Multi-source sensing fusion agricultural monitoring method and system

Abstract

The invention relates to the technical field of agricultural information perception and decision making, in particular to a multi-source perception fused agricultural monitoring method and system. The method comprises the following steps: converting a multi-source heterogeneous agricultural sensing signal into a space-time tensor, constructing a semantic resonance field to simulate nonlinear coupling between modals, and driving multi-modal data to adaptively aggregate by using a gravitational evolution mechanism to form a fused semantic field; calculating non-linear response to generate an agricultural state emergence index, and according to the index, identifying a potential risk area and constructing a binary risk map; for the risk area, semantic disturbance is mapped into an agricultural variable disturbance vector through a modal decoupling matrix, a minimum intervention strategy is generated in combination with sparse optimization of an operation response matrix, and feasible operation suggestions are output after verification of an agricultural knowledge graph; a drift potential energy function is constructed based on strategy execution feedback, strategy parameters are dynamically updated through gradient descent, and closed-loop self-evolution optimization is achieved in combination with trend prediction. According to the invention, full-link adaptive optimization from multi-source sensing to regulation and control decision is realized.

Description

Technical Field

[0001] The present invention relates to the field of agricultural information perception and decision-making technology, and in particular to an agricultural monitoring method and system for multi-source perception fusion. Background Art

[0002] The current agricultural monitoring field is facing the technological evolution needs of the continuous expansion of the scale of multi-source heterogeneous perception data and the deepening of complex interactive characteristics. The dynamic coupling mechanism of multimodal data such as meteorology, soil moisture, and crop physiology in the temporal and spatial dimensions urgently needs to be quantitatively modeled to support accurate decision-making.

[0003] The Chinese invention patent with announcement number CN118709789B discloses a crop growth prediction method and system based on artificial intelligence and crop growth models, including: constructing a three-dimensional perception network for farmland and collecting multi-source heterogeneous data, performing preprocessing, semantic alignment, fusion, generating initial feature representation, performing dimensionality reduction embedding, and obtaining a comprehensive feature representation; constructing the key concepts, relationships, and attributes of the crop growth domain ontology, obtaining crop growth process knowledge and constructing a crop growth knowledge graph, embedding the crop growth process knowledge into the crop growth knowledge graph, constructing a crop growth simulator, and for each crop, identifying key environmental factors and updating the crop growth knowledge graph; constructing a crop growth prediction model, extracting latent variable representation, performing structural optimization, adding the comprehensive feature representation to the crop growth prediction model, obtaining the yield prediction probability distribution, and generating a causal exploration path diagram.

[0004] As smart agriculture develops towards real-time and adaptive directions, the technical requirements of agricultural production systems for early identification of hidden risks, dynamic resource regulation and closed-loop optimization of strategies have increased significantly. It is urgent to build a full-link fusion framework from multi-source perception to intelligent decision-making through innovative methods, so as to achieve a breakthrough improvement in resource utilization efficiency while ensuring system stability. Summary of the Invention

[0005] The purpose of the present invention is to address the problems existing in the background technology and propose an agricultural monitoring method and system with multi-source perception fusion.

[0006] The technical solution of the present invention is a multi-source sensing fusion agricultural monitoring method, which includes the following specific implementation steps: S1. By normalizing and mapping multi-source heterogeneous agricultural perception signals into spatiotemporal semantic fourth-order tensors, an agricultural semantic embedding field is constructed to identify potential event excitation sources, and a semantic resonance field is defined to simulate nonlinear interactions between modalities, driving the generation of coupling strength tensors. S2. Construct a dynamic gravitational tensor field based on modal semantic differences and resonance strength, calculate the net force acting on each modality, and iteratively update the tensor state of the semantic particles in the direction of multimodal consensus. Define a convergence judgment indicator and output the fusion result when the average change in the norm is lower than the convergence threshold. S3. By introducing structured perturbations into the semantic tensor field to simulate external stimuli, the normalized response intensity is calculated to measure the degree of change after the perturbation, which is converted into an agricultural state emergence index with enhanced spatial consistency. Based on the threshold, a risk binary map is constructed to identify potential mutation areas. S4. After detecting the risk area, the target steady-state semantic tensor is constructed, the difference is calculated to form a semantic disturbance tensor, which is mapped into a variable disturbance vector through the modal decoupling matrix. Sparse optimization is used to generate the minimum intervention operation vector, and consistency verification is performed through the agricultural knowledge graph to ensure feasibility and agronomic consistency; S5. After executing the control suggestions, the feedback difference tensor is collected, and the drift potential function is constructed to drive the gradient descent to update the strategy parameter tensor. The target state is dynamically corrected by combining the memory averaging mechanism and trend prediction, and the self-evolution controller is encapsulated to realize closed-loop iterative optimization.

