We are in the process of building our custom carrier board for the Jetson Nano. We want to give the Nano vision capabilities. We have been looking into the cameras offered by the NVIDIA partners but we haven’t been able to zero down on the cameras that satisfy our requirements.
We are currently using the Yocto build system. The meta-tegra layer in use provides the L4T 32.4.3 release.
It seems that the Raspberry Pi Camera Module 2 is supported by default on the Jetson Nano Devkit. It uses the IMX219 sensor which has driver support built into the kernel. This camera looks like a promising option right now.
The Raspberry Pi Camera Module 2 is a 2 lane module. We are also looking to test a 4 lane camera module.
The kernel defconfig in use has the following sensors listed:
# # NVIDIA overlay Encoders, decoders, sensors and other helper chips # CONFIG_VIDEO_IMX185=m CONFIG_VIDEO_IMX219=y CONFIG_VIDEO_IMX268=y CONFIG_VIDEO_IMX274=m CONFIG_VIDEO_IMX318=y CONFIG_VIDEO_LC898212=y CONFIG_VIDEO_OV5693=m CONFIG_VIDEO_OV9281=y CONFIG_VIDEO_OV10823=y CONFIG_VIDEO_OV23850=y CONFIG_I2C_IOEXPANDER_PCA9570=y CONFIG_VIDEO_TC358840=y CONFIG_I2C_IOEXPANDER_SER_MAX9295=y CONFIG_I2C_IOEXPANDER_DESER_MAX9296=y CONFIG_VIDEO_IMX390=y
Based on the info provided above, I have some questions:
How is the support for Raspberry Pi Camera Module 2 added to the Nano? Is it only driver level support, or is there a device tree overlay that is added so that the camera module is detected?
If there is indeed device tree overlay added, would modifications to it be possible so that the custom carrier board supports the module?
Are there any other sensors / modules that have driver support in the L4T 32.4.3 Linux Kernel 4.9, just like the IMX219?