Fruit and vegetable transportation state monitoring system and method

Abstract

The invention provides a fruit and vegetable transportation state monitoring system and method, and relates to the technical field of fruit and vegetable transportation monitoring, and the fruit and vegetable transportation monitoring method mainly comprises the steps: an environment sensing assembly is responsible for collecting fruit and vegetable transportation state data and covering various key information, and a processor carries out the monitoring of the fruit and vegetable transportation state through a pre-trained fruit and vegetable state prediction model; and processing the collected data to obtain a fruit and vegetable physiological index prediction result. According to the invention, collection of fruit and vegetable transportation state data and preliminary state prediction functions are realized, the environment sensing assembly can collect various data, a basis is provided for subsequent analysis, fruit and vegetable physiological indexes are predicted through the pre-training model, the limitation that only current state information can be obtained in traditional monitoring is broken through, and the monitoring accuracy is improved. State changes of the fruits and vegetables in the transportation process can be known in advance, and measures can be taken in time to guarantee the quality of the fruits and vegetables.

Description

Technical Field

[0001] The present invention mainly relates to the technical field of fruit and vegetable transportation monitoring, and in particular to a fruit and vegetable transportation status monitoring system and method. Background Art

[0002] Traditional monitoring methods in the fruit transportation sector face numerous challenges. Firstly, they rely heavily on manual inspections, requiring staff to conduct regular checks. However, due to long transportation distances and complex environments, manual inspections have significant time gaps, making it difficult to detect sudden environmental changes during transportation. Secondly, simple sensors have limited data processing capabilities, resulting in low data accuracy. Traditional analysis methods cannot fully reflect environmental changes and the condition of the fruit during transportation. This makes it difficult for transporters to plan response strategies in advance and adjust the transportation environment in a timely manner. In long-distance transport, without advance knowledge of changes in the fruit's condition, the fruit may already be rotten or spoiled by the time it reaches its destination, resulting in significant economic losses for businesses and a waste of resources. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a fruit and vegetable transportation status monitoring system and method in response to the deficiencies in the existing technology.

[0004] The present invention solves the above technical problems with the following technical solutions: A fruit and vegetable transport status monitoring system comprises a housing, a status analysis component, and an environment perception component. The housing is a hollow cavity shell structure, and the status analysis component and the environment perception component are respectively installed inside the housing;

[0005] The environmental perception component is used to collect fruit and vegetable transportation status data;

[0006] The state analysis component is used to perform fruit and vegetable state prediction processing on the fruit and vegetable transportation state data using a pre-trained fruit and vegetable state prediction model to obtain fruit and vegetable physiological index prediction results.

[0007] Another technical solution of the present invention to solve the above technical problem is as follows: a method for monitoring the transportation status of fruits and vegetables, applied to the above-mentioned fruit and vegetable transportation status monitoring system, comprising the following steps:

[0008] The environmental perception component collects data on the transportation status of fruits and vegetables;

[0009] The state analysis component performs fruit and vegetable state prediction processing on the fruit and vegetable transportation state data through a pre-trained fruit and vegetable state prediction model to obtain fruit and vegetable physiological index prediction results.

[0010] The beneficial effects of the present invention are: it realizes the collection of fruit and vegetable transportation status data and preliminary status prediction functions. The environmental perception component can collect multiple data to provide a basis for subsequent analysis, and predict the physiological indicators of fruits and vegetables through pre-trained models. This prediction function breaks the limitation of traditional monitoring that can only obtain current status information, helps to understand the status changes of fruits and vegetables during transportation in advance, and facilitates timely measures to ensure the quality of fruits and vegetables. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 A schematic diagram showing the connection of functional modules of a fruit and vegetable transportation status monitoring system according to an embodiment of the present invention;

[0012] Figure 2 A schematic diagram of the flow of frequency acquisition processing provided by an embodiment of the present invention;

[0013] Figure 3 This is a schematic flow chart of a method for monitoring the transport status of fruits and vegetables provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0014] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

[0015] Example 1: Figure 1 As shown, an embodiment of the present invention provides a fruit and vegetable transportation status monitoring system, comprising a housing, a status analysis component, and an environment perception component. The housing is a hollow cavity shell structure, and the status analysis component and the environment perception component are respectively installed inside the housing;

[0016] The environmental perception component is used to collect fruit and vegetable transportation status data;

[0017] The environmental sensing component is specifically used to monitor the carbon dioxide concentration around fruits and vegetables according to the set first collection frequency to obtain carbon dioxide concentration data;

[0018] Also used for monitoring the temperature around fruits and vegetables according to the first acquisition frequency to obtain temperature data;

[0019] Also used for monitoring the humidity around fruits and vegetables according to the first collection frequency to obtain humidity data;

[0020] It is also used to monitor the motion state of the fruit and vegetable transport vehicle according to the set second collection frequency to obtain motion state data.

