About mipi unpack

I have a Jetson Origin Nano board and an IMx219 camera. Currently, IMx219 is RAW10, with 10 bits written to 16 bits by default. How to modify the driver program to directly store 10 bits .

The actual meaning is actually an image of 1920 * 1080, with a saved size of 4147200. How can I make it smaller

Can’t do it due to VI HW design. You may need handle it after VI to memory.

Thanks

Is there a similar case? What should I do specifically

Reference to below topic for it.

hello, my camera module is imx219
v4l2-ctl --set-fmt-video=width=1920,height=1080,pixelformat=RG10 --set-ctrl bypass_mode=1 --stream-mmap --stream-count=1 -d /dev/video0 --stream-to=imx219.raw
if i don’t --set-ctrl bypass_mode=0 ,i can’t save picture
imx219 does not support bypass mode?

I saved an image according to what you sent and executed it
hexdump imx219_nano.raw | head -n 10

图片

For example, the LSB at this position is not 0. Is that why? Did I misunderstand

Orin Nano should be use T_R16 instead of T_R16_I

Can I use GPU to complete the task of displacement per frame? I think using CPU to process data displacement per frame and save it may not be able to handle it. Do you think my understanding is correct? Do you have any opinions on using GPU for data processing

Yes, you can reference to MMAPI sample code like 12_v4l2_camera_cuda for how to using GPU to process frame data.

Or another method, can we use GPU to compress each frame of data? Is this fast? Is there a similar case? Which method do you think is better? Actually, our goal is to reduce the size of the data.

You can try cudaBayerDemosaic to check the performance although it’s argus interface instead of v4l2.

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