A central kitchen food material whole-process traceability and freshness intelligent management and control method based on an internet of things

Abstract

This invention belongs to the field of intelligent food ingredient management technology, specifically a method for intelligent traceability and freshness control of central kitchen ingredients based on the Internet of Things (IoT). The method includes the following steps: assigning a unique RFID tag to each batch of ingredients, containing supplier information, quarantine reports, production dates, and ingredient categories. This invention utilizes a combination of RFID tags and blockchain technology to construct a consortium blockchain covering all key nodes of the supply chain. Smart contracts ensure that data undergoes rigorous verification before being uploaded to the blockchain and remains tamper-proof afterward, thus achieving highly reliable traceability throughout the entire process from source to end. Secondly, by integrating NFC tags with miniature gas sensors with an environmental sensor network, and combining edge computing and food spoilage prediction models, the system can dynamically sense and calculate the real-time freshness index of ingredients, achieving a fundamental shift from relying on static shelf-life to dynamic and accurate quality assessment.

Description

Technical Field

[0001] This invention relates to the field of intelligent food ingredient management technology, specifically to a method for intelligent management and control of the entire process of food ingredient traceability and freshness in a central kitchen based on the Internet of Things. Background Technology

[0002] Central kitchens are the core of standardized and large-scale production in the modern catering industry. However, they have a wide variety of ingredients and complex circulation processes. Ingredient traceability often uses barcode or QR code technology to record basic information such as the source and production date of the ingredients. This technology has disadvantages such as small information capacity, easy damage, inability to read in batches, and easy data tampering, resulting in incomplete traceability information and low credibility. At the same time, the previous management of ingredient freshness relied on regular manual inspections and subjective experience judgment, which could not achieve real-time and accurate perception and early warning of the process of food spoilage.

[0003] In addition, management systems are usually single-function, focusing either on traceability or environmental monitoring, and have failed to form a closed-loop management solution that integrates real-time monitoring, intelligent early warning, decision optimization and reliable traceability from suppliers to end users. Therefore, we propose an IoT-based method for full-process traceability and intelligent freshness management of central kitchen ingredients to solve the above problems. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] To address the shortcomings of existing technologies, this invention provides a method for intelligent control of the traceability and freshness of ingredients in a central kitchen based on the Internet of Things, thus solving the problems mentioned in the background section.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, the present invention specifically adopts the following technical solution:

[0008] A method for end-to-end traceability and intelligent freshness control of ingredients in a central kitchen based on the Internet of Things includes the following steps:

[0009] S1: Assign a unique RFID tag to each batch of ingredients, containing supplier information, quarantine report, production date, and ingredient category; attach an NFC flexible tag integrating a miniature gas sensor to the individual packaging of the ingredients; read the RFID tag and NFC flexible tag information through the IoT terminal at the procurement end, and simultaneously collect the initial quality data of the ingredients, including the component content detected by spectral analysis and the shape integrity captured by AI vision; encrypt the tag information and the initial quality data and upload them to the edge computing node;

[0010] S2: Install temperature and humidity sensors and gas sensors for detecting ethylene and ammonia concentrations inside the cold storage or refrigeration cabinet of the central kitchen to collect storage environment data in real time; the edge computing node performs fusion analysis on the storage environment data and the gas concentration data of the food surface microenvironment fed back by the NFC flexible tag, and calculates the real-time freshness index based on the pre-trained food spoilage prediction model.

[0011] S3: An AI visual inspection module and a spectral re-inspection module are set up on the processing line to perform a second verification of the freshness of the ingredients before processing; if the freshness index is lower than the preset threshold, a local warning is triggered and processing is suspended; at the same time, the operator, equipment number, and processing time data of the processing link are bound to the RFID tag of the ingredients and uploaded to the blockchain module.

[0012] S4: Deploy GPS positioning modules and on-board temperature and humidity sensors on delivery vehicles to upload location information and vehicle environment data in real time; construct a virtual mapping of delivery vehicles through a digital twin module, and send a graded warning to the control terminal if environmental data exceeds the standard or deviates from the preset route;

[0013] S5: Users can scan the RFID or NFC tags of ingredients through the barcode terminal and call the full-process data stored in the blockchain module to achieve forward and reverse traceability; the digital twin module simulates different scheduling schemes based on historical data and outputs the optimized management strategy that minimizes the loss rate of ingredients.

