Community Tip - Did you get called away in the middle of writing a post? Don't worry you can find your unfinished post later in the Drafts section of your profile page. X
Hello, Using vuforia model target ,I want to detect a U shaped test tube and overlay some virtual objects in order to explain the chemical process happening inside the test tube.
But, as the test tube is transparent, its difficult to create a 3d model of it. Can anyone help me with this?
Hi @AC_9685392
In generally transparent object and or shining object are not appropriate to be used a model target. We need very sharp edges or tangency edges. Especially with object where there are no real edges but only tangent /silhouette edges ( tubes , cone, cylinder etc. ) and transparencies this will lead to problems
I think in such case we can use different models for target and for the augmentations. When for example you need to display some transparent tubes (test glass, tube) you can use for the target e.g. the non-transparent mount assembly etc.. In this case the 3D model (model widget) and the target model could be a different things.
The following information could be in this case helpful
Model Target Generator User Guide:
Model Targets Supported Objects & CAD Model Best Practices
Optimizing Model Target Tracking
Another appraoch is to generate a target from the stand. As @RolandRaytchev said transparent parts are difficult for computer vision to track.
We did recently release "car mode" for large shiny objects but unfortunately this won't fit your use case
OK- car model , this seem to be a new feature in the model target widget sinse 8.5.13. Thanks for the info , this a great enhancement!
thanks for the idea but the stand looks big but the object scanning space where we place our real object and create 3d model is very less.Can i know how to handle this problem.
When i place object inside the object scanning space ,its big and i guess the entire object can't be scanned.
did you ty to scan only a small object / also inside the transperent object - as already mentioned the scan model /model target could be complete different from the model which you want to augment / 3d Model - for this you can add a second auxiliary model widget which is only used to generate the model target (but it will not be displayed later)
Thanks for the idea and do you have any documentation or the video demos about what are you talking about as I feel it would be much easier to understand.
Unfortunately, no. I did not have this particular use case. I had a case where I used only a small part of the whole model - this part was simplified cut out and then I used it in a second model widget to generate model target. If you have a sample data I could check for you.
Possibly another option - you can use an image target – locate the an image which is near to the geometry some kind of “camouflage” picture located near to the model or inside the model