PIL报错：TypeError: Cannot handle this data type: 1, 1, 3 ＜f4及解决Image.fromarray保存后的结果是纯黑的图片

1.问题背景

在使用深度学习进行图像分类时，有时候需要将内存中的ndarray保存为本地图像，我这边使用了PIL的Image.fromarray函数，具体代码如下：

from PIL import Image
import os
import uuid

img_file = Image.fromarray(images_array_list[_index])
img_file.save(os.path.join(images_save_path, "{}-{}.jpg".format(TIME_STAMP, uuid.uuid4())))

却发生报错：
TypeError: Cannot handle this data type: (1, 1, 3), <f4：

具体报错信息：

Traceback (most recent call last):
  File "C:\Users\Anaconda3\envs\tf1.7\lib\site-packages\PIL\Image.py", line 2828, in fromarray
    mode, rawmode = _fromarray_typemap[typekey]
KeyError: ((1, 1, 3), '<f4')
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
  File "C:\Users\Anaconda3\envs\tf1.7\lib\site-packages\IPython\core\interactiveshell.py", line 3343, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-2-13650c2b0b93>", line 1, in <module>
    runfile('E:/Code/Python/keras不使用generator批量预测图像.py', wdir='E:/Code/Python')
  File "C:\Program Files\JetBrains\PyCharm 2020.1\plugins\python\helpers\pydev\_pydev_bundle\pydev_umd.py", line 197, in runfile
    pydev_imports.execfile(filename, global_vars, local_vars)  # execute the script
  File "C:\Program Files\JetBrains\PyCharm 2020.1\plugins\python\helpers\pydev\_pydev_imps\_pydev_execfile.py", line 18, in execfile
    exec(compile(contents+"\n", file, 'exec'), glob, loc)
  File "E:/Code/Python/keras不使用generator批量预测图像.py", line 44, in <module>
    img_file = Image.fromarray(images_list[_index])
  File "C:\Users\Anaconda3\envs\tf1.7\lib\site-packages\PIL\Image.py", line 2830, in fromarray
    raise TypeError("Cannot handle this data type: %s, %s" % typekey) from e
TypeError: Cannot handle this data type: (1, 1, 3), <f4

2.解决办法

2.1.解决Image.fromarray()保存图片报错

原因是Image.fromarray() 要求输入的numpy数据类型不支持float32类型的数据，但我这边输入的images_array_list数据类型是 float32，便造成了上述报错！此时，只需要将类型转换为Image.fromarray() 支持的类型即可（本文转为uint8类型），如下所示：

from PIL import Image
import os
import uuid
import numpy as np

img_file = Image.fromarray(np.uint8(images_list[_index]))
img_file.save(os.path.join(images_save_path, "{}-{}.jpg".format(TIME_STAMP, uuid.uuid4())))

运行上述代码，没有报错且成功保存了图像，但为啥图像是这样的：

这又是咋回事？

咱们接着往下看：

2.2.解决保存后的结果是纯黑的图片

这主要是因为我们的图片中的像素值被预处理之后，其值在[-1,1]之间，而图片的像素值取值范围一般是[0,255]，所以我们只需要将像素值由[-1,1]缩放到[0,255]即可！因为我这边使用的是Keras框架，里面自带了array_to_img函数，可以方便的转换图像：

def array_to_img(x, data_format=None, scale=True):
    """Converts a 3D Numpy array to a PIL Image instance.

    # Arguments
        x: Input Numpy array.
        data_format: Image data format.
        scale: Whether to rescale image values
            to be within [0, 255].

    # Returns
        A PIL Image instance.

    # Raises
        ImportError: if PIL is not available.
        ValueError: if invalid `x` or `data_format` is passed.
    """
    if pil_image is None:
        raise ImportError('Could not import PIL.Image. '
                          'The use of `array_to_img` requires PIL.')
    x = np.asarray(x, dtype=K.floatx())
    if x.ndim != 3:
        raise ValueError('Expected image array to have rank 3 (single image). '
                         'Got array with shape:', x.shape)

    if data_format is None:
        data_format = K.image_data_format()
    if data_format not in {'channels_first', 'channels_last'}:
        raise ValueError('Invalid data_format:', data_format)

    # Original Numpy array x has format (height, width, channel)
    # or (channel, height, width)
    # but target PIL image has format (width, height, channel)
    if data_format == 'channels_first':
        x = x.transpose(1, 2, 0)
    if scale:
        x = x + max(-np.min(x), 0)
        x_max = np.max(x)
        if x_max != 0:
            x /= x_max
        x *= 255
    if x.shape[2] == 3:
        # RGB
        return pil_image.fromarray(x.astype('uint8'), 'RGB')
    elif x.shape[2] == 1:
        # grayscale
        return pil_image.fromarray(x[:, :, 0].astype('uint8'), 'L')
    else:
        raise ValueError('Unsupported channel number: ', x.shape[2])

我们直接使用上述函数即可，即：

img_file = Image.fromarray(np.uint8(_images_list[_index]))

修改为

img_file = array_to_img(images_list[_index])

可以看到图像显示正常了：

来源：Jayce~