[0007] Preferably, the coupling strength tensor construction process is: The multi-source heterogeneous agricultural sensing data is standardized and mapped to generate a fourth-order modal tensor containing the dimensions of channel number, time length, spatial height and width; By weighted fusion of semantic feature vectors of each modal tensor slice, a three-dimensional agricultural semantic vector field is constructed. The L2 gradient norm of the semantic state vector is calculated and its extreme value is found to locate the excitation source of agricultural events. The full-field semantic resonance intensity field is defined by weighting the Frobenius norm deviation of each modal tensor from its mean through the modal response coefficient. The absolute value of the mixed partial derivative of its function with respect to the dual-modal tensor measures the semantic resonance coupling strength between the modalities, that is, constructing the coupling strength tensor.

[0008] Preferably, the iterative update process for iteratively updating the tensor state is: A dynamic gravitational field is constructed based on modal semantic differences and resonance strength, which is inversely proportional to the square of the modal tensor difference, driving the energy flow between modalities for adaptive aggregation; By calculating the net force tensor of the gravitational forces exerted on any modality by other modalities, its semantic tensor is driven to iteratively update in the direction of multimodal consensus with an update amplitude, thus achieving cross-modal collaborative fusion and semantic consistency convergence: ; ; in, represents the resultant force of all other modal attractions on modal j at (x, y, t), that is, the sum of all semantic guidance effects from other modalities; Represents the semantic tensor of modality j at this spatiotemporal position; Indicates the magnitude of a semantic tensor update; Represents the semantic gravitational vector exerted by the i-th modality on the j-th modality at the three-dimensional space-time coordinate (x, y, t); (x, y) is the spatial position; t is the timestamp.

[0009] Preferably, the convergence judgment index is: ; If satisfied: , it is considered that the modal aggregation has reached a stable state and the final fusion tensor is output; in, represents the average change amplitude of mode j during the k-th evolution process; Represents the tensor state of mode j after the kth iteration; Represents the tensor state at the previous iteration; N is the total number of all (x, y, t) spatiotemporal positions involved in the fusion; Indicates the set convergence threshold.

[0010] Preferably, the risk binary map construction process is: By superimposing a structured perturbation term simulating external stimulation on the fused semantic tensor field, the perturbed modal semantic tensor is constructed: ; in, Represents the semantic tensor of modality j after adding perturbation; Represents the original fused semantic tensor of modality j at position (x, y) and time t; represents the disturbance term; The structural response strength is quantified by calculating the Frobenius norm difference of the semantic tensor before and after perturbation and performing unit perturbation normalization: ; in, represents the response amplitude tensor, that is, the structural response intensity of mode j to the disturbance at point (x, y, t); represents the intensity of the disturbance term itself; represents the stability factor; The modal weights are used to weight the amplitude components of each response, and the spatial Laplace operator is used to enhance the aggregation characteristics to generate the agricultural state emergence index: ; in, represents the agricultural status emergence index; represents the modal weight; represents the adjustment coefficient of the Laplace term; represents the spatial Laplacian operator; Construct a binary risk map based on the emergence index map : ; in, Represents the agricultural status risk marker map, marking whether the point (x, y, t) is a potential key emergence area, 1 means an abnormality is detected, and 0 means normal; Represents the emergence index threshold.

