[Feature] Yolo-based obstacle detection

[chrehall68]

• Combine depth camera and YOLO bounding boxes to produce a list of obstacles.
• For now, use a standard pretrained YOLO model
• See this link for how to use the camera with depth
• This should be its own package

[GT-96]

So I got the sample code linked running. Aside from setting up a new package, is there anything more to this feature? With basic object/no object detection, it seems like it might be sufficient