Intelligent refrigerator monitoring and early warning system and method based on big data

Abstract

The present application relates to the field of refrigerator monitoring technology, and specifically discloses a smart refrigerator monitoring and early warning system and method based on big data, which collects temperature data inside the refrigerator in real time through a temperature sensor, and uses big data analysis technology to perform time series analysis on the temperature data, digs out the time series change pattern of the temperature inside the refrigerator, and intelligently identifies abnormal temperature states. Then, when abnormal changes in the temperature inside the refrigerator are detected, infrared images of food inside the refrigerator are collected and analyzed through infrared sensing technology and image processing technology to detect whether there is hot food inside the refrigerator, so that when abnormal changes in the temperature inside the refrigerator are detected and there is no hot food, a corresponding early warning prompt is issued. In this way, the user can be reminded to take corresponding measures in time when the refrigerator is working abnormally, and unnecessary false alarms are effectively avoided, thereby improving the accuracy of refrigerator monitoring and user experience.

Description

Technical Field

[0001] The present application relates to the field of refrigerator monitoring technology, and more specifically, to an intelligent refrigerator monitoring and early warning system and method based on big data. Background Art

[0002] With the advancement of technology and the improvement of living standards, people have higher and higher requirements for the functionality and intelligence of household appliances. As one of the indispensable household appliances, the main function of refrigerators is to provide a low-temperature environment for food to extend its shelf life.

[0003] Currently, although most refrigerator products on the market are equipped with temperature sensors to monitor the temperature inside the refrigerator, they can often only achieve basic temperature display and alarm functions based on temperature thresholds, and lack comprehensive monitoring and intelligent analysis capabilities of the internal environment of the refrigerator.

[0004] For example, when a user puts hot food in the refrigerator, under normal conditions, the temperature inside the refrigerator will rise rapidly, and then gradually drop under the action of the refrigeration system and return to the set temperature range. However, the traditional refrigerator status abnormality alarm system based on temperature threshold cannot distinguish whether the temperature change inside the refrigerator is caused by the addition of hot food or a refrigerator malfunction, which may lead to false alarms and affect the user experience.

[0005] Therefore, a smart refrigerator monitoring and early warning system and method based on big data is proposed. Summary of the invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides an intelligent refrigerator monitoring and early warning system and method based on big data, which collects the temperature data inside the refrigerator in real time through a temperature sensor, and uses big data analysis technology to perform time series analysis on the temperature data, and mines the time series change pattern of the temperature inside the refrigerator to intelligently identify abnormal temperature conditions. Then, when an abnormal change in the temperature inside the refrigerator is detected, infrared sensing technology and image processing technology are used to collect and analyze infrared images of food inside the refrigerator to detect whether there is hot food inside the refrigerator, so that when an abnormal change in the temperature inside the refrigerator is detected and there is no hot food, a corresponding early warning prompt is issued. In this way, the user can be reminded to take corresponding measures in time when the refrigerator is working abnormally, and unnecessary false alarms can be effectively avoided, thereby improving the accuracy of refrigerator monitoring and user experience.

[0007] Accordingly, according to one aspect of the present application, a smart refrigerator monitoring and early warning system based on big data is provided, which includes: a refrigerator internal temperature monitoring module, which is used to collect a data set of the refrigerator internal temperature through a temperature sensor; a temperature data transmission module, which is used to upload the refrigerator internal temperature data set to a smart refrigerator monitoring and early warning cloud through a wireless communication module; a refrigerator status monitoring module, which is used to input the refrigerator internal temperature data set into a refrigerator status monitoring big model in the smart refrigerator monitoring and early warning cloud to obtain a status monitoring result; a hot food monitoring instruction generation module, which is used to generate a hot food monitoring instruction in response to the status monitoring result that the refrigerator internal temperature timing pattern is abnormal. , generating and sending a hot food monitoring instruction inside the refrigerator to the smart refrigerator; a food infrared image acquisition module, used to respond to receiving the hot food monitoring instruction inside the refrigerator, turn on the camera to collect the infrared image of the food inside the refrigerator through the camera; a food infrared image transmission module, used to upload the infrared image of the food inside the refrigerator to the smart refrigerator monitoring and early warning cloud through the wireless communication module; a food infrared image analysis module, used to perform image analysis on the infrared image of the food inside the refrigerator in the smart refrigerator monitoring and early warning cloud to obtain an image analysis result; a refrigerator status abnormal warning module, used to respond to the image analysis result that there is no hot food, and generate a refrigerator status abnormal warning signal.

[0008] In the above-mentioned big data-based intelligent refrigerator monitoring and early warning system, the refrigerator status monitoring module includes: a data regularization unit, which is used to regularize the data set of the refrigerator internal temperature according to the time dimension to obtain the time series of the refrigerator internal temperature; a singular spectrum analysis unit, which is used to perform singular spectrum analysis on the time series of the refrigerator internal temperature to obtain a set of refrigerator internal temperature subsequences; a subsequence timing analysis unit, which is used to respectively extract the timing pattern features of each refrigerator internal temperature subsequence in the set of refrigerator internal temperature subsequences to obtain a set of refrigerator internal temperature timing pattern feature vectors; a dynamic characteristic concentration unit, which is used to perform dynamic characteristic concentration on the set of refrigerator internal temperature timing pattern feature vectors to obtain a refrigerator internal temperature timing pattern core feature vector; a status monitoring result generation unit, which is used to generate the status monitoring result based on the refrigerator internal temperature timing pattern core feature vector.

[0009] In the above-mentioned big data-based intelligent refrigerator monitoring and early warning system, the singular spectrum analysis unit includes: a time series matrix construction subunit, which is used to arrange the time series of the internal temperature of the refrigerator into a refrigerator internal temperature time series matrix according to a preset sliding window size; a singular value decomposition subunit, which is used to perform singular value decomposition on the refrigerator internal temperature time series matrix to obtain a sequence of left singular eigenvectors inside the refrigerator and a sequence of right singular eigenvectors inside the refrigerator; a time series reconstruction subunit, which is used to perform time series reconstruction on each group of corresponding left singular eigenvectors inside the refrigerator and right singular eigenvectors inside the refrigerator in the sequence of left singular eigenvectors inside the refrigerator and the sequence of right singular eigenvectors inside the refrigerator to obtain a set of refrigerator internal temperature subsequences.

[0010] In the above-mentioned big data-based smart refrigerator monitoring and early warning system, the subsequence timing analysis unit is used to: input each refrigerator internal temperature subsequence in the set of refrigerator internal temperature subsequences into the temperature timing pattern feature extractor based on the forward LSTM model to obtain a set of refrigerator internal temperature timing pattern feature vectors.

[0011] In the above-mentioned big data-based intelligent refrigerator monitoring and early warning system, the dynamic characteristic concentration unit includes: an initial aggregation representation subunit, which is used to calculate the positional mean vector of the set of characteristic vectors of the refrigerator internal temperature time series pattern to obtain the initial aggregation representation vector of the refrigerator internal temperature time series pattern; a characteristic energy potential dynamic analysis subunit, which is used to calculate the energy potential factor of each refrigerator internal temperature time series pattern feature vector in the set of refrigerator internal temperature time series pattern feature vectors based on the initial aggregation representation vector of the refrigerator internal temperature time series pattern to obtain a set of superposition state energy potential factors of the refrigerator internal temperature time series pattern; a weighting subunit, which is used to input the set of superposition state energy potential factors of the refrigerator internal temperature time series pattern into the energy potential factor gated weighting network to obtain a set of characteristic energy potential weight coefficients of the refrigerator internal temperature time series pattern; a core feature convergence subunit, which is used to calculate the positional weighted sum of the set of characteristic energy potential weight coefficients of the refrigerator internal temperature time series pattern to obtain the core feature vector of the refrigerator internal temperature time series pattern.

