Align modeled objects to imported mesh objects

[daviesj]

Hi,

I've been banging my head against this problem for two days now and I'm well and truly stuck.

I am trying to create a model of metal pins placed into a bone acquired from a CT scan. The pins consist of a set of two 1/8" steel pins (3.175mm) passing through 10mm hollow steel spheres. I need to maintain the relative location and alignment of these pins/balls as they will be used to attach a printed jig that will allow for accurate cuts in the underlying bone.

So far, I have taken the original scans in meshmixer and used the fit primitive tool to align a modeled sphere and cylinder to the pins. I have then imported these 4 modeled objects into fusion 360. I can remesh these down and make them solids to create a model around them, but it is a bit slow on my computer to manipulate the model when I do this (recent macbook pro). It feels like there should be a better way to import a meshmixer primitive into fusion than as a mesh... any suggestions?

Aside from me wanting to work with models instead of meshes, My problem is that I need to create a cutting plane on the midline of both spheres/cylinders so I can model a bivalved clamping section around them, but I can't seem to find a way to align anything with the mesh derived solids. 

So my question is, is there a way to accurately align modeled spheres and cylinders to the imported mesh in Fusion360 (Ideally) or is there a way to figure out the centerline of the mesh derived cylinder?

I have attached a screenshot of the mesh model and the .obj file.

Any help is greatly appreciated!

Cheers,

John

[laughingcreek]

Is there a reason you can't model the pins and spheres in fusion to begin with?  This operation would be incredibly simple just using native fusion geometry.

[daviesj]

Yes, I could model them directly in fusion, but I have to maintain the spatial orientation of the scanned pins and spheres.

Is there a way to align a modeled sphere and pin in fusion to the mesh? I used a best fit option in meshmixer but I haven’t come across anything like this in fusion

Thanks so much,
-J

[laughingcreek]

Sorry, I misread your original post.  I thought just the bone was scanned geometry.  There isn't really a function like that in fusion.

So i understand the accuracy issue, and why you want to use the fit primitive functionality.  Something to consider is that meshes are imprecise models.  And ever time you perform an operation a a mesh in mesh mixer, you will lose a little bit of accuracy. (it's like recording analog audio over and over.  you lose a little bit of quality each time).

It may seem counter initiative,  but you  may actually gain  some accuracy by bringing  the original mesh into  fusion, and modeling the pins and spheres there by eyeballing.  You can take create sketches of cross sections of meshes that are very accurate (there's a specific tool for this, with curve fitting functionality).  Using these to "eyeball" the geometry will give you pretty good accuracy.  Instead of introducing some amount of unknown error with each mesh operation, You'll only introducing error once. 

[etfrench]

This thread may help.

[daviesj]

Thank you so much for your help,

I didn't think about multiple errors introduced with multiple mesh manipulations 😞  I was playing around with this last night, and was able to create cross section sketches with the spline tool... is there a way to select the resulting spline and find it's center? (I'm still trying to establish a centerline plane) Or should I be trying to best fit a circle or an elipse to the mesh cross section?

Cheers,

John

[daviesj]

Thank you etfrench,

That arbitray points script looks really interesting - I will also look into that and see if I can make it work.

Cheers,

-John

[daviesj]

Thanks to the help of both of you, I have figured out how to do this. It is a bit cumbersome and takes about me about 5 minutes to do each pin, so if there are any shortcuts you can recommend I am all ears as I need to do four of these/model.

So the way I did it was to select the mesh of the pin and take two mesh sections both proximally and distally. I then fit an elipse to each mesh section with the mesh->fit curve to mesh section tool. I then create an axis between two points and select the centerpoints of both elipses. This gives me the long axis of the pin. I then create a plane at an angle along this axis 
(Pick 0 degrees) and select this plane and draw a line on the plane between the two points. I then create a pipe with the base being the line I created. (It helps that I know the exact dimensions of the sphere and pin in real life!) To model the sphere, I again take a mesh section of the sphere as it crosses the plane I created for the pipe and fit a circle to the resulting section. This gives me the centerpoint from which I create a sphere.

P.S. I think this works well as these mesh objects were created primitives from meshmixer. I don't think it would work well if I used the original scanned data as it is pretty chunky with a fair amount of metal artifact.

Cheers,

John