• Hardware Platform: Jetson Xavier NX
• DeepStream Version: 5.0
• JetPack Version: 4.4.1
• Issue Type: questions
I’m trying to run deepstream-segmentation-test with USB Camera. I use jetpack 4.4.1 and Docker image (nvcr.io/nvidia/deepstream-l4t:5.0.1-20.09-samples) and use Jetson Xavier NX.
I have modified the code and am able to recognize the USB Camera (MJPEG) and import the video, but the video stops right after startup.
My modified code is as follows(deepstream_segmentation_app.c)
static GstElement *
create_source_bin (guint index, gchar * uri)
{
GstElement *bin = NULL;
GstCaps *caps = NULL;
gchar bin_name[16] = { };
g_snprintf (bin_name, 15, "source-bin-%02d", index);
bin = gst_bin_new (bin_name);
GstElement *source, *decoder;
g_print ("Now loading:");
g_print (" %s,", uri);
source = gst_element_factory_make ("v4l2src", "src_elem");
g_object_set (G_OBJECT (source), "device", uri, NULL);
GstElement *cap_filter = gst_element_factory_make ("capsfilter", "src_cap_filter");
decoder = gst_element_factory_make ("nvv4l2decoder", "nvv4l2-decoder");
caps = gst_caps_new_simple ("image/jpeg",
"width", G_TYPE_INT, 1920,
"height", G_TYPE_INT,1080,
NULL);
g_object_set (G_OBJECT (cap_filter), "caps", caps, NULL);
g_object_set (G_OBJECT (source), "io-mode", 2, NULL);
g_object_set (G_OBJECT (decoder), "mjpeg", 1, NULL);
gst_bin_add_many (GST_BIN (bin), source, cap_filter, decoder, NULL);
gst_element_link_many (source, cap_filter, decoder, NULL);
if (!gst_element_add_pad (bin, gst_ghost_pad_new_no_target ("src",
GST_PAD_SRC))) {
g_printerr ("Failed to add ghost pad in source bin\n");
return NULL;
}
GstPad *srcpad = gst_element_get_static_pad (decoder, "src");
if (!srcpad) {
g_printerr ("Failed to get src pad of source bin. Exiting.\n");
return NULL;
}
GstPad *bin_ghost_pad = gst_element_get_static_pad (bin, "src");
if (!gst_ghost_pad_set_target (GST_GHOST_PAD (bin_ghost_pad),
srcpad)) {
g_printerr ("Failed to link decoder src pad to source bin ghost pad\n");
}
return bin;
}
The results of running V4l2-ctl are as follows(Use OV4689)
v4l2-ctl -d /dev/video0 --list-formats-ext
ioctl: VIDIOC_ENUM_FMT
Index : 0
Type : Video Capture
Pixel Format: 'MJPG' (compressed)
Name : Motion-JPEG
Size: Discrete 640x360
Interval: Discrete 0.004s (260.004 fps)
Size: Discrete 1280x720
Interval: Discrete 0.008s (120.000 fps)
Size: Discrete 1920x1080
Interval: Discrete 0.017s (60.000 fps)
The screen came up only the first time, but after that, it locked up and would not update the video.
I change another camera(MJPEG), and I try another Gstreamer pipeline, images are not being updated at all.
The execution is as follows
./deepstream-segmentation-app dstest_segmentation_config_semantic.txt /dev/video0
Execution result. It is running normally.
INFO: [Implicit Engine Info]: layers num: 2
0 INPUT kFLOAT data 3x512x512
1 OUTPUT kFLOAT final_conv/BiasAdd 4x512x512
0:00:29.761102557 964 0x55c5fa9e30 INFO nvinfer gstnvinfer_impl.cpp:313:notifyLoadModelStatus:<primary-nvinference-engine> [UID 1]: Load new model:dstest_segmentation_config_semantic.txt sucessfully
Running...
NvMMLiteOpen : Block : BlockType = 277
NVMEDIA: Reading vendor.tegra.display-size : status: 6
NvMMLiteBlockCreate : Block : BlockType = 277
in videoconvert caps = video/x-raw(memory:NVMM), format=(string)RGBA, framerate=(fraction)60/1, width=(int)512, height=(int)512
I made and tried Pipeline for YUYC and Sample for CSI Camera, but I could not update the video.
If you can give me some hints, that would be great.