[0014] Furthermore, the RFID tag in S1 is a UHF RFID tag, which supports batch reading within a 10-meter range; the miniature gas sensor integrated in the NFC flexible tag is used to detect the concentration of micro-environment gas generated by the respiration of food, and feeds the data back to the reading terminal through near-field communication.

[0015] Furthermore, the food spoilage prediction model in S2 is obtained through machine learning training: using historical storage data of different types of food as training samples, a classification model is constructed using the random forest algorithm. The historical storage data includes temperature and humidity, storage time and gas concentration. The higher the freshness index value output by the model, the better the freshness.

[0016] Furthermore, the AI ​​visual inspection module in S3 uses an AI visual camera to collect images of food ingredients and identifies mold, rot, or insect infestation defects through the YOLOv8 algorithm, achieving high detection accuracy. The spectral re-inspection module uses a near-infrared spectrometer to determine freshness by analyzing changes in the moisture and sugar content of the food ingredients, with extremely short detection time for a single food item.

[0017] Furthermore, the graded early warning in S4 is a three-level early warning, and its triggering conditions are as follows: Level 1 reminder, triggered when environmental data exceeds the threshold by less than 5% and the duration is less than 5 minutes; Level 2 intervention, triggered when environmental data exceeds the threshold by 5%-10% and the duration is 5-10 minutes, the system automatically adjusts the vehicle temperature control equipment; Level 3 interception, triggered when environmental data exceeds the threshold by more than 10% and the duration is more than 10 minutes, the system notifies the control center and delivery drivers to suspend delivery.

[0018] Furthermore, the blockchain module in S5 adopts a consortium blockchain architecture, whose nodes include suppliers, central kitchens, testing institutions, regulatory authorities, and terminal stores; each node interacts with data through smart contracts: the food quarantine reports uploaded by suppliers must be verified by the testing institution node before they can be stored on the blockchain; the processing data uploaded by the central kitchen can be audited in real time by the regulatory authority node.

[0019] Furthermore, in S5, the digital twin module uses an improved genetic-ant colony fusion algorithm for processing scheduling optimization: using the freshness index of ingredients as weights, the initial processing order and equipment allocation scheme are first generated using a genetic algorithm, and then the time nodes in the scheme are optimized using an ant colony algorithm to avoid high-freshness ingredients from waiting too long, ultimately achieving the optimization goal of significantly reducing the processing loss rate.

[0020] A central kitchen food ingredient traceability and freshness intelligent control system based on the Internet of Things, comprising:

[0021] The perception layer includes RFID tags for identifying ingredients and NFC flexible tags with integrated miniature gas sensors, temperature and humidity sensors and gas sensors for collecting environmental data, an AI visual inspection module and a spectral re-inspection module for collecting quality data, and a GPS module for positioning.

[0022] The transport layer includes edge computing nodes, a 5G communication module, and a LoRa gateway; the edge computing nodes are used to perform preliminary processing on the data collected by the perception layer, and upload the processed data through the 5G communication module and the LoRa gateway.

[0023] The platform layer includes a blockchain module, a digital twin module, and a database module; the blockchain module is used to store tamper-proof, end-to-end traceability data; the digital twin module is used to build a virtual model of the central kitchen and synchronize the physical entity status; the database module is used to store historical data and model parameters.

[0024] The application layer includes intelligent early warning terminals, scheduling optimization modules, traceability query terminals, and a regulatory platform, which are used to realize early warning information push, processing and distribution scheduling optimization, traceability information query, and regulatory audit functions.

[0025] (III) Beneficial Effects

[0026] Compared with existing technologies, this invention provides a method for full-process traceability and intelligent freshness control of ingredients in a central kitchen based on the Internet of Things, which has the following beneficial effects:

[0027] This invention utilizes RFID tags and blockchain technology to construct a consortium blockchain covering all key nodes of the supply chain. Smart contracts ensure that data undergoes rigorous verification before being uploaded to the blockchain and remains tamper-proof afterward, thus achieving highly reliable traceability throughout the entire process from source to end. Secondly, by integrating NFC tags with miniature gas sensors with an environmental sensor network, and combining edge computing and a food spoilage prediction model, the system can dynamically sense and calculate the real-time freshness index of ingredients, achieving a fundamental shift from relying on static shelf-life to dynamic and accurate quality assessment. Thirdly, the system deploys AI visual inspection and digital twin monitoring in key stages such as processing and distribution, enabling a tiered early warning mechanism based on preset thresholds, achieving pre-emptive warnings and proactive intervention during the process, effectively preventing problematic flows. Finally, the digital twin module uses simulation and improved optimization algorithms to make processing and distribution scheduling decisions with the real-time freshness index as a key weight, thereby minimizing food waste. Attached Figure Description

[0028] Figure 1 This is a flowchart of the IoT-based method for full-process traceability and intelligent freshness control of ingredients in a central kitchen according to the present invention.