[0012] In the above-mentioned big data-based smart refrigerator monitoring and early warning system, the characteristic energy potential dynamic analysis subunit includes: a static energy potential calculation secondary subunit, which is used to calculate the static energy potential factor of each refrigerator internal temperature time series pattern feature vector in the set of refrigerator internal temperature time series pattern feature vectors to obtain a set of refrigerator internal temperature time series pattern static energy potential factors; a dynamic energy potential calculation secondary subunit, which is used to calculate the dynamic energy potential factor of each refrigerator internal temperature time series pattern feature vector in the set of refrigerator internal temperature time series pattern feature vectors relative to the initial aggregation representation vector of the refrigerator internal temperature time series pattern to obtain a set of refrigerator internal temperature time series pattern dynamic energy potential factors; an energy potential superposition secondary subunit, which is used to determine the superposition state energy potential factor of each refrigerator internal temperature time series pattern feature vector in the set of refrigerator internal temperature time series pattern feature vectors based on the set of refrigerator internal temperature time series pattern static energy potential factors and the set of refrigerator internal temperature time series pattern dynamic energy potential factors to obtain a set of refrigerator internal temperature time series pattern superposition state energy potential factors.

[0013] In the above-mentioned big data-based smart refrigerator monitoring and early warning system, the static energy potential calculation secondary subunit is used to: calculate the fourth-order central moment of the characteristic vector of the refrigerator internal temperature time series pattern divided by the fourth power of its characteristic variance to obtain the static energy potential factor of the refrigerator internal temperature time series pattern.

[0014] In the above-mentioned big data-based smart refrigerator monitoring and early warning system, the dynamic energy potential calculation secondary subunit is used to: calculate the position-weighted sum of the refrigerator internal temperature time series pattern feature vector and the refrigerator internal temperature time series pattern initial aggregation representation vector to obtain the refrigerator internal temperature time series pattern dynamic energy potential representation vector; calculate the square of the L2 norm of the point sum vector of the refrigerator internal temperature time series pattern dynamic energy potential representation vector and the bias vector to obtain the refrigerator internal temperature time series pattern dynamic energy potential factor.

[0015] In the above-mentioned big data-based intelligent refrigerator monitoring and early warning system, the state monitoring result generating unit is used to: input the core feature vector of the refrigerator internal temperature time series pattern into a classifier-based state identifier to obtain the state monitoring result.

[0016] According to another aspect of the present application, a smart refrigerator monitoring and early warning method based on big data is provided, which includes: collecting a data set of the temperature inside the refrigerator through a temperature sensor; uploading the data set of the temperature inside the refrigerator to a smart refrigerator monitoring and early warning cloud through a wireless communication module; in the smart refrigerator monitoring and early warning cloud, inputting the data set of the temperature inside the refrigerator into a refrigerator status monitoring big model to obtain a status monitoring result; in the smart refrigerator monitoring and early warning cloud, in response to the status monitoring result being that the temperature timing pattern inside the refrigerator is abnormal, generating and sending a refrigerator internal hot food monitoring instruction to the smart refrigerator; in response to receiving the refrigerator internal hot food monitoring instruction, turning on a camera to collect an infrared image of food inside the refrigerator through the camera; uploading the infrared image of food inside the refrigerator to the smart refrigerator monitoring and early warning cloud through the wireless communication module; in the smart refrigerator monitoring and early warning cloud, performing image analysis on the infrared image of food inside the refrigerator to obtain an image analysis result; in response to the image analysis result being that there is no hot food, generating a refrigerator status abnormality early warning signal.

[0017] Compared with the prior art, the intelligent refrigerator monitoring and early warning system and method based on big data provided by the present application collects the temperature data inside the refrigerator in real time through the temperature sensor, and uses the big data analysis technology to perform time series analysis on the temperature data, digs out the time series change pattern of the temperature inside the refrigerator, and intelligently identifies the abnormal temperature state. Then, when the abnormal temperature change inside the refrigerator is detected, the infrared image of the food inside the refrigerator is collected and analyzed through infrared sensing technology and image processing technology to detect whether there is hot food inside the refrigerator, so that when the abnormal temperature change inside the refrigerator is detected and there is no hot food, a corresponding early warning prompt is issued. In this way, the user can be reminded to take corresponding measures in time when the refrigerator is working abnormally, and unnecessary false alarms are effectively avoided, thereby improving the accuracy of refrigerator monitoring and user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0019] Figure 1 It is a block diagram of a smart refrigerator monitoring and early warning system based on big data according to an embodiment of the present application.

[0020] Figure 2 It is a block diagram of a refrigerator status monitoring module in a smart refrigerator monitoring and early warning system based on big data according to an embodiment of the present application.

[0021] Figure 3 This is a schematic diagram of data flow of a refrigerator status monitoring module in a smart refrigerator monitoring and early warning system based on big data according to an embodiment of the present application.

[0022] Figure 4 It is a block diagram of a singular spectrum analysis unit in a smart refrigerator monitoring and early warning system based on big data according to an embodiment of the present application.

[0023] Figure 5 It is a block diagram of a dynamic characteristic concentration unit in a smart refrigerator monitoring and early warning system based on big data according to an embodiment of the present application.

[0024] Figure 6 This is a flow chart of a smart refrigerator monitoring and early warning method based on big data according to an embodiment of the present application. DETAILED DESCRIPTION

[0025] Below, the embodiments of the present application will be described in more detail in conjunction with the accompanying drawings, and the above and other purposes, features and advantages of the present application will become more apparent. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the example embodiments described herein.

[0026] In response to the technical problems described in the above background technology, the present application proposes a smart refrigerator monitoring and early warning system based on big data. Figure 1 FIG. 1 is a block diagram of a smart refrigerator monitoring and early warning system based on big data according to an embodiment of the present application. Figure 1As shown, according to the embodiment of the present application, the smart refrigerator monitoring and early warning system 100 based on big data includes: a refrigerator internal temperature monitoring module 110, which is used to collect a data set of the refrigerator internal temperature through a temperature sensor; a temperature data transmission module 120, which is used to upload the data set of the refrigerator internal temperature to the smart refrigerator monitoring and early warning cloud through a wireless communication module; a refrigerator state monitoring module 130, which is used to input the data set of the refrigerator internal temperature into the refrigerator state monitoring big model in the smart refrigerator monitoring and early warning cloud to obtain a state monitoring result; a hot food monitoring instruction generation module 140, which is used to generate a hot food monitoring instruction in response to the state monitoring result that the refrigerator internal temperature timing pattern is abnormal in the smart refrigerator monitoring and early warning cloud. And send the hot food monitoring instruction inside the refrigerator to the smart refrigerator; the food infrared image acquisition module 150 is used to respond to the receipt of the hot food monitoring instruction inside the refrigerator, turn on the camera to collect the infrared image of the food inside the refrigerator through the camera; the food infrared image transmission module 160 is used to upload the infrared image of the food inside the refrigerator to the smart refrigerator monitoring and early warning cloud through the wireless communication module; the food infrared image analysis module 170 is used to perform image analysis on the infrared image of the food inside the refrigerator in the smart refrigerator monitoring and early warning cloud to obtain an image analysis result; the refrigerator state abnormal warning module 180 is used to generate a refrigerator state abnormal warning signal in response to the image analysis result that there is no hot food.

