(663d) Development of a framework for real-time quality monitoring of fresh produce

One-third of the food produced for human consumption is wasted globally. Food wastage can be mitigated through real-time monitoring of quality, which helps in taking informed decisions. As the quality degrades, several physical and biochemical parameters of the food undergo change. Our solution is based on potato case study. The change in the quality of potatoes occurs due to weight loss, sprouting, sugar accumulation, fungal attack, etc. Previous literature shows models for individual parameters such as weight, sugar, etc. but these are only point solutions in contrast to a holistic solution which can be deployed in a real-life food warehouse or a retail store. In the current study, a framework employing seven important parameters (weight, reducing sugar, total sugars, starch, pH, sprout status, disease status) for quality monitoring of potatoes has been developed by using different kinds of models (kinetic, mathematical, and image-based). The models provide prediction of current value as well as forecast for numerical parameters at each time instant which can help user take corrective actions. The framework allows continuous quality monitoring of stored potatoes and shelf-life estimation for different environment conditions using coupled kinetic and image-based models. The models predicted parameters with good accuracy, for e.g., weight with Mean Absolute Percent Error of 0.13-5%, reducing sugar with R2 of 0.7-0.98, image attributes with a classification accuracy of 75-95%. The proposed framework can be employed by multiple stakeholders in the food supply chain (consumer, retailer, etc.) for minimizing food wastage.