[0029] Figure 2 This is a structural diagram of the IoT-based central kitchen food ingredient full-process traceability and freshness intelligent control system of the present invention.

[0030] Figure 3 This is a traceability business process diagram of the IoT-based central kitchen food ingredient full-process traceability and freshness intelligent control system of the present invention.

[0031] Figure 4 This is a schematic diagram of the regional chain module architecture of the present invention. Detailed Implementation

[0032] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0033] Example

[0034] like Figure 1 , Figure 2 , Figure 3 and Figure 4 As shown in the figure, an embodiment of the present invention proposes a method for full-process traceability and intelligent freshness control of ingredients in a central kitchen based on the Internet of Things, which includes the following steps:

[0035] S1: Assign a unique RFID tag to each batch of ingredients, containing supplier information, quarantine report, production date, and ingredient category; attach an NFC flexible tag integrating a miniature gas sensor to the individual packaging of the ingredients; read the RFID tag and NFC flexible tag information through the IoT terminal at the procurement end, and simultaneously collect the initial quality data of the ingredients, including the component content detected by spectral analysis and the shape integrity captured by AI vision; encrypt the tag information and the initial quality data and upload them to the edge computing node;

[0036] S2: Temperature and humidity sensors and gas sensors for detecting ethylene and ammonia concentrations are installed inside the cold storage or refrigerated display cases in the central kitchen to collect real-time storage environment data. Edge computing nodes fuse and analyze the storage environment data with the gas concentration data of the food surface microenvironment fed back by NFC flexible tags. Based on a pre-trained food spoilage prediction model, a real-time freshness index is calculated. The food spoilage prediction model in S2 is obtained through machine learning training: using historical storage data of different types of food as training samples, a classification model is constructed using the random forest algorithm. The historical storage data includes temperature, humidity, storage time, and gas concentration. The higher the freshness index value output by the model, the better the freshness. During the process, key data on different types of food during historical storage are collected, including the temperature and humidity of the storage environment, storage time, and the concentration of gases produced by the respiration or spoilage of the food. At the same time, the actual freshness of the food at the corresponding time point is recorded to form labeled training samples. Then, the random forest algorithm is used to train the sample data. The algorithm captures the nonlinear relationship between features and freshness by randomly sampling multiple sets of features and learning multiple decision trees in parallel. The trained model receives real-time input, such as the current temperature and humidity, storage time, and real-time gas concentration data. Through voting or averaging calculations of multiple decision trees, a quantitative freshness index is output to achieve dynamic prediction of the current freshness status of the food.

[0037] S3: An AI visual inspection module and a spectral re-inspection module are set up on the processing line to perform a second verification of the freshness of the ingredients before processing; if the freshness index is lower than the preset threshold, a local warning is triggered and processing is suspended; at the same time, the operator, equipment number, and processing time data of the processing link are bound to the RFID tag of the ingredients and uploaded to the blockchain module.

[0038] S4: Deploy GPS positioning modules and onboard temperature and humidity sensors on delivery vehicles to upload location information and vehicle environment data in real time; construct a virtual mapping of delivery vehicles through a digital twin module. If environmental data exceeds the standard or deviates from the preset route, a tiered warning is sent to the control terminal. The tiered warning in S4 is a three-level warning, with the following triggering conditions: Level 1 alert, triggered when environmental data exceeds the threshold by less than 5% for less than 5 minutes; Level 2 intervention, triggered when environmental data exceeds the threshold by 5%-10% for 5-10 minutes, and the system automatically adjusts the onboard temperature control equipment; Level 3 interception, triggered when environmental data exceeds the threshold by more than 10% for more than 10 minutes, and the system notifies the control center and delivery drivers to suspend delivery. In use, the dual-dimensional judgment of the exceedance range and duration distinguishes between occasional fluctuations (such as a slight temperature rise caused by a brief opening of the vehicle door) and substantial risks (such as sustained high temperatures caused by a refrigeration failure), avoiding false alarms or excessive intervention (such as suspending delivery for slight fluctuations) that may be caused by a single threshold triggering mechanism, thereby achieving a balance between control accuracy and delivery efficiency.

