AI-Enabled Efficient and Safe Food Supply Chain

Interpretation and explanation of computer vision classification of carambola (Averrhoa carambola L.) according to maturity stage

The classification of carambola, also known as starfruit, according to quality parameters is usually conducted by trained human evaluators through visual inspections. This is a costly and subjective method that can generate high variability in results. As an alternative, computer vision systems (CVS) combined with deep learning (DCVS) techniques have been introduced in the industry as a powerful and an innovative tool for the rapid and non-invasive classification of fruits. However, validating the learning capability and trustworthiness of a DL model, aka black box, to obtain insights can be challenging. To reduce this gap, we propose an integrated eXplainable Artificial Intelligence (XAI) method for the classification of carambolas at different maturity stages. We compared two Residual Neural Networks (ResNet) and Visual Transformers (ViT) to identify the image regions that are enhanced by a Random Forest (RF) model, with the aim of providing more detailed information at the feature level for classifying the maturity stage. Changes in fruit colour and physicochemical data throughout the maturity stages were analysed, and the influence of these parameters on the maturity stages was evaluated using the Gradient-weighted Class Activation Mapping (Grad-CAM), the Attention Maps using RF importance. The proposed approach provides a visualization and description of the most important regions that led to the model decision, in wide visualization follows the models an importance features from RF. Our approach has promising potential for standardized and rapid carambolas classification, achieving 91 % accuracy with ResNet and 95 % with ViT, with potential application for other fruits.

The Coordination of Multi-Stage Discounts in a Dual Channel Fresh Agricultural Produce Supply Chain: Minimizing the Loss of Quantity and Quality

This paper explores the coordination of the agricultural cooperative to supermarket or E-commerce supply chain, under the condition of quantity loss with a mixed decay function of exponential and logistical distribution. The nature of this process is analyzed, and the corresponding demand and supply functions with single- and multi-stage discount strategies are constructed, respectively, to create a working model. The optimal discount ratios for supermarkets and agricultural cooperatives in decentralized and centralized decision-making modes coupled with single- and multi-stage discounts are calculated, respectively. Finally, a universal optimal strategy is designed, which can be applied to various quantity decay scenarios and makes the discount strategy more generalized. The results show that discounts can coordinate supply chains more effectively; not only would fresh agricultural produce be sold out before it starts to rot, but also the benefit conflicts arising from both supermarkets vs. cooperatives and traditional vs. E-commerce channels could be equilibrated. Further, multi-stage discounts are more difficult to coordinate than single-stage ones, the corresponding optimal discount ratios rely on the initial quantity of fresh agricultural produce in the supply chain, its market share in the traditional distribution channel, the potential market size, retail price, the price sensitivity coefficient of the channel, the cross-elasticity coefficient of prices between different channels, and the properties of the quantity loss. It is concluded that, regardless of what kind of quantity and quality loss occurs, whether decentralized or centralized decision making is selected, or whether the supermarket’s or agricultural cooperative’s discount ratio is considered, a universal price discount consisting of a fixed term and a drift term could both maximize supply chain profit and coordinate this dual-channel supply chain.

Can a Byte Improve Our Bite? An Analysis of Digital Twins in the Food Industry

The food industry faces many challenges, including the need to feed a growing population, food loss and waste, and inefficient production systems. To cope with those challenges, digital twins that create a digital representation of physical entities by integrating real-time and real-world data seem to be a promising approach. This paper aims to provide an overview of digital twin applications in the food industry and analyze their challenges and potentials. Therefore, a literature review is executed to examine digital twin applications in the food supply chain. The applications found are classified according to a taxonomy and key elements to implement digital twins are identified. Further, the challenges and potentials of digital twin applications in the food industry are discussed. The survey revealed that the application of digital twins mainly targets the production (agriculture) or the food processing stage. Nearly all applications are used for monitoring and many for prediction. However, only a small amount focuses on the integration in systems for autonomous control or providing recommendations to humans. The main challenges of implementing digital twins are combining multidisciplinary knowledge and providing enough data. Nevertheless, digital twins provide huge potentials, e.g., in determining food quality, traceability, or designing personalized foods.

Digitalization and Future Agro-Food Supply Chain Management: A Literature-Based Implications

Achieving transition towards sustainable and resilient food systems is a critical issue on the current societal agenda. This study examined the potential contribution of digitalization of the food system to such transition by reviewing 76 relevant journal articles, indexed on the Scopus database, using the integrative literature review approach and descriptive content analysis with MAXQDA 2020 software. ‘Blockchain’ was the top hit among keywords and main concepts applied to the food system. The UK as a country and Europe as a continent were found to lead the scientific research on food system digitalization. Use of digital technologies such as blockchain, the Internet of Things, big-data analytics, artificial intelligence, and related information and communications technologies were identified as enablers. Traceability, sustainability, resilience to crises such as the COVID-19 pandemic, and reducing food waste were among the key benefit areas associated with digitalization for different food commodities. Challenges to practical applications related to infrastructure and cost, knowledge and skill, law and regulations, the nature of the technologies, and the nature of the food system were identified. Developing policies and regulations, supporting infrastructure development, and educating and training people could facilitate fuller digitalization of the food system.