Get image buffer from sipl

Please provide the following info (check/uncheck the boxes after creating this topic):
Software Version
DRIVE OS Linux 5.2.0
DRIVE OS Linux 5.2.0 and DriveWorks 3.5
NVIDIA DRIVE™ Software 10.0 (Linux)
NVIDIA DRIVE™ Software 9.0 (Linux)
other DRIVE OS version
other

Target Operating System
Linux

Hardware Platform
NVIDIA DRIVE™ AGX Xavier DevKit (E3550)

SDK Manager Version
1.7.1.8928
other

Host Machine Version
native Ubuntu 18.04

I want to get image buffer from sipl , and send the buffer to consumer(such as cuda).
I have read the source code “nvsipl_camera”, and tried some methods , but it did not work.
Here is my questions:
1.How to get “INvSIPLNvMBuffer *buffer” from SIPL.
2.How to send the buffer to consumer.

Dear @FYJ,
May I know the DRIVE release version.
Have you checked nvsipl_sample and nvscistream samples for this use case? You may have combine these two samples.
Also, if you are using DW already, You can check gmsl camera sample and Image streamer sample to know the workflow.

Dear @SivaRamaKrishnaNV , thanks for your replay.
I used DRIVE OS 5.2.0 and no DW.
I combined nvsipl_sample and nvsipl_camera code, and i used api “SetPipelineConfig()” to get the structure isp0CompletionQueue. and then I used api “isp0CompletionQueue->Get(pBuffer, IMAGE_QUEUE_TIMEOUT_US)” in a thread to get the pBuffer, but at this time the frame rate is 0;

1.So,is my method a the correct way to get the buffer?
2.Also,do you mean this is one of the correct to combine nvsipl_sample and nvscistream samples?(I prefer to use the nvsipl_camera)

Dear @FYJ,

I used api “isp0CompletionQueue->Get(pBuffer, IMAGE_QUEUE_TIMEOUT_US)” in a thread to get the pBuffer, but at this time the frame rate is 0

I did not get this. How this would translate to framerate?

Are you asking for a way to access NvMediaImage buffer data? If so, NvMediaImage is an abstract datatype and accessing image buffer data from NvMediaImage is not possible.
I recommend to make use NvSci APIs for interoperability between NvMedia Image and CUDA Image. So you need to use NvSciStream to send NvMediaImage data to CUDA without any extra copy. Hence I recommended to check NvSciStream sample and APIs first as a starting point for your use case.