对来自各种“图像”的 numpy 数组进行标准化物体

Question

请考虑这个可重现的示例：

from PIL import Image
import numpy as np
import scipy.misc as sm
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
import matplotlib.cbook as cbook
import urllib

datafile = cbook.get_sample_data('lena.jpg')
lena_pil = Image.open(datafile)
lena_pil_np = np.asarray(lena_pil)

lena_scipy = sm.lena()

lena_tmp = open('lena_tmp.png', 'wb')
lena_tmp.write(urllib.urlopen('http://optipng.sourceforge.net/pngtech/img/lena.png').read())
lena_tmp.close()

lena_mpl = mpimg.imread('lena_tmp.png')

sm.info(lena_pil_np)
sm.info(lena_scipy)
sm.info(lena_mpl)

输出为：

>>> sm.info(lena_pil_np)
class:  ndarray
shape:  (512, 512, 3)
strides:  (1536, 3, 1)
itemsize:  1
aligned:  True
contiguous:  True
fortran:  False
data pointer: 0xb707e01cL
byteorder:  little
byteswap:  False
type: uint8

>>> sm.info(lena_scipy)
class:  ndarray
shape:  (512, 512)
strides:  (2048, 4)
itemsize:  4
aligned:  True
contiguous:  True
fortran:  False
data pointer: 0xb6f7d008L
byteorder:  little
byteswap:  False
type: int32

>>> sm.info(lena_mpl)
class:  ndarray
shape:  (512, 512, 3)
strides:  (6144, 12, 4)
itemsize:  4
aligned:  True
contiguous:  True
fortran:  False
data pointer: 0xb6c7b008L
byteorder:  little
byteswap:  False
type: float32

因此所有数组都具有不同的形状和类型。

为了进行额外的处理，我希望这个数组在最后一个变量 lena.mpl 中表示，或者只是将数组值转换为其标准化的 [0..1] float32 类型。

最好的方法是什么？

Answer 1

def normalize(arr):
    arr=arr.astype('float32')
    if arr.max() > 1.0:
        arr/=255.0
    return arr