Opencv works wrong with YOLO

Autonomous Machines Jetson & Embedded Systems Jetson TX1
1

Hi
I am trying to make camera detection with YOLO on the TX1.
When i am doing it with OpenCV4Tegra,it makes wrong like this:

HIGHGUI ERROR: V4L2: Pixel format of incoming image is unsupported by Opencv
Couldn’t connect to webcam.
: Bad file descriptor
(Sorry, i don’t know how to uoload a picture.)

Then, i try to change the OpenCV4Tegra to Opencv3.1.0,and it has no such problem.
But, the new problem is that i can’t get a window with something, the window is always black.

What’s wrong with this?
Can anybody tell me?
Thanks a lot!

2

Were you able to fix it. I have the same issue.

Thanks,
-P

3

Sorry, i have no idea.
Maybe you can try follow the http://dev.t7.ai/jetson/opencv/.
I have tried and i can’t success, because when i compile the opencv,i find the disk space is not enough.
Maybe you can get one bigger HDD,and then have a try.:)

4

Hi
Does anybody can help me to deal with the problem?
Or, does anybody have the same problem with me, maybe we can discuss the problem and find some way.

5

Hello,

I had a similar problem, YOLO wasn’t working with the CSI camera.
So I just changed the YOLO code to have it capturing frames using VisionWorks.
Thus I used the FrameSource class which works with both CSI camera and USB webcam.

But you may make it work using OpenCV4Tegra and a USB webcam.
I didn’t try OpenCV3.1.0, but a colleague did and managed to do the work with VideoGrabber.