The embedded linux devices used throughout the food computer product family are over-spec'd for their immediate application (reading sensors and setting actuators) but are in place so that we can build for the future without having significant infrastructure re-works.

With all this processing power at the device level (the "edge"), we can do interesting computationally heavy tasks that span computer vision and deep learning applications. For example detecting number of leaves or foliage height from the cameras then reporting that as a data point and NOT just an image blog allows for the device to incorporate that computational insight back into the control loops to create the best possible control throughout the cultivation.

This area in general is a bit under-explored and many interesting features could arise out of this pursuit.