[0021] Preferably, the system further comprises a wireless communication module and a remote control terminal, wherein the wireless communication module is mounted on the circuit board and is wirelessly connected to the remote control terminal; the wireless communication module may be a LoRa chip;

[0022] The state analysis component is further configured to send the fruit and vegetable transportation state data and the fruit and vegetable physiological index prediction results to the remote control terminal via the wireless communication module;

[0023] The remote control terminal is further used to compare the fruit and vegetable transportation status data with a preset threshold value, and to issue an alarm notification if the preset threshold value is exceeded.

[0024] Preferably, the device further comprises a power supply module, which is mounted on the circuit board and is used to provide power to the environment sensing component and the state analysis component.

[0025] The advantages of Example 1 are: the outer shell adopts a hollow cavity shell structure, providing safe and stable physical protection for the internal circuit board, state analysis component, and environmental sensing component. It can effectively resist external interference such as collision, vibration, dust, etc. during transportation, ensuring the normal operation of each component. The circuit board is installed in the hollow cavity of the outer shell, with a compact and reasonable layout, facilitating line connection and signal transmission between components, and also facilitating the installation and maintenance of the entire device.

[0026] The environmental sensing component integrates multiple sensors to comprehensively collect data on the transport status of fruits and vegetables, covering key indicators such as carbon dioxide concentration, temperature, humidity, and vehicle motion. This data reflects real-time changes in the transport environment and the status of the fruits and vegetables themselves, providing a rich and accurate foundation for subsequent status prediction and transport management.

[0027] The state analysis component collects data on the transport status of fruits and vegetables and provides preliminary status predictions, breaking the limitations of traditional monitoring, which only captures current status information. The environmental perception component collects a variety of data to inform subsequent analysis. Predicting fruit and vegetable physiological indicators through pre-trained models helps predict changes in their status during transport, facilitating timely action to ensure quality.

[0028] The device integrates a wireless communication module and remote control terminal, enabling remote data transmission and threshold alarms. This allows managers to monitor transport status in real time, address anomalies promptly, and improve transport management efficiency. The power module ensures stable operation of all components and continuous monitoring.

[0029] Example 2: The state analysis component further includes a process of pre-training the fruit and vegetable state prediction model.

[0030] Specifically, the state analysis component includes a digital twin module, a prediction model building module, and a data processing module.

[0031] The digital twin module is used to collect fruit and vegetable transportation status data through the environmental perception component and use it as training data, and perform data preprocessing on the training data.

[0032] The prediction model construction module is used to construct a fruit and vegetable state prediction model based on the deep learning model, and import the preprocessed training data into the fruit and vegetable state prediction model for model training to obtain a pre-trained fruit and vegetable state prediction model.

[0033] The data processing module is used to perform fruit and vegetable status prediction processing on the fruit and vegetable transportation status data through the pre-trained fruit and vegetable status prediction model to obtain fruit and vegetable physiological index prediction results.

[0034] It should be understood that the process of pre-training the fruit and vegetable status prediction model is completed by the digital twin module and the model building module.

[0035] The advantage of Example 2 is that it integrates the model pre-training process into the state analysis component, enabling the device to self-optimize and learn. By collecting actual transportation status data for training, it can continuously adapt to different transportation scenarios and changes in fruit and vegetable varieties, improving the accuracy and applicability of the prediction model, laying the foundation for more accurate subsequent fruit and vegetable status predictions.

