but that may have been a result of my choice of settings. The output model is highly optimised which may be OK for buildings but not necessarily for objects.
https://community.acute3d.com/more-challenging-project (https://community.acute3d.com/more-challenging-project). There's a big jump from "basic" to the "advanced" version which on the surface looks to be the equivalent of Photoscan Pro, although still has an input image limit and an annual license fee (or very expensive perpetual license). I've run projects with Photoscan Basic that would have required the Ultimate version of Smart3DCapture. Only limits with PS are the available computer resources.
500 megapixel in regard to an obj ? what does that mean?
I have no idea,if it said 500mb or 500k polycount ok,but megapixel?
Quote from: j meyer on August 25, 2014, 12:13:21 PM
500 megapixel in regard to an obj ? what does that mean?
I have no idea,if it said 500mb or 500k polycount ok,but megapixel?
Just a guess. But is that saying the photos used are limited in the pixels? so no image with more than 500mp can be fed to the software?
It's the total area of imagery to be processed. There are two limits on the free version, total amount of imagery it will process (5 gigapixels) a limit for imagery processed that it will allow you to export a model (0.5 gigapixels)
The benchmark scene I'm using is 50 x 12megapixel images = 600megapixels (0.6gigapixels), so no object export available. Fewer images = less overlap so not really an option. Smaller images = less resolution. I'm working with a 21mp camera and there's no way I can create a 3D model with only 23 images. It will still generate a 3D model but you'll only be able to export it in their proprietary format, teasing you to purchase an upgrade to the basic version.
From there the next limit is the total amount of imagery that it will process at all. e.g. I'd be limited to projects ~230 images before having to shell out 2,500 euros for an annual license to double that limit. Photoscan has no limits, is cheaper, it's only a functional difference between the standard and pro versions and is (IMO) better for objects. But there are free versions of both programs so test them out for yourself to be sure.
Please tell us ben,
Of the three or four softs you have been testing, which is the simplest vs quality?
What I mean is, to me (so far) 123 catch looks the easiest to use. But are there limitations that would make a harder to use soft a better option?
So If I think 123 is easy, you would say sure, but your results would not be as good as...
I have some objects I have been collecting over time to use this tech on, but I have other things to do first. So I just let them sit until I see what you and others post, to help me decide what is the best way forward.
All of the objects I have are small enough to move and place by hand, table top size stuff. If that makes a difference to you?
I have been using Photoscan and I think there is a serious limitation on using photogrammetry to create 3D models. Areas with uniform color don't give the software enough information to create the necessary geometry in those areas and it is a serious limitation to have to scan only highly patterned models. So I have decided to wait for an application that uses the Kinect 2 camera. Kinect 2 for Windows is already out but I am waiting on Newtek to release support for this camera in NevronMotion, their MoCap plugin for Lightwave. Because of the way the Kinect camera works it should be able to accurately create 3D models from even uni-colored objects; something that photogrammetry does poorly.
Of the photogrammetry software I've tested Photoscan is certainly the best for quality of output and ease of use whilst still providing access to an assortment of controls and tweaking for more advanced users, including scripting for automating more complex workflows.
At it's most basic level you simply drag and drop images into it and then step through the Workflow menu in order. Each stage has a few basic options to adjust quality/speed. You can also queue up all of the steps in one big batch. There is also a reasonably active and very helpful forum, just like here.
The Pro version provides georeferencing support and the addition of control points, but the retail price is up there with the rest of them at an arm and a leg (and maybe even your first born child) unless you qualify for the educational discount. There's not too much software in this field that won't cost you $3000+, even for just the point cloud processing.
Things Photoscan doesn't do so well:
1. With the standard version there is no control over scaling so you have to size your model in another application (easy enough in Meshlab which is free)
2. UV Mapping is not very efficient (but the photo blending is pretty good)
vs. 123D Catch (I've repocessed a lot of projects now)
1. Better image alignment out of the box
2. Greater accuracy of model, fewer artefacts
3. You get the maximum resolution your images are capable of providing
4. Size of project only limited by your computer (and you've beefed it up for TG, right? ;) )
As PabloMack rightly points out there are limitations to photogrammetry, just as there are limitations to laser scanning and most other methods of 3D data capture. For capturing the geometry of smoothly coloured objects we use the same trick for laser scanning glass/reflective surfaces... coat the object with a fine powder. We use 21mp cameras and I find that it doesn't take much fine detail in a surface to register, but it is always a potential problem.
Nearly all methods of 3D capture use triangulation so narrow cracks/crevices always present a problem. The larger museums etc... will use a combination of technologies to combine the best data/capabilities of each method.
For a lot of stuff that's come across my desk, a handheld scanner would be nice but from what we've seen you're looking at $20-30K for decent quality and $30-50K to add coloured textures. I've also tested various options with Kinects and while they have their niche the resolution is still a limitation (PabloMack: have you tried http://reconstructme.net/ (http://reconstructme.net/)).
