IoT-based Cost-Effective Fruit Quality Monitoring System using Electronic Nose

Post-harvest losses due to subjective quality assessment cause significant damage to the economy and food safety, especially in countries like Bangladesh. To mitigate such damages, objective decision-making backed by scientific methods is necessary. An IoT-based, cost-effective quality monitoring system can provide a solution by going beyond subjective quality monitoring and decision-making practices. Here, we propose a low-power, cost-effective fruit quality monitoring system with an array of MQ gas sensors, which can be used as an electronic nose. We track the volatile gas emissions, specifically ethanol, methane, and ammonia, encompassing both ripening and decomposition for a set of bananas. Based on the gas concentration thresholds, we develop a mathematical model to accurately assess fruit quality. We also integrate this information into a dashboard for prompt decision-making and monitoring to make it useful to the farmers. This approach has the potential to reduce economic losses, enhance food safety, and provide scalable solutions for the supply chain.

💡 Research Summary

This paper presents the design, implementation, and evaluation of a low-cost, IoT-based fruit quality monitoring system that functions as an electronic nose (E-nose). The research addresses the critical issue of post-harvest losses in developing economies like Bangladesh, where subjective quality assessment leads to significant economic and food security challenges.

The proposed system centers around an array of affordable metal-oxide semiconductor (MOS) gas sensors: MQ-3 (targeting ethanol), MQ-4 (methane), and MQ-135 (ammonia). These sensors, along with a DHT22 temperature/humidity sensor, are connected to an ESP32 microcontroller with built-in WiFi. The entire hardware setup, mounted on a sealed container, costs approximately $18 (USD). Bananas were chosen as the test fruit due to their high post-harvest loss rate and distinct volatile emission profile during ripening and decomposition. Data (gas concentration, temperature, humidity) was collected every minute for over three days.

A key contribution of the work is the development of a mathematical model for quality assessment, avoiding computationally intensive machine learning to ensure simplicity and cost-effectiveness. The model is based on gas concentration thresholds. Through visual inspection, the researchers identified specific concentration levels (in ppm/kg) for each target gas at two key stages: “ripe” and “decomposed.” Using these thresholds, they formulated a normalized quality index (Q_gas) for each gas, designed to drop from near-perfect (0.98) at the ripening threshold to zero beyond the decomposition threshold. Separate indices were also created to account for the impact of suboptimal storage temperature (Q_temp) and relative humidity (Q_RH). The final, comprehensive quality score (Q_total) is a weighted average of all five indices (three gases + temp + RH). The weights are assigned based on the observed rate of change of each factor and sensor reliability, giving higher importance to gases like methane and ammonia that show more dynamic changes during spoilage.

To ensure data credibility from the low-cost sensors, the team performed a rigorous signal quality analysis, calculating metrics like Signal-to-Noise Ratio (SNR), residual noise, and autocorrelation. This analysis confirmed that the MQ-135 (ammonia) sensor provided the most reliable and stable readings, while the MQ-3 (ethanol) sensor was more susceptible to noise.

The experimental results demonstrated the system’s functionality. Under non-ideal, warm storage conditions (32°C, 97% RH), the banana shelf life was short. Sensor readings showed a clear increase in all gas concentrations over time, with methane and ammonia rising sharply between days 2 and 3, coinciding with the onset of decomposition visually observed. The calculated combined quality index successfully reflected this deterioration, showing a marked drop during the same period.

Finally, to make the technology accessible to its intended end-users—rural farmers—the system includes a real-time web dashboard. This dashboard displays current sensor readings, historical trends, the calculated quality score, and a simple five-level quality categorization (Excellent, Good, Moderate, Bad, Rotten). It is available in both Bengali and English.

In discussion, the authors emphasize that the primary achievement is creating a cost-effective (sub-$20), scalable, and practical solution that moves quality assessment from subjective judgment to objective, data-driven decision-making. This approach has the potential to significantly reduce economic losses, enhance food safety, and be adapted for various fruits and points in the supply chain.