[0036] Embodiment 3: In the state analysis component, data preprocessing is performed on the training data, including:

[0037] Obtaining motion state data from the training data, performing feature expansion processing on the motion state data through a complementary filter to obtain vehicle driving state data;

[0038] The carbon dioxide concentration data, temperature data, and humidity data obtained from the training data are processed at a unified acquisition frequency based on a resampling strategy, and the vehicle driving state data, carbon dioxide concentration data, temperature data, and humidity data with unified acquisition frequencies are constructed into a multivariate time series;

[0039] Based on the delivery and receipt time of fruits and vegetables, the data length of the multivariate time series is clipped according to the set time unit, and the clipped data are constructed into a multivariate time series dataset;

[0040] Data augmentation is performed on the multivariate time series dataset to obtain preprocessed training data.

[0041] Specifically, the process of feature expansion is as follows:

[0042] Motion state data includes six-axis sensor data, which includes accelerometer data in three directions Gyroscope data First, the accelerometer data and the gyroscope data are fused to obtain the tilt angle. The specific calculation formula is as follows:

[0043]

[0044] α is the weight parameter of the complementary filter, usually 0.98; f s is the acquisition frequency per second, and the calculation formulas for horizontal acceleration and velocity are as follows:

[0045]

[0046] Calculate three features, speed, displacement, and distance:

[0047] speed t =||v t ||,

[0048]

[0049] distance t =||position t ||,

[0050]

[0051] The specific calculation process is:

[0052] Input: Data matrix D, including acceleration Angular velocity Output: displacement d, speed S, distance d total .

[0053] S1. Initialize the sampling frequency (collect 5 samples per second) and complementary filter parameters;

[0054] S2, initialization speed v t ←(0,0,0), position p t ←(0,0,0), tilt angle θ t ←(0,0,0);

[0055] S3. Initialize the displacement list d←[], speed list S←[], and distance list d total ←[];

[0056] S4. For each time point t=0 to len(D)-1, perform the following operations:

[0057] Use complementary filtering to estimate the tilt angle,

[0058] Calculate the horizontal acceleration at the current time point,

[0059] Use complementary filtering to estimate the tilt angle,

[0060]

[0061] Update the velocity vector,

[0062] Calculate the speed at the current time point, S t ←||v t ||,

[0063] The instantaneous speed S t Add to speed list S,

[0064] Update the position, calculate the displacement at the current time point, and add the instantaneous displacement to the displacement list,

[0065]

[0066] Calculate the displacement at the current time point, d t =||p t ||,

[0067] The instantaneous displacement d t Add to the displacement list d,

[0068] Accumulated distance;

[0069] S5. If t>0, then

[0070] S6, end the loop;

[0071] S7. Return the calculated displacement, speed, and distance.

[0072] Specifically, this embodiment adopts a resampling strategy. Compared with the simple processing method of directly deleting or copying data points, resampling technology can more effectively maintain the complete information content of the original data, thereby ensuring data quality. For indicators with an acquisition frequency higher than this standard, a downsampling operation is performed, reducing the number of data points by averaging aggregation to reduce data density. Conversely, for indicators with an acquisition frequency lower than the target value, an upsampling strategy is adopted, using linear interpolation technology to increase data points to improve the temporal resolution of the data.

[0073] The process of unified acquisition frequency processing is as follows: Figure 2 As shown:

[0074] Downsampling: For sensor metrics that collect data more frequently than the target frequency (recording data every 10 minutes), average aggregation is used. For example, if a sensor originally collects data once per second, the data within a certain interval (such as 10 minutes) is averaged to obtain a new data point, thereby reducing the number of data points.

[0075] Upsampling: For metrics whose collection frequency is lower than the target frequency, linear interpolation is used. For example, if a metric is originally collected every hour and the frequency is increased to every 10 minutes, new data points are inserted between adjacent collection times based on a linear relationship, increasing the number of data points. A consistency check is then performed to ensure that the data is logically reasonable in the time series.

[0076] For example, the environmental and exercise indicators in the collected data set have different collection frequencies. The frequency and time of collection for these indicators are shown in Table 1, so the frequencies need to be standardized. Given the large data volume and limited processing capabilities, the target collection frequency is set to record data every 10 minutes. Table 1 shows the collection frequency table.

[0077] Table 1

[0078]

[0079]

[0080] In Table 1, the environmental indicators are temperature, humidity, and carbon dioxide concentration data.

[0081] The motion index is motion state data, that is, six-axis sensor data, which includes accelerometer data and gyroscope data in three directions.