[0011] Preferably, the variable disturbance vector generation process is: When a state emergence is detected in a certain area, the expected value of each modal semantic tensor in the normal area with a risk mark of zero is calculated through the conditional expectation operator, and the target steady-state semantic tensor is constructed to define the ideal agricultural state: ; in, The semantic tensor representing the desired agricultural system, i.e., at the spatial location (x, y) and time point t; represents the conditional expectation operator; Indicates the spatial position of the jth mode , the semantic tensor at time point t; A binary value representing the agricultural status risk marker map; The difference between the target state tensor and the current actual state tensor is calculated to construct the modal semantic disturbance tensor, which is mapped into a variable disturbance vector through the modal decoupling matrix trained based on agricultural domain knowledge and historical data: ; ; in, represents the semantic perturbation tensor under modality j; Representing the modal decoupling matrix, mapping the modal semantic tensor space to the specific agricultural control variable space, based on agricultural domain knowledge and historical data training, characterizing the correspondence between each dimension of modal semantics and agricultural variables; represents the disturbance vector of agricultural variables, that is, the agricultural variable value that should be adjusted for mode j at point (x, y, t), that is, the disturbance response value of the p agricultural variables associated with mode j.

[0012] Preferably, the verification process for consistency verification through the agricultural knowledge graph is: By mapping the operation variable vector through the agricultural operation response matrix, the optimization problem of minimizing the sum of squared response errors and the L1 norm regularization of the operation vector is solved to generate the minimum intervention strategy operation vector with sparsity constraints: ; in, represents the agricultural operation vector; It represents the agricultural operation response matrix, describing the impact intensity of each operation on agricultural variables; represents the vector of operation variables; represents the regularization parameter, which controls the weight of the sparse constraint term; Represents the 1-norm of the operation vector; Consistency verification is performed through the agricultural knowledge graph, and judgment functions are used to screen whether the operation plan complies with agricultural regulations: ; in, Represents the agricultural knowledge graph; Indicates final agricultural practice recommendations after consistency verification and revision; Used to determine whether the operation plan complies with agricultural knowledge and standards; Indicates adjustments to unreasonable strategies.

[0013] Preferably, the updating process of the policy parameter tensor is: According to regulatory recommendations After execution, during the delay Then collect new perception tensors Compare and build the feedback difference tensor: ; in, Indicates that during policy execution The multi-source perception tensor is then recaptured; represents the feedback difference tensor; Construct the state drift potential energy function: ; ; in, represents the stability score of the agricultural system at spatial location x, y and time t; Indicates the tensor fluctuation amplitude of the specified area in the past K2 time units; K2 represents the length of the time window used to calculate historical fluctuations; Represents the multi-source perception semantic tensor at the spatial position (x, y) at the time point tk; represents the average tensor state within the time window [t-K2, t-1]; Represents the state drift potential energy function value; represents the stability adjustment coefficient; The potential energy calculated in the previous step is used to adjust the next round of strategy output tensor, and the strategy parameters are expressed as tensors , the following response gradient descent is used: ; in, represents the policy parameter tensor, i.e., the control model weights or feature map parameters used to generate recommendations at the t-th round of decision making; represents the control model weight used to generate recommendations in the t+1th round of decision making; represents the learning rate; represents the gradient of the drift potential with respect to the policy parameters.

[0014] Preferably, the correction process of dynamically correcting the target state is: The memory averaging mechanism uses a decay factor to fuse the new strategy weights with historical weights to avoid oscillation. At the same time, the trend prediction module uses historical observation tensor sequence modeling to model multi-channel coupling characteristics, dynamically decode and predict future target states to achieve dynamic correction: ; ; in, represents the strategy memory decay factor; represents the time trend prediction function; represents the observation sequence within the time window; Indicates the target agricultural status of a new round.

