Dynamic fresh-keeping system for fruits and vegetables

Abstract

The invention discloses a dynamic fruit and vegetable fresh-keeping system, and relates to the technical field of fruit and vegetable fresh-keeping. The system comprises an intelligent monitoring subsystem, the intelligent monitoring subsystem integrates various sensor modules to collect temperature, humidity and gas component data in the fresh-keeping equipment in real time, and the intelligent monitoring subsystem is connected with a cloud computing platform through a microcontroller and regulates and controls operation parameters of the fresh-keeping equipment according to platform instructions; the strategy making subsystem is combined with information collected by the intelligent monitoring subsystem and the visual identification subsystem to automatically adjust a personalized fresh-keeping strategy and environment parameters, a block chain tracing technology is adopted, and according to the change of the variety, maturity and storage time of fruits and vegetables, the fresh-keeping strategy and the environment parameters are automatically adjusted; according to the fruit and vegetable fresh-keeping management system, fresh-keeping strategies and storage environment parameters are automatically adjusted, precise, dynamic and informatization management of fruit and vegetable fresh-keeping is achieved, the fresh-keeping period is effectively prolonged, the quality is improved, food safety is guaranteed, and an all-round innovative solution is provided for the fruit and vegetable fresh-keeping industry.

Description

Technical Field

[0001] This invention relates to the technical field of computer vision and remote sensing image processing, and in particular to a dynamic preservation system for fruits and vegetables. Background Technology

[0002] Change detection, as one of the core technologies of remote sensing Earth observation, plays an irreplaceable role in various fields such as disaster monitoring and assessment, urban monitoring, and environmental perception. With the rapid advancement of aerospace information technology, Synthetic Aperture Radar (SAR) image data has become increasingly abundant, making it an indispensable data source in the field of change detection. However, due to its unique electromagnetic active imaging mechanism, SAR images face complex problems such as speckle noise and false changes. These issues significantly reduce the separability between changed areas and the background, severely impacting the accuracy and reliability of change detection.

[0003] Currently, matrix factorization-based methods are favored due to their strong noise suppression capabilities and high computational efficiency. These methods primarily focus on utilizing the sparsity properties of changing regions by constructing low-rank matrix factorization models to extract sparse components representing the changes. However, the effectiveness of these methods largely depends on the precondition that the changing region must constitute a very small proportion of the entire observation area, which is often difficult to meet in practice. Furthermore, when faced with complex and variable scenarios, these methods struggle to accurately separate sparse components, leading to significant errors in the extraction of changing regions.

[0004] Therefore, designing a more general, robust, and adaptable change detection method to overcome the above limitations and improve the accuracy and reliability of change detection is an important issue that urgently needs to be addressed. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the present invention aims to provide a dynamic fruit and vegetable preservation system to solve the problem of accurately extracting changing areas in complex scenarios. To achieve the above-mentioned objectives and other advantages of the present invention, a dynamic fruit and vegetable preservation system is provided, comprising:

[0006] The intelligent monitoring subsystem integrates multiple sensor modules to collect real-time data on temperature, humidity, and gas composition inside the preservation equipment. Furthermore, the intelligent monitoring subsystem connects to a cloud computing platform via a microcontroller and adjusts the operating parameters of the preservation equipment according to platform instructions. The intelligent monitoring subsystem also provides real-time fault warning functionality.

[0007] The visual recognition subsystem includes a high-definition camera, an intelligent lighting module, and a visual recognition module. The high-definition camera is used to accurately plan the installation site based on the optical environment of the preservation equipment and the layout of the fruit and vegetable stacking. The intelligent lighting module is used to dynamically supplement the light according to the ambient light intensity. The visual recognition module is used to accurately identify the fruit and vegetable type, maturity, and surface condition using deep learning algorithms from the captured photos, providing key information for customizing preservation strategies.

[0008] The strategy formulation subsystem combines information collected by the intelligent monitoring subsystem and the visual recognition subsystem to automatically adjust personalized preservation strategies and environmental parameters. After collecting data on fruit and vegetable types, maturity, and storage time, the strategy formulation subsystem uses a deep learning model to formulate preservation strategies. These strategies include adjustments to various environmental parameters, enabling personalized customization of preservation strategies for different types and maturity levels of fruits and vegetables. Furthermore, this subsystem also includes: using a deep learning model for predictive judgment, analyzing environmental change trends, and adjusting storage conditions in advance to avoid the impact of unforeseen external factors on fruits and vegetables.

