XYZ into Blender 2.5

Support Other Software
1

Hi,
is it possible to import terrain coordinate info (as comma separated values) into Blender 2.5? When imported I need to skin the points with a quad mesh.
Any insights would be much appreciated.

Thanks

James

2

It’s possible, but you’d need some python scripting for the import.

what’s the format of the csv? does it give columns/rows or is it face definitions?

3

Hi, thanks for your response.
Here is a random sample of the data:

169700,850950,220
169700,850900,224
169700,850850,226
169700,850800,228
169700,850750,234
169700,850700,239
169700,850650,242

all values are in metres, it’s a 50m grid Eastings, Northings, Altitude. The file I’m looking to use is 30MB. There are no headers or footers in the file as far as I can tell.
I guess the scripting may be reasonably straightforward?

4

well, pseudocoding it might help get you started:

open the file

#read in the lines

for each file line, make a vertex

#for every 4 vertecies, make a face :smiley: - this might need some tweaking

#close file

you might want to start with a small segment and make sure it works first, then test a larger segment. I’m guessing it’s going to read like a page:
Line 1(start here->): ============
Line 2: *****************
Line 3: ---------------------
I’m not sure how you would tell if you’re starting a new “line” of faces though - maybe see if the space between current vertex and next vertex are greater than 5 ish?
This is how to make a quad face from vertecies:
http://blenderartists.org/forum/showthread.php?t=197978&highlight=manipulating+geometry&page=2

and this is my stab at reading in the file/ making a face - it almost works(see the console for vertex output):


import bpy
from mathutils import Vector
import sys

file = open("c:\	xt.txt")

vcount = 8
fcount = 6
count = 0
for i in file.readlines():
    #print(i)
    a = i.split(",")
    scale = .00001
    x=float(a[0])
    y=float(a[1])
    z=float(a[2])
    newVertex= Vector((x*scale,y*scale,z*scale))
    cube = bpy.data.objects['Cube']
    cube.data.vertices.add(1)
    cube.data.vertices[vcount].co = newVertex
    print (vcount,newVertex)
    vcount +=1
    count+=1
    if(count %4 ==0):
        print("let's make a face! count is:", count)
        cube.data.faces.add(1)
        cube.data.faces[fcount].vertices_raw = [(vcount-3),(vcount-2),(vcount-1),vcount]
        fcount+=1
    cube.data.update(calc_edges=True)

#memory block tnorms corrupt
    
# 

#
#"0 thru 7 are cube data objects"
#cube.data.vertices[8].co = Vector((0,0,3))
#cube.data.vertices[9].co = Vector((1,0,3))
#cube.data.vertices[10].co = Vector((1,1,3))
#cube.data.vertices[11].co = Vector((0,1,3))
#
#cube.data.faces.add(1)
#cube.data.faces[6].vertices_raw = [8,9,10,11]
#cube.data.update(calc_edges=True)
#

I keep getting an error in the console “memory block tnorms corrupt”
which means I’m setting something in memory that i’m not supposed to. My guess is that the vertices_raw might have something to do with it. anyways, your vertices are imported correctly
-amdbcg

5

Thanks for this information AMDBCG :), I’m looking at this now.

cheers

James

6

Meshlab http://meshlab.sourceforge.net now imports xyz files. But I don’t know anything about the format so I don’t know how useful this info might be.

7

Meshlab works fine with Blender.
Open XYZ file, Filters > Remeshing > Surface Reconstruction Ball Pivoting, save as STL or PLY.
Removing doubles in Blender is recommended.

8

One tip: don’t mess about directly with those vertices and faces attributes in the mesh data. Instead, use the from_pydata method to load the vertices, edges and faces from arrays that you supply in one go. It knows the peculiarities of the required formats to put into those attributes.