[Question] Collision Detection object placement

[googlesamples]

[Question] Collision Detection object placement

[ruofeidu]

Thank you Robert for your interests in DepthLab!

You are right, our anchoring may be more stable than the ARCore traditional anchors.

How is the anchoring of the object achieved

In the oriented reticle and collision checking demo, the algorithm anchors a virtual object to a vertex in a 3D space; and we used a low-pass filter on the reticle / object to make it more stable. Traditional anchors only rely on RGB features, which runs on lower-end phones that do not support depth.

Has someone a workaround to get a more stable anchoring

I don't have a quantitative comparison between the two solutions though, you may choose either way that works best for you :) The key API of our anchoring is https://github.com/googlesamples/arcore-depth-lab/blob/master/Assets/ARRealismDemos/Common/Scripts/DepthSource.cs#L367

Cheers,
Ruofei

[RGregat]

Hello Ruofei,
thanks for you response. I will take a look into the mentioned file.
I did on Friday a small test, where I placed on the ARCore Session several anchors every 2 meter based on the results of the Depth API. The result was promising, I hit almost always 2 meters at the end.

[ruofeidu]

Glad to know that! We recently released new updates, please feel free to check it out: https://github.com/googlesamples/arcore-depth-lab/releases