What is the latest best way of converting a mesh of a vintage tractor brought in from Agisoft metashape into surfaces, I then intend to construct the shapes with new curves, extrude surfaces etc, its just going to give me the shapes I need to construct in normal Rhino ways.
lay cross sections through it, create profiles, sweep etc etc.
I export from Metashape as a .obj file which I import directly into Rhino as a mesh and save in a .3dm file. I scale and align the mesh as needed and save.
Then I open a new file in Rhino with the appropriate units, tolerances, grid, etc for modeling the object. The file with the mesh is attached using Worksession (Rhino for Windows only). I don’t convert the mesh to surfaces but use the mesh as the reference geometry.
I study the mesh and think about how I want to create a NURBS model terms of surface arrangement, and lines, curves and points needed to create the surfaces. Only after having an initial strategy for creating the model do I start the model. I set up an initial layer structure, and then start creating geometry.
I can use the mesh its made to act as a basis to create everything using normal geometry creation tools. Its given me enough to recreate it all in surfaces curves etc. I certainly would not want to use any of the mesh as a final item itself.
I am doing likewise. Thanks for the .obj file type advice.
Now I didnt know of that Worksession method as this is my first time at this. I need to get cross sections, have my tools snap to the item, etc, all must latch onto it, not end up going through like some hologram ! So having it as surfaces seemed the way to go, everything I did would behave as I wish, redraw entirely and export out as a set of Rhino layers to the final file where I can add in anything else from photos and adjust build to suit 3D printing at different scales.
So you use mesh as reference geometry, neeed to know quite what that entails, do all your tools snap to it as normal, or do you mean you get it into ‘right’ view and ‘trace’ visually a shape ? then front view and trace then use the traces to extrude or sweep etc ?
If for example the cab pillars are being traced in a side view I wouldnt know nearside from offisde so might end up tracing one and in fact have traced both in a slight offset way so end up thicker then reality, for example, whilst a section through one and a line along its outer edge, both then refined, would give me an accurate rail and profile for a sweep.
Ditto, I know in my minds eye how I intend to construct from scratch the tractor.
The tyre treads I will need a section of the front as its simple rings around circumference, and rear has zig zag tread so with mesh as surfaces I can trace along the edges of the raised parts, refine and then array around centre a final zig-zag. I couldnt do that without having a surface to follow and snap to.
Whenever I work with meshes I cant snap easily to them as it snaps to anything front or back and its a ‘target rich’ environment to snap to, far too many. if a surface I can have it follow surface as if a magnet on a compound curve steel shape.
is there a video anywhere I wonder of your reference geometry method ?
sent you a PM by the way the other day as metashape is puzzling me in some areas. forum nothing like this one, wait and wait and not progress, no grid for alignment is a pain. no undo if happen to rotate model independent of grid another, knowing what method and when and why , deciding if mask is required etc and which method to use, is a BIG need. My tractor came out very good, glazing covered in reflections didnt get in the way at all, then since the PM the cab interior has now rejected quite a few images, and they were accepted in the first chunk and gave me the cab interior ! I like mask from depth map, far less bits. Can one align pics and then go mask from depth maps and skip all the other stages I now wonder.
Contour works with meshes.
Section works with meshes.
OnMesh is a one shot Osnap on a mesh.
PersistentOnMesh / POnMesh is a persistent Osnap on a mesh.
The above OSnaps are available by holding down the Ctrl key with the cursor over the Osnap control bar. Rhinoceros Help
ClippingPlane works with meshes. I use it frequently
There are numerous Mesh specific commands including MeshIntersect
A mesh can be converted to a polysurface using MeshToNURB. It creates a new surface for each mesh face. This is not recommended for large meshes.
hope these are in V5, win10 rebuild is ‘around the corner’ now.
So from such I take it you are tracing over the mesh, this is what reference geometry means, having curves etc latch onto the meshes, and with the massive file and lag that a large item would see happen with Mesh2NURB, this is the way fwd.
To see a video of someone doing such would be great.
ClippingPlane very useful, thats good, I tried that for the first time on a normal Rhino object and couldnt get it to do as per video but will try again.
I model from scan data for a living,
I find working with pointclouds far easier than meshes in Rhino, I generally use CloudCompare (Free) to extract a pointcloud from the mesh, then work with that.
I use the following techiniques;
More recently, I have been using Rhino 7.0 to convert the input mesh to a Quad mesh, then convert that surface to SubD’s …it works beautifully.