We developed a novel servo system to control micro-droplets. Our machine vision algorithms estimate droplet size and feedback that signal to control a pump’s flow-rate. The closed-loop transient response below shows that droplet size is quite controllable.
Detecting different sized droplets as well as merging two droplets into a larger-single droplet is important. The droplet detection algorithm detects and classifies droplets of various shapes and sizes, an example image below shows two different sized droplets. Large droplets are colored in orange and blue droplets are colored in cyan by the droplet detection algorithm. Note: the image input into the algorithm is grey-scale, the colors are applied by the algorithm in post processing. The algorithm detects each droplet morphology and then classifies each droplet as “Large” or “Small” based on that morphology.
Another servo loop controls the pairing of the two droplet sizes, each orange droplet needs to pair with a single blue droplet. The histogram on the left shows the open loop pairing ratio (#cyan droplets/#orange droplets) and the histogram on the right shows the closed loop pairing ratio. The closed loop pairing ratio has a tighter distribution, thus the process is much more stable and controllable.