Low-cost VIS/NIR range hand-held and portable photospectrometer and evaluation of machine learning algorithms for classification performance

In this study, the electronic design of a low-cost and portable spectrophotometer device capable analyzing in visible-near infrared region was established. The C#.NET-based user-friendly control software development machine learning algorithms for data classification as well comparison results were presented. When implementation studies are reviewed literature, two groups subjects become prominent: (i) new fabrication, (ii) solution approaches to current problems by combining commercial spectrometer systems devices with artificial intelligence applications. This work encompasses both groups, supportive approach has been followed on how transform theoretical knowledge into practice help from production. Three spectral sensors, each six photodiode arrays, adopted spectrophotometer. Thus, 18 features belonging sample acquired optical 410 nm 940 band range. analyses conducted 9 different food types powder or flake structures. A Support Vector Machines (SVM) Convolutional Neural Network (CNN) employed classification. As result, SVM CNN achieved 97% 95% accuracies, respectively. Moreover, we provided measurement data, circuit designs, API files containing graphical user interface (GUI).