Custom model object detection with two camera inputs

General
1

Hello,
i successfully run custom hef file object detection wtih single camera on raspberry pi with hailo AI Kit.

I tried multiple setups, topics and chatbots regarding two camera input but with no luck.

i need to be able to scan detected object from two sides at the same time. Is there any easy way how to add second input or model twice at same time?
Using two mipi raspberry pi cameras but can switch to USB’s.

Thank you for any help

2

Maybe the TAPPAS Multistream example can give you some ideas:

GitHub - Tappas - Multistream Detection

3

Thank you for recommendation but TAPPAS multistream example is one of things i already tried but getting errors i was unable to solve.

when follow official guide for manual TAPPAS installation i am getting error message about missing gcc9 and g++ 9 that i was unable to find solution for.

20250324_12h48m45s_grim
20250324_12h48m45s_grim1920×1080 126 KB

20250321_12h16m38s_grim
20250321_12h16m38s_grim1920×1080 188 KB

when i am following seeed benchmark example for multistream i successfully installed tappas but getting error about missing “hailoroundrobin”