[0015] The technical solution of the present invention is a multi-source sensing fusion agricultural monitoring system, which is used to execute the above-mentioned multi-source sensing fusion agricultural monitoring method, including: The multi-source perception semantic tensor construction module is responsible for receiving and fusing multi-source heterogeneous data, and then processing it through spatial and temporal fusion and resonance field initialization to construct a high-dimensional semantic tensor, completing the deep semantic expression of agricultural environment and crop growth information. The multimodal aggregation module is used to analyze the coupling of multimodal semantic tensors. By building a gravitational tensor model between devices, it organically aggregates different modal and time series data. The agricultural status emergence detection module is used to dynamically analyze aggregated agricultural status data, identify potential anomalies, emergent behaviors, and nonlinear evolution characteristics, and provide a scientific basis for risk warning and status diagnosis; The agricultural regulation suggestion generation module is used to automatically generate targeted agricultural regulation strategies and operational suggestions based on the detected agricultural status changes and emergent features; The agricultural strategy dynamic feedback and self-evolution module is used to combine multi-source perception data after the execution of control measures to construct feedback difference tensors and state drift potential energy, and dynamically adjust and optimize agricultural strategies by combining policy gradient updates and trend predictions.

[0016] Compared with the prior art, the above technical solution of the present invention has the following beneficial technical effects: The present invention designs an agricultural monitoring method and system with multi-source perception fusion. By constructing a multimodal data fusion mechanism driven by a physical model, the processing efficiency and accuracy of agricultural perception data are significantly improved: first, the nonlinear coupling strength between modes is quantified by using spatiotemporal tensor modeling and semantic resonance field, and gravitational evolution is combined to drive the adaptive aggregation of multi-source data, overcoming the shortcomings of traditional methods in modeling the interaction of heterogeneous data; second, an agricultural state emergence index is generated through a structured disturbance response mechanism to accurately locate potential risk areas such as disease outbreaks or drought centers, greatly improving the sensitivity and accuracy of mutation warnings; third, based on the modal decoupling matrix, the semantic disturbance is mapped to a variable disturbance vector, and the minimum intervention strategy is generated through sparse optimization of the operation response matrix, and combined with the agricultural knowledge graph verification to ensure the agronomic feasibility and safety of the control suggestions; finally, a closed-loop self-evolution system is constructed, and the drift potential energy is calculated based on the feedback difference tensor to drive the gradient update of the strategy parameters, and the historical memory and trend prediction are integrated to dynamically adjust the target state, realizing full-link adaptive optimization from perception to decision-making, thereby comprehensively improving the real-time performance, stability and resource regulation efficiency of the agricultural system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flowchart of a multi-source sensing fusion agricultural monitoring method proposed by the present invention; Figure 2 This is a system architecture diagram of a multi-source perception fusion agricultural monitoring system proposed in the present invention. DETAILED DESCRIPTION

