Support Website Contact Support Blog

10000000 Triangles are not enough.

I successfully merged two projects.  The first project was a 12 acre field that had a house on the property; the second project was a model of the house.  Separately, the triangle mesh models look great:  the topography of the field shows all the details that I was after, and the model of the house showed tons of details.  I merged the models, got everything in the correct location,  but the triangle mesh of the house in the merged model looked like garbage compared to the separate house model.  All the details were gone.  I tried many different settings, but nothing helped.  So, I looked at the densified point cloud: all the details of the house were in the point cloud.  By switching back and forth between the triangle mesh and the point cloud, I could see that there just weren’t enough triangles to create the detailed mesh.  Sharp corners of the mesh were rounded; detailed lines were blurred; and the general appearance was blah.

also interested how to deal with this. when talking about point clouds i often do not even “merge” the projects, i just copy to pointclounds together which works great. but for meshes i have no idea how to do this. also i would love to “slice” a project (clipping box) with the mesh enabled. but in fact the meshes are not filtered.

so in general dealing with meshes in pix4d isn’t perfect so far for me…

 

best, thomas

Jamie,

It sounds like you’ll need to generate the mesh of both projects separately, and then use third-party software to merge the two different meshes together. By doing so, you basically give yourself 20000000 triangles to work with instead of 10000000.

 

Thomas,

Which point clouds are you talking about, the tie points or the densified? Assuming densified, either before or after you copy one into the other, you may want to try going in and deleting some of the points and then on the menu, Process -> Generate 3D Mesh. This is the same method I would use for “slicing” a project, as the 3D mesh is generated using the densified point cloud.