[0039] S5: Users can scan the RFID or NFC tags of ingredients through the barcode terminal and call the full-process data stored in the blockchain module to achieve forward and reverse traceability; the digital twin module simulates different scheduling schemes based on historical data and outputs the optimized management strategy that minimizes the loss rate of ingredients.

[0040] like Figure 2 As shown, in some embodiments, the RFID tag in S1 is a UHF RFID tag, supporting batch reading within a 10-meter range; the NFC flexible tag integrates a miniature gas sensor to detect the concentration of gases in the microenvironment generated by the respiration of food ingredients, and feeds the data back to the reading terminal through near-field communication; in use, the UHF RFID tag has a unique electronic code built in it, and communicates with the reader non-contactly through radio frequency signals. When the UHF RFID reader of the IoT terminal at the purchasing end emits a radio frequency signal of a specific frequency, the tag receives the signal and is activated, and then feeds back the stored supplier information, quarantine report, production date and other data to the reader through radio frequency waves, realizing the simultaneous reading of multiple tags within a 10-meter range without the need for individual alignment and scanning; the NFC flexible tag integrates a miniature gas sensor, and when it is attached to the surface of the individual food packaging, the sensor directly contacts the microenvironment formed by the respiration of the food ingredients, and detects changes in gas concentration in real time; when the reading terminal is in close contact with the tag, the tag establishes communication with the terminal through electromagnetic induction coupling, and then transmits the gas concentration data collected by the sensor to the terminal, completing near-field data interaction.

[0041] like Figure 2As shown, in some embodiments, the AI ​​vision inspection module in S3 uses an AI vision camera to acquire images of food ingredients and identifies mold, rot, or insect-infested defects using the YOLOv8 algorithm, achieving high detection accuracy. The spectral re-inspection module uses a near-infrared spectrometer to determine freshness by analyzing changes in the moisture and sugar content of the food ingredients, with extremely short detection time for a single item. During image acquisition, the AI ​​vision camera is equipped with a high-resolution lens and light source to capture real-time images of the food ingredients, obtaining clear surface images to ensure that defects such as mold spots, rotten areas, and insect holes are completely captured. The acquired images are transmitted to the backend system and finally processed by the YOLOv8 algorithm. The YOLOv8 algorithm adopts a target detection + bounding box localization mode. First, it extracts image features through the backbone network, such as the dark areas of mold spots and the soft, collapsed texture of rotten food. Then, it enhances the recognition ability of defects such as tiny insect holes through a feature fusion layer. Finally, it outputs the defect category as mold, rot, or insect infestation, and simultaneously outputs the location coordinates, achieving high-precision real-time detection.

[0042] like Figure 2 and Figure 4 As shown, in some embodiments, the blockchain module in S5 adopts a consortium blockchain architecture, whose nodes include suppliers, central kitchens, testing institutions, regulatory authorities, and terminal stores. Each node interacts with data through smart contracts: food quarantine reports uploaded by suppliers must be verified by the testing institution node before being stored on the blockchain; processing data uploaded by the central kitchen can be audited in real time by the regulatory authority node. During use, the supplier, central kitchen, testing institution, regulatory authority, and terminal store nodes in the consortium blockchain must be authenticated before joining the network. Each node has an independent encrypted identity, ensuring the traceability of the data interaction entities. Each node uploads data for the corresponding stage according to its permissions, such as suppliers uploading food quarantine reports. The central kitchen uploads processing records, etc. The data is encrypted and formed into blocks. The blocks need to be verified by the pre-set consensus nodes within the alliance, such as testing institutions and regulatory departments, before they can be synchronously recorded by all nodes. This achieves one-time on-chain storage and network-wide evidence preservation, and no node can unilaterally tamper with the data. Data interaction rules are defined by pre-written smart contracts. For example, the quarantine report uploaded by the supplier needs to trigger the testing institution's verification contract. Only when the testing institution node confirms the report's authenticity and validity through private key signing will the contract allow the report to be stored on the chain. The processing data uploaded by the central kitchen automatically triggers the regulatory audit contract. Regulatory department nodes can read the on-chain data in real time and complete the compliance review without manual application.

