Hi, so the callback function is called on every completed request being returned from the camera. Each request will contain (if configured as shown) a "main" and a "lores" image, you are free to draw boxes on one or other, both or neither images. So if your callback function draws only on the "main" image, then you can use a second encoder for the other image, thereby obtaining one image stream with boxes, and one without. Does that help at all?
It should be easy enough to adapt the imx500 example with the code snippets I posted, and then you would be able to try out the two streams.
It should be easy enough to adapt the imx500 example with the code snippets I posted, and then you would be able to try out the two streams.
Statistics: Posted by therealdavidp — Mon Dec 22, 2025 12:35 pm