[0009] Compared with existing technologies, the beneficial effects of this invention are: by using a deep learning model to customize preservation strategies, it can achieve both targeting and efficiency, organically unify and combine technologies from multiple fields to realize a complete fruit and vegetable preservation system, and achieve in-depth development in the field of fruit and vegetable preservation. Attached Figure Description

[0010] Figure 1 This is a schematic diagram of the structural framework of the dynamic fruit and vegetable preservation system according to the present invention;

[0011] Figure 2 The overall circuit integration design diagram of the intelligent monitoring subsystem of the fruit and vegetable dynamic preservation system according to the present invention;

[0012] Figure 3 A flowchart illustrating the data recording and traceability subsystem architecture of the fruit and vegetable dynamic preservation system according to the present invention;

[0013] Figure 4 This is a schematic diagram of the heat and mass transfer principle of the dynamic preservation system for fruits and vegetables according to the present invention.

[0014] Figure 5 The ACF diagram involved in calculating the order determination parameters of the ARIMA model of the dynamic fruit and vegetable preservation system according to the present invention;

[0015] Figure 6 PACF diagrams are used to calculate the order parameters of the ARIMA model of the dynamic fruit and vegetable preservation system according to the present invention. Detailed Implementation

[0016] 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.

[0017] Reference Figure 1 A dynamic fruit and vegetable preservation system, comprising:

[0018] The intelligent monitoring subsystem integrates multiple sensor modules to collect real-time data on temperature, humidity, and gas composition within the preservation equipment. This subsystem connects to a cloud computing platform via a microcontroller, adjusting the equipment's operating parameters according to platform instructions. The subsystem includes gas sensor modules, humidity sensor modules, and temperature sensor modules. Data collected by the sensor modules is transmitted to the microprocessor for preprocessing. The microprocessor program determines whether each environmental parameter exceeds a threshold and whether an alarm is triggered. Simultaneously, the preprocessed sensor data is uploaded to the cloud platform and a data recording and traceability system for further processing.

[0019] Furthermore, such as Figure 2 As shown, the sensor module of the intelligent monitoring subsystem functions to acquire environmental parameters. The sensor module consists of sensors and related circuits, forming a complete sensor circuit module that can be triggered by environmental stimuli. For the gas sensor module, given that most gases involved in fruit and vegetable preservation are CO2 or organic gases, optical interferometry or infrared absorption-scattering gas sensors can be selected, with matching circuits configured according to the actual sensor type. For the humidity sensor module, since the system's preservation strategy adjustment needs to be timely, the sensor module's response time should be as short as possible; resistive humidity sensors and ceramic resistive humidity sensors are preferred. For the temperature sensor module, the temperature control required in fruit and vegetable preservation is in the low-temperature range of -5 to approximately 15 degrees Celsius. The sensor module used can be a fast-responding and sensitive thermistor as the core temperature sensor, and the specific matching circuitry will vary significantly depending on the type of thermistor. The accuracy of sensor data is the foundation for the subsequent decision-making system to make personalized preservation strategies. To achieve this, the sensor circuit response is suitable for efficient processing via an analog-to-digital converter and then by a digital filter, and is also suitable for efficient computation by a microprocessor.

[0020] The intelligent monitoring subsystem's main circuit is based on a microcontroller, preferably an STMicroelectronics series processor. It transmits data collected by temperature, humidity, and gas sensor modules to a cloud platform via a network module. Simultaneously, the microcontroller receives control commands from the cloud platform, controlling the temperature control and gas transmitter modules within the preservation equipment to adjust storage environment parameters. Furthermore, the core microcontroller should include an alarm module to promptly issue alerts and automatically adjust strategies when environmental parameters are abnormal, ensuring the safe storage of fruits and vegetables. The microcontroller circuit should also include a power supply module to provide a stable power supply for the entire system. This module converts AC mains power to a suitable DC voltage and performs voltage regulation.

[0021] The visual recognition subsystem includes a high-definition camera, an intelligent lighting module, and a visual recognition module. The high-definition camera is used to accurately plan installation sites based on the optical environment of the preservation equipment and the layout of the fruit and vegetable stacking. The intelligent lighting module is used to dynamically supplement lighting according to the ambient light intensity. The visual recognition module is used to accurately identify the type, maturity, and surface condition of fruits and vegetables from the captured photos using a deep learning algorithm, providing key information for customizing preservation strategies. After image acquisition and preprocessing, the visual recognition subsystem uses a deep learning model to accurately identify the type of fruits and vegetables and uploads the identification results to a cloud platform for subsequent operations.

