strange plot surface of matplotlib

Question

I'm trying to generate a 3D height figure, I have a regular grid, the height data collected by the sensor and data store in a file which name is "data.txt". data stored one data per line. the file link on github

import numpy as np
import matplotlib.pyplot as pit

from mpl_toolkits.mplot3d import Axes3D
from scipy.stats import multivariate_normal
from matplotlib import cm

x = np.linspace(0,350,18)
y = np.linspace(0,350,15)
z = np.loadtxt('data.txt')

xx,yy = np.meshgrid(x,y)
fig = pit.figure()
ax = fig.add_subplot(111,projection='3d')
ax.scatter(xx,yy,z)

use the code above, I got a scatter. It looks good! I found this , I want convert the figure to surface, than I add the code below, but it looks very strange

xa = np.reshape(xx, (18,15))
ya = np.reshape(yy, (18,15))
za = np.reshape(z, (18,15))
surf=ax.plot_surface(xa,ya,za,cmap="summer",linewidth=0,antialiased=False, alpha=0.5)
fig.colorbar(surf)
pit.show()

the image

i don't know what happened, it look too strange! Should i smooth it?

Answer 1

You need to use xx and yy defined earlier and reshape z to the same shape as xx:

za = z.reshape(xx.shape)
fig = pit.figure()
ax = fig.add_subplot(111,projection='3d')
surf=ax.plot_surface(xx,yy,za,cmap="summer",linewidth=0,antialiased=False, alpha=0.5)
fig.colorbar(surf)
pit.show()

Note that I have rotate the chart for better clarity.

Answer 2

I think you need scipy.griddata. Try this code:

from scipy.interpolate import griddata
za = griddata(x, y, z, (xx, yy), method='linear')
surf=ax.plot_surface(xx,yy,za,cmap="summer",linewidth=0,antialiased=False, alpha=0.5)
fig.colorbar(surf)
plt.show()