pipeline-server
pipeline-server copied to clipboard
Published
20 hours ago •
dlstreamer
dlstreamer
pipeline shows error with splitmuxsink
@nnshah1 I'm running yolov5 on pipeline-server without splitmuxsink. Its shows the proper result. But if i add splitmuxsink in pipeline. the pipeline shows error like this...
{"levelname": "DEBUG", "asctime": "2022-11-10 07:10:05,093", "message": "Setting model-proc to /home/pipeline-server/models/object_detection/coco_yolov5_tiny_608to416_FP32/head_yolov5_tiny_608to416_FP32.json for element detection", "module": "gstreamer_pipeline"}
{"levelname": "DEBUG", "asctime": "2022-11-10 07:10:05,095", "message": "Gst launch string is only for debugging purposes, may not be accurate", "module": "gstreamer_pipeline"}
{"levelname": "DEBUG", "asctime": "2022-11-10 07:10:05,095", "message": "gst-launch-1.0 urisourcebin uri=rtsp://aivid:[email protected]:554/cam/realmonitor?channel=6&subtype=0 ! tee name=t ! decodebin ! video/x-raw ! videoconvert qos=True ! gvadetect model=/home/pipeline-server/models/object_detection/coco_yolov5_tiny_608to416_FP32/FP32/yolov5s.xml model-instance-id=detection_e4e789cc60b511edad0a0242ac110002 model-proc=/home/pipeline-server/models/object_detection/coco_yolov5_tiny_608to416_FP32/head_yolov5_tiny_608to416_FP32.json threshold=0.10000000149011612 ! gvametaconvert tags={\"sensor\": \"sensor\", \"location\": \"location\", \"algorithm\": \"algorithmName\"} ! gvametapublish file-path=/tmp/results_objects.txt file-format=<enum each line is valid JSON with inference results per frame of type gstreamer_pipeline.GvaMetaPublishFileFormat> ! gvapython module=/home/pipeline-server/server/pplcount_copy.py class=PeopleCount ! appsink sync=False emit-signals=True", "module": "gstreamer_pipeline"}
{"levelname": "DEBUG", "asctime": "2022-11-10 07:10:05,185", "message": "Called Status", "module": "gstreamer_pipeline"}
{"levelname": "DEBUG", "asctime": "2022-11-10 07:10:08,185", "message": "Called Status", "module": "gstreamer_pipeline"}
{"levelname": "DEBUG", "asctime": "2022-11-10 07:10:11,187", "message": "Called Status", "module": "gstreamer_pipeline"}
{"levelname": "ERROR", "asctime": "2022-11-10 07:10:11,753", "message": "**Error on Pipeline** b392bf4860c611ed97990242ac110002: **gst-stream-error-quark**: Internal data stream error. (1): ../gst/rtsp/gstrtspsrc.c(6252): gst_rtspsrc_loop (): /GstPipeline:pipeline1/GstURISourceBin:source/GstRTSPSrc:rtspsrc0:\nstreaming stopped, reason not-linked (-1)", "module": "gstreamer_pipeline"}
{"levelname": "DEBUG", "asctime": "2022-11-10 07:10:11,753", "message": "Setting Pipeline b392bf4860c611ed97990242ac110002 State to ERROR", "module": "gstreamer_pipeline"}
{"levelname": "DEBUG", "asctime": "2022-11-10 07:10:14,187", "message": "Called Status", "module": "gstreamer_pipeline"}
{"levelname": "DEBUG", "asctime": "2022-11-10 07:10:17,188", "message": "Called Status", "module": "gstreamer_pipeline"}
{"levelname": "DEBUG", "asctime": "2022-11-10 07:10:20,188", "message": "Called Status", "module": "gstreamer_pipeline"}
{"levelname": "DEBUG", "asctime": "2022-11-10 07:10:23,189", "message": "Called Status", "module": "gstreamer_pipeline"}
here is the pipeline
{
"type": "GStreamer",
"template": ["{auto_source} ! tee name=t ! decodebin ! video/x-raw ! videoconvert name=videoconvert",
" ! gvadetect model={models[object_detection][coco_yolov5_tiny_608to416_FP32][network]} name=detection",
" ! gvametaconvert name=metaconvert ! gvametapublish name=destination",
" ! gvapython class=PeopleCount module=/home/pipeline-server/server/pplcount.py name=object-counter",
" ! appsink name=appsink",
" t. ! queue ! qtdemux ! splitmuxsink name=splitmuxsink"
],
"description": "Person Vehicle Bike Detection based on person-vehicle-bike-detection-crossroad-0078",
"parameters": {
"type": "object",
"properties": {
"detection-properties": {
"element": {
"name": "detection",
"format": "element-properties"
}
},
"detection-device": {
"element": {
"name": "detection",
