Hi,
We are working on designing a custom application with an RGB-IR image sensor. We have implemented a graph as shown below.
RGB-IR-Camera --> Split ---> VISS RGB -> AEWB RGB ->LDC RGB ->|
| --->|
| --> MOSAIC -> Display
| --->|
---> VISS IR -> AEWB IR -> LDC IR ->|
The above implementation is divided into 4 graphs.
- Capture graph containing the capture node and RGB-IR split node
- RGB graph containing the VISS RGB, AEWB RGB and LDC RGB nodes
- IR graph containing the VISS IR, AEWB IR and LDC IR nodes
- Display graph containing Mosaic and Display nodes
During the graph creation, we have defined RGB and IR LDC node outputs as graph parameters. We have also defined Mosaic node input as graph parameters.
When running the graph, we are en-queuing empty output buffers to RGB and IR LDC outputs. We check if they are filled using "vxGraphParameterCheckDoneRef" and dequeue the filled buffer. Once a buffer is dequeued, we would then enqueue it to mosaic node and wait for it to be released. If released, we would dequeue it and then re enqueue the buffer to RGB/ IR LDC.
Srikanth V.