[0027] Specifically, the big data-based smart refrigerator monitoring and early warning system described in this application integrates multiple functions such as temperature monitoring, data transmission, temperature time series state analysis, food infrared monitoring, food infrared image analysis, and early warning prompts, realizing comprehensive monitoring of the internal environment of the refrigerator and timely early warning of abnormal situations.

[0028] In the above-mentioned big data-based intelligent refrigerator monitoring and early warning system 100, the refrigerator internal temperature monitoring module 110 is used to collect a data set of the refrigerator internal temperature through a temperature sensor. In an embodiment of the present application, the temperature sensor can be a high-precision digital temperature sensor to ensure the accuracy and real-time performance of the temperature data.

[0029] In the above-mentioned big data-based smart refrigerator monitoring and early warning system 100, the temperature data transmission module 120 is used to upload the data set of the internal temperature of the refrigerator to the smart refrigerator monitoring and early warning cloud through the wireless communication module. That is, the collected internal temperature data set of the refrigerator is transmitted using wireless communication technology to ensure that the temperature data can be quickly and stably uploaded to the smart refrigerator monitoring and early warning cloud for data analysis. In an embodiment of the present application, the wireless communication module can use existing wireless network technology, such as Wi-Fi, Bluetooth or cellular network technology, to achieve efficient data transmission.

[0030] In the above-mentioned big data-based smart refrigerator monitoring and early warning system 100, the refrigerator status monitoring module 130 is used to input the data set of the refrigerator internal temperature into the refrigerator status monitoring big model in the smart refrigerator monitoring and early warning cloud to obtain the status monitoring result. In the technical solution of the present application, the refrigerator status monitoring big model is constructed based on a deep learning algorithm, and is used to perform time series analysis on the received data set of the refrigerator internal temperature to identify the time series change pattern of the refrigerator internal temperature and determine whether there is an abnormal temperature state.

[0031] Figure 2 It is a block diagram of a refrigerator status monitoring module in a smart refrigerator monitoring and early warning system based on big data according to an embodiment of the present application. Figure 3 FIG. 1 is a schematic diagram of data flow of a refrigerator status monitoring module in a smart refrigerator monitoring and early warning system based on big data according to an embodiment of the present application. Figure 2 and Figure 3 As shown, the refrigerator state monitoring module 130 includes: a data regularization unit 131, which is used to perform data regularization on the data set of the refrigerator internal temperature according to the time dimension to obtain a time series of the refrigerator internal temperature; a singular spectrum analysis unit 132, which is used to perform singular spectrum analysis on the time series of the refrigerator internal temperature to obtain a set of refrigerator internal temperature subsequences; a subsequence timing analysis unit 133, which is used to respectively extract the timing pattern features of each refrigerator internal temperature subsequence in the set of refrigerator internal temperature subsequences to obtain a set of refrigerator internal temperature timing pattern feature vectors; a dynamic characteristic concentration unit 134, which is used to perform dynamic characteristic concentration on the set of refrigerator internal temperature timing pattern feature vectors to obtain a refrigerator internal temperature timing pattern core feature vector; a state monitoring result generation unit 135, which is used to generate the state monitoring result based on the refrigerator internal temperature timing pattern core feature vector.

[0032] Specifically, the data regularization unit 131 is used to regularize the dataset of the temperature inside the refrigerator according to the time dimension to obtain the time series of the temperature inside the refrigerator. It should be understood that since the change of the temperature inside the refrigerator is a dynamic process that evolves over time, it has a certain continuity and trend. Therefore, in the technical solution of the present application, by regularizing the dataset of the temperature inside the refrigerator according to the time dimension, the time sequence relationship of the data can be maintained, which helps to better understand and analyze the temporal variation law of the temperature inside the refrigerator, and then accurately identify the abnormal pattern of the temperature inside the refrigerator.

[0033] Specifically, the singular spectrum analysis unit 132 is used to perform singular spectrum analysis on the time series of the temperature inside the refrigerator to obtain a set of subsequences of the temperature inside the refrigerator. It should be understood that considering that the original time series data of the temperature inside the refrigerator usually has complex nonlinear characteristics and may be affected by random factors such as door opening and sensor noise, it becomes difficult to extract the characteristics of its temperature change pattern. Therefore, in order to optimize the time series analysis effect of the temperature inside the refrigerator, the present application further uses a singular spectrum analysis method to process the time series of the temperature inside the refrigerator to separate the noise components and useful information in the original time series. Those of ordinary skill in the art should know that singular spectrum analysis (SSA) as a powerful time series analysis tool can effectively extract the main dynamic components from complex time series, decompose the complex time series of the temperature inside the refrigerator into multiple subsequences with different frequencies, and reveal the periodicity or trend characteristics in the data, thereby helping to deeply understand the nature of temperature changes, simplify subsequent processing, and enhance the system's abnormality detection capabilities.

[0034] Figure 4 FIG. 1 is a block diagram of a singular spectrum analysis unit in a smart refrigerator monitoring and early warning system based on big data according to an embodiment of the present application. Figure 4 As shown, the singular spectrum analysis unit 132 includes: a time series matrix construction subunit 1321, which is used to arrange the time series of the internal temperature of the refrigerator into a refrigerator internal temperature time series matrix according to a preset sliding window size; a singular value decomposition subunit 1322, which is used to perform singular value decomposition on the internal temperature time series matrix of the refrigerator to obtain a sequence of left singular eigenvectors inside the refrigerator and a sequence of right singular eigenvectors inside the refrigerator; a time series reconstruction subunit 1323, which is used to perform time series reconstruction on each group of corresponding left singular eigenvectors inside the refrigerator and right singular eigenvectors inside the refrigerator in the sequence of left singular eigenvectors inside the refrigerator and the sequence of right singular eigenvectors inside the refrigerator to obtain a set of subsequences of the internal temperature of the refrigerator.

[0035] Specifically, first, the time series of the temperature inside the refrigerator is arranged into a refrigerator internal temperature time series matrix according to a preset sliding window size, wherein the form of the refrigerator internal temperature time series matrix is ​​as follows: ; Wherein, each column of the refrigerator internal temperature time series matrix is ​​a local subsequence in the original refrigerator internal temperature time series, L is the preset sliding window length, N is the total length of the refrigerator internal temperature time series, and L is less than N, K=N-L+1.

[0036] Next, the singular value decomposition (SVD) of the refrigerator internal temperature time series matrix is ​​performed to obtain a set of singular values ​​and the left singular vectors and right singular vectors corresponding to each singular value, that is, the sequence of the left singular eigenvectors and the sequence of the right singular eigenvectors inside the refrigerator, so as to reveal the main dynamic components in the refrigerator internal temperature time series. The singular value decomposition formula is: , is the left singular vector matrix composed of the sequence arrangement of the left singular eigenvectors inside the refrigerator, is a diagonal matrix of singular values, It is a right singular vector matrix composed of the sequence arrangement of the right singular eigenvectors inside the refrigerator.