[0082] This paper adopts a resampling strategy. Compared with the simple processing method of directly deleting or copying data points, resampling technology can more effectively maintain the complete information content of the original data, thereby ensuring data quality. For indicators with an acquisition frequency higher than this standard, a downsampling operation is performed, reducing the number of data points by averaging aggregation to reduce data density. Conversely, for indicators with an acquisition frequency lower than the target value, an upsampling strategy is adopted, using linear interpolation technology to increase data points to improve the temporal resolution of the data.

[0083] Specifically, data enhancement processing is performed on multivariate time series datasets, including:

[0084] Each data in the multivariate time series data set is represented as X = [x1, x2, ..., x n], Gaussian noise is added to each data in the multivariate time series data set. The data after adding Gaussian noise is expressed as:

[0085]

[0086] Among them, x i is the original i-th data in the multivariate time series dataset, ε i is Gaussian noise and obeys the normal distribution ε i ~N(0,σ 2 ), σ is the standard deviation of noise, σ=noise_level×σ X , determined by the noise intensity noise_level and the standard deviation of the data. The noise intensity noise_level controls the amplitude of the noise;

[0087] Use a window size of k to smooth the data in the multivariate time series data set after adding Gaussian noise. The smoothed data is expressed as:

[0088]

[0089] in, x j is time series data, i.e., a collection of observations arranged in chronological order. k is typically a small value, set between 0.005 and 0.02 in this example. The window size determines the degree of smoothing. The larger the window, the stronger the smoothing effect.

[0090] After the above two multivariate time series data enhancement methods, the size of the dataset is expanded to 948, and the features and value ranges are shown in Table 2. Table 2 is a feature introduction.

[0091] Table 2

[0092]

[0093] In Table 2, Firmness, SSC, and Titrable Acidity are the physiological quality parameters of fruits and vegetables collected in advance.

[0094] The advantages of Example 3 are: It employs multiple optimization methods in data preprocessing to comprehensively improve data quality. Complementary filters are used to expand motion state data features to obtain more vehicle driving information; a resampling strategy unifies the acquisition frequency to avoid data inconsistencies; data set pruning ensures consistency in the temporal dimension of the data; and data augmentation processing expands the data set size and enhances model robustness, enabling the trained model to better cope with complex and changing transportation environments and improving prediction accuracy and stability.

[0095] Example 4: In the state analysis component, a fruit and vegetable state prediction model is constructed based on a deep learning model, and pre-processed training data is imported into the fruit and vegetable state prediction model for model training to obtain a pre-trained fruit and vegetable state prediction model, including:

[0096] A fruit and vegetable status prediction model is constructed based on the deep learning model BasenjiModel, which includes a data processing layer, a local feature extraction module, a dilated convolution module and a prediction layer.

[0097] The data processing layer is used to input the physiological quality parameters of fruits and vegetables and perform normalization processing on the physiological quality parameters of fruits and vegetables and the pre-processed training data respectively;

[0098] The local feature extraction module includes multiple convolution layers and multiple pooling layers, and a pooling layer is connected between every two convolution layers. The local feature extraction module is used to perform local feature extraction training on the normalized fruit and vegetable quality parameters and the normalized training data;

[0099] The dilated convolution module includes a plurality of convolutional layers connected in sequence, and is used to extract and train features that are far from the shipping time point (i.e., the initial time point) in the normalized training data;

[0100] The prediction layer is used to perform feature prediction training on all extracted features through a multi-task Poisson regression model to obtain a pre-trained fruit and vegetable state prediction model.

[0101] Specifically, the physiological quality parameters of fruits and vegetables include fruit firmness, fruit soluble solid content, acidic substances, etc. as listed in Table 2.

[0102] Specifically, the physiological quality parameters of fruits and vegetables and the pre-processed training data were normalized and calculated using the normalization formula:

[0103]

[0104] Among them, x represents the original data, x min and x max are the minimum and maximum values ​​in the data set, respectively, norm is the normalized data value.

[0105] In this embodiment, the local feature extraction module is mainly composed of convolutional layers and pooling layers. Each convolutional layer has four parts. The GELU activation function is mainly used to add nonlinear factors to the network; the convolution block is used to capture local features; batch normalization is used to normalize the data in each training batch, aiming to accelerate training and improve the stability and performance of the model; Dropout prevents overfitting by changing the values ​​of some hidden layer nodes to 0. This module performs multi-layer convolution and pooling, with a total of 6 layers, including 3 convolutional layers and 3 pooling layers. Pooling is performed between every two convolutional layers, with pooling rates of 2, 2, 2, and 4, respectively, for a total of 8 times; through these convolutional layers, patterns in multivariate time series are observed and matrix data is converted into a set of vectors, where each vector represents a sample.

