Need help about inference on webcam stream

Robotics & Edge Computing Jetson & Embedded Systems Jetson TX2
1

Hi. I am trying to use my webcam as a stream source. I’ve changed

camera-v4l2-dev-node=0

parameter to 1 placed under [source0] section of the PGIE-FP16-CarType-CarMake-CarColor.txt file. After this, when I start the nvgstiva-app-gui with

$ nvgstiva-app-gui -c /home/nvidia/configs/PGIE-FP16-CarType-CarMake-CarColor.txt -g /home/nvidia/configs/gui_cfg.txt

command, it opens the gui and my webcam’s LED blinks for 1 sec and goes off with the GUI. I think that i should increase the stream buffer but i don’t know how to do. How can I feed DeepStream with the live webcam stream? Thanks in advance.

Here is the log on terminal aftrer everything goes off:

** INFO: <bus_callback:123>: Pipeline running

libv4l2: error got 4 consecutive frame decode errors, last error: v4l-convert: error Invalid argument
Error from src_elem: Failed to allocate a buffer
Debug info: gstv4l2src.c(851): gst_v4l2src_create (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin0/GstV4l2Src:src_elem
Error from src_elem: Internal data flow error.
Debug info: gstbasesrc.c(2948): gst_base_src_loop (): /GstPipeline:pipeline/GstBin:multi_src_bin/GstBin:src_sub_bin0/GstV4l2Src:src_elem:
streaming task paused, reason error (-5)
** INFO: <bus_callback:162>: Received EOS. Exiting …

Quitting
Exiting from gst_nv_caffe_inference_thread thread
Exiting from gst_nv_caffe_inference_thread thread
Exiting from gst_nv_caffe_inference_thread thread
Exiting from gst_nv_caffe_inference_thread thread
App run failed

2

Hi,

Please notice that this is DeepStream for Tesla board.

Could you share more information about the device you used?
Do you run DeepStream on the Jetson platform?

Thanks.

3

Sorry for wrong topic placement. Yes I am using TX2. I’ve changed csi sensor id in the config file and problem is gone. But I still have questions about how to use custom networks and how to create a config file for these networks (like sample config file).

4

Hi,

You can find the corresponding information in our document:

[i][deepstream folder]/nviva_docs/index.html

Application Configuration[/i]

Thanks.