AR0820 GMSL Camera RAW format memory layout

Please provide the following info (tick the boxes after creating this topic):
Software Version
DRIVE OS 6.0.8.1
DRIVE OS 6.0.6
DRIVE OS 6.0.5
DRIVE OS 6.0.4 (rev. 1)
DRIVE OS 6.0.4 SDK
other

Target Operating System
Linux
QNX
other

Hardware Platform
DRIVE AGX Orin Developer Kit (940-63710-0010-300)
DRIVE AGX Orin Developer Kit (940-63710-0010-200)
DRIVE AGX Orin Developer Kit (940-63710-0010-100)
DRIVE AGX Orin Developer Kit (940-63710-0010-D00)
DRIVE AGX Orin Developer Kit (940-63710-0010-C00)
DRIVE AGX Orin Developer Kit (not sure its number)
other

SDK Manager Version
1.9.3.10904
other

Host Machine Version
native Ubuntu Linux 20.04 Host installed with SDK Manager
native Ubuntu Linux 20.04 Host installed with DRIVE OS Docker Containers
native Ubuntu Linux 18.04 Host installed with DRIVE OS Docker Containers
other

Can we get some information regarding the memory layout of RAW12 camera images? I found a related discussion, but it was about a different sensor: AR0231 GMSL Camera RAW format memory layout

Based on our testing, every pixel value (RAW12) is saved in 16bits, and 4 LSBs seem to be replicas of 4 MSBs for padding purpose. Is this correct? Thanks!

Yes, your understanding is correct. In RAW12 images, each pixel value is stored in 16 bits, and the 4 least significant bits (LSBs) are typically used as padding, replicating the 4 most significant bits (MSBs).

May we know why you need this information or if there’s anything specific you’d like to achieve with the RAW12 camera images?

@VickNV, Thanks! We are experimenting to use the RAW12 images from ICP (without ISP), so we need to figure out how to get accurate value of every pixel.

Thank you for the update! To assist you effectively, could you please share more details on why you need to obtain accurate values for every pixel from the RAW12 images when bypassing the ISP (Image Signal Processor)? Understanding the specific use case or application can help us provide more tailored guidance and support.

Sure. We are exploring the possibility to train NN models for computer vision purposes with the RAW images instead of RGB images.

With the AR0820 sensor, the image frame from ICP is in RAW12 format by default. Can we modify the sample code to change the format? For example, RAW16?

The document (SIPL Guidance on Output Image Formats | NVIDIA Docs) says “ICP output formats are dependent on the sensor used”. Does that mean we cannot change the format by changing the code?

Yes, camera module development is out of the forum support scope. For this matter, I recommend reaching out to your NVIDIA representative for guidance and support.

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