为什么灰色图像代表3D张量？在我的代码中

Question

我将图像更改为Numpy数组格式。我以为是黑白图像（288,432），但我得到了288,432,4。为什么会这样？

import librosa
import librosa.display
import matplotlib.pyplot as plt
import librosa.featur
import numpy as np
import PIL.Image as pilimg
from PIL import Image
from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas

w1,sr  =librosa.load('test1.wav')
s1= librosa.feature.melspectrogram(y=w1, sr=sr, n_mels=128)
librosa.display.specshow(librosa.power_to_db(s1, ref=np.max), 
cmap='gray', sr=sr, y_axis='mel', x_axis='time')

[出现图片]

fig = plt.Figure()
canvas = FigureCanvas(fig)
ax = fig.add_subplot(111)
p = librosa.display.specshow(librosa.power_to_db(s1, ref=np.max), ax=ax, 
y_axis='mel', x_axis='time',cmap='gray')
fig.savefig('spec1.png')

image = pilimg.open('sepc1.png')
image = np.array(image)
print(image.shape)

（288,432,4）

print(image[250][255])
print(image[250][257])
print(image[251][255])

[236 236 236 255]

[230 230 230 255]

[222 222 222 266]