使用 matplotlib 从 2D 矩阵生成 3D 曲面

Question

Using a 2d matrix in python, how can I create a 3d surface plot, where columns=x, rows=y and the values are the heights in z?

I can't understand how to creat 3D surface plot using matplotlib. Maybe it's different from MatLab.

example:

from pylab import *
from mpl_toolkits.mplot3d import Axes3D

def p(eps=0.9, lmd=1, err=10e-3, m=60, n=40):

    delta_phi = 2 * np.pi / m
    delta_lmd = 2 / n
    k = 1
    P0 = np.zeros([m + 1, n + 1])
    P = np.zeros([m + 1, n + 1])
    GAP = 1

    while GAP >= err:
        k = k + 1
        for i in range(0, m):
            for j in range(0, n):
                if (i == 1) or (j == 1) or (i == m + 1) or (i == n + 1):
                    P[i,j] = 0
                else:
                    A = (1+eps*np.cos((i+1/2)*delta_phi))**3
                    B = (1+eps*np.cos((i-1/2)*delta_phi))**3
                    C = (lmd*delta_phi/delta_lmd)**2 * (1+eps*np.cos((i)*delta_phi))**3
                    D = C
                    E = A + B + C + D
                    F = 3*delta_phi*((1+eps*np.cos((i+1/2)*delta_phi))-(1+eps*np.cos((i-1/2)*delta_phi)))
                    P[i,j] = (A*P[i+1,j] + B*P[i-1,j] + C*P[i,j+1] + D*P[i,j-1] - F)/E
                    if P[i,j] < 0:
                        P[i,j] = 0
        S = P.sum() - P0.sum()
        T = P.sum()
        GAP = S / T
        P0 = P.copy()
    return P, k

def main():
    start = time.time()
    eps = 0.9
    lmd = 1
    err = 10e-8
    m = 60
    n = 40
    P, k = p()

    fig = figure()
    ax = Axes3D(fig)
    X = np.linspace(0, 2*np.pi, m+1)
    Y = np.linspace(-1, 1, n+1)
    X, Y = np.meshgrid(X, Y)
    #Z = P[0:m, 0:n]
    #Z = Z.reshape(X.shape)
    ax.set_xticks([0, np.pi/2, np.pi, np.pi*1.5, 2*np.pi])
    ax.set_yticks([-1, -0.5, 0, 0.5, 1])
    ax.plot_surface(X, Y, P)
    show()


if __name__ == '__main__':
    main()

ValueError: shape mismatch: objects cannot be broadcast to a single shape

And the pic pic by matplotlic

And I also use MatLab to generate,the pic: pic by MatLab

Answer 1

I should think this is a problem of getting the notaton straight. A m*n matrix is a matrix with m rows and n columns. Hence Y should be of length m and X of length n , such that after meshgridding X , Y and P all have shape (m,n) .

At this point there would be no need to reshape of reindex and just plotting

ax.plot_surface(X, Y, P) 

would give your the desired result.

Answer 2

Let's assume if you have a matrix mat .

import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits import mplot3d


h, w = mat.shape
plt.figure(figsize=(16, 8))
ax = plt.axes(projection='3d')
X, Y = np.meshgrid(np.arange(w), np.arange(h))
ax.plot_surface(X, Y, mat, rstride=1, cstride=1, cmap='viridis', edgecolor='none', antialiased=False)