@nkos Our driver code is developed based on Radxa's BSP, and the BSP development documentation can be found at the following URL:
https://docs.radxa.com/en/zero/zero3/low-level-dev/kernel

To address the issue you raised, we attempted to set up a Yocto development environment for the Radxa ZERO 3. Unfortunately, we were unsuccessful in this endeavor.

Later, we noticed Radxa's documentation regarding building Linux images with Yocto:
https://github.com/radxa/yocto-manifests

In the "List of Radxa Boards supported" section, the Radxa ZERO 3 is not listed among the supported development boards.

Therefore, we do not recommend using Yocto to build a Linux image for the Radxa ZERO 3.

@veye_xumm
Thank you, I take a some time for tests

@veye_xumm
I'm experiencing issues with capturing a still image from my RAW-MIPI-IMX462M camera.

Current image processing pipeline:

v4l2-ctl -d 0 -v width=1920,height=1080,pixelformat=XY12 --stream-mmap --stream-count 1 --stream-to captured_frame.raw magick -size 1920x1080 -depth 16 -endian LSB gray:captured_frame.raw -depth 8 frame.png

The problem:

First capture after boot produces correct image:frame_good.png

All subsequent captures yield corrupted images:
frame_bad.png

Camera reboot restores proper functionality.

./mv_mipi_i2c_new.sh -w reboot -b 9

Troubleshooting performed:

Tested both XY10 and XY12 pixel formats

Used different conversion tools (ImageMagick, ffmpeg)

Tried 8-bit and 16-bit grayscale output

Applied pixel bit shifting (both pixel_layer_convert and custom scripts)

Compared camera settings pre/post capture using:

v4l2-ctl -d 0 -L Read internal camera registers via ./mv_mipi_i2c.sh

No configuration differences were detected between "working" and "failing" camera states.

What could cause this single-good-frame behavior, and how can it be resolved?

@nimadibaj

We are analyzing the output signal. It seems that between each video line, the CSI-2 output from your module enters Low-Power (LP) mode, which causes the video stream to continuously drop in and out. The module is completely new, and we have not sent any I²C commands to it. It appears that the module should operate in High-Speed (HS) mode, but it does not.

@lozinski Sorry, we currently do not have products that meet your requirements. The VEYE-LVDS-327 has been discontinued for a long time.