TY - JOUR
AU - Castro Neto, Jarbas C.
AB - The chemical composition of coffee directly influences the quality of the beans. Obviously, the type of roasting impacts the sensory tastes, flavors, and other attributes. Currently, score analysis is evaluated by humans, introducing a subjective aspect. However, Raman spectroscopy is a technique that can obtain information about vibrational behavior in the second other. Due to these technical points, spectrum acquisition provides more information about the samples compared to first-order fluorescence. This research acquired spectra of two coffee types: Arabica (Coffee Arabica) and Conilon (Coffea Canephora). After pre-treatment, the spectra obtained only Raman effects. For data collection, a spectrometer was used. The apparatus consists of a 785 nm diode laser that illuminates the samples, collects scattering data as well fluorescence, and sends it to the spectrometer. Subsequently, the SVM method was applied to the Raman spectra of green coffee beans, achieving an accuracy of around 88 %. These results demonstrate the potential of applying artificial intelligence methods to classify coffee bean samples. Our goal is to develop a device capable of evaluating these aspects without altering the physical-chemical properties of the coffee beans, providing an impartial assessment.
TI - Raman spectral analysis and quality classify: aided by artificial intelligence
JF - Proceedings of SPIE
DO - 10.1117/12.3055052
DA - 2025-03-19
UR - https://www.deepdyve.com/lp/spie/raman-spectral-analysis-and-quality-classify-aided-by-artificial-2bVSJ502eq
SP - 133570J
EP - 133570J-7
VL - 13357
IS - 
DP - DeepDyve
ER -