TypeError: rescale() got an unexpected keyword argument ‘multichannel‘

代码实例：

# %% DOWNSAMPLE TRAINING IMAGES

num_train_images = len(listdir(train_hr_path))
for i in tqdm(range(num_train_images)):
    img_name = listdir(train_hr_path)[i]
    img = io.imread('{}{}'.format(train_hr_path, img_name))

    scaled_image = rescale(img, (1/scale), multichnnel=True,anti_aliasing=1)
    # scaled_image=rgb2gray(scaled_image)

    scaled_image =(scaled_image * 255).astype(np.uint8)

    io.imsave('{}{}'.format(train_lr_path, img_name), scaled_image)

运行出现错误TypeError: rescale() got an unexpected keyword argument 'multichannel'

原因：

scikit-image新版本已经废弃了multichannel函数

解决方案：

将multichannel=True改成channel_axis=-1

channel_axis：int 或无，可选

如果为 None，则假定图像是灰度(单通道)图像。否则，此参数指示数组的哪个轴对应于通道。