solved Facet Index = null when using getClosestFacetAtCoordinates

[Bladetrick]

Hello!

I was noticing that .getClosestFacetAtCoordinates results in index == null given the parameters below.  In what situation can this occur?  The nearestMesh is valid from what I can see.  It occurs fairly often in this set of data i'm using.

    nearestpoint = BABYLON.Vector3.Zero();
    var index = nearestMesh.getClosestFacetAtCoordinates(-20.5, 115, 226, nearestpoint);  // If i change the second parameter to 115 I get a facet index. 

    if (index != null) {
        worldPos = nearestMesh.getFacetPosition(index);
    }

 

I noticed that if i switch the second paramter's value by just 2, I get an index back.  Does this mean that the distance between the given point and a mesh facet has a limit for it to work?

Thanks for the help!

[Sebavan]

@Bladetrick it would be amazing to have a PG with a repro ?

Pinging at @jerome the facet master.

[Sebavan]

and @Bladetrick this could happen if your position is outside the block partitioning use for facet.

From the doc: 

The method returns just the index of the closest facet, if any.
Actually the world coordinates (x, y, z) are internally transformed to local coordinates in the mesh local system.
If these local coordinates aren't in any block or if there's no facet in this block, it returns null.
This method can be called as many times you need, even in the render loop.

You might try to tweak the partitioning using : https://doc.babylonjs.com/how_to/how_to_use_facetdata#tweaking-the-partitioning

[jerome]

of course, there's a limit for it to work, in order to avoid to compute things when not necessary (too far from the mesh).

It depends mainly on the size of the global box (bouding box) containing the partitioning : https://doc.babylonjs.com/how_to/how_to_use_facetdata#mesh-partitioning

maybe, you could change the partiotining Bbox ratio

mesh.partitioningBBoxRatio = 1.08;   // 8% bigger than the mesh

to get a bigger bbox

[jerome]

@Sebavan

it looks like something doesn't work in the PG examples in the partitioning section in the doc because we can't see any longer the plane depicting the partitioning around the skull https://www.babylonjs-playground.com/#UZGNA#1

Maybe this has something to do with the modification of the BBox implementation in the last BJS version, not sure ...

[Sebavan]

IT is because the private fields have been replaced by public properties: https://www.babylonjs-playground.com/#UZGNA#6

[Sebavan]

I have updated the doc.

[Bladetrick]

Thanks, that's giving me a good starting point to work with.  I'm reading on partitioning more carefully and playing with the updated playground.   Question, though:

When using 

var index = mesh.getClosestFacetAtCoordinates(dot.position.x, dot.position.y, dot.position.z, nearestpoint);

on the playground I get the following error:

Line 38:22 - Cannot read property 'getClosestFacetAtCoordinates' of undefined

even though the intellisense picked it up fine.  am I doing something wrong here?

Thanks again.

[jerome]

Just check your code : the mesh is probably not loaded yet when you call getClosestEtc()

[Bladetrick]

You're right, Jerome.  I rearranged it and got it to load properly; however, my point (a sphere) is close as heck and it still gives a value of null.

https://www.babylonjs-playground.com/#UZGNA#13

Not sure what I'm doing wrong.  I'm just trying to cast a ray to the closest facet/point

[Bladetrick]

i used getClosestFaceAtLocalCoordinates and it works, though the point it came up with is not the closest that I can see.  i don't know.  I'll do some more reading/testing and let you know my results.  Thanks all, for the help/ideas!

[Bladetrick]

From the links suggested I figured out how things worked; however, I have another problem on my end.  I started a new topic here in case anyone wants to take a looksee:

Again thanks for all the advice up to this point