Read an image with OpenCV and display it with Tkinter

Question

I have a very simple program on Ubuntu 14.04 LTS to read and display an image using OpenCV:

import cv2 #import OpenCV

img = cv2.imread('picture.jpg') #read a picture using OpenCV
cv2.imshow('image',img) # Display the picture
cv2.waitKey(0) # wait for closing
cv2.destroyAllWindows() # Ok, destroy the window

My problem:

How can I keep reading the picture in OpenCV but display it using Tkinter ?

I ask this because I want to make an interface for my program but OpenCV is not able to do it so I need Tkinter for this. However, all the image processing I must do it on the background using OpenCV. Only displaying the results must be done using Tkinter.

EDIT:

From the answer above, I change the line:

im = Image.open('slice001.hrs').convert2byte()

To:

im=cv2.imread() # (I imported cv2)

But I got an error.

I would appreciate any hints.

score 21 · Accepted Answer · edited Feb 13 '23 at 11:14

21

You might want to take a look at this one. Here is something works for me:

import numpy as np
import cv2
import Tkinter 
from PIL import Image, ImageTk

# Load an color image
img = cv2.imread('img.png')

#Rearrang the color channel
b,g,r = cv2.split(img)
img = cv2.merge((r,g,b))

# A root window for displaying objects
root = Tkinter.Tk()  

# Convert the Image object into a TkPhoto object
im = Image.fromarray(img)
imgtk = ImageTk.PhotoImage(image=im) 

# Put it in the display window
Tkinter.Label(root, image=imgtk).pack() 

root.mainloop() # Start the GUI

edited Feb 13 '23 at 11:14

Andreas Haferburg

5,189
3
37
63

answered Feb 23 '15 at 10:25

Ha Dang

1,218
1
10
12

Thank you. But I do not see where did you read the image using OpenCV ? – Feb 23 '15 at 10:39
1

Just edit the answer. But guys, why don't you give it a try yourself? DIY - that's the best way of learning. – Ha Dang Feb 23 '15 at 11:27
I tried (see my edit). Your second solution works almost perfectly. Only one drawback, however: my image is almost orange, but now it is displayed almost in blue :( – Feb 23 '15 at 11:33
I added the RGB mode for reading it like this `im = Image.fromarray(img,'RGB')`, but same result :( – Feb 23 '15 at 11:37
it's because color channel order is BBGR, not RGB. You can add the two following lines into the code: b,g,r = cv2.split(img) img = cv2.merge((b,g,r)) – Ha Dang Feb 23 '15 at 11:40
Still same result even if your intention, I think, is right (to split and merge the colors). You made so much efforts for me. – Feb 23 '15 at 11:47
Hey, pay attention to the order of the cv2.merge parameters. In my comment, it's not in good order. Try with the code in my answer. – Ha Dang Feb 23 '15 at 11:48
Yes, I did an error in ordering the colors. I do not not know how to thank you man. – Feb 23 '15 at 11:51
Instead of using `cv2.merge`, you can simply use `numpy` indexing to rearrange the third dimension of the image: `out = img[:,:,::-1]`. Just a suggestion, but merge is the accepted way when referring to the OpenCV docs. – rayryeng Feb 23 '15 at 17:30
1

It's `from PIL import Image, ImageTk`, not `import Image, ImageTk`. – makeworld Jul 14 '21 at 13:22

score 2 · Answer 2 · edited Aug 03 '22 at 11:54

2

For me both answers above did not work but were close. The following code did the trick for me (I also want to use place instead of pack):

from PIL import ImageTk, Image

image = cv2.cvtColor(self.image, cv2.COLOR_BGR2RGB)
image = ImageTk.PhotoImage(image=Image.fromarray(image))
label_image = Label(self.detection, image=image)
label_image.image = image
label_image.place(x=0, y=0, anchor="w")

edited Aug 03 '22 at 11:54

Christoph Rackwitz

11,317
4
27
36

answered Mar 31 '20 at 15:48

Jop Knoppers

676
1
10
22

2

this is obviously the most up to date and working answer :) and please remember to add the library imports too: from PIL import ImageTk,Image – SoajanII Jul 24 '22 at 17:00

score 1 · Answer 3 · answered Aug 20 '17 at 20:15

For Python3 I had to modify @Ha Dang answer:

from tkinter import *
from PIL import Image, ImageTk
import cv2
import numpy as np

image_name = 'bla.jpg'

image = cv2.imread(image_name)

#Rearrang the color channel
b,g,r = cv2.split(image)
img = cv2.merge((r,g,b))

# A root window for displaying objects
root = Tk()  

# Convert the Image object into a TkPhoto object
im = Image.fromarray(img)
imgtk = ImageTk.PhotoImage(image=im) 

# Put it in the display window
Label(root, image=imgtk).pack() 

root.mainloop() # Start the GUI

Requirements were:

pip3

numpy==1.13.1
opencv-python==3.3.0.9
Pillow==4.2.1

brew

python3
tcl-tk

Read an image with OpenCV and display it with Tkinter

3 Answers3

Linked