[0018] Example 1, as Figure 1 As shown, the agricultural monitoring method of multi-source perception fusion proposed in the present invention includes the following specific implementation steps: S1. Extract tensor representations that can embed semantic structures from heterogeneous agricultural perception signals from different sources. By introducing the physical abstraction of semantic resonance fields, a dynamically driven data fusion preparation state is constructed to provide a unified tensor representation and semantic interaction field for subsequent coupled perception mechanisms. The specific implementation process is as follows: S11. Standardize and map heterogeneous agricultural perception signals into a fourth-order tensor with spatiotemporal and semantic distribution properties: The original data source set is denoted as: , each data source passes through the modal sensor Perform preliminary nested encoding to generate modal tensors : ; in, Represents the high-order tensor representation of the m-th modality perception data after embedding mapping; Represents the semantic embedding function corresponding to the mth modality, which is used to convert the original data into a standard tensor form; represents the original m-th agricultural perception data; M is the total number of modalities of agricultural perception data; C m represents the number of channels of the mth modal tensor; T represents the length of the time dimension; H is the height dimension of the spatial dimension in the tensor, corresponding to the vertical division unit in the agricultural area grid map; W is the width dimension of the spatial dimension in the tensor, corresponding to the horizontal division unit in the grid map; S12. Constructing agricultural semantic embedding field based on tensor mapping results , which describes the semantic expression state at different times and geographical locations. The source points of potential important agricultural events are identified through the gradient evolution of this field: ; ; in, express Represents the semantic state vector of the coordinate point (x, y, t) in the three-dimensional semantic vector field, reflecting the agricultural meaning of the point at the current spatiotemporal position (e.g., "humidity is too high" or "crops are growing well"). represents a function that maps a slice of the mth modal tensor (i.e., the perceptual data at point (x, y, t)) to a semantic feature vector; Representing a tensor The slice at the coordinate point (x, y, t), that is, all channel data at this point under this mode (dimension: C m ); represents the semantic fusion weight matrix of the mth modality. In this embodiment, the mutual information adaptive estimation is used; The spatial location and time point of the semantic excitation source point, that is, the point where the semantic field changes most dramatically in the agricultural area, may represent the point where a disease is about to break out, the center of drought development, etc. The gradient vector representing the semantic vector field, i.e., the rate of change of semantics in local space and time (i.e., semantic volatility); represents the vector norm. In this embodiment, the L2 norm (Euclidean length) is used to measure the rate of semantic change. Indicates finding the value of the variable that maximizes the expression; (x, y) is the spatial position; t is the timestamp; S13. Define a semantic resonance field to simulate the nonlinear interaction between perceptual data due to semantic proximity (similar to the phase resonance phenomenon in physics). This field provides the driving force for the subsequent self-aggregation of multimodal tensors. Specifically: Define the resonance field: ; To measure the semantic resonance coupling strength between modes i and j, a coupling strength tensor is constructed: ; in, Represents the semantic resonance intensity field value generated by all modal perception data at the point (x, y, t), which is used to represent the excitation energy that produces coupled resonance behavior between perceptions; represents the modal response coefficient, i.e., the degree of participation (response strength) of the mth mode in the semantic resonance behavior. In this embodiment, it is obtained by learning the modal volatility and coupled empirical entropy in historical data; Represents the Frobenius norm, that is, the channel norm of the tensor at this slice position, which is used to measure the difference from the mean; Represents the average tensor value of all modes at this point, which is used as the resonance reference center; represents the semantic resonance coupling strength between mode i and mode j at point (x, y, t); Based on this: a tensor network of inter-modal response relations is constructed through tensor partial derivatives, and coupled physical modeling is introduced into the agricultural semantic perception system.

[0019] S2. Obtaining multi-source perception semantic tensors After determining the semantic coupling strength between them, a dynamic evolution mechanism is constructed. With the help of the analogy model of physical gravitational evolution, the adaptive aggregation between modalities is driven and a coupled semantic structure is constructed to achieve efficient integrated modeling of agricultural multimodal perception information. The specific implementation process is as follows: S21. Construct a tensor gravitational effect determined by modal semantic differences and resonance strength, and construct an evolvable coupled gravitational tensor field in the space-time-semantic domain. This gravitational field acts on the tensor space position and its internal semantic distribution structure, thereby forming a dynamic energy flow between modalities. Specifically: ; in, Represents the semantic gravitational force vector (tensor) exerted by the i-th modality on the j-th modality at the three-dimensional space-time coordinate (x, y, t), determining whether the semantic tensor of modality j should be attracted to modality i. The larger the gravitational force value, the stronger the willingness to merge. Represents the semantic attraction coefficient between modalities, a global adjustment factor that controls the magnitude of semantic attraction; Indicates the stability factor to prevent numerical instability caused by division by zero or too small a value. In this embodiment, it is set to 1e-6; S22. Construct a gravity-driven tensor behavior system, modeling each modality as a semantic particle. Its fusion behavior is dynamically regulated by a global coupled tensor field, and provides an evolutionary cross-modal collaboration mechanism: Define the total gravitational field force tensor for mode j: ; Define the status update formula: ; in, represents the net force of all other modal attractions on modality j at (x, y, t), that is, the sum of all semantic guidance effects from other modalities, which is used to guide the modality to move closer to the multimodal consensus; The semantic tensor representing modality j at that spatiotemporal position gradually moves closer to the consensus direction of other modalities after each update, achieving iterative convergence of semantic consistency; Indicates the amplitude of a semantic tensor update, controlling the influence of the gravity tensor on the modal tensor; S23. Define convergence criteria: ; If satisfied: , it is considered that the modal aggregation has reached a stable state and the final fusion tensor is output; in, It represents the average change amplitude of mode j in the k-th evolution process, which is a key measure to measure whether the fusion is stable. If the change is very small, it means that the mode state is basically stable and the fusion can be considered complete. Represents the tensor state of mode j after the kth iteration; Represents the tensor state at the previous iteration. The difference between the two is used to determine whether there is sufficient change. N is the total number of all (x, y, t) spatiotemporal positions involved in the fusion. Indicates the set convergence threshold.

