Hello,
I am trying to use my official Raspberry Pi HQ camera with the Sony IMX477 sensor on my jetson nano orin L4T 36.4.7 jetpack6.2. I modified the settings in jetson-io, enabling both the IMX477-dual and IMX477-A modes. Both modes correctly detect my camera using the following command:
v4l2-ctl --list-devices
Output obtained:
NVIDIA Tegra Video Input Device (platform:tegra-camrtc-ca):
/dev/media0
vi-output, imx477 9-001a (platform:tegra-capture-vi:1):
/dev/video0
Then, I start a video stream using the following GStreamer command:
gst-launch-1.0 nvarguscamerasrc ! \
'video/x-raw(memory:NVMM),width=3820,height=2464,framerate=21/1,format=NV12' ! \
nvvidconv flip-method=0 ! \
'video/x-raw,width=960,height=616' ! \
nvvidconv ! nvegltransform ! nveglglessink -e
Unfortunately, the output is really poor despite trying several parameters.
Do you have any solution to this problem?
Hello @vincent.specht
Could you please try testing by using a pipeline similar to the following?
FRAMERATE=30
gst-launch-1.0 nvarguscamerasrc sensor-id=$SENSOR_ID ! "video/x-raw(memory:NVMM),width=3840,height=2160,framerate=$FRAMERATE/1" ! nvvidconv ! "video/x-raw(memory:NVMM),width=1920,height=1080,framerate=$FRAMERATE/1" ! nvoverlaysink
To see if the poor image persists
Also, I see that in your current pipeline that you are trying to convert from NVMM memory to CPU memory and then you convert to NMM again. Try not to perform that conversion and change your pipeline to the following for testing:
gst-launch-1.0 nvarguscamerasrc ! \
'video/x-raw(memory:NVMM),width=3820,height=2464,framerate=21/1,format=NV12' ! \
nvvidconv flip-method=0 ! \
'video/x-raw(memory:NVMM),width=960,height=616' ! \
nvegltransform ! nveglglessink -e
Please, let us know your results!
Regards!
Eduardo Salazar
Embedded SW Engineer at RidgeRun
Contact us: support@ridgerun.com
Developers wiki: https://developer.ridgerun.com/
Website: www.ridgerun.com
Confirm if already applied below patch.
[Argus][JP-6.2] Resolve blurry capture results, improve image quality
https://forums.developer.nvidia.com/t/320681/15/
I tried your corrections, but they didn’t change the image quality. I also tested modifying the IMX477-C parameter and switching the camera to the MIPI/CSI 1 port. Then, I used nvgstcapture-1, which solved the image issues. Thank you for your input.