cv2.imshow(img) is crashing the kernel

[epignatelli]

As per title cv2.imshow(img) is crashing the server.

If we avoid cv2.waitForKey() and cv2.closeAllWindows(), and keep the windows open, the notebook will continue running.

Python version:

Python 3.5.6 |Anaconda custom (64-bit)| (default, Aug 26 2018, 16:05:27) [MSC v.1900 64 bit (AMD64)]

Kernel about:

Python 3.5.5 |Anaconda, Inc.| (default, Mar  9 2018, 12:39:44) [MSC v.1900 64 bit (AMD64)]
Type 'copyright', 'credits' or 'license' for more information
IPython 6.2.1 -- An enhanced Interactive Python. Type '?' for help.

To reproduce: (coming)

[ah3243]

I'm having a similar problem, the imshow() window is displayed it crashes when it would normally be destroyed. Even it's the last of 10 windows created and destroyed in a loop the last one will always freeze as the program moves on. Gives no errors when running identical code in a terminal.

OSx Mojave 10.14.2
Python 3.5
opencv-python 4.0.0.12

[nvt-ak]

You can use this solution if you're using google colab:
from google.colab.patches import cv2_imshow
cv2_imshow(img)
It works fine!

[mitesh98]

image is appearing too big.I use below code
import matplotlib.pyplot as plt
plt.figure(figsize=(18,18))
plt.imshow(image)
plt.axis('off')
plt.show()

[Oshana]

how to convert this to cv2_imshow?

cv2.imshow('object detection', cv2.resize(image_np, (800, 600)))

[shivam-kotwalia]

@Oshana just say -
cv2_imshow(cv2.resize(image_np, (800, 600)))
cv2_imshow only takes one input parameter, https://github.com/googlecolab/colabtools/blob/8708337447efdef2732cca9764ede400da6c051f/google/colab/patches/__init__.py#L14

[epignatelli]

You can also consider using matplotlib rather than IPython.display, although, I would not recommend to mix the two.

import matplotlib.pyplot as plt
import cv2
def cv2_imshow(a, **kwargs):
    a = a.clip(0, 255).astype('uint8')
    # cv2 stores colors as BGR; convert to RGB
    if a.ndim == 3:
        if a.shape[2] == 4:
            a = cv2.cvtColor(a, cv2.COLOR_BGRA2RGBA)
        else:
            a = cv2.cvtColor(a, cv2.COLOR_BGR2RGB)

    return plt.imshow(a, **kwargs)

[kyteinsky]

image is appearing too big.I use below code
import matplotlib.pyplot as plt
plt.figure(figsize=(18,18))
plt.imshow(image)
plt.axis('off')
plt.show()

You can use this insted:
plt.figure(figsize=(10,18))
for resizing the image.

[kyteinsky]

This fix also works for opening image using imshow() function from OpenCV::

cv.namedWindow(winname='my')
while True:
cv.imshow('my',img)
if cv.waitKey(1) & 0xFF == 27:
break
cv.destroyAllWindows()

[PrasadNR]

You can use this solution if you're using google colab:
from google.colab.patches import cv2_imshow
cv2_imshow(img)
It works fine!

This is just showing some black blank image.

[antopolskiy]

Copying the mentioned function cv2_imshow from colabtools and using it locally worked well for me. It displays the image inline.

[kilicsedat]

This approach worked for me, too.

[ozrite]

hi guys, here is the solution of the problem.

If you are using a 64-bit machine, you will have to modify
k = cv2.waitKey(0)
line as follows :
k = cv2.waitKey(0) & 0xFF

The full solution is like this:

```
cv2.imshow('image',img)
k = cv2.waitKey(0) & 0xFF
if k == 27:         # wait for ESC key to exit
    cv2.destroyAllWindows()

Now the kernel will not be stuck or crash. You just need to press esc to quit.

[KrishnaChaitanya-Gopaluni]

You can also consider using matplotlib rather than IPython.display, although, I would not recommend to mix the two.

import matplotlib.pyplot as plt
import cv2
def cv2_imshow(a, **kwargs):
    a = a.clip(0, 255).astype('uint8')
    # cv2 stores colors as BGR; convert to RGB
    if a.ndim == 3:
        if a.shape[2] == 4:
            a = cv2.cvtColor(a, cv2.COLOR_BGRA2RGBA)
        else:
            a = cv2.cvtColor(a, cv2.COLOR_BGR2RGB)

    return plt.imshow(a, **kwargs)

Thank you.

[kevin-bates]

This is not a notebook issue and should probably have been redirected to ipykernel or ipython. However, the last few comments have been success stories, so, on that account, I'm closing this issue.

[szelesaron]

You can use this solution if you're using google colab:
from google.colab.patches import cv2_imshow
cv2_imshow(img)
It works fine!

This is just showing some black blank image.

If you have this issue, check if the numpy array is of type float, if so then multiply the array by 255.

https://docs.opencv.org/2.4/modules/highgui/doc/user_interface.html?highlight=imshow#imshow