Hi Luxonis Team, I am currently using the OAK-D Pro PoE in standalone mode, and I’ve set up a TCP stream with the RGB camera, which is working well. Below are two code samples: one for setting up the RGB stream and another for the TCP socket communication. I have two questions related to expanding the functionality of my setup: 1. **AI Model Integration**: I would like to display an AI model result (e.g., the age-gender model from [link for model](https://github.com/luxonis/depthai-experiments/tree/master/gen2-age-gender)) within the same stream. How can I run an AI inference model while maintaining the current RGB stream and show the results in the output? 2. **Additional Sensors**: Currently, I’m streaming only from the RGB camera. How can I also utilize the other two sensors on the OAK-D Pro, such as the stereo depth and IR cameras, and include them in the stream? Are there any examples or recommendations on how to configure and stream multiple sensors simultaneously? oak code: host code: Thank you for your help and guidance! Best regards, Eden

@"edensheiko"#p26431 1. On RVC2 it is generally not possible to draw the results over the image since Script node does not support the OpenCV library. I'd suggest looking through [this](https://github.com/luxonis/depthai-experiments/tree/master/gen2-cumulative-object-counting/standalone) example to see how it is done. 2. I recommend sending the streams through different ports/endpoints for different streams. Eg. for TCP, RGB is on PORT1 and DISPARITY is on PORT2. Make sure you are not exceeding the bandwidth of the network (generally 1gbps). Thanks, Jaka

OAK-D Pro PoE Standalone: AI Model Integration & Multi-Sensor Streaming

edensheiko

Hi Luxonis Team,

I am currently using the OAK-D Pro PoE in standalone mode, and I’ve set up a TCP stream with the RGB camera, which is working well. Below are two code samples: one for setting up the RGB stream and another for the TCP socket communication. I have two questions related to expanding the functionality of my setup:

AI Model Integration: I would like to display an AI model result (e.g., the age-gender model from link for model) within the same stream. How can I run an AI inference model while maintaining the current RGB stream and show the results in the output?
Additional Sensors: Currently, I’m streaming only from the RGB camera. How can I also utilize the other two sensors on the OAK-D Pro, such as the stereo depth and IR cameras, and include them in the stream? Are there any examples or recommendations on how to configure and stream multiple sensors simultaneously?

oak code:

https://privatebin.net/?f44fde23e8eecb13#6tRYuBbDhVLHDeJM3auM9LAYNMFJXh3sU8Bto5hnmdd2

host code:

https://privatebin.net/?ee6156d3b02e2599#269FptWVvNHWpuhV7Uh6XNzXCVgCnxBeKqhvrZE59ZVn

Thank you for your help and guidance!

Best regards,
Eden

jakaskerl

edensheiko

On RVC2 it is generally not possible to draw the results over the image since Script node does not support the OpenCV library. I'd suggest looking through this example to see how it is done.
I recommend sending the streams through different ports/endpoints for different streams. Eg. for TCP, RGB is on PORT1 and DISPARITY is on PORT2. Make sure you are not exceeding the bandwidth of the network (generally 1gbps).

Thanks,
Jaka