0

I'm trying to get a Luxonis camera to run a custom model. I have been successful at running the Luxonis depthai examples on a virtual environment on my computer (MacBook Pro) with the Oak-D. I was also able to create a model "Oak-D ball detection" that does a pretty good job at identifying balls. I was able to check this in the example web app and it works well enough.

However, I cannot seem to deploy it to the Luxonis camera. Following Roboflow's steps in my virtual environment (https://docs.roboflow.com/inference/luxonis-oak#using-the-oak-inference-api), I make it to step 4, "run the server with the following command.” I think it doesn’t like this because when I go to http://localhost:9001/validate the webpage says: "OAK device connection not validated, follow docs.roboflow.com to debug: No available devices.” This seems strange since I am able to run example code.

How do I remedy this? I do not see anything in the documentation about the OAK device failing to connect.

4

1 回答 1

0

根据我的经验,如果您可以运行 Luxonis 示例 python 脚本,那么 Luxonis docker 也会找到该设备......

作为临时步骤,您可以确保他们的 docker 在您的主机上运行

docker pull luxonis/depthai-library
docker run --rm \
    --privileged \
    -v /dev/bus/usb:/dev/bus/usb \
    --device-cgroup-rule='c 189:* rmw' \
    -e DISPLAY=$DISPLAY \
    -v /tmp/.X11-unix:/tmp/.X11-unix \
    luxonis/depthai-library:latest \
    python3 /depthai-python/examples/rgb_preview.py
于 2021-10-28T19:12:16.860 回答