[0022] Furthermore, high-definition cameras possess high resolution to capture clear images of fruits and vegetables. Installed in suitable locations within the preservation equipment, these cameras comprehensively cover the storage area, ensuring effective image capture for each item. Multiple cameras are installed to avoid blind spots. The cameras need excellent light adaptability, functioning normally and acquiring clear images under various lighting conditions. This can be achieved by equipping the cameras with automatic aperture and automatic gain control. The cameras connect to the control computer via a network interface, transmitting the captured image data to the computer for processing in real time. If shooting is impossible due to insufficient ambient light, an intelligent lighting module illuminates the area during the shooting process.

[0023] Furthermore, the visual recognition algorithm for the visual recognition subsystem can be the YOLOv8 algorithm. Image acquisition and preprocessing: Image acquisition is performed using OpenCV's cv2 library. Image acquisition functions are called to obtain images, and image preprocessing automatically corrects orientation to ensure the correct image orientation. To achieve model scale invariance and simplify network input requirements, all images are uniformly resized to 640*640 pixels. Model loading and prediction: Assuming a deep learning-based model library is used, the specific implementation needs to be tailored to the selected model framework in actual applications. The model loading function and prediction function predict the fruit category. The prediction results are parsed to obtain the fruit category and the final prediction result.

[0024] The YOLOv8 algorithm can be selected for assessing the freshness of fruits and vegetables. Image acquisition and preprocessing: Image acquisition is performed using the OpenCV library. Image acquisition functions are called to obtain images, and automatic orientation correction and size uniformity are performed. Model loading and prediction: The predict method is the core of model prediction. It takes the preprocessed image as input, calls the YOLO model to perform object detection, and returns the detection results.

[0025] Furthermore, the principle behind this application for formulating a multi-parameter preservation strategy is as follows:

[0026] First, a modeling and analysis of the heat and mass transfer process in fruits and vegetables is performed. Figure 5 Based on the principles of heat and mass transfer in fruits and vegetables, lower temperatures can inhibit the respiration rate of fruits and vegetables, slowing down the process of water and weight loss. Temperature and humidity are the main control factors, with concentration as an auxiliary variable. The parameter ranges for temperature, humidity, and gas concentration are set according to the fruit tree species, maturity level, and storage requirements. The temperature, humidity, and gas concentration in the cold storage are monitored in real time by sensors, and adjustments are made accordingly.

[0027] Analysis using mathematical models:

[0028] In fruit and vegetable preservation environments, the thermal balance is as follows:

[0029] Q a =Q c +Q Y +Q w +Q p +Q b +Q L +Q e

[0030] Among them, Q a The change in heat of the air per unit time in the preservation environment, expressed in J / s; Q Y The heat generated during liquid nitrogen filling, including latent heat and sensible heat, J / s; Q c Q is the heat exchange rate of the evaporator, in J / s; w To maintain the structural heat transfer rate, J / s; Q p The heat exchange rate of fruits and vegetables is expressed in J / s; Q b Q is the heat exchange rate of the water storage tank, in J / s; L Q represents the total latent heat of vaporization of liquid water in the preservation environment per unit time, in J / s; e The heat transfer rate caused by container leakage, J / s

[0031] Q c This refers to the amount of heat exchanged between the evaporator and the air per unit time when the refrigeration system is working. It can be calculated using the following formula:

[0032] Qc =k c A c (T a -T c )

[0033]

[0034] Where, k c The heat transfer coefficient of the evaporator is initially taken as 20.63 W / (m²). 2 K); A c Evaporator heat exchange area, m 2 ;T a For air temperature, K; T c Evaporator evaporation temperature, K; T ao K represents the air temperature at the evaporator outlet; k1 is the fin heat transfer coefficient, which is equal to the initial heat transfer coefficient of the evaporator, in W / (m²). 2 K); L1 is the fin thickness, m, taken as 0.002 m; k2 is the frost layer heat transfer coefficient W / (m²). 2 K), take 2.22; L2 is the frost layer thickness, m.