[0106] To share information over long distances, this example employs multiple layers of densely connected dilated convolutions. These layers consist of four convolutional layers. The primary purpose of using dilated convolutions in this model is to observe distant regulatory elements, extracting and combining features that are far removed from the initial time point.

[0107] Finally, a convolutional layer of width 1 is applied to parameterize a multi-task Poisson regression that computes the normalized count of reads aligned to that region for each provided dataset. The total number of parameters in this model is 404,874. By combining these components, this model effectively predicts the firmness and sugar content of fruits and vegetables, providing reliable quality assessment.

[0108] The advantages of Example 4 are: the fruit and vegetable status prediction model, built based on the deep learning model BasenjiModel, has clear division of labor and collaborative work at each layer. The data processing layer normalizes data processing to improve model convergence speed; the local feature extraction module effectively captures local data features; the dilated convolution module extracts long-range features to obtain more comprehensive transportation information; and the prediction layer uses multi-task Poisson regression to accurately predict fruit and vegetable quality, providing a reliable basis for fruit and vegetable quality assessment and facilitating transportation strategy optimization.

[0109] Example 5: Figure 3 As shown, an embodiment of the present invention further provides a method for monitoring the transport status of fruits and vegetables, which is applied to the above-mentioned system for monitoring the transport status of fruits and vegetables, and includes the following steps:

[0110] The environmental perception component collects data on the transportation status of fruits and vegetables;

[0111] The state analysis component performs fruit and vegetable state prediction processing on the fruit and vegetable transportation state data through a pre-trained fruit and vegetable state prediction model to obtain fruit and vegetable physiological index prediction results.

[0112] Preferably, the method further includes the steps of pre-training the fruit and vegetable status prediction model:

[0113] The environment perception component collects fruit and vegetable transportation status data as training data, and performs data preprocessing on the training data;

[0114] A fruit and vegetable state prediction model is constructed based on the deep learning model, and the preprocessed training data is imported into the fruit and vegetable state prediction model for model training to obtain a pre-trained fruit and vegetable state prediction model.

[0115] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0116] In the several embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of units is merely a logical functional division. In actual implementation, other division methods may be used, such as combining or integrating multiple units or components into another system, or ignoring or not implementing certain features.

[0117] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected based on actual needs to achieve the objectives of the embodiments of the present invention.

[0118] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A fruit and vegetable transportation status monitoring system, characterized in that: It includes a shell, a state analysis component and an environment perception component, wherein the shell is a hollow cavity shell structure, and the state analysis component and the environment perception component are respectively installed inside the shell; The environmental perception component is used to collect fruit and vegetable transportation status data; The state analysis component is used to perform fruit and vegetable state prediction processing on the fruit and vegetable transportation state data using a pre-trained fruit and vegetable state prediction model to obtain fruit and vegetable physiological index prediction results.

2. The fruit and vegetable transportation status monitoring system according to claim 1, characterized in that: The environmental sensing component is specifically used to monitor the carbon dioxide concentration around fruits and vegetables according to the set first collection frequency to obtain carbon dioxide concentration data; Also used for monitoring the temperature around fruits and vegetables according to the first acquisition frequency to obtain temperature data; Also used for monitoring the humidity around fruits and vegetables according to the first collection frequency to obtain humidity data; It is also used to monitor the motion state of the fruit and vegetable transport vehicle according to the set second collection frequency to obtain motion state data.

3. The fruit and vegetable transportation status monitoring system according to claim 1, characterized in that: The state analysis component also includes a process of pre-training the fruit and vegetable state prediction model: The environment perception component collects fruit and vegetable transportation status data as training data, and performs data preprocessing on the training data; A fruit and vegetable state prediction model is constructed based on the deep learning model, and the preprocessed training data is imported into the fruit and vegetable state prediction model for model training to obtain a pre-trained fruit and vegetable state prediction model.

