使用Matplotlib绘制多数组频谱图时遇TypeError错误求助

错误原因 #

你遇到的TypeError由两个核心问题导致：

• plt.specgram()默认仅处理一维时间序列数据，你传入的是10×10的二维数组，不符合函数输入要求。
• 调用specgram时未指定Fs（采样频率）参数，函数内部默认值为None，导致计算pad_xextent时出现int / NoneType的非法运算。

修正方案 #

方案1：绘制单通道频谱图 #

从二维数组中提取一维时间序列（示例取第一行），同时指定采样频率Fs（根据数据实际采样率设置，示例设为100）：

import matplotlib.pyplot as plt
import numpy as np

# 你的二维数组
values = np.array([[ 0.         , 35.34028168, 36.09317546, 20.88056919, 35.65257548, 36.39835111,
  29.89568243, 34.94770207, 36.03834023, 36.51054007],
 [35.34028168,  0.         , 36.31999524, 34.455879  , 35.46084735, 35.99851703,
  35.23984878, 35.98339627, 35.87905754, 36.08155069],
 [36.09317546, 36.31999524,  0.         , 35.68280382, 36.11374003, 35.30407355,
  35.97087692, 36.20023994, 35.75066441, 36.01957597],
 [20.88056919, 34.455879  , 35.68280382,  0.         , 24.76405132, 36.31596819,
  30.47948816, 36.27410762, 36.30076717, 35.76478691],
 [35.65257548, 35.46084735, 36.11374003, 24.76405132,  0.         , 36.17666549,
  29.51602419, 36.33520073, 36.21391866, 36.19753349],
 [36.39835111, 35.99851703, 35.30407355, 36.31596819, 36.17666549,  0.        ,
  36.22994817, 36.32591538, 36.05558867, 36.20784639],
 [29.89568243, 35.23984878, 35.97087692, 30.47948816, 29.51602419, 36.22994817,
   0.         , 36.14432269, 36.19523145, 35.40707841],
 [34.94770207, 35.98339627, 36.20023994, 36.27410762, 36.33520073, 36.32591538,
  36.14432269,  0.         , 36.05260398, 36.38549268],
 [36.03834023, 35.87905754, 35.75066441, 36.30076717, 36.21391866, 36.05558867,
  36.19523145, 36.05260398,  0.         , 35.92987568],
 [36.51054007, 36.08155069, 36.01957597, 35.76478691, 36.19753349, 36.20784639,
  35.40707841, 36.38549268, 35.92987568,  0.        ]])

# 提取第一行作为一维时间序列，指定采样频率Fs=100
plt.specgram(values[0], Fs=100)
plt.title("单通道频谱图")
plt.xlabel("时间 (s)")
plt.ylabel("频率 (Hz)")
plt.colorbar(label="功率谱密度 (dB/Hz)")
plt.show()

方案2：绘制所有通道的频谱图（子图布局） #

如果需要可视化二维数组中所有行的频谱，可通过循环生成子图：

import matplotlib.pyplot as plt
import numpy as np

values = np.array([[ 0.         , 35.34028168, 36.09317546, 20.88056919, 35.65257548, 36.39835111,
  29.89568243, 34.94770207, 36.03834023, 36.51054007],
 [35.34028168,  0.         , 36.31999524, 34.455879  , 35.46084735, 35.99851703,
  35.23984878, 35.98339627, 35.87905754, 36.08155069],
 [36.09317546, 36.31999524,  0.         , 35.68280382, 36.11374003, 35.30407355,
  35.97087692, 36.20023994, 35.75066441, 36.01957597],
 [20.88056919, 34.455879  , 35.68280382,  0.         , 24.76405132, 36.31596819,
  30.47948816, 36.27410762, 36.30076717, 35.76478691],
 [35.65257548, 35.46084735, 36.11374003, 24.76405132,  0.         , 36.17666549,
  29.51602419, 36.33520073, 36.21391866, 36.19753349],
 [36.39835111, 35.99851703, 35.30407355, 36.31596819, 36.17666549,  0.        ,
  36.22994817, 36.32591538, 36.05558867, 36.20784639],
 [29.89568243, 35.23984878, 35.97087692, 30.47948816, 29.51602419, 36.22994817,
   0.         , 36.14432269, 36.19523145, 35.40707841],
 [34.94770207, 35.98339627, 36.20023994, 36.27410762, 36.33520073, 36.32591538,
  36.14432269,  0.         , 36.05260398, 36.38549268],
 [36.03834023, 35.87905754, 35.75066441, 36.30076717, 36.21391866, 36.05558867,
  36.19523145, 36.05260398,  0.         , 35.92987568],
 [36.51054007, 36.08155069, 36.01957597, 35.76478691, 36.19753349, 36.20784639,
  35.40707841, 36.38549268, 35.92987568,  0.        ]])

# 设置子图布局：5行2列
fig, axes = plt.subplots(5, 2, figsize=(12, 15))
axes = axes.flatten()
Fs = 100  # 采样频率

for i in range(len(values)):
    axes[i].specgram(values[i], Fs=Fs)
    axes[i].set_title(f"通道 {i+1} 频谱图")
    axes[i].set_xlabel("时间 (s)")
    axes[i].set_ylabel("频率 (Hz)")

plt.tight_layout()
plt.show()

关键说明 #

• Fs参数必须指定，它代表数据的采样频率（单位：Hz），需根据数据采集的实际情况调整。
• specgram的输入必须是一维数组，若为二维数组，需明确处理每一个维度的序列。

内容的提问来源于stack exchange，提问作者Aditya Bhandwalkar