Basic Weed Detection Model
- Given images, the model must accurately determine where the weed is located. It must handle varying weed density, lighting conditions and growth stages of the weeds. Assume the Camera is positioned so it gives a full view of what is in front of the Robot (From the top view).
- Determine the type of model (transfer learning or building it from scratch)
- Find a Reliable Weed detection dataset (We reccomend finding an additional small dataset you can overfit to for debugging purposes)
- Build the model and test on your dataset
The model should be done by the second meeting in June as we will be doing field testing in June.
References: https://www.mdpi.com/2077-0472/15/21/2296