[0043] like Figure 2As shown, in some embodiments, the digital twin module in S5 uses an improved genetic-ant colony fusion algorithm for processing scheduling optimization: using the freshness index of ingredients as weights, the genetic algorithm is first used to generate an initial processing sequence and equipment allocation scheme, and then the ant colony algorithm is used to optimize the time nodes in the scheme to avoid high-freshness ingredients waiting too long, ultimately achieving the optimization goal of significantly reducing the processing loss rate. In use, the algorithm automatically prioritizes high-freshness ingredients to enter the processing stage, avoiding their rapid decline in freshness due to waiting too long, thereby reducing spoilage losses caused by time delays from the scheduling level. Secondly, the genetic algorithm optimizes equipment allocation to avoid equipment idleness or excessive congestion; the ant colony algorithm optimizes time nodes to reduce redundant waiting in process connections. The combination of the two makes the processing flow more compact, improves equipment utilization, and shortens the overall processing cycle. In addition, the virtual simulation capability of the digital twin module can respond in real time to sudden situations such as changes in ingredient freshness and equipment failures. Through the fusion algorithm, the scheduling scheme is quickly iterated to ensure that a low loss rate can still be maintained in a dynamic environment, enhancing the resilience of the central kitchen processing stage.

[0044] A central kitchen food ingredient traceability and freshness intelligent control system based on the Internet of Things, comprising:

[0045] The sensing layer includes RFID tags for identifying ingredients and NFC flexible tags integrating miniature gas sensors, temperature and humidity sensors and gas sensors for collecting environmental data, AI visual inspection modules and spectral re-inspection modules for collecting quality data, and a GPS module for positioning. The temperature and humidity sensors can be SHT30 / SHT31, the gas sensors can be ethylene sensors of model SGXSensortechMQ-2 and ammonia sensors of model WinsenMQ-137, the AI ​​visual inspection module can be DH-IPC-HF8231, the spectral re-inspection module can be FelixInstrumentsF-750 or Ocean InsightS TS-NIR, and the GPS module can be UBloxNEO-M8N or QuectelL80-R.

[0046] The transport layer includes edge computing nodes, 5G communication modules, and LoRa gateways. Edge computing nodes are used to perform preliminary processing on the data collected by the perception layer and upload the processed data through 5G communication modules and LoRa gateways. Edge computing nodes can be models such as PowerEdge IPC-1002V1, 5G communication modules can be models such as RG620T, and LoRa gateways can be models such as H3CIG4500-L base station type.

[0047] The platform layer includes a blockchain module, a digital twin module, and a database module. The blockchain module is used to store tamper-proof, end-to-end traceability data. The digital twin module is used to build a virtual model of the central kitchen and synchronize the physical entity status. The database module is used to store historical data and model parameters. The blockchain module can be an Antimacassars model, the digital twin module can be an SHYX-A01Y07DZ1 model, and the database module can be a Kingbase ESV8 database management system model.

[0048] The application layer includes intelligent early warning terminals, scheduling optimization modules, traceability query terminals, and a regulatory platform. These are used to realize functions such as early warning information push, processing and distribution scheduling optimization, traceability information query, and regulatory audit. The intelligent early warning terminals can be handheld (e.g., Urovo i6310Plus), vehicle-mounted (e.g., Hikvision DS-M5501), or desktop (e.g., Huawei IPPhone7960). The scheduling optimization module can use Huawei Cloud Campus scheduling optimization components plus a Dell Power Edge T440 server. The traceability query terminal can be a smartphone or barcode scanner. The regulatory platform can use Alibaba Cloud Computing Platform (Regulatory Version).