[0035]

[0036] Q YT =m Y C N (T a -77.35)

[0037] In the formula, Q Y Q represents the total heat exchanged between liquid nitrogen vaporization and air per unit time, expressed in kJ / s. YL Q is the total latent heat of vaporization of liquid nitrogen per unit time, expressed in kJ / s. YT The heat absorbed by nitrogen gas per unit time as its temperature rises, expressed in kJ / s; m Y The flow rate for liquid nitrogen charging is taken as 0.00416 kg / s; L Y The latent heat of vaporization of liquid nitrogen is taken as 2.7928 kJ / mol; Molar mass of nitrogen, kg / k·mol.

[0038] In a food preservation environment, the humidity content of the air is mainly affected by air temperature, humidification source, and evaporation source. The air humidity balance can be expressed by the following formula:

[0039]

[0040] m c =m aw +m Y

[0041] maw (t)=m ai [w ai (t-1)-w ao (t-1)]

[0042] In the formula, w a air humidity; m t Dew point separation; m c Change in moisture content upon condensation, kg / s; m h Humidification and evaporation produce, kg / s; m v Surface condensation and evaporation, kg / s; m e Changes in moisture content due to internal and external exchange, kg / s; m aw Changes in moisture content during evaporator frosting, kg; w Y Changes in moisture content during liquid nitrogen charging; m ai Air mass entering the evaporator, kg; w ai Moisture content before entering the evaporator; w ao Moisture content after exiting the evaporator.

[0043] Liquid nitrogen filling and venting, fruit and vegetable respiration, and air leakage can all alter the gas volume fraction in the preservation environment. Oxygen volume fraction:

[0044]

[0045] φ o Oxygen volume fraction, N leakage coefficient, φ aeo External oxygen volume fraction, φ fo Oxygen levels change during respiration.

[0046] Carbon dioxide volume fraction:

[0047]

[0048] φ o Carbon dioxide volume fraction; φ aec External carbon dioxide volume fraction; φ fc Oxygen levels change during respiration.

[0049] Changes in the volume fraction of CO2 in the environment due to respiration, taking lychee as an example:

[0050]

[0051] R CO2 The respiration rate of litchi fruit is expressed in mL / (kg·h); m p V represents the total mass of lychees, in kg; a m is the free volume of air in the preservation environment. 3 .

[0052] The strategy formulation subsystem combines information collected by the intelligent monitoring subsystem and the visual recognition subsystem to automatically adjust personalized preservation strategies and environmental parameters. This subsystem uses data such as fruit and vegetable type, maturity, and storage time, along with environmental parameters, and employs a deep learning model to formulate preservation strategies that include adjustments to various environmental parameters. Furthermore, the strategy formulation subsystem uses a deep learning model for predictive judgment, analyzing environmental change trends to adjust storage conditions in advance and prevent fruits and vegetables from being affected by unforeseen external factors.

[0053] The strategy formulation subsystem also includes a preservation effect evaluation model Q, where Q represents the degree of preservation. This strategy introduces four indicators to evaluate the degree of preservation: weight loss rate L, decay index RI, and vitamin C content (Vit). C Titrable acid content (TA).

[0054]

[0055] (Where M0 is the original weight; M1 is the current weight)

[0056] RI=ΣRL×m÷(RL max ×N)

[0057] (where RL represents the decay level; m represents the number of fruits at that level; RL max (This represents the highest level of decay; N is the total number of fruits)

[0058]

[0059] Where V is the volume of KIO3 solution consumed in the titration; B is the volume of sample solution taken during titration; a is the final volume of the sample; b is the sample mass; and 0.088 is the vitamin conversion factor, which means that 1 ml of 0.001N solution is equivalent to the amount of ascorbic acid.

[0060]

[0061] Where V is the total volume of sample dilution (ml); V1 is the volume of sample taken during titration; C is the molar concentration of sodium hydroxide standard solution consumed; N is the stoichiometric coefficient of the reaction with acid; W is the sample weight; k is the conversion factor, which depends on the types of common organic acids in food; and the preservation effect Q can be a weighted sum of these indicators:

[0062] Q = w1L + w2R + w3W ic +w4A

[0063] Among them, w1, w2, w3, and w4 are the weight coefficients of the corresponding indicators, which are determined according to the degree of importance that different fruits and vegetables attach to each indicator.

[0064] Based on the above model analysis, in preservation strategies, these models can be used to adjust environmental parameters such as temperature and gas composition to reduce the respiration of fruits and vegetables and extend their shelf life. At the same time, these models also help to study the sensitivity of different fruits and vegetables to environmental changes, providing theoretical support for the development of personalized preservation solutions. Depending on the different influencing factors and their weights for different fruits, corresponding preservation parameter settings can be added for more fruit categories, and preservation strategies can be selected based on the identification results.