[0020] S3. By introducing a structured perturbation term into the semantic tensor field and calculating the nonlinear response of the tensor in the spatiotemporal structure before and after the perturbation, we can identify latent regions that are extremely sensitive to the perturbation. These regions are potential mutation points or critical evolution sources of agricultural trends. The specific implementation process is as follows: S31, the fused semantic tensor field outputted in step S2 (Mode j) performs disturbance construction to simulate weak external stimuli that agricultural ecosystems may face, such as microclimate changes, water fluctuations, and pathogen signal infiltration: Construct the perturbation tensor: ; in, Represents the semantic tensor of the mode j after the perturbation, that is, the structural state of the mode after the perturbation at the space-time point (x, y, t); Represents the original fused semantic tensor of modality j (such as image, weather, soil, etc.) at position (x, y) and time t; represents the disturbance term, corresponding to the structural disturbance of mode j at that point, simulating minor changes in climate, humidity, pathogens, etc. S32. After the perturbation is introduced, the response strength of the system state in the semantic space is focused on. The specific measurement is the unit perturbation normalization result of the tensor difference before and after the perturbation: ; in, represents the response amplitude tensor, that is, the structural response strength of mode j to the disturbance at point (x, y, t), that is, the degree to which the semantic structure changes after being disturbed; Indicates the intensity of the disturbance term itself, and is used as a normalization term to eliminate the influence of different disturbance amplitudes on the response value; Indicates a stability factor (such as 1e-6), which is used to prevent instability caused by the denominator being 0 or too small; S33. Introducing spatial consistency analysis, the high response points in the response amplitude field are further converted into emergence indices with the possibility of evolutionary mutations for spatial clustering assessment: ; in, It represents the agricultural state emergence index, that is, whether the point has the possibility of mutation in the multimodal disturbance response field, and is a measure of the potential critical state of the system; represents the modality weight, reflecting the relative importance of modality j in the judgment of agricultural status; represents the adjustment coefficient of the Laplace term, controlling the effect of spatial clustering on the overall emergence index; represents the spatial Laplacian operator; S34. Construct a binary risk map based on the emergence index map , used for subsequent decision-making in agricultural systems: ; in, Represents the agricultural status risk labeling map, a binary output map, marking whether the point is a potential key emergence area in the agricultural system; Represents the emergence index threshold, which is used to determine whether there is a critical state mutation.

