Project Video Scanner

juju · 24 Jul 2015, 08:56

http://www.aboveallhouseplans.com/PVS.zip

don't forget small exe updates on that page and the Xvid codec.

Thanks Kris, appreciated. Looks like the download was finally successful from the google+ source.

Yep, got the codecs (Xvid and HuffyUV) already, looks like the HuffyUV (v2.1.1 - don't get 2.2 as it seems to be problematic) codec has 64-bit and multi-threaded variants, got those as well to test.

I have a number of questions on this system, hopefully I can figure it out along the way. I'm hoping the product is good enough to offer this as a service.

Remember to download the programs referred to in the blog post from 2 April 2014.

I wonder when the version 2 will see the light of day, in the blog post of 20 July 2015 there is a reference to it.

juju · 24 Jul 2015, 12:27

for information, PLY files can be read with Blender and Meshlab (amongst others) as well

FYI new blogpost, seems like PVS version 2 actively being worked on, seems like some of the requests are things I'd like to see as well...

kaas · 24 Jul 2015, 16:15

I realized there might be one 'small' problem.. these kind of applications create meshes with a very high polycount. Something like 100k to several millions of triangles for single scan objects. Importing those meshes into Sketchup will at least slow Sketchup down bigtime or will even fail.
Too bad Trimble Scan Explorer only reads their own file format and no commonly used point cloud file types.

edit: I see JuJu already said the same in the top post.

JQL · 24 Jul 2015, 17:20

How did I miss this post!

Do you think a polygon reducing app like Simplygon could be used with this or is it simply too much polygons for that to handle?

Krisidious · 24 Jul 2015, 18:01

You'll work in sections to either rebuild using it as a base or if you have the right tools to perhaps reduce and rebuild the mesh automatically. I would be happy just to get the rough model.

kaas · 30 Jul 2015, 13:53

Excited about these techniques, I did some tests with VisualSFM & CMPMVS.
You don't need a stereo camera and a normal camera will do just fine. Just need a lot of overlapping pictures. Take a video and export every 10th frame as a picture worked fine as well.

Tried it on the exterior of a nearby mill. Calculation of the 3d model using these programs took several hours - 123D created the 3d model in several minutes but it had less detail (64k vertices vs 1.1M vertices) and the textures looked blurry.

The end result resembles the building but it's not really useful to me (yet):

Nothing is flat in the 3d model - even if the real thing has flat surfaces. That will make rebuilding a low poly object in SketchUp difficult - how to decide which bump/point of the model to snap to??
It's hard to get the (textured) model into SketchUp. Meshlab opens the model in seconds. Blender in a minute, SketchUp's ply importer is already busy for 3 hours and I'm not sure if it will ever finish. Maybe have to split up the large model in Blender into smaller ones and rebuild it in SketchUp?!

Also:

reflective surfaces don't work well (found that one out 'scanning' a table with some objects on it)
the software needs 'anchors' to match the individual frames. For more large uniform planes (ceiling, white walls etc) the software can't find matches so that results in large gaps.

Still, it's fun to see the software build something 3d out of pictures.

Krisidious · 30 Jul 2015, 17:56

You're blowing my doors off...

kaas · 31 Jul 2015, 05:02

@unknownuser said:

You're blowing my doors off...
I don't think thats a good thing or...? Kristoff, I'm sorry if I hijacked your thread. I just thought, these programs and their results were related and thus relevant.

juju · 31 Jul 2015, 11:15

@kaas said:

Excited about these techniques, I did some tests with VisualSFM & CMPMVS.
You don't need a stereo camera and a normal camera will do just fine. Just need a lot of overlapping pictures. Take a video and export every 10th frame as a picture worked fine as well.

Tried it on the exterior of a nearby mill. Calculation of the 3d model using these programs took several hours - 123D created the 3d model in several minutes but it had less detail (64k vertices vs 1.1M vertices) and the textures looked blurry.

The end result resembles the building but it's not really useful to me (yet):

Nothing is flat in the 3d model - even if the real thing has flat surfaces. That will make rebuilding a low poly object in SketchUp difficult - how to decide which bump/point of the model to snap to??

It's hard to get the (textured) model into SketchUp. Meshlab opens the model in seconds. Blender in a minute, SketchUp's ply importer is already busy for 3 hours and I'm not sure if it will ever finish. Maybe have to split up the large model in Blender into smaller ones and rebuild it in SketchUp?!

Also:

reflective surfaces don't work well (found that one out 'scanning' a table with some objects on it)

the software needs 'anchors' to match the individual frames. For more large uniform planes (ceiling, white walls etc) the software can't find matches so that results in large gaps.

Still, it's fun to see the software build something 3d out of pictures.

can you share some details?

which cameras used (also, FOV lense specs)
what resolution
size of *.ply files
how long did these take to process
what is you system specification
anything else relevant that you have not previously mentioned...

kaas · 31 Jul 2015, 13:05

Hi Juju, here's some answers:

which cameras used (also, FOV lense specs) -> I tried a Olympus SP600UZ and that worked fine. For the mill I used a 1080p video, made with a Samsung Galaxy S3 - used Blender to convert every 10th frame of the video to a single image.
what resolution -> 1920x1080
size of *.ply files -> high quality file was 95MB, low quality was 8.9 MB
how long did these take to process -> I think it was four hours for both programs
what is you system specification -> Win 7 X64 - 3.4Ghz (I5 3570) - 8GB - GTX560TI
anything else relevant that you have not previously mentioned -> its fun. Just try shooting some pictures and feed it into those programs and off you go.

to get you going, maybe have a look at some youtube videos. I don't have the link at the moment but there are some showing these two programs after another.