Color Sensor for the Determination of the Ripening Stage of Cherry Tomatoes

Dataset, images and code about a article proposal : A low-cost wireless colorimetric sensor node (USD 27.80) was validated against a benchtop Konica Minolta CM-700d spectrophotometer across 130 cherry tomatoes. A multivariate linear calibration pipeline reduced Euclidean color error by 89.6%, achieving high agreement in CIELAB space. Integrated with an on-device TinyML Decision Tree, the system accurately classifies USDA ripening stages with 93.08% accuracy (97.40% precision for ripe fruit) and a sub-28 µs inference footprint on an 8-bit ATmega328P microcontroller. Operating via 915 MHz RF telemetry, this prototype offers an accessible, distributed solution for real-time postharvest monitoring and smart cold-chain logistics.