IoT-based Cocoa Bean Fermentation Classification System using TCS34725 Colour Sensor and NodeMCU ESP8266

Agriculture is a vital sector in Indonesia, with cocoa representing a major commodity for both domestic processing and international trade. The quality of cocoa beans is strongly influenced by fermentation, yet traditional evaluation methods based on visual inspection are inefficient and prone to human error. This study proposes an Internet of Things (IoT)-based fermentation classification system using the TCS34725 colour sensor integrated with a NodeMCU ESP8266 microcontroller. The device captures colour intensity values of cocoa beans, processes the data through NodeMCU, and transmits results to the Blynk application for real-time monitoring on mobile devices. Experimental testing involved 40 trials across four categories: unfermented–unroasted, unfermented–roasted, fermented–unroasted, and fermented–roasted beans. The system achieved an overall classification accuracy of 65%, with performance influenced by sensor distance and ambient light. The findings demonstrate the feasibility of employing low-cost IoT-enabled colour sensors to support cocoa fermentation monitoring, offering farmers more consistent postharvest quality control and reducing reliance on subjective manual inspection.