Reflectance intensity dataset of roast coffee beans measured with AS7265X sensor

Published: 5 November 2025| Version 1 | DOI: 10.17632/zpv2kjcyh7.1
Contributor:
Diang Sagita

Description

This dataset contains reflectance intensity measurements of roasted coffee beans collected using the AS7265X multispectral sensor module. The sensor provides 18 discrete spectral bands across the visible to near-infrared region (410-940 nm), capturing spectral responses that correspond to color development and chemical characteristics of roast coffee. The dataset includes samples representing various roasting levels quantified by Agtron value, allowing for detailed analysis of the relationship between roasting degree (Agtron) and multispectral reflectance characteristics. Each sample entry consists of label (type and origin of Indonesian coffee), Agtron value reference, and the intensity values obtained from the AS7265X sensor for all available wavelength channels. The reflectance intensity values are original sensor outputs, which can be used directly or further calibrated depending on the analytical objectives. This dataset is suitable for applications in coffee quality assessment, machine learning–based prediction of roast degree, spectral feature analysis, and development of real-time roasting monitoring systems. It provides a foundation for exploring non-destructive sensing techniques to classify and quantify coffee roasting characteristics. Key Features: Multispectral reflectance data from 18 wavelength channels in arbitrary unit (a.u.). Multiple roast degrees of Arabica and Robusta coffee beans. Original sensor output values for flexible processing and modeling. Enables development of predictive and classification models for roast degree.

Files

Institutions

Institut Pertanian Bogor Departemen Teknik Mesin dan Biosistem, Badan Riset dan Inovasi Nasional Republik Indonesia

Categories

Spectroscopy, Agricultural Engineering, Agricultural Machinery, Engineering Property of Food, Coffee

Licence