You can also get some good results out of the free SFM applications but they require some relatively complex set up for best results so ease of use puts it out of the range of casual users.
Thanks for testing and spreading the info Ben,much appreciated.
I still don't understand the mega/gigapixel thing in regard to a model.
How big is a model with 0.5 gigapixels in mb or polygons?
Anybody know if ProForma is still aound ? Looks like some interesting tech.
http://www.shapeways.com/blog/archives/332-Interview-with-Qi-Pan-about-his-Webcam-3D-scanner-proForma.html (http://www.shapeways.com/blog/archives/332-Interview-with-Qi-Pan-about-his-Webcam-3D-scanner-proForma.html)
Quote from: j meyer on August 26, 2014, 10:49:57 AM
Thanks for testing and spreading the info Ben,much appreciated.
I still don't understand the mega/gigapixel thing in regard to a model.
How big is a model with 0.5 gigapixels in mb or polygons?
It's not the size of the model, it's a limit on the amount of images that are used to create the model expressed as a total pixel count. Too many images and it won't let you export a model in a reusable format, but too few images (or low res images) and you lower the quality of model. e.g. 8megapixel iphone images = 62 images max (500/8) which is probably OK for a model, but with a 21mp camera your down to 23 which is barely enough for a single row of images around an object.
Not for no reason, but there are cameras where you can operate in video mode and each frame is high res photo. So for example 24/30 photos per second. at least one of the RED cameras is good for this.
And in the iphone (others too?) you can just pan your camera by hand and get a pano image.
So my question is cant we in any of these softs just use video to get the capture? I mean rather than move around an object taking however many shots, is there a way to just use a video mode and walk around the subject a couple of times. And lastly just up load a single file to the soft?
effectively using the camera as a scanner of a kind?
Seems like thats how it should be. So I imagine someone out there is working on it.
Actually, wouldn't that sort of discribe mocap?
I suppose you can use reduced image sizes, as long as the basic shapes are recorded. You won't need 3x3k pictures, but maybe 1x1k will do just as nicely.
It all depends on your end usage. For making models i probably wouldn't go below 8mp, but once you shoot the same object at 21mp it's very hard to go back. In theory using video is possible, but compression artefacts will create noise in your data and from a software perspective it will still be treated as a series of stills. There is a lot of processing involved with photogrammetry so it's not the sort of thing that can be done on the fly at significant resolution. The images also need to be sharp so excessive camera movement can cause problems, and before you suggest increasing the ISO, that also introduces noise.
I've done a few reconstructions from short HD video clips pulled from Youtube and it does work to a reasonable extent.
Motion capture is essentially the same thing, as is determining the camera positions in a video for matching perspective in renders is essentially stage 1 of reconstruction via photogrammetry.
Since Smart3DCapture is essentially the same software as 123D Catch I'd just use 123D Catch out of the free options as it's less limited for exporting
Speaking of HD video resolution.... Here's the Melbourne CBD recreated from 300 screen grabs of Google Earth (couldn't get the OpenGL or DirectX capture tools to work)
Thanks again Ben!
300 screen grabs :o I would say (but I'm no expert) that for this sort of resolution only 10-15 would suffice, half from a lower altitude circling and half from a higher altitude.
Overkill for the sake of being sure. 60 probably would have sufficed. I "flew" 5 grids over the city at different camera orientations: Vertically down, and oblique facing N, S, E & W. 10-15 would have been enough for a photo texture, but there wouldn't have been enough overlap for a 3D reconstruction. Image quality and inadequate overlap are the main problems that people have.
I ripped the images out of the first of the Photosynths on this page: http://www.xrez.com/case-studies/cultural-heritage/sunken-treasures-of-the-nile/ (http://www.xrez.com/case-studies/cultural-heritage/sunken-treasures-of-the-nile/) to see if I could do a 3D reconstruction from them, and it worked quite well except for a few small holes where there wasn't overlap between images. But these images were captured to provide a photo texture for applying to laser scans.
If you know anyone who has a Fuji Real3D camera, Photoscan makes use of the stereo jpeg format (I picked mine up secondhand, and the 3D screen's really cool for when you're taking group selfies at the pub, as well as it fits in your pocket). You also get much better results if you take the time to mask your subject, and I do like the reconstructed cameras being in the exported Collada file as a camera texture reprojection safety. Also, for a giggle, try rendering a TG scene as a walkaround and run it through Photoscan (a logical progressive walkaround seems to be fairly important to Photoscan). The results are actually pretty good and you can end up with a fairly nice UV mapped model with baked lighting.
The latter is an interesting idea...
I just remembered doing a test on a terrain without shadows and flat lit, which also worked out not too badly because there's still detail in the surfaces. And make sure there are no reflections or specularity.