如何使用 matlib function plt.imshow(image) 显示多个图像？

Question

def exiacc():
    global my_img, readimg , im, list_name, exis_img, tkimage
    print('Opening Existing Account...')
    topl = Toplevel()  
    readimg = './facedata/'
    list_img = os.listdir(readimg)
    row, col = 0, 0
    k = 0
    for fn in os.listdir(readimg):
        if fn.endswith('.jpg'):
            list_name.append(fn[:fn.index('_')])
            im = Image.open(os.path.join(readimg, fn))
            im2 = im.resize((200, 200), Image.LANCZOS)
            tkimage = ImageTk.PhotoImage(im2)
            exis_img = Label(topl, image=tkimage)
            exis_img.grid(row = row + 1, column = col + 1, padx=2, pady=2)
            exis_name = Label(topl, text = list_name[k] , font = ("consolas", 12, "bold"))
            exis_name.grid(row=row + 2, column = col + 1, padx=2, pady=2)
            col += 1
            k +=1
            if col == 5:
                row += 2
                col = 0

My results show that only the last processed image is shown effectively overwriting the other images.

Answer 1

plt.imshow has a parameter extent=[x0, x1, y0, y1] that can position the image anywhere in a plot. Annoying is that imshow forces the aspect ratio to be 'equal' (so distances in x and in y are forced to be the same number of pixels). Even more annoying is that imshow also forcefully sets the limits in x and in y. To have more than one image, the limits need to be set explicitly (later than the call to imshow ). If the 'equal' aspect ratio isn't wanted (it often is desired for images) set_aspect('auto') frees the aspect ratio again.

Here is an example, using different colormaps to get differently colored images.

import matplotlib.pyplot as plt
import numpy as np

fig, ax = plt.subplots()
cmaps = ['Greys', 'Purples', 'Blues', 'Greens', 'Oranges', 'Reds',
         'YlOrBr', 'YlOrRd', 'OrRd', 'PuRd', 'RdPu', 'BuPu',
         'GnBu', 'PuBu', 'YlGnBu', 'PuBuGn', 'BuGn', 'YlGn']
img = ((np.arange(100) % 10) * np.arange(100)).reshape(10, 10)
cols = 4
for i, cmap in enumerate(cmaps):
    plt.imshow(img, extent=[i % cols, i % cols +1, i // cols, i // cols +1], cmap=cmap)
plt.xlim(0, cols)
plt.ylim(0, (len(cmaps) - 1) // cols + 1)
plt.gca().set_aspect('auto') # allow a free aspect ratio, to fully fit the plot figure
plt.show()

Answer 2

A straight forward solution would be to use plt.subplots instead:

import numpy as np
import matplotlib.pyplot as plt

lst_cmaps = ['Greys', 'Purples', 'Blues', 'Greens', 'Oranges', 'Reds',
         'YlOrBr', 'YlOrRd', 'OrRd', 'PuRd', 'RdPu', 'BuPu',
         'GnBu', 'PuBu', 'YlGnBu', 'PuBuGn']

data = ((np.arange(100) % 10) * np.arange(100)).reshape(10, 10)

nxn = int(np.sqrt(len(lst_cmaps)))

plt, ax = plt.subplots(nxn, nxn)

for i, ax_ in enumerate(ax.flatten()):
    ax_.imshow(data, cmap=lst_cmaps[i])

This gives you an image like this: