Germination process control method and system of germinated food

Abstract

The invention discloses a germination process control method and system for germinated food, and the method comprises the steps: building historical germination based on the growth conditions of the germinated food in each growth stage, and storing the reference growth data of the germinated food in each growth stage; collecting growth data of the germinated food in each stage in real time, and obtaining the change rate of each growth parameter of the germinated food; comparing the real-time sprouting food growth data with historical sprouting data of a corresponding growth cycle in a historical sprouting database to obtain an error value; constructing a growth deviation analysis model, and analyzing and acquiring growth deviation factors of each germinated food based on the calculated error values; and generating a germination process influence factor regulation instruction, and dynamically regulating the germination influence factors to control the germination process. The method has the advantages that through real-time monitoring, historical data comparison and LSTM neural network analysis, the growth process of germinated food is accurately regulated and controlled, growth conditions are optimized, and stable food quality is ensured.

Description

Technical Field

[0001] The present invention relates to a germination process control technology, in particular to a germination process control method and system for germinated food. Background Art

[0002] Sprouted foods are widely used in health foods, baby foods, and functional foods due to their rich, easily absorbed nutrients and high bioactivity. They can also be used to improve gastrointestinal health and boost immunity. With consumers' increasing emphasis on healthy eating, market demand for sprouted foods is steadily increasing. Therefore, precise control of the sprouting process not only improves product quality but also drives the development of the sprouted food industry.

[0003] Current commercially available control methods and systems for the sprouting process of sprouted foods lack precise real-time monitoring and dynamic regulation capabilities. They often rely on fixed empirical data or manual intervention, making it difficult to achieve precise control and automatic adjustment of the growth process. Secondly, many existing systems fail to fully utilize advanced machine learning algorithms, such as LSTM neural networks, for deviation analysis and process optimization. Consequently, they struggle to respond promptly and effectively to complex growth deviations. Furthermore, traditional control methods often focus on a single growth parameter or stage, lacking comprehensive cross-stage regulation capabilities. This can lead to unforeseen impacts during the growth process and an inability to achieve precise multi-factor regulation. Furthermore, current systems lack the flexibility and intelligence of their control strategies. They often lack robust dynamic adjustment mechanisms, making it impossible to update and optimize control schemes in real time. This can easily lead to imbalances or suboptimal results at certain stages of the growth process. Summary of the Invention

[0004] In order to improve the existing germination process control method and system for sprouted food, a germination process control method and system for sprouted food are provided. The method realizes precise control of the growth process of sprouted food, optimizes growth conditions, and ensures stable food quality through real-time monitoring, historical data comparison, and LSTM neural network analysis.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] A method for controlling the germination process of a germinated food, comprising:

[0007] Based on the growth status of sprouted foods at each growth stage, a historical sprouting database is constructed to store reference growth data of sprouted foods at each growth stage;

[0008] Collect growth data of sprouted food at each stage in real time and obtain the change rate of each growth parameter of sprouted food;

[0009] Based on the acquired real-time sprouted food growth data, the data is compared with the historical sprout data of the corresponding growth period in the historical sprout database, and the error value between the real-time growth data and the reference data is calculated;

[0010] Based on the historical germination database, the LSTM neural network model was trained to build a growth deviation analysis model. Based on the calculated error values, the growth deviation factors of each sprouted food were analyzed and obtained.

[0011] Based on the acquired growth deviation factors, control instructions for factors affecting the germination process are generated, and the germination process is controlled by dynamically adjusting the germination influencing factors.

[0012] Preferably, the construction of a historical germination database based on the growth conditions of the sprouted food at each growth stage, and storing reference growth data of the sprouted food at each growth stage specifically includes:

[0013] The growth cycle of sprouted food is divided into germination period, elongation period, leaf expansion period and maturity period, and the growth cycle is identified by the morphology of sprouted food;

[0014] Morphological dynamics feature extraction, environmental response feature extraction and metabolic marker extraction were performed based on the sprouted foods in each growth cycle;

[0015] The acquired parameter data are stored in a historical germination database and classified based on the growth cycle of the germinated food.

[0016] Preferably, the real-time collection of growth data of the sprouted food at each stage and obtaining the change rate of each growth parameter of the sprouted food specifically includes:

[0017] Use near-infrared cameras to detect the growth of radicles and plumules;

[0018] Biomass is measured using high-precision strain gauge sensors, and the interference of irrigation water weight is eliminated through dynamic weighing compensation algorithms;

[0019] Use the camera to shoot the bud crown at a low angle and obtain leaf expansion data in real time;

[0020] Based on different sampling periods for each growth cycle, the rate of change of each growth parameter is calculated by differential method, and the growth turning point is identified by the second-order derivative to judge the growth cycle.

[0021] Preferably, the real-time sprouted food growth data obtained is compared with the historical sprout data of the corresponding growth period in the historical sprout database, and the error value between the real-time growth data and the reference data is calculated and obtained, specifically comprising:

[0022] Generate real-time growth curves from the growth data of various sprouted foods, align them with the historical benchmark curves in the historical sprout database through a dynamic time warping algorithm, and establish a stage mapping relationship table;

[0023] Based on the curves of each stage of various types of growth data, the stratified error index is calculated and the error value is obtained;

[0024] Based on the obtained error values, an error adjacency matrix is ​​constructed to quantify the influence relationship between various parameters;

[0025] Based on the obtained error values ​​of various parameters and the quantitative relationship obtained from the error adjacency matrix, a comprehensive error value is calculated to obtain the true error value between the real-time growth data and the reference data.

[0026] Preferably, the growth deviation analysis model is constructed by training the LSTM neural network model based on the historical germination database, and the growth deviation factors of each sprouted food are analyzed and obtained based on the calculated error value, specifically including:

[0027] Convert the time series data in the historical database into a three-dimensional tensor structure, including the number of germination batches, time steps, and feature dimensions;

[0028] Define error labels based on various influencing factors and quantify evaluation criteria;

[0029] Design an LSTM-attention hybrid architecture through the input layer, bidirectional LSTM layer, temporal attention layer, and feature attention layer for model training;

[0030] Based on the trained growth deviation analysis model, various error values ​​obtained by calculation are input to obtain the growth deviation factors of each sprouted food;

[0031] By calculating the contribution score of each time step and feature dimension, the key impact period of various influencing factors can be obtained;

[0032] The LSTM hidden layer state and growth parameters are used as nodes, and the influence direction is used as the edge to construct an impact causal map. It is updated in real time based on the results of the growth deviation analysis model. When a new type of deviation is determined, a new deviation pattern node is automatically created and the key identification features are marked.

[0033] Preferably, the generating of the control instructions for factors affecting the germination process based on the acquired growth deviation factors, and controlling the germination process by dynamically adjusting the factors affecting the germination process specifically include:

[0034] Based on the causal map, a three-level regulatory rule base is constructed, including single-factor linear compensation, multi-factor coupling regulation, and cross-stage predictive regulation;

[0035] Based on the two-dimensional evaluation model of urgency and impact, a control priority matrix was constructed to regulate the germination process under different circumstances.

[0036] Dynamic control strategy adjustments are made based on the real-time updated impact causal map.

[0037] Furthermore, a sprouting process control system for sprouted food is proposed, comprising:

[0038] Historical germination database module: The historical germination database module is responsible for storing and managing reference growth data of sprouted foods at various growth stages, and supports growth cycle and stage classification;

[0039] Data acquisition module: The data acquisition module is used to collect real-time growth data of sprouted food at various stages, such as radicle length, embryo height, biomass, leaf expansion, etc., and calculate the rate of change;

[0040] Error calculation module: The error calculation module is used to compare the real-time collected growth data with the reference data in the historical database, calculate the error value, and generate a growth curve;

[0041] Model training module: The model training module is used to train historical data through LSTM neural network, build a growth deviation analysis model, and analyze deviation factors;

[0042] Graph construction module: The graph construction module is used to construct and update the causal graph of the growth process based on the analysis results of the LSTM model;

[0043] Regulation and control module: The regulation and control module is used to generate regulation instructions according to growth deviation factors and dynamically adjust the factors affecting the germination process;

[0044] Processor: The processor is mainly used for the calculation process of each formula and the construction calculation process of each model.

[0045] Compared with the prior art, the advantages of the present invention are:

[0046] By integrating historical databases, real-time growth data collection, LSTM neural network models, and dynamic control strategies, the germination process can be precisely controlled to ensure the quality and consistency of food growth. First, by establishing reference growth data from a historical germination database and comparing it with real-time growth data, the growth status of sprouted foods can be monitored in real time, allowing for the timely detection and analysis of growth deviations. The application of the LSTM neural network model makes deviation analysis more precise. By identifying key influencing factors, the germination environment can be dynamically adjusted to optimize conditions at each growth stage. Furthermore, by influencing the causal map and hierarchical control rule base, combined with the dynamic adjustment of the priority matrix, this method makes the control of the germination process more intelligent and flexible, enabling rapid response to varying environments and needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A schematic diagram of the method proposed in the present invention;

[0048] Figure 2 A schematic diagram of the construction of the historical germination database proposed in the present invention;

[0049] Figure 3 This is a schematic diagram of the growth parameter change rate proposed by the present invention;

[0050] Figure 4 This is a schematic diagram of the growth data error value proposed by the present invention;

[0051] Figure 5 This is a schematic diagram of the growth deviation factors proposed by the present invention;

[0052] Figure 6 This is a schematic diagram of the dynamic control instructions proposed by the present invention;

[0053] Figure 7 This is a diagram of the architecture of the electronic equipment in this solution;

[0054] Figure 8 This is a schematic diagram of the computer-readable storage medium structure in this solution. DETAILED DESCRIPTION

[0055] The following description is intended to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are merely examples, and those skilled in the art may conceive of other obvious variations.

[0056] A sprouting process control system for sprouted food, comprising:

[0057] Historical germination database module: The historical germination database module is responsible for storing and managing reference growth data of sprouted foods at various growth stages, and supports growth cycle and stage classification;

[0058] Data acquisition module: The data acquisition module is used to collect real-time growth data of sprouted food at various stages, such as radicle length, embryo height, biomass, leaf expansion, etc., and calculate the rate of change;

[0059] Error calculation module: The error calculation module is used to compare the real-time collected growth data with the reference data in the historical database, calculate the error value, and generate a growth curve;

[0060] Model training module: The model training module is used to train historical data through LSTM neural network, build a growth deviation analysis model, and analyze deviation factors;

[0061] Graph construction module: The graph construction module is used to construct and update the causal graph of the growth process based on the analysis results of the LSTM model;

[0062] Regulation and control module: The regulation and control module is used to generate regulation instructions according to growth deviation factors and dynamically adjust the factors affecting the germination process;

[0063] Processor: The processor is mainly used for the calculation process of each formula and the construction calculation process of each model.

[0064] See Figure 1 As shown, a method for controlling the germination process of a sprouted food comprises:

[0065] Step 1: Based on the growth status of sprouted foods at each growth stage, a historical sprouting database is constructed to store reference growth data of sprouted foods at each growth stage;

[0066] Step 2: Real-time collection of growth data of sprouted food at each stage, and obtaining the change rate of each growth parameter of the sprouted food;

[0067] Step 3: Based on the acquired real-time sprouted food growth data, compare it with the historical sprout data of the corresponding growth period in the historical sprout database, and calculate the error value between the real-time growth data and the reference data;

[0068] Step 4: Based on the historical germination database, the LSTM neural network model is trained to build a growth deviation analysis model. Based on the calculated error values, the growth deviation factors of each sprouted food are analyzed and obtained;

[0069] Step 5: Generate control instructions for factors affecting the germination process based on the acquired growth deviation factors, and control the germination process by dynamically adjusting the germination influencing factors.

[0070] See Figure 2 As shown, based on the growth status of sprouted foods at each growth stage, a historical sprouting database is constructed to store reference growth data of sprouted foods at each growth stage, specifically including:

[0071] The growth cycle of sprouted food is divided into germination period, elongation period, leaf expansion period and maturity period, and the growth cycle is identified by the morphology of sprouted food;

[0072] Morphological dynamics feature extraction, environmental response feature extraction and metabolic marker extraction were performed based on the sprouted foods in each growth cycle;

[0073] The acquired parameter data are stored in a historical germination database and classified based on the growth cycle of the germinated food.

[0074] Specifically, the germination stage is characterized by the beginning of germination, when the seed absorbs water and expands, and the embryo emerges from the seed coat, which is marked by the cracking of the seed coat and the initial growth of the root tip and embryo; the elongation stage is characterized by the rapid elongation of the embryo and the beginning of root differentiation, which is marked by the beginning of root growth and the extension of cotyledons; the leaf expansion stage is characterized by the full expansion of leaves and the beginning of photosynthesis, which is marked by the expansion of leaves and the beginning of independent plant growth; the maturity stage is characterized by the plant reaching its final form, stable photosynthesis, and slow growth, which is marked by the complete development of all parts of the plant and the cessation of rapid growth;

[0075] Morphodynamic characteristics mainly include the rate and direction of change of plant morphology during growth. The growth rate is described by measuring the rate of change of plant root length, stem length or leaf area. The growth direction is identified by analyzing the distribution of morphology in space. Edge detection algorithms can be used to identify the directionality of morphology.

[0076] Environmental response characteristics mainly refer to the reaction of sprouted food to environmental changes such as temperature, humidity, and light. Data is collected through environmental sensors and then correlated with plant growth data for analysis.

[0077] Metabolic markers refer to chemical substances related to growth and development produced by plants during different growth stages. Changes in metabolic markers can help identify the metabolic state of plants in different growth cycles, thereby assisting in determining their growth stage.

[0078] See Figure 3 As shown, real-time collection of growth data of sprouted food at each stage and acquisition of change rates of various growth parameters of sprouted food specifically include:

[0079] Use near-infrared cameras to detect the growth of radicles and plumules;

[0080] Biomass is measured using high-precision strain gauge sensors, and the interference of irrigation water weight is eliminated through dynamic weighing compensation algorithms;

[0081] Use the camera to shoot the bud crown at a low angle and obtain leaf expansion data in real time;

[0082] Based on different sampling periods for each growth cycle, the rate of change of each growth parameter is calculated by differential method, and the growth turning point is identified by the second-order derivative to judge the growth cycle.

[0083] Specifically, since the weight of irrigation water will interfere with biomass measurement, a dynamic weighing compensation algorithm is used to eliminate this influence, and the total weight measured by the sensor is Including plant biomass and the weight of irrigation water , by the weight of irrigation water For modeling purposes, the weight of water can be subtracted from the total weight to obtain the net biomass of the plant: ,Using dynamic compensation algorithm, the weight estimation of irrigation water is updated in real time, and the biomass data is adjusted in real time;

[0084] The leaf spread is calculated by analyzing the morphological changes of the leaf, such as leaf area and angle. At time t, the area of ​​the leaf is A(t). The leaf spread can be defined as the percentage relative to the maximum spread area:

[0085]

[0086] Among them, A max is the maximum expansion area, and D(t) is the percentage of the maximum expansion area;

[0087] In order to accurately identify the turning point in plant growth, the second-order derivative of the rate of change of the growth parameter can be calculated. The change in the sign of the second-order derivative can be used as a sign of the growth turning point. When the second-order derivative changes from a positive value to a negative value, it means that the growth rate begins to slow down, which may be a turning point in the growth cycle.

[0088] By analyzing the second-order derivatives of each growth stage, the turning point of growth is determined, and then different growth cycles are divided.

[0089] See Figure 4 As shown, based on the acquired real-time sprouted food growth data, it is compared with the historical sprout data of the corresponding growth period in the historical sprout database, and the error value between the real-time growth data and the reference data is calculated and obtained, specifically including:

[0090] Generate real-time growth curves from the growth data of various sprouted foods, align them with the historical benchmark curves in the historical sprout database through a dynamic time warping algorithm, and establish a stage mapping relationship table;

[0091] Based on the curves of each stage of various types of growth data, the stratified error index is calculated and the error value is obtained;

[0092] Based on the obtained error values, an error adjacency matrix is ​​constructed to quantify the influence relationship between various parameters;

[0093] Based on the obtained error values ​​of various parameters and the quantitative relationship obtained from the error adjacency matrix, a comprehensive error value is calculated to obtain the true error value between the real-time growth data and the reference data.

[0094] Specifically, the real-time growth curve is P(t) and the historical benchmark curve is Q(t). The lengths of the two curves are N and M respectively. The cumulative distance matrix D(i, j) is defined as follows:

[0095]

[0096] Where D(i, j) is the cost of aligning the i-th point of the real-time curve with the j-th point of the reference curve;

[0097] The update rule of the cumulative distance matrix is:

[0098]

[0099] Among them, the minimum cumulative cost is selected to update the value of each position;

[0100] The DTW algorithm can be used to obtain a pair of optimal matching paths, representing the best alignment between the real-time growth curve and the historical benchmark curve. The corresponding growth stages of the real-time growth curve and the historical curve are found at different time points. By analyzing these alignment points, a stage mapping relationship table can be created to record the corresponding stages of the real-time data and the historical data.

[0101] The growth curve is divided into different growth stages, the corresponding error value is calculated for each stage, and the errors of all stages are integrated to calculate the overall error value;

[0102] The error adjacency matrix is ​​used to quantify the influence relationship between various growth parameters. There may be interdependence between multiple growth parameters such as root length, stem length, and leaf area, which can be represented by the error adjacency matrix. The error adjacency matrix reflects the error propagation relationship between growth parameters. The higher the value, the stronger the error correlation between the two parameters.

[0103] See Figure 5 As shown in the figure, based on the historical germination database, the LSTM neural network model is trained to build a growth deviation analysis model. Based on the calculated error values, the growth deviation factors of each sprouted food are analyzed and obtained, including:

[0104] Convert the time series data in the historical database into a three-dimensional tensor structure, including the number of germination batches, time steps, and feature dimensions;

[0105] Define error labels based on various influencing factors and quantify evaluation criteria;

[0106] Design an LSTM-attention hybrid architecture through the input layer, bidirectional LSTM layer, temporal attention layer, and feature attention layer for model training;

[0107] Based on the trained growth deviation analysis model, various error values ​​obtained by calculation are input to obtain the growth deviation factors of each sprouted food;

[0108] By calculating the contribution score of each time step and feature dimension, the key impact period of various influencing factors can be obtained;

[0109] The LSTM hidden layer state and growth parameters are used as nodes, and the influence direction is used as the edge to construct an impact causal map. It is updated in real time based on the results of the growth deviation analysis model. When a new type of deviation is determined, a new deviation pattern node is automatically created and the key identification features are marked.

[0110] Specifically, in order to evaluate the performance of the model, we define error labels. The quantitative evaluation criteria of error labels can be classified according to the error size. Based on different error ranges, we can assign a deviation level to each data point.

[0111] Use the trained LSTM-attention model to predict real-time growth data and calculate deviations. Deviations are obtained by calculating the error labels for each growth cycle. Based on the time attention coefficient and feature attention coefficient, the contribution of each time step and each feature dimension to the growth deviation can be calculated. By analyzing the contribution scores, the time periods and features that have a greater impact on growth deviation can be identified.

[0112] Using LSTM hidden states and growth parameters as nodes and influence directions as edges, we construct a causal impact graph. Each node represents a growth parameter or hidden state, and edges represent the mutual influence between different parameters. When the model detects a new deviation type, it automatically creates a new deviation pattern node and annotates key identifying features. By tracking changes in different growth cycles, the causal graph can be updated in real time.

[0113] See Figure 6 As shown, based on the acquired growth deviation factors, the control instructions of the factors affecting the germination process are generated. By dynamically adjusting the factors affecting the germination, the germination process is controlled specifically including:

[0114] Based on the causal map, a three-level regulatory rule base is constructed, including single-factor linear compensation, multi-factor coupling regulation, and cross-stage predictive regulation;

[0115] Based on the two-dimensional evaluation model of urgency and impact, a control priority matrix was constructed to regulate the germination process under different circumstances.

[0116] Dynamic control strategy adjustments are made based on the real-time updated impact causal map.

[0117] Specifically, single-factor linear compensation is applicable to the independent adjustment of each growth parameter, and is usually used to compensate when certain parameters exceed the expected range. Multi-factor coupling regulation considers the interaction between multiple factors and is applicable to situations where multiple related growth parameters need to be adjusted simultaneously. Cross-stage predictive regulation considers the impact and parameter changes between different growth stages, and makes advance adjustments to future stages based on historical data and prediction models.

[0118] Based on the two-dimensional assessment of urgency and impact, a regulatory priority matrix is ​​constructed to sort the priorities at different time points. Based on the sorting results, the most urgent and impactful deviations are adjusted;

[0119] Based on real-time growth data and adjusted growth parameters, the feedback mechanism updates the causal map and dynamically adjusts the control rules. Using real-time data and the latest priority matrix, the control strategy is adjusted to cope with the new growth stage.

[0120] Furthermore, the method according to the embodiment of the present application can also be used with the aid of Figure 7 The electronic device architecture shown in FIG. Figure 7 As shown, the electronic device 500 may include a bus 501, one or more CPUs 502, a read-only memory (ROM) 503, a random access memory (RAM) 504, a communication port 505 connected to a network, an input / output component 506, a hard disk 507, etc. The storage device in the electronic device 500, such as the ROM 503 or the hard disk 507, may store a method and system for controlling the sprouting process of a sprouted food provided in the present application. The electronic device 500 may also include a terminal interface 508. Of course, Figure 7 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 7 One or more components of an electronic device are shown.

[0121] Figure 8 This is a schematic diagram of the computer-readable storage medium structure provided by an embodiment of the present application. Figure 81 shows a computer-readable storage medium 600 according to one embodiment of the present application. Computer-readable instructions are stored on the computer-readable storage medium 600. When the computer-readable instructions are executed by a processor, the sprouting process control method and system for sprouted food according to the embodiment of the present application described with reference to the above figures can be executed. The storage medium 600 includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory. Non-volatile memory may include, for example, read-only memory (ROM), a hard disk, flash memory, etc.

[0122] It should be noted that the order in which the embodiments of the present invention are described above is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0123] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on the differences from other embodiments.

[0124] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for controlling the germination process of germinated food, characterized in that: include: Based on the growth status of sprouted foods at each growth stage, a historical sprouting database is constructed to store reference growth data of sprouted foods at each growth stage; Collect growth data of sprouted food at each stage in real time and obtain the change rate of each growth parameter of sprouted food; Based on the acquired real-time sprouted food growth data, the data is compared with the historical sprout data of the corresponding growth period in the historical sprout database, and the error value between the real-time growth data and the reference data is calculated; Based on the historical germination database, the LSTM neural network model was trained to build a growth deviation analysis model. Based on the calculated error values, the growth deviation factors of each sprouted food were analyzed and obtained. Based on the acquired growth deviation factors, control instructions for factors affecting the germination process are generated, and the germination process is controlled by dynamically adjusting the germination influencing factors.

2. The method for controlling the germination process of a germinated food according to claim 1, wherein: The method of constructing a historical germination database based on the growth conditions of the sprouted food at each growth stage and storing reference growth data of the sprouted food at each growth stage specifically includes: The growth cycle of sprouted food is divided into germination period, elongation period, leaf expansion period and maturity period, and the growth cycle is identified by the morphology of sprouted food; Morphological dynamics feature extraction, environmental response feature extraction and metabolic marker extraction were performed based on the sprouted foods in each growth cycle; The acquired parameter data are stored in a historical germination database and classified based on the growth cycle of the germinated food.

3. The method for controlling the germination process of a germinated food according to claim 1, wherein: The real-time collection of growth data of the sprouted food at each stage and obtaining the change rate of each growth parameter of the sprouted food specifically includes: Use near-infrared cameras to detect the growth of radicles and plumules; Biomass is measured using high-precision strain gauge sensors, and the interference of irrigation water weight is eliminated through dynamic weighing compensation algorithms; Use the camera to shoot the bud crown at a low angle and obtain leaf expansion data in real time; Based on different sampling periods for each growth cycle, the rate of change of each growth parameter is calculated by differential method, and the growth turning point is identified by the second-order derivative to judge the growth cycle.

4. The method for controlling the germination process of a germinated food according to claim 1, wherein: The method of comparing the acquired real-time sprouted food growth data with the historical sprout data of the corresponding growth period in the historical sprout database and calculating the error value between the real-time growth data and the reference data specifically includes: Generate real-time growth curves from the growth data of various sprouted foods, align them with the historical benchmark curves in the historical sprout database through a dynamic time warping algorithm, and establish a stage mapping relationship table; Based on the curves of each stage of various types of growth data, the stratified error index is calculated and the error value is obtained; Based on the obtained error values, an error adjacency matrix is ​​constructed to quantify the influence relationship between various parameters; Based on the obtained error values ​​of various parameters and the quantitative relationship obtained from the error adjacency matrix, a comprehensive error value is calculated to obtain the true error value between the real-time growth data and the reference data.

5. The method for controlling the germination process of a germinated food according to claim 1, wherein: The method is based on the historical germination database and is trained by the LSTM neural network model to construct a growth deviation analysis model. Based on the calculated error value, the growth deviation factors of each sprouted food are analyzed and obtained, specifically including: Convert the time series data in the historical database into a three-dimensional tensor structure, including the number of germination batches, time steps, and feature dimensions; Define error labels based on various influencing factors and quantify evaluation criteria; Design an LSTM-attention hybrid architecture through the input layer, bidirectional LSTM layer, temporal attention layer, and feature attention layer for model training; Based on the trained growth deviation analysis model, various error values ​​obtained by calculation are input to obtain the growth deviation factors of each sprouted food; By calculating the contribution score of each time step and feature dimension, the key impact period of various influencing factors can be obtained; The LSTM hidden layer state and growth parameters are used as nodes, and the influence direction is used as the edge to construct an impact causal map. It is updated in real time based on the results of the growth deviation analysis model. When a new type of deviation is determined, a new deviation pattern node is automatically created and the key identification features are marked.

6. The method for controlling the germination process of a germinated food according to claim 1, wherein: The generating of the control instructions of the factors affecting the germination process based on the acquired growth deviation factors, and controlling the germination process by dynamically adjusting the factors affecting the germination process specifically include: Based on the causal map, a three-level regulatory rule base is constructed, including single-factor linear compensation, multi-factor coupling regulation, and cross-stage predictive regulation; Based on the two-dimensional evaluation model of urgency and impact, a control priority matrix was constructed to regulate the germination process under different circumstances. Dynamic control strategy adjustments are made based on the real-time updated impact causal map.

7. In combination with a sprouting process control system of a sprouted food, it is used to implement a sprouting process control method of a sprouted food according to any one of claims 1 to 6, characterized in that: include: Historical germination database module: The historical germination database module is responsible for storing and managing reference growth data of sprouted foods at various growth stages, and supports growth cycle and stage classification; Data acquisition module: The data acquisition module is used to collect real-time growth data of sprouted food at various stages, such as radicle length, embryo height, biomass, leaf expansion, etc., and calculate the rate of change; Error calculation module: The error calculation module is used to compare the real-time collected growth data with the reference data in the historical database, calculate the error value, and generate a growth curve; Model training module: The model training module is used to train historical data through LSTM neural network, build a growth deviation analysis model, and analyze deviation factors; Graph construction module: The graph construction module is used to construct and update the causal graph of the growth process based on the analysis results of the LSTM model; Regulation and control module: The regulation and control module is used to generate regulation instructions according to growth deviation factors and dynamically adjust the factors affecting the germination process; Processor: The processor is mainly used for the calculation process of each formula and the construction calculation process of each model.

8. An electronic device, characterized in that: include: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method for controlling a sprouting process of a sprouted food according to any one of claims 1 to 6.

9. A computer-readable storage medium storing computer-readable instructions, characterized in that: When the computer-readable instructions are executed by a processor, a method for controlling a sprouting process of a sprouted food according to any one of claims 1 to 6 is implemented.

Images