Perception Subsystem
For perception we utilize a CSI camera to perform dense image segmentation. We use a CNN to classify every pixel into "vein" or "not vein". These labels serve as target points for the our planner and controller, which actuate the toolhead to the vein.