NIR monochrome imaging for monitoring of apple drying process: Light-emitting diode and Band-pass filter imaging techniques

• DOI: 10.1016/j.fbio.2023.102898
• Publication Date: 2023-06-29T01:20:52Z
• Authors: Arman Arefi; Barbara Sturm; Oliver Hensel; Sharvari Raut
• Keywords: Shrinkage; Digital Imaging

The convergence of drying technology and Digital Twins is enabling a transition from traditional methods to new phase where the process can be continuously monitored. A crucial aspect development digital model kinetic reactions. In this context, hyperspectral imaging has been widely utilized, however, it limitations due its high cost computational intensity, which restrict application lab-scale settings. To address limitation, study introduces technically simple cost-effective techniques: Light-emitting diode (LED) Band-pass Filters (BPF). These techniques were implemented at 980 nm 1450 measure moisture content, Soluble Solids Content (SSC), shrinkage apple slices undergoing 60 °C. Images during captured train Gaussian Process Regression (GPR) models. Moisture content achieved highest accuracy, with GPR models yielding R-squared values 0.995 0.993, RMSE 1.85% 2.45% for LED BPF, respectively. Similarly, SSC (R-squared ≥0.874 ≤7.85%) ≥0.979 ≤4.29) well predicted. Furthermore, accurate prediction results external apples demonstrated models' reproducibility. line concept twins-based smart drying, LEDs embedded into an experimental dryer exhibited accuracy ≥0.968 ≤4.6%) in inline content. This represents significant step towards dryers using sensors.