PhotoModeling

OldDog

Active Member
#23
friend took some pictures and i did a test unknown focal length...
handy-cam, auto detected focal length 8.10 can't be over 90% accuracy but still..
 

Attachments

EquiNOX

Active Member
#25
looks interesting. you´re saying it is buggy, is it actualy too buggy to use, or is it bearable :)
Bearable or not? it depends on individual's patience tolerance. Insight3d, crashes a lot for unknown reasons, it doesn't have autosave features. You'd have to manually save it everytime before it would unexpectedly crash. So annoying. But it's very nice program, UI friendly, easy to use.

it wont hurt to try, you'll be the judge ;)
 

Falango

Active Member
#26
It is kind of buggy, and doesn't always give accurate results. Autodesk discontinued ImageModeler after the 2009 release. IMO, it just didn't seem like it was developing enough.
 
#27
IMO, it just didn't seem like it was developing enough.
IM does its job very well. I don't know what else can be added to it.
Autodesk started project butterfly (123d at this time), so i guess because of it the discontinued IM.
 

Jpaang

Active Member
#28
okay, now that I have time I will start writing my tutorials...there will be two separate techniques one with Imagemodeler/MAX, and the other Insight3D/Blender. Expect it here in two weeks!
 
E

ettore

Guest
#31
Darn! That's good, bloodyzyy. I've downloaded it but haven't tried it yet. Where has Jpaang disappeared to?
 

Jpaang

Active Member
#33
123D Catch is quite the innovation. but reflections upon surfaces in your image ruins the mesh in my case. I think its a great tool that still needs some bug fixes.

P.S. tutorials are almost complete
 
E

ettore

Guest
#34
Jpaang, Run the images through Photoshop and kill the reflections. I'll be trying it this weekend. Will report here on my experience. Looking forward to your tutorials.
 

TwoOneOne

Administrator
#35
Jpaang, Run the images through Photoshop and kill the reflections. I'll be trying it this weekend. Will report here on my experience. Looking forward to your tutorials.
uhh good luck with that. painting out all the reflections from all photos sounds like a huge task. Plus you´re most likely getting rid of shading and surface details which will falsify your result (as the reflections do).
so far about my theory. Now if you really try it and succeed, please do share your process and result, since this would be kinda cool.

cheers
 
E

ettore

Guest
#36
I tried 123d catch for the first time this afternoon. I do not have photos of the car I want to model, so I downloaded about 30 pics from the net. I arranged the photos as if I walked around the car, however, 123d catch rearranged them according to their source - 3 or 4 photos from one website, 1 from another, etc. It did make a 3d image using several photos taken near each other in front of the car, but I was not able to stitch them together with photos taken from the side. I'll try again. I do have a set of photos taken at a concours that I think will work and I'll try them soon. I think this process has an excellent potential.
 
E

ettore

Guest
#38
At Amelia Island Concours, I photographed n Osca MT4, hoping to model it using Insight3D. I took photos, walking around about 60% of it.This morning I loaded it into 123D Catch and tried it, but was unsuccessful at producing a mesh. It is really necessary to make several circuits of the car, photographing from a low angle and a higher angle for this to work properly. In spite of this failure, I think with the correct number pf pics, this could work very well. In the meantime, I look forward to Jpaang's tutorials on IM and Insight 3D.
 
#39
123D Catch is not a good tool for car modeling...
for some non-reflective objects, which you can take picture of by your self - yes...
but cars using photos from internet - bad idea
IM of IS are the best for this purpose.
 

Falango

Active Member
#40
I would expect 123D to do a better job if you were able to take pictures of a car with a matte wrap on it (no reflections). This is how 3D scans are done on cars, because the reflections really mess with what the laser picks up in terms of surface distance. The fewer reflections and shadows you get, the better the result would be.