[0037] Then, according to the size of each singular value, the singular values ​​and their corresponding left singular eigenvectors inside the refrigerator and right singular eigenvectors inside the refrigerator are grouped, and the singular values ​​and singular vectors of each group are reconstructed to generate a reconstructed time series matrix: ,in, is the left singular vector matrix composed of the arrangement of the left singular eigenvectors inside each refrigerator in the same group, is the right singular vector matrix composed of the right singular eigenvectors of each refrigerator in the same group, is a singular value diagonal matrix composed of the permutations of the singular values ​​of the same group, To reconstruct the refrigerator internal temperature time series sub-matrix.

[0038] Finally, by calculating the column average of each group of reconstructed refrigerator internal temperature time series submatrices, each reconstructed refrigerator internal temperature time series submatrix is ​​converted back to the time series form, thereby obtaining a set of refrigerator internal temperature subsequences. In this way, the noise components in the original data can be effectively filtered out, and a more concise and regular temperature subsequence can be generated, which greatly simplifies the subsequent feature extraction and pattern recognition process, helps to accurately capture the main trends and periodic characteristics of temperature changes, and thus provides effective data support for accurate judgment of the refrigerator status.

[0039] Specifically, the subsequence timing analysis unit 133 is used to extract the timing pattern features of each refrigerator internal temperature subsequence in the set of refrigerator internal temperature subsequences to obtain a set of refrigerator internal temperature timing pattern feature vectors. In a specific example of the present application, the subsequence timing analysis unit 133 is used to: input each refrigerator internal temperature subsequence in the set of refrigerator internal temperature subsequences into a temperature timing pattern feature extractor based on a forward LSTM model to obtain a set of refrigerator internal temperature timing pattern feature vectors. That is, in order to capture the temperature change pattern contained in each refrigerator internal temperature subsequence, the present application uses a forward LSTM model (long short-term memory network) widely used in the field of timing analysis to extract features from each refrigerator internal temperature subsequence. It should be understood that the forward LSTM model is a special recurrent neural network, which, based on its internal loop structure and gating mechanism, can effectively process and memorize long-term dependencies in sequence data, so that it can fully understand the dynamic characteristics of temperature changes, capture the long-term temperature change trends and periodic characteristics in each refrigerator internal temperature subsequence, and obtain a set of refrigerator internal temperature time series pattern feature vectors.

[0040] Specifically, the dynamic characteristic concentration unit 134 is used to perform dynamic characteristic concentration on the set of refrigerator internal temperature time series pattern feature vectors to obtain the refrigerator internal temperature time series pattern core feature vector. It should be understood that in order to further extract the core features that can better reflect the essential characteristics of the data from the set of refrigerator internal temperature time series pattern feature vectors to improve the accuracy of subsequent refrigerator status monitoring, the present application proposes a dynamic characteristic concentration method based on feature dynamic analysis, which dynamically aggregates the set of refrigerator internal temperature time series pattern feature vectors by considering the information density carried by each refrigerator internal temperature time series pattern feature vector and the correlation between each refrigerator internal temperature time series pattern feature vector, and abstracts a highly generalized core representation form therefrom.

[0041] Figure 5 FIG. 1 is a block diagram of a dynamic characteristic concentration unit in a smart refrigerator monitoring and early warning system based on big data according to an embodiment of the present application. Figure 5As shown, the dynamic characteristic concentration unit 134 includes: an initial aggregation representation subunit 1341, which is used to calculate the positional mean vector of the set of characteristic vectors of the refrigerator internal temperature time series pattern to obtain the initial aggregation representation vector of the refrigerator internal temperature time series pattern; a characteristic energy potential dynamic analysis subunit 1342, which is used to calculate the energy potential factor of each refrigerator internal temperature time series pattern characteristic vector in the set of refrigerator internal temperature time series pattern characteristic vectors based on the initial aggregation representation vector of the refrigerator internal temperature time series pattern to obtain a set of superposition state energy potential factors of the refrigerator internal temperature time series pattern; a weighting subunit 1343, which is used to input the set of superposition state energy potential factors of the refrigerator internal temperature time series pattern into the energy potential factor gated weighting network to obtain a set of characteristic energy potential weight coefficients of the refrigerator internal temperature time series pattern; a core feature convergence subunit 1344, which is used to calculate the positional weighted sum of the set of characteristic energy potential weight coefficients of the refrigerator internal temperature time series pattern to obtain the core feature vector of the refrigerator internal temperature time series pattern.

[0042] Specifically, the initial aggregate representation subunit 1341 is used to calculate the position mean vector of the set of feature vectors of the refrigerator internal temperature time series pattern to obtain the initial aggregate representation vector of the refrigerator internal temperature time series pattern. The process can be expressed as follows: ;in, represents the set of characteristic vectors of the time series pattern of the temperature inside the refrigerator, and respectively represent the first, second, and third in the set of the refrigerator internal temperature time series pattern feature vectors and The characteristic vector of the time series pattern of the refrigerator internal temperature, is the number of characteristic vectors of the time series pattern of the internal temperature of the refrigerator, The initial aggregate representation vector for the time series pattern of the temperature inside the refrigerator.

[0043] It should be understood that the process of calculating the positional mean vector of the set of feature vectors of the refrigerator internal temperature time series pattern to obtain the initial aggregate representation vector of the refrigerator internal temperature time series pattern is essentially a comprehensive processing of the temperature data collected at multiple time points. This process adds and averages the values ​​of each feature vector at each time point (i.e., position) to generate a new vector that represents the overall behavior of temperature changes in all sampling time periods.

[0044] In this way, a simplified representative vector that reflects the overall trend can be obtained. This vector not only regularizes and simplifies the original temperature change sequence, which may be very complex and high-dimensional, but also better captures the general outline of the evolution of the temperature inside the refrigerator over time. This method helps to identify long-standing patterns or anomalies, such as the gradual decline in refrigeration system performance, and can effectively reduce the impact of random noise on data analysis, because the averaging process can offset these accidental factors to a certain extent, making the final feature vector more stable and reliable.

[0045] In addition, this initial aggregate representation vector can be used as a benchmark to measure the degree of difference between other individual feature vectors and it. For example, when calculating the dynamic energy potential factor in the subsequent steps, this benchmark is needed for comparative analysis to evaluate whether the temperature distribution at a specific moment deviates from the normal range. This step also enhances the robustness and generalization ability of the model. Even in the face of diverse temperature changes caused by different user habits, it can maintain good adaptability and prediction accuracy, and provide a solid foundation for more complex feature extraction, ensuring the effective operation of subsequent links.

[0046] More specifically, the characteristic energy potential dynamic analysis subunit 1342 includes: a static energy potential calculation secondary subunit, which is used to calculate the static energy potential factor of each refrigerator internal temperature time series pattern feature vector in the set of refrigerator internal temperature time series pattern feature vectors to obtain a set of refrigerator internal temperature time series pattern static energy potential factors; a dynamic energy potential calculation secondary subunit, which is used to calculate the dynamic energy potential factor of each refrigerator internal temperature time series pattern feature vector in the set of refrigerator internal temperature time series pattern feature vectors relative to the initial aggregate representation vector of the refrigerator internal temperature time series pattern to obtain a set of refrigerator internal temperature time series pattern dynamic energy potential factors; an energy potential superposition secondary subunit, which is used to determine the superposition state energy potential factor of each refrigerator internal temperature time series pattern feature vector in the set of refrigerator internal temperature time series pattern feature vectors based on the set of refrigerator internal temperature time series pattern static energy potential factors and the set of refrigerator internal temperature time series pattern dynamic energy potential factors to obtain a set of refrigerator internal temperature time series pattern superposition state energy potential factors.

