Proper hygienic conditions are major issues in the industrial food manufacturing process to provide products with prolonged shelf life in the absence of health risks arising from (pathogenic) microorganisms or other residues. Additionally, sustainability in terms of limited use of water and chemicals is becoming increasingly important. Therefore, a cleaning process on-demand with reduced use of chemicals would be beneficial, also for saving time and lowering costs. To implement a need-based washing mechanism, an appropriate inline sensor system for a detection of residues has to be found and installed. Microbiological measurements are not applicable for contamination control due to the long growth time (bacteria up to three days and molds up to five days). Chemical determination methods (e.g. Kjeldahl for protein or Soxlet for fat analysis) are also time-consuming and therefore not useable directly in the production line.

However, optical measurements seem to be particularly suitable for this task, as they are fast and non-invasive. For example, contaminations of deli slicers were detected using fluorescence excitation at 405 nm and emissions at 475, 520 and 675 nm, where residues of cheese and ham were visible (Beck, Lefcourt, Lo, & Kim, 2015). In another study, inline UV–Vis spectroscopy was used to control cleaning-in-place cycles and optimize concentrations of cleaning agents (Berg, Ottosen, van den Berg, & Ipsen, 2017). An early diagnosis of biofouling could be determined via surface-enhanced Raman spectroscopy (SERS), together with a distinction of the involved bacteria species (Chen, Cui, & Zhang, 2015). These spectroscopic techniques are highly sensitive and substance-specific, but also afford high instrumental costs and require skilled personnel for application and data analysis.

The measurement of L*a*b*-color values is simple, fast and economical and therefore widely used in industrial production. Color sorting can be done on the one hand by transforming the reflectance of the product surface into a signal, which varies due to the reflectance shift after a color change and on the other hand by taking black and white pictures where different gray levels represent the desired color or a discoloration. As the lack of color information provokes different limitations, better resolutions are acquired by using red-green-blue (RGB) cameras or sensors with subsequent control algorithms for the data analysis. In the food processing, color sorting can be used e.g. for wheat (Pearson, Brabec, & Haley, 2008), rice (Nitka & Sioma, 2018), peanuts (Pelletier & Reizner, 1992) and bulgur (Bayram & Öner, 2006). For fruits and vegetables, finding of deterioration or quality grading can be performed via color values (Chherawala, Lepage, & Doyon, 2006).

Together with a reduction of water and chemicals due to a demand-oriented approach, also other decontamination agents could be used as substitute for the ecologically harmful chemicals. Plasma-processed water (PPW) could be an alternative, as its reactive substances are also naturally occurring and degrade without environmentally hazardous traces. For the preparation of PPW, plasma-processed air is injected into distilled or tap water. PPW consists of many reactive nitrogen and oxygen species (e.g. HNO2, HNO3, NO*, OH*, O3), which altogether have a high antimicrobial efficiency (Schnabel et al., 2019).

In this study, the use of PPW as alternative disinfectant for conveyor belts was determined in comparison to two common washing solutions (foam detergent and amine-based disinfectant) and their combinations. Residues of minced meat and Braeburn apple on two belt materials (silicone and PVC) were used as examples. Concomitantly, an RGB color sensor system was tested for automated inline detection of food residues on the belts.