BACKGROUND

An accurate AI regression model will be constructed to predict tomato mold severity and likelihood of load rejection at processing facilities based on 10 years of data available from the Processing Tomato Advisory Board (PTAB) and the digital model of mold growth on the fruit based on physiological and environmental factors. A digital model of mold growth and potential risk of spoilage will be modeled. Meanwhile, an accurate AI regression model will be structured to predict tomato spoilage and load rejection based on the given fruit properties and sourcing metadata. The resultant data set will be used to retrain the initial multimodal AI model in order to determine if the additional predictors improve the accuracy of predicting tomato damage and spoilage.

GOALS

• Develop a database of processing tomato quality at the point of delivery to processing facilities and associated metadata for harvesting and hauling;
• Develop digital model to predict mold growth based on quality parameters and agricultural meta-data;
• Develop an AI model of tomato damage, spoilage, and serum leakage based on said meta data;
• Validate model to determine prediction accuracy

IMPACT

Identifying predictors for tomato damage during transportation and developing a digital model to predict potential for mold growth and product rejection will enable control strategies that prevent the hauling of loads that are likely to be rejected, improve processing economics, avoid loss of embedded resources in damaged fruit, and decrease the environmental challenges posed by discharging high strength wastewater.