"MOS-Based Gas and NIR Spectral Sensors Data for Quality Analysis of Fruits and Vegetables"

"AbstractThis dataset provides sensor readings specifically designed for quality assessment, freshness detection, and ripeness prediction of agricultural produce. The data was collected as part of the research project \"Multi-Modal AI System for Detecting and Predicting the Freshness and Ripeness of Produce,\" the dataset fuses chemical gas sensing data with Near-Infrared (NIR) spectroscopy readings.The data covers 10 distinct types of produce (5 fruits and 5 vegetables) monitored across various stages of decay. Multiple specimens under each fruit and vegetable have been used to ensure the correct variations. The data acquisition system utilizes an array of gas sensors to detect surface-level gaseous emissions, paired with an NIR spectrometer to detect internal compositional changes:MQ4 Sensor: Targeted detection of Methane (CH4) emissions.MQ135 Sensor: Targeted detection of Carbon Dioxide (CO2) and Ammonia (NH3).TGS2602 Sensor: Targeted detection of various Volatile Organic Compounds (VOCs) and odorous gases.AS7263 Spectrometer: NIR spectral readings focused on the 860nm band to identify internal structural variations as they produce transitions from fresh\/unripe to rotten.Raw analog voltage readings from the gas sensors have been converted to digital readings using ADS1015 Analog to Digital Converter and pre-processed into Parts Per Million (PPM) concentration values using standard sensitivity curves. The dataset is organized by produce type and labelled according to distinct quality stages: Unripe, Ripe, and Rotten for fruits, and Fresh, Intermediate Fresh, and Rotten for vegetables. This resource is suitable for researchers developing non-destructive testing (NDT) methods, time-series classification models, and sensor fusion algorithms for food quality assurance."