[0065] Predictive analytics: Data collection and preprocessing involves acquiring historical data on the fruit and vegetable preservation environment from sensors, including time-series data on temperature, humidity, and gas concentrations. Outlier detection and processing are performed, and statistical methods are used to calculate the data mean. and standard deviation Data point x i and The comparison is performed, where k is a set multiple; data outside this range are considered outliers. Missing values ​​are handled by filling in the missing data with the median, i.e., x. i =median(x). Performs a first-order difference operation on the data:

[0066] Δx t =x t+1 -x t

[0067] Second-order difference:

[0068] Δ 2 x t =Δ(Δx) t )=Δx t+1 -Δx t =(x t+2 -x t+1 )-(x t+1 -x t )=x t+2 -2x t+1 +x t

[0069] General order difference:

[0070]

[0071] in These are binomial coefficients. Trends are observed by plotting line graphs of the data, and seasonality components are analyzed using seasonal decomposition methods. Then, based on the analysis results, appropriate differencing is performed to transform the non-stationary series into a stationary one. The model order is determined, and the autocorrelation function (ACF) and partial autocorrelation function (PACF) of the data are calculated. ACF calculation formula:

[0072]

[0073] Where Cov(x) t ,x t-k ) is x t and x t-k covariance, Var(x) t ) is x t The variance. The order (p,d,q) of the fruit and vegetable preservation prediction model is determined based on the characteristics of the ACF and PACF plots, where p is the number of autoregressive terms, q is the number of moving average terms, and d is the time. For the fruit and vegetable preservation prediction (p,d,q) model, its prediction formula is the number of differencing operations performed when the sequence becomes stationary.

[0074] The prediction formula for the fruit and vegetable preservation prediction (p,d,q) model is as follows:

[0075]

[0076] in It is the predicted value for the next h steps, μ is a constant term, and γ is the predicted value for the next h steps. i It is the autoregressive coefficient, θ i It is the moving average coefficient, ∈ T +hi is the error term.

[0077] Observing the ACF plot, if the ACF suddenly drops to near 0 after a certain lag order and fluctuates within a certain range, it may indicate the moving average order. Observing the PACF plot, if the PACF suddenly drops to near 0 after a certain lag order and fluctuates within a certain range, it may indicate the autoregressive order. The differencing order is determined based on the stationarity test results of the data. A unit root test is used to determine if the data is stationary; if not, differencing is performed until the data is stationary. Model training and evaluation: The fruit and vegetable preservation prediction model with the determined order is trained using historical data. During training, the model parameters are estimated by minimizing the prediction error function. Optimization algorithms can be used for parameter estimation. The trained model is applied to the validation dataset, and evaluation metrics are calculated to assess the model's predictive performance. If the evaluation metrics are unsatisfactory, the model order or other parameters can be adjusted, and training and evaluation can be repeated. Prediction and decision-making: The trained fruit and vegetable preservation prediction model is used to predict future fruit and vegetable preservation environmental parameters. During prediction, the predicted values ​​for future times are calculated based on the model's formula and estimated parameters, combined with historical data.

[0078] Furthermore, the system includes a cloud computing platform, and the cloud platform connection algorithm uses the MQTT protocol to connect the microcontroller to the cloud computing platform. MQTT is a lightweight messaging protocol suitable for communication between IoT devices and the cloud platform. The microcontroller acts as an MQTT client, connecting to a designated MQTT server (cloud platform). During connection, parameters such as server address, port number, client ID, username, and password need to be set. Sensor data is encapsulated into MQTT messages and published to designated topics using the publish function. The cloud platform subscribes to these topics and receives sensor data.

[0079] Cloud platform connection and interaction process: Sensors collect data such as temperature, humidity, and gas composition inside the preservation equipment in real time, convert analog signals into digital signals, and transmit them to the microcontroller; the microcontroller sends the data to the cloud platform via the MQTT protocol; after receiving the data, the cloud platform stores, analyzes, and processes it; based on the preset preservation strategy and algorithm, the cloud platform calculates the optimal storage environment parameter settings and sends control commands back to the microcontroller via the MQTT protocol; the microcontroller, based on the received control commands, controls the corresponding execution modules inside the preservation equipment to adjust the storage environment parameters, thereby achieving intelligent control of the fruit and vegetable preservation environment.