"property": "device"
},
"type": "string",
"default": "{env[DETECTION_DEVICE]}"
},
"detection-model-instance-id": {
"element": {
"name": "detection",
"property": "model-instance-id"
},
"type": "string"
},
"inference-interval": {
"element": "detection",
"type": "integer",
"minimum": 1,
"maximum": 4294967295,
"default": 1
},
"max-size-time": {
"element": "splitmuxsink",
"type": "integer",
"minimum": 1,
"maximum": 200000000000,
"default": 2000000000
},
"recording_prefix": {
"type": "string",
"element": {
"name": "splitmuxsink",
"property": "location"
},
"default": "/home/pipeline-server"
},
"threshold": {
"element": "detection",
"type": "number"
}
}
}
}
Can you confirm the rtsp stream is compatible with mp4 (qtdemux / qtmux)? splitmuxsink uses qtmux as default - if the incoming stream is not compatible with qtmux / qtdemux - additional modifications would be needed
Another experiment you could do to narrow down the problem is check that the record part of the record_playback sample works. Then change to using your media and let us know what happens.
@nnshah1 @whbruce
With the given pipeline I can run with splitmuxsink and its also generating the videos. But how can I get the timestep in generated video name? working pipeline:
{
"type": "GStreamer",
"template": ["urisourcebin name=source ! decodebin ",
" ! gvadetect model={models[object_detection][coco_yolov5_tiny_608to416_FP32][network]} name=detection",
" ! gvametaconvert name=metaconvert ! queue ! gvametapublish name=destination",
" ! gvapython class=PeopleCount module=/home/pipeline-server/server/pplcount.py name=object-counter",
" ! gvawatermark ! videoconvert ! x264enc ! splitmuxsink muxer=avimux location=\"/tmp/temp-%d.mp4\" max-size-time=30000000000"
],
"description": "Person Vehicle Bike Detection based on person-vehicle-bike-detection-crossroad-0078",
"parameters": {
"type": "object",
"properties": {
"detection-properties": {
"element": {
"name": "detection",
"format": "element-properties"
}
},
"detection-device": {
"element": {
"name": "detection",
"property": "device"
},
"type": "string",
"default": "{env[DETECTION_DEVICE]}"
},
"detection-model-instance-id": {
"element": {
"name": "detection",
"property": "model-instance-id"
},
"type": "string"
},
"inference-interval": {
"element": "detection",
"type": "integer",
"minimum": 1,
"maximum": 4294967295,
"default": 1
},
"recording_prefix": {
"type": "string",
"element": {
"name": "splitmuxsink",
"property": "location"
},
"default": "/home/pipeline-server"
},
"threshold": {
"element": "detection",
"type": "number"
}
}
}
}
to get the timestamp in the video segments do
- remove the element block in the recording prefix parameter
- add a name=splitmuxsink to the splitmuxsink element
{
"type": "GStreamer",
"template": ["urisourcebin name=source ! decodebin ",
" ! gvadetect model={models[object_detection][coco_yolov5_tiny_608to416_FP32][network]} name=detection",
" ! gvametaconvert name=metaconvert ! queue ! gvametapublish name=destination",
" ! gvapython class=PeopleCount module=/home/pipeline-server/server/pplcount.py name=object-counter",
" ! gvawatermark ! videoconvert ! x264enc ! splitmuxsink muxer=avimux name=splitmuxsink max-size-time=30000000000"
],
"description": "Person Vehicle Bike Detection based on person-vehicle-bike-detection-crossroad-0078",
"parameters": {
"type": "object",
"properties": {
"detection-properties": {
"element": {
"name": "detection",
"format": "element-properties"
}
},
"detection-device": {
"element": {
"name": "detection",
"property": "device"
},
"type": "string",
"default": "{env[DETECTION_DEVICE]}"
},
"detection-model-instance-id": {
"element": {
"name": "detection",
"property": "model-instance-id"
},
"type": "string"
},
"inference-interval": {
"element": "detection",
"type": "integer",
"minimum": 1,
"maximum": 4294967295,
"default": 1
},
"recording_prefix": {
"type": "string",
"default": "/home/pipeline-server"
},
"threshold": {
"element": "detection",
"type": "number"
}
}
}
}