TDA4VM: Two video streams combine one CSI to TDA4. When TDA4 executes app_multi_cam Demo, the Frame Drop Count is accumulated

Hi,

May I know if you are using 2 Capture nodes for this usecase?

In Multi-cam usecase, We use only 1 Capture node, as the multi sensor output is assumed to be of the same resolution (as this is required by the object array).

May I know how have you modified the multi-cam demo to cater to your usecase?

Regards,

Nikhil

yes，i used 1 node，how do i add one new node， thanks for your reply

Hi,

Currently, you would be having 1 Capture node + 1 Display node.

You could refer the single cam demo and create 2 sets of Capture and Display node and pass each camera's input to each capture node individually.

Regards,

Nikhil

dear TI expert

         thanks for your reply, I had a problem before I created the new node,After solving this problem, I will create  new nodes according to your suggestion

Regards

Hi,

Sure, Awaiting for your observations here.

Let me know if you are facing any problems here.

Regards,

Nikhil

dear TI expert

How do I convert a data type to NV12 before mosaic processes it? mosaic module input type is Vx_object_array and output is vx_imag, but tivxDisplayM2MNode is vx_image in input and output

Regards

Hi,

You could replicate the tivxDisplayM2MNode using the API vxReplicateNode() and send the 0th instance to the displayM2Mnode. 
This way the framework takes care of all the images in the object array.

Regards,

Nikhil

thanks nikhil

The data type problem has been solved. i have anther problem about create new capture node，capture nodes for cameras with different resolutions，How should the code be modified?

Here's the log at run time

dearTI expert

Whether two cameras with different resolutions can use the same graph?When I use only one Graph, the following error occurs

Hi,

It would be recommended to use 2 separate graphs running on two tasks each containing a capture and a display node (as you have YUV422 image format).

QW Wang said:

capture nodes for cameras with different resolutions，How should the code be modified?

I see that you are using same captureObj for both nodes. Is that correct? If yes, please use different struct for both capture nodes as the configuration parameters are different for both capture nodes.

Regards,

Nikhil

Thanks for your reply，I will try your recommendation