Interpolate irregular 3d data from a XYZ file to a regular grid

Question

I have a xyz file containing a lot of 3D coordinates like so:

 370373.771    6535261.431      2.908       
 370373.788    6535261.441      2.911       
 370373.787    6535261.442      2.909       
 370373.809    6535261.449      2.908       
 370373.810    6535261.439      2.909       
 370373.743    6535261.466      2.922       
 370373.584    6535261.455      2.915       
 370373.559    6535261.471      2.898       
 370373.544    6535261.559      2.887       
 370373.552    6535261.538      2.866       
 370373.797    6535261.486      2.876       
 370373.795    6535261.557      2.892       
 ..........    ...........      .....

This file is quite dense and irregular, and I would like to interpolate these coordinates on a regular grid with a point each 5m for example.

Here is what I tried so far:

import numpy as np
from scipy.interpolate import griddata

coord_x = []
coord_y = []
coord_z = []
coord_xy = []

xyzfile = open("xyzfile.txt")
for line in xyzfile:
    x,y,z = line.split()
    coord_x.append(float(x))
    coord_y.append(float(y))
    coord_xy.append([float(x),float(y)])
    coord_z.append(float(z))
xyzfile.close()

lon = np.linspace(min(coord_x), max(coord_x), 200)
lat = np.linspace(min(coord_y), max(coord_y), 200)
X, Y = np.meshgrid(lon, lat)

grid = griddata(np.array(coord_xy), coord_z, (X, Y), method='nearest')

Unfortunately I get an error :

TypeError: only integer arrays with one element can be converted to an index

I'm also not sure how to get a 5m spacing between each point in the regular grid. How can I do this ?

Second question, say I have 6 known points on this grid, how to extract the z value on this regular 5m spaced grid for these 6 points.

Thank you

What exactly do you mean by 'interpolate these coordinates on a regular grid with a point each 5m'? Is this a grid on the x-y plane and you do want to interpolate the z-value for every grid-point from the surrounding points? — Jupiter, Nov 25 '15 at 15:16

score 4 · Answer 1 · answered Nov 25 '15 at 15:59

4

The coord_z argument passed in must also be an array:

grid = griddata(np.array(coord_xy), np.array(coord_z), (X, Y), method='nearest')

answered Nov 25 '15 at 15:59

pv.

33,875
8
55
49

Thanks, but that is only partial answer, I would like to have the grid 5m spaced, and also to extract z value of 6 known xy coordinates points on the grid. – GeoffreyB Nov 25 '15 at 16:02
Wow, thanks, man! Just got stuck with this problem and saw your answer. – Israel Unterman Sep 04 '16 at 15:43

Interpolate irregular 3d data from a XYZ file to a regular grid

1 Answers1