Coca Cola seeks to develop an optical method to measure the sugar concentration of a beverage for quality control. A methodology for optical data collection as well as a comprehensive software package for data analysis has been developed for this purpose.

Using images gathered from a planar laser induced fluorescence (PLIF) system, the software package features edge detection algorithms, convolutional neural networks, and sugar concentration image analysis to analyze the mixing process from start to end.

The system is purely optical and can analyze in real-time both how well the solution is mixing as well as the overall sugar content of the final solution, providing Coca Cola an easy to implement, non-invasive method to monitor the quality of soda produced both from their FreeStyle machines and potentially factories worldwide.