4. The fruit and vegetable transportation status monitoring system according to claim 3, characterized in that: In the state analysis component, data preprocessing is performed on the training data, including: Obtaining motion state data from the training data, performing feature expansion processing on the motion state data through a complementary filter to obtain vehicle driving state data; obtaining carbon dioxide concentration data, temperature data, and humidity data from the training data, performing unified acquisition frequency processing on the vehicle driving state data, carbon dioxide concentration data, temperature data, and humidity data based on a resampling strategy, and constructing the vehicle driving state data, carbon dioxide concentration data, temperature data, and humidity data with unified acquisition frequencies into a multivariate time series; Based on the delivery and receipt time of fruits and vegetables, the data length of the multivariate time series is clipped according to the set time unit, and the clipped data are constructed into a multivariate time series dataset; Data augmentation is performed on the multivariate time series dataset to obtain preprocessed training data.

5. The fruit and vegetable transportation status monitoring system according to claim 4, characterized in that: The data enhancement processing of the multivariate time series dataset includes: Each data in the multivariate time series data set is represented as X = [x1, x2, ..., x n ], Gaussian noise is added to each data in the multivariate time series data set. The data after adding Gaussian noise is expressed as: X′=x i +e i , Among them, x i is the original i-th data in the multivariate time series dataset, ε i is Gaussian noise and obeys the normal distribution ε i ~N(0,σ 2 ), σ is the standard deviation of noise, σ=noise_level×σ X , determined by the noise intensity noise_level and the standard deviation of the data. The noise intensity noise_level controls the amplitude of the noise; Use a window size of k to smooth the data in the multivariate time series data set after adding Gaussian noise. The smoothed data is expressed as: in, or x j For time series data.

6. The fruit and vegetable transportation status monitoring system according to claim 3, characterized in that: In the state analysis component, a fruit and vegetable state prediction model is constructed based on a deep learning model, and pre-processed training data is imported into the fruit and vegetable state prediction model for model training to obtain a pre-trained fruit and vegetable state prediction model, including: A fruit and vegetable status prediction model is constructed based on the deep learning model BasenjiModel, which includes a data processing layer, a local feature extraction module, a dilated convolution module and a prediction layer. The data processing layer is used to input the physiological quality parameters of fruits and vegetables and perform normalization processing on the physiological quality parameters of fruits and vegetables and the pre-processed training data respectively; The local feature extraction module includes multiple convolution layers and multiple pooling layers, and a pooling layer is connected between every two convolution layers. The local feature extraction module is used to perform local feature extraction training on the normalized fruit and vegetable quality parameters and the normalized training data; The dilated convolution module includes a plurality of convolutional layers connected in sequence, and is used to extract and train features that are far from the shipping time point in the normalized training data; The prediction layer is used to perform feature prediction training on all extracted features through a multi-task Poisson regression model to obtain a pre-trained fruit and vegetable state prediction model.

7. The fruit and vegetable transportation status monitoring system according to claim 1, characterized in that: The device further comprises a wireless communication module and a remote control terminal, wherein the wireless communication module is mounted on the circuit board and is wirelessly connected to the remote control terminal; The state analysis component is further configured to send the fruit and vegetable transportation state data and the fruit and vegetable physiological index prediction results to the remote control terminal via the wireless communication module; The remote control terminal is further used to compare the fruit and vegetable transportation status data with a preset threshold value, and to issue an alarm notification if the preset threshold value is exceeded.

8. The fruit and vegetable transportation status monitoring system according to any one of claims 1 to 7, characterized in that: The device further includes a power module, which is mounted on the circuit board and is used to provide power to the environment sensing component and the state analysis component.

9. A method for monitoring the transport status of fruits and vegetables, applied to the fruit and vegetable transport status monitoring system according to any one of claims 1 to 8, characterized in that: The steps include: The environmental perception component collects data on the transportation status of fruits and vegetables; The state analysis component performs fruit and vegetable state prediction processing on the fruit and vegetable transportation state data through a pre-trained fruit and vegetable state prediction model to obtain fruit and vegetable physiological index prediction results.

10. The method for monitoring the transportation status of fruits and vegetables according to claim 9, characterized in that: It also includes the steps to pre-train the fruit and vegetable status prediction model: The environment perception component collects fruit and vegetable transportation status data as training data, and performs data preprocessing on the training data; A fruit and vegetable state prediction model is constructed based on the deep learning model, and the preprocessed training data is imported into the fruit and vegetable state prediction model for model training to obtain a pre-trained fruit and vegetable state prediction model.

Images