Jetson nano nvbuf_utils: Could not get EGL display connection

Autonomous Machines Jetson & Embedded Systems Jetson Nano
1

Hello there,
I have Jetson nano 4Gb
Version: 4.6.1-b110
set
nvidia-l4t-core 32.7.1-20220219090432
by instuction from this link : https://developer.nvidia.com/embedded/learn/get-started-jetson-nano-devkit

after :
sudo apt update
sudo apt install xrdp
sudo systemctl enable xrdp

set Vino, ok with Remote dekctop.
set nomachine

set jtop :
sudo apt install python3-pip
sudo pip3 install -U jetson-stats

by

set docker and set up usb camera (sudo apt install v4l-utils)

on test
./video-viewer /dev/video0

i have error from start :
nvbuf_utils: Could not get EGL display connection

and later
[cuda] cudaGraphicsGLRegisterBuffer(&interop, allocDMA(type), cudaGraphicsRegisterFlagsFromGL(flags))
[cuda] unknown error (error 999) (hex 0x3E7)
[cuda] /jetson-inference/utils/display/glTexture.cpp:360
[cuda] cudaGetLastError()
[cuda] unknown error (error 999) (hex 0x3E7)
[cuda] /jetson-inference/utils/cuda/cudaYUV-YV12.cu:119
[cuda] cudaI420ToRGB(input, (uchar3*)output, width, height)
[cuda] unknown error (error 999) (hex 0x3E7)
[cuda] /jetson-inference/utils/cuda/cudaColorspace.cpp:53
[cuda] cudaConvertColor(latestYUV, mFormatYUV, nextRGB, format, mOptions->width, mOptions->height)
[cuda] unknown error (error 999) (hex 0x3E7)
[cuda] /jetson-inference/utils/codec/gstBufferManager.cpp:445
[gstreamer] gstBufferManager – unsupported image format (rgb8)
[gstreamer] supported formats are:
[gstreamer] * rgb8
[gstreamer] * rgba8
[gstreamer] * rgb32f
[gstreamer] * rgba32f
[gstreamer] gstCamera::Capture() – an error occurred retrieving the next image buffer
video-viewer: shutting down…

and other test (./imagenet images/jellyfish.jpg images/test/jellyfish.jpg) the same problem!

UPD : if I test camera with nvgstcapture-1.0 --camsrc=0 --cap-dev-node=/dev/video0

i have the same error - nvbuf_utils: Could not get EGL display connection
and no photo on screen - empty

Help please!

3

Hi,
It seems like xrdp does not enable EGL libs, or there is hard-coded path while loading the EGL libs. So nveglglessink plugin does not work properly in the environment. You would need to customize the sample to use other renderer such as xvimagesink plugin.

4

Hi, I didn’t understand your answer. Let’s start from begin. I want to have jetson nano with remote control with fastest graphical interface from Windows PC. Can you write link to make this ?
I use this intructions - https://www.youtube.com/watch?v=DVfmw3Ftu3I&list=PLv8Cp2NvcY8AkXRldCAYCvFxRUs0h5JJF&ab_channel=CodeWithAarohi

5

Hi,
For remote control, we suggest use VNC. Please refer to
Accessing Jetson NANO with VNC - #3 by WayneWWW

6

As far as I understand, when using Headless Mode, error “Could not get EGL display connection” occurs. When I connected the monitor, keyboard and mouse, everything works. That is, using remote access it is impossible to obtain video or photo information? I’m right ?

7

Hi,
If Jetson Nano is at remote location, we generally login through ssh and run commands in console. If we need to check video or photo, we will run scp to copy the files into local and check. For checking HDMI output, We would suggest access the device in person. It may be some constraint in checking display output remotely.

8

Now in “headless” mode there is another problem. All examples GitHub - dusty-nv/jetson-inference: Hello AI World guide to deploying deep-learning inference networks and deep vision primitives with TensorRT and NVIDIA Jetson. of working with a USB camera do not work, since the result is displayed on the screen. And if you add the monitor, then everything works, but if it’s in “headless” mode, then it doesn’t. In my case, I don’t need the result to be displayed on the screen. It will be enough to process the image from the USB camera, and save the result, for example, as a jpeg file or save recognition result to sql base. Is there a solution or example?

9

Hi,
We would suggest use DeepStream SDK. Please install it through SDKManager and give it a try. For Jetson Nano the latest version is 6.0.1:
NVIDIA Metropolis Documentation

For USB camera, you can change to v4l2 source in source group:
DeepStream Reference Application - deepstream-app — DeepStream 6.0.1 Release documentation

And change to saving to a file in sink group:
DeepStream Reference Application - deepstream-app — DeepStream 6.0.1 Release documentation

11

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.