Train an Axon model to identify key field elements like cargo

[sarango2158]

We can use the Axon tool to train a model for object detection, and offload some decision-making from the drive team. For example, we could train it to recognize cargo on the field and feed that information back to the robot to orient it for intake.

This issue is only concerned with training a model that will recognize the different colored cargo on the field. If the model is reliable, we can create future issues to incorporate that model into our robot logic.

Documentation:
https://docs.wpilib.org/en/stable/docs/software/wpilib-tools/axon/index.html
Some troubleshooting other teams have done:
https://www.chiefdelphi.com/t/any-success-with-axon/399934/4