[0047] In a specific example of the present application, the static energy potential calculation secondary subunit is used to: calculate the fourth-order central moment of the characteristic vector of the refrigerator internal temperature time series pattern divided by the fourth power of its characteristic variance to obtain the static energy potential factor of the refrigerator internal temperature time series pattern.

[0048] In a specific example of the present application, the dynamic energy potential calculation secondary subunit is used to: calculate the position-weighted sum of the characteristic vector of the refrigerator internal temperature time series pattern and the initial aggregate representation vector of the refrigerator internal temperature time series pattern to obtain the dynamic energy potential representation vector of the refrigerator internal temperature time series pattern; calculate the square of the L2 norm of the point sum vector of the refrigerator internal temperature time series pattern dynamic energy potential representation vector and the bias vector to obtain the dynamic energy potential factor of the refrigerator internal temperature time series pattern.

[0049] In a specific example of the present application, the energy potential superposition secondary subunit is used to: use the squares of the static energy potential factor and the dynamic energy potential factor of the characteristic vector of the refrigerator internal temperature time series pattern as exponents, calculate the exponential function value with base e to obtain the indexed static energy potential factor and the indexed dynamic energy potential factor, and add the indexed static energy potential factor and the indexed dynamic energy potential factor to obtain the superposition state energy potential factor of the refrigerator internal temperature time series pattern.

[0050] Specifically, the process can be expressed as:

[0051]

[0052] .

[0053] in, Indicates the The first characteristic vector of the time series pattern of the refrigerator internal temperature eigenvalues, and Respectively represent the The characteristic mean and characteristic variance of the characteristic vector of the time series pattern of the internal temperature of the refrigerator, represents the length of the characteristic vector of the time series pattern of the temperature inside the refrigerator, and Respectively represent the The static potential factor, dynamic potential factor and superposition potential factor of the characteristic vector of the time series pattern of the internal temperature of the refrigerator, represents the bias vector, and represents different weight parameters, represents dot product, It means calculating the square of the L2 norm of a vector.

[0054] That is, for each refrigerator internal temperature time series pattern feature vector, by calculating its static energy potential factor, the information density carried by the feature itself is revealed, and the inherent properties and information importance of each refrigerator internal temperature time series pattern feature vector when it is independent of other vectors are evaluated. Specifically, the static energy potential factor reflects the characteristics of each refrigerator internal temperature time series pattern feature vector itself, that is, it is independent of the information density carried by the data at other time points. By calculating the fourth-order central moment divided by the fourth power of its characteristic variance, it can be measured whether the time series segment has abnormal fluctuations or extreme values. For smart refrigerators, this helps to identify short-term temperature mutations that may be caused by user behavior (such as putting in a large amount of hot food at one time) or other non-fault factors. Feature vectors with higher static energy potential factors indicate that the temperature changes during this time period are more drastic, which may be a time period that requires special attention. At the same time, considering that each refrigerator internal temperature time series pattern feature vector has a certain correlation interaction with the overall change pattern of the data set, the dynamic energy potential factor of each refrigerator internal temperature time series pattern feature vector is further calculated to capture the relative relationship between the data and measure its interaction and influence with the overall data set, so as to have a deeper understanding of how the local characteristic structure of temperature affects its overall layout. Specifically, the dynamic energy potential factor considers the changes of each time series fragment relative to the average performance of the entire data set (i.e., the initial aggregate representation vector). This process involves calculating the weighted sum between each feature vector and the initial aggregate representation vector, and further calculating the L2 norm square of this sum and the bias vector. The dynamic energy potential factor reveals the degree of difference between the temperature change in a specific time period and the global trend, helping to distinguish between real anomalies caused by problems with the refrigerator itself (such as reduced refrigeration efficiency) and short-term fluctuations under normal operation. If the dynamic energy potential factor of a feature vector is high, it means that the temperature change in that time period has deviated significantly from the historical average level, which may be a signal of potential problems. Next, in order to comprehensively consider the static and dynamic properties of the characteristic vectors of the internal temperature time series pattern of each refrigerator, the static energy potential factor and the dynamic energy potential factor of the characteristic vectors of the internal temperature time series pattern of each refrigerator are further integrated to form a superposition energy potential factor. This comprehensive indicator is used to reflect the composite influence of the characteristic vectors of the internal temperature time series pattern of each refrigerator in the entire set.

[0055] In a specific example of the present application, the weighting subunit 1343 is used to: after the set of superposition state energy potential factors of the refrigerator internal temperature time series pattern is input into the sigmoid function for normalization, the set of normalized superposition state energy potential factors of the refrigerator internal temperature time series pattern is masked and inactivated based on the mask threshold to obtain the set of characteristic energy potential weight coefficients of the refrigerator internal temperature time series pattern. Specifically, the process can be expressed as follows: .

[0056] in, Indicates A normalized refrigerator internal temperature time series pattern superposition state energy potential factor, is the mask function, represents the mask threshold, For the The characteristic energy potential weight coefficient of the time series pattern of the internal temperature of the refrigerator.

[0057] That is, the set of superposition state potential factors of the refrigerator's internal temperature time series pattern is input into the sigmoid function for normalization, and the normalized factor set is masked and inactivated based on the mask threshold, and finally the set of characteristic potential weight coefficients of the refrigerator's internal temperature time series pattern is obtained. This process is the key link to ensure that the system can accurately distinguish between normal and abnormal temperature changes. First, by inputting the superposition state potential factors into the sigmoid function, nonlinear normalization can be achieved. The sigmoid function maps each superposition state potential factor to a continuous interval between 0 and 1, which not only maintains the proportional relationship of the original data, but also enables the temperature changes at different time points to be compared on the same scale. The normalized potential factors more intuitively reflect the importance of each time series segment to the overall temperature change trend, that is, those time periods with higher potential factors still maintain higher values ​​after normalization, indicating that they may be important indicators of abnormal conditions. Next, the normalized potential factor set is masked and inactivated based on the set mask threshold. The masking mechanism is introduced here to filter out data points that contribute little to the overall pattern or are considered to be noise. Specifically, if a normalized potential factor is lower than the preset mask threshold, it is considered that the temperature change in this time period is not representative enough or may be caused by random factors, so it is set to invalid (usually assigned a value of 0). On the contrary, the potential factor that is higher than or equal to the mask threshold is retained and used as valid information for subsequent analysis. This method effectively reduces the possibility of false alarms and improves the accuracy of the system in detecting anomalies. Finally, after the above processing, a set of potential weight coefficients of the characteristic time series pattern of the temperature inside the refrigerator is obtained. These weight coefficients directly reflect the relative importance of each time series segment in the entire temperature change process. A high weight means that the temperature change in this time period is more worthy of attention and may be an early signal of a potential problem; while a low weight means that the data in this time period is relatively ordinary and is unlikely to point to an abnormal situation. Such a weight distribution helps to highlight those time periods that really need attention, while suppressing unnecessary interference information, thereby helping the system make more accurate state judgments and early warning decisions.

