I'm currently using a Raspberry Pi 4 with a Google Coral usb accelerator to run object detection on a Intel Realsense 435i. As of right now im running Mobilenet_SSD_V2 but will begin to experiment with different models in an effort to increase speed and accuracy. With my current setup Realsense_Object_Detection.py is running ~25 fps, but Realsense_Object_Detection_Align.py is running ~20 fps with the depth map disabled and ~8 fps with it enabled. Code optimization needs to be explored.
If you have any suggestions or features that you would like to see, please let me know!