[0080] Furthermore, the system also includes a data recording and traceability subsystem. This subsystem utilizes blockchain technology to comprehensively collect information on fruit and vegetable production locations, harvesting times, supplier details, preservation equipment operation logs, preservation processing records, and transportation trajectory data. Its working principle is as follows:

[0081] In a fruit and vegetable preservation system, relevant information about fruits and vegetables, including type, origin, harvest time, storage time, changes in storage environment parameters, and preservation treatment records, is recorded on a blockchain. Each data block contains the hash value of the previous data block, forming a chain structure. When new data is generated, a data block is added to the blockchain through a consensus mechanism, such as proof-of-work or proof-of-stake.

[0082] The workflow of the data recording and traceability subsystem is as follows:

[0083] The data is established based on basic information provided by fruit and vegetable suppliers, including variety, origin, and harvest time; environmental parameter data automatically collected by the preservation system; and manually recorded preservation processing information, including the time and dosage of preservatives added. Before entering the blockchain, the data needs to be encoded and encrypted to ensure security and privacy. Appropriate encryption algorithms are used to perform hash operations on the data, generating hash values ​​as identifiers for data blocks. A data index is established to facilitate quick consumer searches. Indexes can be created based on fruit and vegetable type, storage time, etc., to improve search efficiency.

[0084] The number of devices and processing scale described herein are for simplification of the invention. Applications, modifications, and variations of this invention will be readily apparent to those skilled in the art. Although embodiments of the invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. It can be applied to various fields suitable for this invention, and further modifications can be readily implemented by those skilled in the art. Therefore, without departing from the general concept defined by the claims and their equivalents, this invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A dynamic fruit and vegetable preservation system, comprising a cloud computing platform, characterized in that, include: The intelligent monitoring subsystem integrates multiple sensor modules to collect real-time data on temperature, humidity, and gas composition inside the preservation equipment. The intelligent monitoring subsystem is connected to a cloud computing platform via a microcontroller and adjusts the operating parameters of the preservation equipment according to the platform's instructions. The visual recognition subsystem includes a high-definition camera, an intelligent lighting module, and a visual recognition module. The high-definition camera is used to accurately plan the installation site based on the optical environment of the preservation equipment and the layout of the fruit and vegetable stacking. The intelligent lighting module is used to dynamically supplement the light according to the ambient light intensity. The visual recognition module is used to accurately identify the fruit and vegetable type, maturity, and surface condition using deep learning algorithms from the captured photos, providing key information for customizing preservation strategies. The strategy formulation subsystem combines information collected by the intelligent monitoring subsystem and the visual recognition subsystem to automatically adjust personalized preservation strategies and environmental parameters.

2. The dynamic fruit and vegetable preservation system as described in claim 1, characterized in that, The system also includes a data recording and traceability subsystem, which uses blockchain technology to comprehensively collect information on fruit and vegetable production areas, harvesting times, supplier details, operation logs of preservation equipment, preservation processing records, and transportation trajectory data.

3. The dynamic fruit and vegetable preservation system as described in claim 1, characterized in that, The intelligent monitoring subsystem includes a gas sensor module, a humidity sensor module, and a temperature sensor module. The data collected by the sensors is collected by a microprocessor and uploaded to a cloud platform and a data recording and traceability system for subsequent operations.

4. The dynamic fruit and vegetable preservation system as described in claim 1, characterized in that, The strategy formulation subsystem includes a fruit and vegetable preservation prediction model. The fruit and vegetable preservation prediction model acquires historical data of the fruit and vegetable preservation environment from sensors through data collection and preprocessing. The historical data includes time series data of temperature, humidity and gas concentration, and performs outlier detection and processing on the data.

5. A dynamic fruit and vegetable preservation system as described in claim 4, characterized in that, The prediction formula of the fruit and vegetable preservation prediction model is as follows: in It is the predicted value for the next h steps, μ is a constant term, and γ is the predicted value for the next h steps. i It is the autoregressive coefficient, θ i It is the moving average coefficient, ∈ T +hi is the error term.

6. The dynamic fruit and vegetable preservation system as described in claim 1, characterized in that, The system also features a user-friendly interface, allowing consumers to remotely monitor and adjust the system via mobile devices, providing a personalized fruit and vegetable preservation experience.

Images