极坐标等高线图异常问题：轮廓呈现锯齿状

问题说明 #

用模拟数据时，极坐标散点和基于meshgrid生成的等高线都能正常显示；但使用真实的50个离散样本点时，散点绘制正常，用同样方法调用contourf生成的等高线却呈现锯齿状异常，无法得到预期的同心轮廓效果。

模拟数据代码（正常运行） #

import numpy as np
import math
import matplotlib.pyplot as plt

# 模拟数据准备
phi = np.linspace(0, math.pi*2, 40)  # 完整圆周角度
phi = np.concatenate([phi, phi, phi, phi])  # 重复4次
rho = np.array([0,1,2,3])  # 四个固定半径
rho = np.repeat(rho, 40)

assert phi.shape == rho.shape

# 绘制散点
plt.figure()
ax = plt.subplot(111, projection='polar')
ax.scatter(phi, rho, c=rho)
ax.set_ylim(0,4)
plt.show()

真实数据样本代码（散点正常） #

import numpy as np
import matplotlib.pyplot as plt

rho = np.array([0.38818333, 0.73367091, 0.42336148, 1.39013061, 0.31064486,
0.34546275, 0.05445943, 0.85551576, 0.55174167, 1.42371249,
0.17644804, 1.76221456, 0.64519126, 0.02408941, 1.43986863,
0.72718428, 0.4262945 , 0.1355583 , 0.86319986, 0.71212376,
0.14891707, 1.01624534, 1.26915981, 1.39384488, 0.09623481,
0.92635469, 1.74757901, 0.15811954, 0.22052651, 0.30784166,
0.92740352, 1.29621377, 0.29832842, 1.04442307, 1.36185399,
0.42979785, 0.94402815, 0.3786981 , 0.75865969, 1.97273479,
0.61140136, 0.71452862, 0.25793468, 1.1751275 , 1.53945948,
0.64150917, 0.09274101, 0.52548715, 0.7932458 , 0.90292444])
phi = np.array([1.04208195, 4.67055389, 3.32909655, 1.18709268, 0.86036178,
5.820191  , 4.30457004, 1.81242968, 0.64295926, 4.85684143,
2.73937709, 3.22891963, 0.25822595, 0.69526782, 0.70709764,
1.92901075, 3.44538869, 5.38541473, 0.95255568, 4.01519928,
0.8503274 , 5.26774545, 4.07787945, 4.51718652, 0.3170884 ,
2.1946835 , 3.12550771, 5.67275731, 1.0000195 , 1.82570239,
5.62578391, 0.81923255, 2.00131474, 0.48190872, 4.78875363,
5.60395833, 2.01674743, 2.13494958, 5.10829845, 0.95324309,
1.59531506, 4.99145225, 6.19873491, 3.32802456, 1.15590926,
0.52989939, 6.02205398, 3.66013508, 4.16276819, 2.60498467])

assert phi.shape == rho.shape # 均为(50,)

# 绘制散点
plt.figure()
ax = plt.subplot(111, projection='polar')
ax.scatter(phi, rho, c=rho)
ax.set_ylim(0,3)
plt.show()

异常的等高线绘制代码 #

# 直接用原始数据生成网格
X, Y = np.meshgrid(phi, rho)

plt.figure(figsize=(10,10))
ax = plt.subplot(111, projection='polar')
ax.scatter(X, Y, c=Y)
CS = ax.contourf(X,Y, Y, 2, alpha=0.4)
ax.set_ylim(0,4)
plt.show()

问题原因 #

模拟数据的phi和rho是规则排列的：每个固定半径对应完整的角度序列，用meshgrid能生成整齐的极坐标网格。但真实数据是无序的离散点，直接用原始的phi和rho生成meshgrid会得到混乱的网格，contourf无法识别出同心轮廓，自然呈现锯齿状。

解决方案 #

需要先将离散的真实数据插值到规则的极坐标网格上，步骤如下：

1. 生成均匀的角度和半径序列，构建规则网格
2. 使用插值方法将原始数据映射到规则网格
3. 基于规则网格绘制等高线

完整修复代码 #

import numpy as np
import matplotlib.pyplot as plt
from scipy.interpolate import griddata

# 真实数据
rho = np.array([0.38818333, 0.73367091, 0.42336148, 1.39013061, 0.31064486,
0.34546275, 0.05445943, 0.85551576, 0.55174167, 1.42371249,
0.17644804, 1.76221456, 0.64519126, 0.02408941, 1.43986863,
0.72718428, 0.4262945 , 0.1355583 , 0.86319986, 0.71212376,
0.14891707, 1.01624534, 1.26915981, 1.39384488, 0.09623481,
0.92635469, 1.74757901, 0.15811954, 0.22052651, 0.30784166,
0.92740352, 1.29621377, 0.29832842, 1.04442307, 1.36185399,
0.42979785, 0.94402815, 0.3786981 , 0.75865969, 1.97273479,
0.61140136, 0.71452862, 0.25793468, 1.1751275 , 1.53945948,
0.64150917, 0.09274101, 0.52548715, 0.7932458 , 0.90292444])
phi = np.array([1.04208195, 4.67055389, 3.32909655, 1.18709268, 0.86036178,
5.820191  , 4.30457004, 1.81242968, 0.64295926, 4.85684143,
2.73937709, 3.22891963, 0.25822595, 0.69526782, 0.70709764,
1.92901075, 3.44538869, 5.38541473, 0.95255568, 4.01519928,
0.8503274 , 5.26774545, 4.07787945, 4.51718652, 0.3170884 ,
2.1946835 , 3.12550771, 5.67275731, 1.0000195 , 1.82570239,
5.62578391, 0.81923255, 2.00131474, 0.48190872, 4.78875363,
5.60395833, 2.01674743, 2.13494958, 5.10829845, 0.95324309,
1.59531506, 4.99145225, 6.19873491, 3.32802456, 1.15590926,
0.52989939, 6.02205398, 3.66013508, 4.16276819, 2.60498467])

# 1. 生成规则的极坐标网格
phi_grid = np.linspace(0, 2*np.pi, 100)  # 均匀角度序列
rho_grid = np.linspace(0, 3, 100)        # 均匀半径序列
PHI, RHO = np.meshgrid(phi_grid, rho_grid)

# 2. 转换为笛卡尔坐标（griddata要求输入笛卡尔坐标）
x = rho * np.cos(phi)
y = rho * np.sin(phi)
x_grid = RHO * np.cos(PHI)
y_grid = RHO * np.sin(PHI)

# 3. 插值：将原始rho值映射到规则网格
z_grid = griddata((x, y), rho, (x_grid, y_grid), method='cubic')

# 4. 绘制极坐标等高线
plt.figure(figsize=(10,10))
ax = plt.subplot(111, projection='polar')
ax.scatter(phi, rho, c=rho, zorder=5)  # 叠加原始散点
CS = ax.contourf(PHI, RHO, z_grid, 10, alpha=0.4)
ax.set_ylim(0,3)
plt.colorbar(CS)
plt.show()

说明 #

• 使用griddata的cubic插值方法能生成平滑的轮廓，也可根据需求选择linear或nearest方法
• 规则网格的分辨率（示例中为100）可调整：数值越大轮廓越平滑，但计算量也会增加

内容的提问来源于stack exchange，提问作者Mitchell van Zuylen