[0049] Working Principle: Sensing Layer: RFID tags assign a unique electronic identity to each batch of food, storing basic information such as supplier and quarantine details, and supporting batch reading within 10 meters; NFC flexible tags integrate miniature gas sensors, adhering to individual food packaging to detect the concentration of gases in the microenvironment generated by respiration in real time, and transmitting data via near-field communication; temperature and humidity sensors and gas sensors are deployed in cold storage, refrigerated cabinets, and delivery vehicles to continuously collect data such as temperature, humidity, and gas concentration in the storage / transportation environment, reflecting the external environmental conditions of the food; the spectral re-inspection module analyzes changes in the food's internal components such as moisture and sugar content using near-infrared light, while the AI ​​vision inspection module captures images and identifies them using the YOLOv8 algorithm. The system double-checks food quality by eliminating appearance defects such as mold and insect infestation. GPS modules are deployed in delivery vehicles to collect real-time location information, supporting delivery trajectory tracking. The transmission layer efficiently and reliably transmits the massive amounts of data collected by the perception layer to the platform layer, while also undertaking preliminary edge processing tasks to reduce the computational burden on the core platform. It performs localized preprocessing on perception layer data, such as filtering outliers and fusing environmental and gas concentration data to reduce invalid data transmission. Through 5G communication modules and LoRa gateways, it can select appropriate transmission paths based on data type, ensuring secure and stable data upload to the platform layer. The platform layer stores, models, and performs in-depth processing on the data uploaded from the transmission layer, building... A digital twin space and a trusted data foundation provide decision support for the application layer. A consortium blockchain architecture receives data uploaded from each stage, verifies it by testing institutions or regulatory departments, and stores it on the blockchain. Distributed ledger characteristics ensure data immutability and traceability, forming a trusted data chain throughout the entire process. A virtual mapping model is built based on the central kitchen and delivery vehicles, synchronously presenting the food storage status, processing progress, and delivery trajectory. An improved genetic-ant colony fusion algorithm is integrated to simulate different processing scheduling schemes in the virtual environment, outputting the optimization strategy with the lowest loss rate. Historical data and model parameters are stored, providing data support for data analysis and model iteration. The application layer transforms the processing results from the platform layer into... The system provides comprehensive functionality, serving users in various roles. It receives abnormal information pushed from the platform layer, such as delivery vehicle environmental conditions exceeding standards or food freshness falling below thresholds, and sends reminders, intervention instructions, or interception notices to drivers and management personnel according to a three-tiered early warning system. It invokes processing / delivery scheduling schemes output by the digital twin module to guide the central kitchen in adjusting processing sequences, equipment allocation, and delivery routes, reducing food waste. It supports users in scanning RFID and NFC flexible tags to access full-process data from the blockchain module, enabling forward and reverse traceability. It provides regulatory authorities with an on-chain data audit entry point to view food processing compliance and the validity of test reports in real time, achieving transparent supervision.Therefore, by collecting data on ingredient identification, environment, quality, and location at the perception layer, and then uploading the processed data via the 5G communication module and LoRa gateway in the transmission layer, the data undergoes blockchain notarization, digital twin modeling, and database storage at the platform layer to realize data value mining. Finally, the application layer transforms the platform output into functions such as early warning, scheduling, traceability, and supervision, ultimately forming a closed-loop system of "data collection-transmission-processing-application," thereby achieving full-process traceability and intelligent freshness control of central kitchen ingredients from procurement to the end consumer.

[0050] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for full-process traceability and intelligent freshness control of ingredients in a central kitchen based on the Internet of Things, characterized in that, Includes the following steps: S1: Assign a unique RFID tag to each batch of ingredients, containing supplier information, quarantine report, production date, and ingredient category; attach an NFC flexible tag integrating a miniature gas sensor to the individual packaging of the ingredients; read the RFID tag and NFC flexible tag information through the IoT terminal at the procurement end, and simultaneously collect the initial quality data of the ingredients, including the component content detected by spectral analysis and the shape integrity captured by AI vision; encrypt the tag information and the initial quality data and upload them to the edge computing node; S2: Install temperature and humidity sensors and gas sensors for detecting ethylene and ammonia concentrations inside the cold storage or refrigeration cabinet of the central kitchen to collect storage environment data in real time; the edge computing node performs fusion analysis on the storage environment data and the gas concentration data of the food surface microenvironment fed back by the NFC flexible tag, and calculates the real-time freshness index based on the pre-trained food spoilage prediction model. S3: An AI visual inspection module and a spectral re-inspection module are set up on the processing line to perform a second verification of the freshness of the ingredients before processing; if the freshness index is lower than the preset threshold, a local warning is triggered and processing is suspended; at the same time, the operator, equipment number, and processing time data of the processing link are bound to the RFID tag of the ingredients and uploaded to the blockchain module. S4: Deploy GPS positioning modules and on-board temperature and humidity sensors on delivery vehicles to upload location information and vehicle environment data in real time; construct a virtual mapping of delivery vehicles through a digital twin module, and send a graded warning to the control terminal if environmental data exceeds the standard or deviates from the preset route; S5: Users can scan the RFID or NFC tags of food ingredients through the barcode scanning terminal and call the full-process data stored in the blockchain module to achieve forward traceability and reverse traceability. The digital twin module simulates different scheduling schemes based on historical data and outputs the optimized management strategy that minimizes food loss.