[0058] In summary, based on the gated mask mechanism, the generated superposition state energy potential factors are normalized and weighted, so as to perform weighted aggregation on the set of characteristic vectors of the time series pattern of the internal temperature of the refrigerator, thereby explicitly marking the key pattern features that have a greater impact on the overall temperature change trend, and filtering out the feature representations that contribute less to the overall pattern, so as to obtain the core characteristic vector of the time series pattern of the internal temperature of the refrigerator.

[0059] Specifically, the core feature convergence subunit 1344 is used to calculate the position-weighted sum of the set of feature vectors of the refrigerator internal temperature time series pattern based on the set of feature potential weight coefficients of the refrigerator internal temperature time series pattern to obtain the core feature vector of the refrigerator internal temperature time series pattern. Specifically, the process can be expressed as follows: ;in, Represents the core feature vector of the time series pattern of the temperature inside the refrigerator.

[0060] It should be understood that the process of calculating the position-weighted sum of the set of feature vectors of the time series pattern of the internal temperature of the refrigerator based on the set of potential weight coefficients of the time series pattern of the internal temperature of the refrigerator, and finally obtaining the core feature vector of the time series pattern of the internal temperature of the refrigerator, is to condense the previous complex data processing and feature extraction results into a highly generalized core representation. This process not only simplifies the input data required for subsequent analysis, but also enhances the understanding and recognition of the temperature change pattern inside the refrigerator. By introducing the potential weight coefficient for weighted summation, the system can explicitly mark the key pattern features that have a greater impact on the overall temperature change trend. Specifically, the feature vector of each time series segment is assigned different weights according to its corresponding potential weight coefficient, which means that those time periods that are considered more important and more likely to reflect abnormal situations will occupy a larger proportion in the core feature vector finally formed. This approach ensures that even in a large amount of data, truly representative and discriminative information will not be overwhelmed, thereby improving the sensitivity and accuracy of anomaly detection.

[0061] Specifically, the state monitoring result generating unit 135 is used to generate the state monitoring result based on the core feature vector of the temperature time series pattern inside the refrigerator. In a specific example of the present application, the state monitoring result generating unit 135 is used to: input the core feature vector of the temperature time series pattern inside the refrigerator into a state identifier based on a classifier to obtain the state monitoring result. That is, a classification algorithm is used to classify the core feature vector of the temperature time series pattern inside the refrigerator to identify and distinguish different refrigerator temperature states. Specifically, the classifier establishes a classification model of the refrigerator temperature state by learning and training a large amount of labeled temperature data, so that it can intelligently judge whether there is an abnormality in the current temperature pattern inside the refrigerator based on the temperature time series change information contained in the core feature vector of the temperature time series pattern inside the refrigerator, thereby realizing real-time monitoring and abnormality detection of the refrigerator operation state.

[0062] Preferably, inputting the core feature vector of the refrigerator internal temperature time series pattern into a state identifier based on a classifier to obtain a state monitoring result comprises: calculating the absolute value sum of all feature values ​​of the core feature vector of the refrigerator internal temperature time series pattern and the square root of the sum of squares to obtain a first refrigerator internal temperature time series pattern core feature space structure value and a second refrigerator internal temperature time series pattern core feature space structure value, that is: ;in, The first core feature vector of the refrigerator internal temperature time series pattern is represented by eigenvalues, represents the core feature space structure value of the time series pattern of the internal temperature of the first refrigerator, Represents the core feature space structure value of the second refrigerator internal temperature time series pattern.

[0063] Determine the total number of eigenvalues ​​of the core eigenvector of the refrigerator internal temperature time series pattern .

[0064] For each eigenvalue of the core eigenvector of the refrigerator internal temperature time series pattern, calculate the core feature space structure value of the first refrigerator internal temperature time series pattern minus the product of the eigenvalue and the total number of eigenvalues ​​to obtain the core feature long-range dependency value of the first refrigerator internal temperature time series pattern ,in, represents the core feature space structure value of the time series pattern of the internal temperature of the first refrigerator, The first core feature vector of the refrigerator internal temperature time series pattern is represented by eigenvalues, The number of eigenvalues ​​representing all eigenvalues ​​of the core eigenvector of the refrigerator internal temperature time series pattern, Represents the long-range dependency value of the core feature of the time series pattern of the internal temperature of the first refrigerator.

[0065] Calculate the square root of the total number of eigenvalues ​​multiplied by the product of the eigenvalues ​​minus the core feature space structure value of the second refrigerator internal temperature time series pattern to obtain the core feature long-range dependency value of the second refrigerator internal temperature time series pattern ,in, represents the core feature space structure value of the second refrigerator internal temperature time series pattern, The first core feature vector of the refrigerator internal temperature time series pattern is represented by eigenvalues, The number of eigenvalues ​​representing all eigenvalues ​​of the core eigenvector of the refrigerator internal temperature time series pattern, Represents the long-range dependency value of the core feature of the second refrigerator's internal temperature time series pattern.

[0066] The index value calculated by taking the long-range dependency value of the core characteristic of the first refrigerator internal temperature time series pattern as the exponent of the natural constant and the inverse of the long-range dependency value of the core characteristic of the second refrigerator internal temperature time series pattern are weighted summed to obtain the optimized characteristic value corresponding to each characteristic value ,in, represents the long-range dependency value of the core feature of the time series pattern of the internal temperature of the first refrigerator, represents the long-range dependency value of the core feature of the second refrigerator internal temperature time series pattern, represents a natural constant, and represents the weighted hyperparameter, Indicates the optimized eigenvalue corresponding to each eigenvalue.

[0067] The optimized refrigerator internal temperature time series pattern core feature vector composed of the optimized feature values ​​is input into a classifier-based state identifier to obtain a state monitoring result.

[0068] Here, in the case where each refrigerator internal temperature time series pattern feature vector in the set of refrigerator internal temperature time series pattern feature vectors represents the time series characteristics of the refrigerator internal temperature in different time periods, after the core features are captured based on the dynamic analysis of the feature force field, the refrigerator internal temperature time series pattern core feature vectors will also have significant cross-domain time series aggregation spatial structure differences due to the differences in the structural dynamics between local time domains, affecting the convergence consistency of the classifier, and thus affecting the accuracy of the state monitoring results obtained by the state identifier based on the classifier.

[0069] Based on this, in view of the possible spatial structure loss of the feature set of the core feature vector of the refrigerator internal temperature time series pattern in the high-dimensional space, which causes the weight matrix of the classifier to implicitly infer the spatial structure information based on the features, resulting in inconsistent convergence. A long-distance feature dependency relationship is established based on the spatial structure representation of the core feature vector of the refrigerator internal temperature time series pattern based on the overall feature scale of the core feature vector of the refrigerator internal temperature time series pattern, so as to establish the feature local connectivity of the core feature vector of the refrigerator internal temperature time series pattern, and capture the spatial ambiguous information of the object feature value through the unstructured feature value point prediction of the core feature vector of the refrigerator internal temperature time series pattern, thereby improving the spatial inductive deviation perception ability of the feature set of the core feature vector of the refrigerator internal temperature time series pattern, improving the convergence consistency of the classifier, and improving the accuracy of the state monitoring result obtained by the state identifier based on the classifier based on the input of the core feature vector of the refrigerator internal temperature time series pattern.

