Unable to computer vision in jupyter lab

[Wen_Jie]

Good days everyone

I was doing my project related to computer vision to detect the car plate. Hence I start by calling out the image and displaying it. I am just using a simple function: cv2.imshow("Image",img). However, it turns out in the kernel died.

How can this be resolved?

[AthiraM_Intel]

Hi,

Thank you for posting in Intel Communities.

Could you please let us know which DevCloud (DevCloud for oneAPI/ DevCloud for edge) you are using?

Thanks

[Wen_Jie]

Hi

I was using  DevCloud for edge.

Thanks

[Wen_Jie]

Hi all

I also have a doubt that: Can DevCloud for edge perform a real-time video stream using a webcam in jupyter lab by calling function  cv2.VideoCapture(0) ? 

Thank you

[JesusE_Intel · Accepted solution]

Hi Wen_Jie,

Thanks for reaching out, using cv2.imshow on Intel Developer Cloud for the Edge will not work as you are submitting jobs to the Edge nodes and not on the node running the Jupyter Notebook. Please take a look at the sample applications, we provide samples to submit a video for inference and save the results in a .webm file that can be viewed using the videoHTML function as follows:

videoHTML('Intel CPU + Intel NCS2', 

     ['results/ncs2/output_{}.webm'.format(id_ncs2)], 'results/ncs2/stats_{}.txt'.format(id_ncs2))

https://www.intel.com/content/www/us/en/developer/tools/devcloud/edge/build/sample-apps.html

Hope this helps!

Regards,

Jesus

[Wen_Jie]

Hi JesusU_Intel

Yes, it helps me to solve my problem.

Thank you very much!!

[JesusE_Intel]

Perfect! If you need any additional information, please submit a new question as this thread will no longer be monitored.