2. The method for intelligent control of the entire process of food ingredient traceability and freshness management in a central kitchen based on the Internet of Things, as described in claim 1, is characterized in that: The RFID tag in S1 is a UHF RFID tag, which supports batch reading within a 10-meter range; the NFC flexible tag integrates a miniature gas sensor to detect the concentration of micro-environment gas generated by the respiration of food, and feeds the data back to the reading terminal through near-field communication.

3. The method for intelligent control of the entire process of food ingredient traceability and freshness management in a central kitchen based on the Internet of Things, as described in claim 1, is characterized in that: The food spoilage prediction model in S2 is obtained through machine learning training: using historical storage data of different types of food as training samples, a classification model is constructed using the random forest algorithm. The historical storage data includes temperature and humidity, storage time and gas concentration. The higher the freshness index value output by the model, the better the freshness.

4. The method for intelligent control of the entire process of food ingredient traceability and freshness management in a central kitchen based on the Internet of Things, as described in claim 1, is characterized in that: The AI ​​visual inspection module in S3 uses an AI visual camera to collect images of food ingredients and identifies mold, rot, or insect damage through the YOLOv8 algorithm, achieving high detection accuracy. The spectral re-inspection module uses a near-infrared spectrometer to determine freshness by analyzing changes in the moisture and sugar content of the food ingredients, with extremely short detection time for a single food item.

5. The method for intelligent control of the entire process of food ingredient traceability and freshness management in a central kitchen based on the Internet of Things, as described in claim 1, is characterized in that: The graded early warning in S4 is a three-level early warning, and its triggering conditions are as follows: Level 1 reminder, triggered when environmental data exceeds the threshold by less than 5% and the duration is less than 5 minutes; Level 2 intervention, triggered when environmental data exceeds the threshold by 5%-10% and the duration is 5-10 minutes, the system automatically adjusts the vehicle temperature control equipment; Level 3 interception, triggered when environmental data exceeds the threshold by more than 10% and the duration is more than 10 minutes, the system notifies the control center and delivery drivers to suspend delivery.

6. The method for intelligent control of the entire process of food ingredient traceability and freshness management in a central kitchen based on the Internet of Things, as described in claim 1, is characterized in that: The blockchain module in S5 adopts a consortium blockchain architecture, with nodes including suppliers, central kitchens, testing institutions, regulatory authorities, and terminal stores. Each node interacts with data through smart contracts: food quarantine reports uploaded by suppliers must be verified by the testing institution node before they can be stored on the blockchain; processing data uploaded by the central kitchen can be audited in real time by the regulatory authority node.

7. The method for intelligent control of the entire process of food ingredient traceability and freshness management in a central kitchen based on the Internet of Things, as described in claim 1, is characterized in that: In S5, the digital twin module uses an improved genetic-ant colony fusion algorithm to optimize processing scheduling: using the freshness index of ingredients as weights, the initial processing order and equipment allocation scheme are first generated using a genetic algorithm, and then the time nodes in the scheme are optimized using an ant colony algorithm to avoid high-freshness ingredients from waiting too long, ultimately achieving the optimization goal of significantly reducing processing loss rate.

8. A smart control system for the traceability and freshness management of central kitchen ingredients based on the Internet of Things (IoT) for implementing the method described in any one of claims 1-7, characterized in that, include: The perception layer includes RFID tags for identifying ingredients and NFC flexible tags with integrated miniature gas sensors, temperature and humidity sensors and gas sensors for collecting environmental data, an AI visual inspection module and a spectral re-inspection module for collecting quality data, and a GPS module for positioning. The transport layer includes edge computing nodes, a 5G communication module, and a LoRa gateway; the edge computing nodes are used to perform preliminary processing on the data collected by the perception layer, and upload the processed data through the 5G communication module and the LoRa gateway. The platform layer includes a blockchain module, a digital twin module, and a database module; the blockchain module is used to store tamper-proof, end-to-end traceability data; the digital twin module is used to build a virtual model of the central kitchen and synchronize the physical entity status; the database module is used to store historical data and model parameters. The application layer includes intelligent early warning terminals, scheduling optimization modules, traceability query terminals, and a regulatory platform, which are used to realize early warning information push, processing and distribution scheduling optimization, traceability information query, and regulatory audit functions.

Images