[0070] In the above-mentioned big data-based smart refrigerator monitoring and early warning system 100, the hot food monitoring instruction generation module 140 is used in the smart refrigerator monitoring and early warning cloud, in response to the status monitoring result that the internal temperature timing pattern of the refrigerator is abnormal, to generate and send the refrigerator internal hot food monitoring instruction to the smart refrigerator. It should be understood that considering that the temperature abnormality inside the refrigerator may be caused by non-fault factors, such as the user putting a large amount of hot food in the refrigerator. Therefore, in order to avoid false alarms, further check the temperature rise phenomenon caused by possible hot food. When the refrigerator status monitoring large model identifies that there is an abnormality in the timing change pattern of the internal temperature of the refrigerator, the present application further generates a refrigerator internal hot food monitoring instruction and sends it to the smart refrigerator to start the hot food monitoring program.

[0071] In the above-mentioned big data-based smart refrigerator monitoring and early warning system 100, the food infrared image acquisition module 150 is used to respond to receiving the hot food monitoring instruction inside the refrigerator, turn on the camera to collect the infrared image of the food inside the refrigerator through the camera. That is, after receiving the instruction, the smart refrigerator will start the built-in infrared camera to capture the infrared image of the food inside the refrigerator. It should be understood that the infrared image of the food inside the refrigerator can provide detailed information about the temperature distribution of the food, so as to further confirm whether the abnormal temperature inside the refrigerator is caused by hot food.

[0072] In the above-mentioned big data-based smart refrigerator monitoring and early warning system 100, the food infrared image transmission module 160 is used to upload the infrared image of the food inside the refrigerator to the smart refrigerator monitoring and early warning cloud through the wireless communication module. Here, the food infrared image transmission module also uses wireless communication technology to upload the captured infrared image of the food inside the refrigerator to the cloud to ensure the timeliness and integrity of information transmission, so as to use the image analysis algorithm deployed in the smart refrigerator monitoring and early warning cloud to conduct in-depth analysis of the infrared image and determine whether there is hot food inside the refrigerator.

[0073] In the above-mentioned big data-based smart refrigerator monitoring and early warning system 100, the food infrared image analysis module 170 is used to perform image analysis on the infrared image of the food inside the refrigerator in the smart refrigerator monitoring and early warning cloud to obtain an image analysis result. In one embodiment of the present application, the food infrared image analysis module uses a convolutional neural network model to perform image analysis on the infrared image of the food inside the refrigerator, wherein the convolutional neural network model includes multiple convolutional layers, pooling layers and fully connected layers, which extracts the thermal radiation features of the food image through multiple layers of convolution, performs feature dimensionality reduction through the pooling layer to prevent overfitting, and classifies the extracted thermal radiation features in combination with the fully connected layer, thereby determining whether the image contains hot food and obtaining the image analysis result.

[0074] In the above-mentioned big data-based smart refrigerator monitoring and early warning system 100, the abnormal refrigerator status early warning module 180 is used to generate an abnormal refrigerator status early warning signal in response to the image analysis result that there is no hot food. It should be understood that when the image analysis result shows that there is no hot food inside the refrigerator, it can be preliminarily determined that there is a problem with the refrigerator itself. The abnormal refrigerator status early warning module will trigger the generation of an abnormal refrigerator status early warning signal, and notify the user through the user interface of the smart refrigerator or the mobile device application connected to the refrigerator, prompting the user to check whether the refrigerator has a malfunction or other problems. Through this comprehensive monitoring and analysis mechanism, false alarms can be effectively reduced, and the accuracy of refrigerator monitoring and the user's experience can be improved.

[0075] In summary, according to the embodiment of the present application, the intelligent refrigerator monitoring and early warning system based on big data is explained, which collects the temperature data inside the refrigerator in real time through the temperature sensor, and uses the big data analysis technology to perform time series analysis on the temperature data, and mines the time series change pattern of the temperature inside the refrigerator to intelligently identify abnormal temperature conditions. Then, when the abnormal temperature change inside the refrigerator is detected, the infrared image of the food inside the refrigerator is collected and analyzed through infrared sensing technology and image processing technology to detect whether there is hot food inside the refrigerator, so that when the abnormal temperature change inside the refrigerator is detected and there is no hot food, a corresponding early warning prompt is issued. In this way, the user can be reminded to take corresponding measures in time when the refrigerator is working abnormally, and unnecessary false alarms are effectively avoided, thereby improving the accuracy of refrigerator monitoring and user experience.

[0076] Figure 6 FIG. 1 is a flow chart of a smart refrigerator monitoring and early warning method based on big data according to an embodiment of the present application. Figure 6 As shown, according to the embodiment of the present application, the smart refrigerator monitoring and early warning method based on big data includes the following steps: S1, collecting a data set of the temperature inside the refrigerator through a temperature sensor; S2, uploading the data set of the temperature inside the refrigerator to the smart refrigerator monitoring and early warning cloud through a wireless communication module; S3, in the smart refrigerator monitoring and early warning cloud, inputting the data set of the temperature inside the refrigerator into the refrigerator state monitoring big model to obtain a state monitoring result; S4, in the smart refrigerator monitoring and early warning cloud, in response to the state monitoring result that the temperature timing pattern inside the refrigerator is abnormal, generating and sending a refrigerator internal hot food monitoring instruction to the smart refrigerator; S5, in response to receiving the refrigerator internal hot food monitoring instruction, turning on the camera to collect an infrared image of the food inside the refrigerator through the camera; S6, uploading the infrared image of the food inside the refrigerator to the smart refrigerator monitoring and early warning cloud through the wireless communication module; S7, in the smart refrigerator monitoring and early warning cloud, performing image analysis on the infrared image of the food inside the refrigerator to obtain an image analysis result; S8, in response to the image analysis result that there is no hot food, generating a refrigerator state abnormality early warning signal.

[0077] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned smart refrigerator monitoring and early warning method based on big data have been referred to above. Figures 1 to 5 The description of the big data-based smart refrigerator monitoring and early warning system has been introduced in detail, and therefore, its repeated description will be omitted.

[0078] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0079] In the above embodiments, the description of each embodiment has its own emphasis. For the parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiment described above is only schematic. For example, the unit division is only a logical function division, and there may be other division methods in actual implementation. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0080] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0081] Finally, it should be noted that the above description has been given for the purpose of illustration and description. In addition, the above embodiments are only used to illustrate the technical solution of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A smart refrigerator monitoring and early warning system based on big data, characterized in that: include: The refrigerator internal temperature monitoring module is used to collect the data set of the refrigerator internal temperature through the temperature sensor; A temperature data transmission module, used to upload the data set of the internal temperature of the refrigerator to the smart refrigerator monitoring and early warning cloud through a wireless communication module; A refrigerator status monitoring module, used to input the data set of the internal temperature of the refrigerator into the refrigerator status monitoring big model in the smart refrigerator monitoring and early warning cloud to obtain a status monitoring result; A hot food monitoring instruction generation module is used to generate and send a hot food monitoring instruction for the refrigerator to the smart refrigerator in response to the state monitoring result indicating that the temperature timing pattern inside the refrigerator is abnormal in the smart refrigerator monitoring and early warning cloud; a food infrared image acquisition module, configured to, in response to receiving the hot food monitoring instruction inside the refrigerator, turn on the camera to acquire an infrared image of the food inside the refrigerator through the camera; A food infrared image transmission module, used to upload the infrared image of the food inside the refrigerator to the smart refrigerator monitoring and warning cloud through the wireless communication module; A food infrared image analysis module, used to perform image analysis on the infrared image of food inside the refrigerator in the smart refrigerator monitoring and early warning cloud to obtain an image analysis result; A refrigerator state abnormality warning module, for generating a refrigerator state abnormality warning signal in response to the image analysis result indicating that hot food does not exist; Wherein, the refrigerator status monitoring module includes: A data regularization unit, used for regularizing the data set of the temperature inside the refrigerator according to the time dimension to obtain a time series of the temperature inside the refrigerator; A singular spectrum analysis unit, used for performing a singular spectrum analysis on the time series of the temperature inside the refrigerator to obtain a set of subsequences of the temperature inside the refrigerator; A subsequence time series analysis unit, used to respectively extract the time series pattern features of each refrigerator internal temperature subsequence in the set of refrigerator internal temperature subsequences to obtain a set of refrigerator internal temperature time series pattern feature vectors; A dynamic characteristic concentration unit, used for performing dynamic characteristic concentration on the set of characteristic vectors of the time series pattern of the temperature inside the refrigerator to obtain a core characteristic vector of the time series pattern of the temperature inside the refrigerator; The state monitoring result generating unit is used to generate the state monitoring result based on the core feature vector of the time series pattern of the internal temperature of the refrigerator.

