Examples of Custom Object Detection Model for ROS2 Wrapper

Software ROS2
1

I’m getting Cuda errors when I am trying to use a custom trained ONNX file. It is crashing with a resizing error. I know this is a new functionality, but I was wondering if there was any specific examples available or tutorials with anyone that has had success with this. I think it could be with the shape of the output of my model but I’m not super familiar with that. I tried to dig through the code but I wasn’t able to follow it too well. I have had success with running my own scripts with a custom object detection model using an .engine file and the zed and publishing that to ROS but I was hoping to move towards using the full wrapper.

Thanks for any help and support.

2

Hi Braden,

Have you used our ROS2 tutorial for object detection : https://www.stereolabs.com/docs/ros2/090_object-detection ?

Best,
Rodolphe Perrin

Stereolabs Support

3

Hi,
Yes I have extensively reviewed that page as well as other pages on object detection. I would just love to see one example of someone that has actually done it. And any more requirements on the input and output of the ONNX yolo model.