Hyperspectral imaging combined with artificial intelligence techniques to explore the drying behavior of natural Lonicerae Japonicae Flos extracts

The process of drying solute-containing droplets can lead to dynamic redistribution of solutes. Tracking morphological changes and obtaining drying kinetics will help optimize the spray drying process, but there are few techniques available for measuring the spatio-temporal concentration of solutes in drying droplets. In this study, hyperspectral imaging was used as a non-invasive method to simultaneously obtain information on the morphology and moisture content changes of droplets, and was applied to investigate the drying process of Lonicerae Japonicae Flos extract. The Faster R-CNN algorithm was employed to locate the target droplet and identify its size through the series of droplet images recorded by a hyperspectrometer. Droplet moisture content prediction model was established using PLS and ANN algorithms, with the ANN showing better prediction ac-curacy. The hyperspectral imaging combined with artificial intelligence algorithms as a promising method can be used for investigating the drying kinetics of solutions with different drying methods.

Automated summary

Hyperspectral imaging was used to investigate the drying process of Lonicerae Japonicae Flos extract, providing information on the morphology and moisture content changes of droplets. The Faster R-CNN algorithm located and identified the target droplets, while the PLS and ANN algorithms established a moisture content prediction model, with the ANN showing better accuracy.