2. According to claim 1, the intelligent refrigerator monitoring and early warning system based on big data is characterized in that: The singular spectrum analysis unit comprises: A time series matrix construction subunit, used for arranging the time series of the temperature inside the refrigerator into a time series matrix of the temperature inside the refrigerator according to a preset sliding window size; A singular value decomposition subunit, used for performing singular value decomposition on the refrigerator internal temperature time series matrix to obtain a sequence of left singular eigenvectors inside the refrigerator and a sequence of right singular eigenvectors inside the refrigerator; A time series reconstruction subunit is used to perform time series reconstruction on each group of corresponding left singular feature vectors inside the refrigerator and right singular feature vectors inside the refrigerator in the sequence of left singular feature vectors inside the refrigerator and the sequence of right singular feature vectors inside the refrigerator to obtain a set of subsequences of the temperature inside the refrigerator.

3. The big data-based intelligent refrigerator monitoring and early warning system according to claim 2 is characterized in that: The subsequence timing analysis unit is used for: Each refrigerator internal temperature subsequence in the set of refrigerator internal temperature subsequences is input into the temperature time series pattern feature extractor based on the forward LSTM model to obtain a set of refrigerator internal temperature time series pattern feature vectors.

4. The big data-based intelligent refrigerator monitoring and early warning system according to claim 3 is characterized in that: The dynamic characteristic concentration unit comprises: An initial aggregate representation subunit, used to calculate a positional mean vector of a set of feature vectors of the refrigerator internal temperature time series pattern to obtain an initial aggregate representation vector of the refrigerator internal temperature time series pattern; A characteristic energy potential dynamic analysis subunit is used to calculate the energy potential factor of each refrigerator internal temperature time series pattern characteristic vector in the set of refrigerator internal temperature time series pattern characteristic vectors based on the initial aggregation representation vector of the refrigerator internal temperature time series pattern to obtain a set of refrigerator internal temperature time series pattern superposition state energy potential factors; A weighting subunit, used for inputting the set of superposition state energy potential factors of the refrigerator internal temperature time series pattern into the energy potential factor gated weighting network to obtain a set of characteristic energy potential weight coefficients of the refrigerator internal temperature time series pattern; The core feature convergence subunit is used to calculate the position-weighted sum of the set of feature vectors of the refrigerator internal temperature time series pattern based on the set of feature potential weight coefficients of the refrigerator internal temperature time series pattern to obtain the core feature vector of the refrigerator internal temperature time series pattern.

5. The big data-based intelligent refrigerator monitoring and early warning system according to claim 4 is characterized in that: The characteristic energy potential dynamic analysis subunit comprises: A static energy potential calculation secondary subunit is used to calculate the static energy potential factor of each refrigerator internal temperature time series pattern feature vector in the set of refrigerator internal temperature time series pattern feature vectors to obtain a set of refrigerator internal temperature time series pattern static energy potential factors; A dynamic energy potential calculation secondary subunit is used to calculate the dynamic energy potential factor of each refrigerator internal temperature time series pattern feature vector in the set of refrigerator internal temperature time series pattern feature vectors relative to the refrigerator internal temperature time series pattern initial aggregation representation vector to obtain a set of refrigerator internal temperature time series pattern dynamic energy potential factors; The potential superposition secondary subunit is used to determine the superposition state energy potential factor of each refrigerator internal temperature timing pattern feature vector in the set of refrigerator internal temperature timing pattern feature vectors based on the set of static energy potential factors of the refrigerator internal temperature timing pattern and the set of dynamic energy potential factors of the refrigerator internal temperature timing pattern to obtain the set of superposition state energy potential factors of the refrigerator internal temperature timing pattern.

6. The big data-based intelligent refrigerator monitoring and early warning system according to claim 5 is characterized in that: The static energy potential calculation secondary subunit is used for: The fourth-order central moment of the characteristic vector of the refrigerator internal temperature time series pattern is calculated and divided by the fourth power of its characteristic variance to obtain the static energy potential factor of the refrigerator internal temperature time series pattern.

7. The big data-based intelligent refrigerator monitoring and early warning system according to claim 6 is characterized in that: The dynamic energy potential calculation secondary subunit is used for: Calculating the position-weighted sum of the characteristic vector of the refrigerator internal temperature time series pattern and the initial aggregation representation vector of the refrigerator internal temperature time series pattern to obtain a dynamic energy potential representation vector of the refrigerator internal temperature time series pattern; The square of the L2 norm of the point-sum vector of the dynamic energy potential representation vector and the bias vector of the refrigerator internal temperature time series pattern is calculated to obtain the dynamic energy potential factor of the refrigerator internal temperature time series pattern.

8. The big data-based intelligent refrigerator monitoring and early warning system according to claim 7 is characterized in that: The status monitoring result generating unit is used to: The core feature vector of the refrigerator internal temperature time series pattern is input into a classifier-based state identifier to obtain the state monitoring result.

9. A smart refrigerator monitoring and early warning method based on big data, using the smart refrigerator monitoring and early warning system based on big data according to claim 1, characterized in that: include: Collect a data set of the internal temperature of the refrigerator through a temperature sensor; Uploading the data set of the internal temperature of the refrigerator to the smart refrigerator monitoring and early warning cloud through the wireless communication module; In the smart refrigerator monitoring and early warning cloud, the data set of the internal temperature of the refrigerator is input into the refrigerator status monitoring big model to obtain the status monitoring result; In the smart refrigerator monitoring and early warning cloud, in response to the status monitoring result indicating that the temperature timing pattern inside the refrigerator is abnormal, a refrigerator internal hot food monitoring instruction is generated and sent to the smart refrigerator; In response to receiving the hot food monitoring instruction inside the refrigerator, turning on the camera to collect infrared images of the food inside the refrigerator through the camera; Uploading the infrared image of the food inside the refrigerator to the smart refrigerator monitoring and warning cloud through the wireless communication module; In the smart refrigerator monitoring and early warning cloud, image analysis is performed on the infrared image of the food inside the refrigerator to obtain an image analysis result; In response to the image analysis result indicating that no hot food exists, a warning signal for abnormal refrigerator status is generated.

Images