Internet of Fruit Things (Ioft): A Real Time Iot Based Multi Sensor Framework For Post-Harvest Fruit Health Monitoring

Abstract

The Internet of Things (IoT) has become a driving force behind the development and deployment of smart, sensor enabled devices capable of addressing real-world challenges across domestic, industrial, and agricultural domains. Despite considerable advances in wireless sensing technologies, effective real-time assessment of the health and quality of perishable plant products, particularly fruits and vegetables, remains a significant challenge, contributing to substantial post-harvest losses. Recent estimates indicate that global fruit and vegetable production is approximately 675 million metric tons annually, of which nearly 0.3 billion metric tons are wasted. In India alone, annual production of 86.602 million metric tons is accompanied by an estimated 5.6 million metric tons of spoilage-related waste, with fruits and vegetables accounting for as much as 18% of total production loss, equivalent to approximately ₹13,300 crore discarded annually; government-supported estimates further indicate that 5- 13% of Indian fruit and vegetable output is lost solely due to post-harvest inefficiencies. To address this problem, the present work proposes a real time IoT based fruit health monitoring system (RTIFHMS), termed the Internet of Fruit Things (IoFT), designed to continuously monitor fruit quality and enable early detection of spoilage. The system employs a NodeMCU ESP8266 microcontroller interfaced with an HW-88 moisture sensor and two gas sensors (MQ-135 and MQ-3) via a CD74HC4067 analog multiplexer, with sensor data transmitted to cloud-based IoT platforms, namely ThingSpeak and Blynk, for real-time visualization and early spoilage warning. The experiment over a 21-day period of storage showed a characteristic non-monotonic response of the moisture sensor: an initial decrease from 93% to about 30%, then a significant increase to 100%. This is due to progressive tissue degradation and cellular leakage, not to a real rehydration. Unsupervised K-means clustering of this sensor stream identified three statistically distinguishable response regimes (silhouette score = 0.507) demonstrating the feasibility of data driven pattern recognition as a complement to fixed threshold detection. The proposed IoFT framework integrates low cost multi-parameter sensing, embedded processing, wireless connectivity and cloud based analytics in order to provide a scalable and economically feasible solution for reducing post harvest losses of fruits and vegetables and improving supply chain efficiency in the agricultural sector.