@毛宪历
这个板子上，请使用如下命令进行测试先：

python3 ./v4l2dev_2_opencv_show_grey.py --width 2432 --height 2048 --fps 50 --i2c 7

经过验证，xavier平台可以通过设置配置prefer_stride为64，分辨率是2432x2048的大小。但是tx2平台不行，还是得要求256的整数倍。

@xu 识别到产品型号了，正常。

@chenhaoyu
这个实现不了，树莓派只有两个mipi口。

@毛宪历 不好意思，现在我们还没有提供成型的SDK。
提供的资料是以标准v4l2接口实现的。
你可以参考下面这个链接的7.4.4 yavta应用范例。
https://wiki.veye.cc/index.php/MV_Camera_on_Firfly_Boards/zh#yavta.E5.BA.94.E7.94.A8.E8.8C.83.E4.BE.8B

@veye_xumm 好的，非常感谢

@victor Sorry for the late reply.
After communicating with the sensor manufacturer, I found that the built-in temperature sensor in the SC130GS is extremely inaccurate and cannot be used properly.

@dangal
你已经搞成了。
由于你只在i2c-30上接入了一个，因此只识别到了一个相机。其他i2c 31-35都报了读取寄存器错误，这是正常的。

@veye_xumm 好的，感谢

@zang 相机是支持这个分辨率配置的。你的整个操作报什么错误吗？
可以用mv_mipi_i2c.sh脚本读取一下相机实际roi确认是否已经配置到camera里面。
另外300不利于对齐，你可以试试304

@skysplitter It is recommended that you turn off the 3D denoise and wdr features if you have a relatively fast moving vehicle in your usage scenario. This is because these two features essentially use two frames of data to calculate.
2d denoise can be turned on.

@vitob said in NVidia Jetson Jetpack 6 support?:

In the meantime, can you help explain this question on tegra-camera.ko and v4l2-core/*.ko? Why they needs to be updated and where can I find the source code of them? Once updating them using the veye release .ko, will that affect other cameras (for example, I want to port/develop driver for another sensor myself)? Thanks.

These driver changes add support for RAW8 RAW10 RAW12 data types on the one hand, and externally triggered modes on the other.
Their code reference:
https://github.com/veyeimaging/nvidia_jetson_veye_bsp/tree/master/kernel_patch_4_mv/l4t-r36.3

Compile step by step reference:：
https://wiki.veye.cc/index.php/VEYE_CS_Camera_source_for_Jetson

@otca I will continue discussing that issue with you in the other post.

Did u solve it ???
Iam using jetson nano B01 Devkit
I have exact problem like u :
1b047cd6-2fcf-43d5-a62e-ded2c5ab38de-image.png

@veye_xumm 感谢感谢

@leibnitz
是的，qv4l2拿到图像，进行转码，搬移，预览，做了很多工作。整个pipeline的冗长和软件效率的问题导致预览无法达到相机的输出能力。

用下面https://wiki.veye.cc/index.php/Mv_series_camera_appnotes_4_jetson/zh#.E5.B8.A7.E7.8E.87.E7.BB.9F.E8.AE.A1
7.3.2 帧率统计
的命令来统计一下实际相机输出的帧率。

@lin 相机触发模式出图和流模式出图，对你的程序接收图像而言，没有区别啊！

@scroll 您好，我使用其他的相机，遇到了和您一样的问题，请问您解决了吗？

@fancong
请参考：
https://github.com/veyeimaging/nvidia_jetson_veye_bsp/releases/tag/v1.29

@alex_ai said in IMX462M pixelformat on RPi5:

I need wait for clear sky to test and provide feedback

Finally tested, for gain value 20 it is looks pretty same with color sensor with ISP, but background no so black as for color sensor.
For me this issue resolved, attached examples to compare (10 seconds stacks)

IMX462 mono
IMX462M_small.jpg

IMX462 color with ISP
IMX462C_small.jpg