Fluence (UV dose) distribution assessment of UV-C light at 254 nm on food surfaces using radiochromic film dosimetry integrated with image processing and convolutional neural network (CNN)

Abstract

Uniform Fluence (UV Dose) distribution on food surfaces is essential for an effective UV process design. In this study, the use of radiochromic films (RCFs) with a computer vision system (CVS) integrating image processing and Convolutional Neural Network (CNN) is proposed as an alternative method to assess Fluence distribution of UV-C light at 254 nm on food surfaces. The color difference of RCFs exposed to different UV irradiance and exposure times was correlated with Fluence. The validity of the developed methodology was proved by applying it to the surface of apple fruits of different shapes and sizes. A linear relationship was found between the color difference of RCF and Fluence. The maximum Fluence to be determined using RCFs was similar to 60 mJ/cm(2). The color of the films after UV irradiation remained stable for up to 15 days in darkness when stored at room and refrigeration temperatures. The results showed that RCF can be used as an alternative UV dosimeter.