[0021] S4. Construct a mechanism for generating agricultural control suggestions based on state emergence. By guiding variable inversion through semantic perturbation and combining modal decoupling and sparse optimization, a reversible mapping mechanism from agricultural state emergence to operational strategies is constructed. Furthermore, knowledge graph correction is used to ensure the feasibility and agronomic consistency of control suggestions, thus achieving dynamic, adaptive, and precise control strategy generation in the smart agricultural environment. The specific implementation process is as follows: S41, when the system detects that a certain region (x, y) exists at time t (i.e. =1), construct the desired state tensor field to define the target steady state: ; in, represents the desired semantic tensor of the agricultural system, i.e., the ideal steady-state modal semantic state of the agricultural system at the spatial location (x, y) and time point t; represents the conditional expectation operator; Indicates the spatial position of the jth mode , the semantic tensor at time point t; It represents the binary value of the agricultural status risk marker map, where 1 indicates that an abnormality is detected and 0 indicates that it is normal; S42, according to the obtained target state tensor and the current actual state tensor , construct the semantic perturbation tensor of the modal layer: ; This tensor represents the deviation of the system's current perceived state from the desired steady state in mode j; in, represents the semantic perturbation tensor under modality j, that is, the difference between the current state and the target steady state; For example, in image mode, it can reflect vegetation color changes, texture density fluctuations, etc.; in meteorological mode, it may indicate humidity drop or abnormal temperature difference. Introducing the modal decoupling matrix , mapping the semantic perturbation tensor to the variable perturbation vector: ; in, Representing the modal decoupling matrix, mapping the modal semantic tensor space to the specific agricultural control variable space, based on agricultural domain knowledge and historical data training, characterizing the correspondence between each dimension of modal semantics and agricultural variables; represents the disturbance vector of agricultural variables, i.e., the agricultural variable value that should be adjusted for mode j at point (x, y, t), i.e., the disturbance response value of the p agricultural variables associated with mode j; It should be noted that the modal decoupling matrix This learning is achieved by constructing a large-scale annotated sample set between multi-source sensory data and corresponding agricultural variables: first, the semantic tensor output of each modality and the actual measured control variable values ​​are simultaneously collected in experimental fields or greenhouses. Combined with the prior feature association information provided by agronomic experts, data-driven methods such as multivariate regression or tensor decomposition are used to fit the mapping relationship between sensory features and variable changes. On this basis, through cross-validation and calibration of the domain knowledge graph, the matrix parameters are continuously optimized to ensure that it can accurately reflect the contribution of each modality to different control variables while also having good generalization and anti-interference capabilities. Ultimately, a modal decoupling matrix is ​​formed that not only conforms to the laws of physical agronomy but also can be efficiently calculated in practical systems. S43. Introducing the Agricultural Operation Response Matrix , calculate the minimum intervention strategy by optimization: ; in, represents the agricultural operation vector, which represents the agricultural intervention operation to be performed (such as fertilizer application amount, irrigation time, etc.); Represents the agricultural operation response matrix, which describes the impact of each operation on agricultural variables. It is constructed based on historical agricultural experimental data, physical models or expert experience and serves as a mapping bridge between control variables and operations. represents the vector of manipulated variables, which contains the intensities or parameters of all possible agricultural interventions; represents the regularization parameter, which controls the weight of the sparse constraint term; Represents the 1-norm of the operation vector, that is, the sum of the absolute values ​​of the operation intensities; S44, after preliminarily obtaining the operation vector After that, you need to go through the agricultural knowledge graph To perform consistency verification: ; in, Represents the agricultural knowledge graph; Indicates final agricultural practice recommendations after consistency verification and revision; Used to determine whether the operation plan complies with agricultural knowledge and standards, that is, to screen for non-compliant or conflicting strategies; It means adjusting unreasonable strategies to meet agricultural standards within the minimum change range; It should be noted that the agricultural knowledge graph The entity set is the core element in the agricultural knowledge graph, including but not limited to crop type entities, soil type entities, climate environment entities, agricultural disease entities, agricultural operation entities and agricultural variable entities; the relationship set describes the semantic relationship between entities, such as <yellow rice leaves, need operation, topdressing nitrogen fertilizer>; the attribute set defines numerical or enumeration attribute labels for entities and relationships.

[0022] S5. Build a dynamic self-evolution method for agricultural strategies based on policy feedback tensor construction, drift potential energy estimation, and response gradient updating. By building a closed-loop perception-control-response system, the agricultural strategy can be adaptively adjusted and evolved over the long term based on actual perception deviations, thereby improving the intelligence and stability of the system. The specific implementation process is as follows: S51. According to the regulation recommendations After execution, during the delay Then collect new perception tensors Compare and build the feedback difference tensor: ,This difference reflects whether the system responds as expected after the policy is executed; in, Indicates that during policy execution The multi-source perception tensor that is then recollected, i.e., the changes in the real agricultural environment; Represents the feedback difference tensor, that is, the difference between the actual effect after the strategy is executed and the expected target; S52. Construct a state drift potential energy function: ; ; in, represents the stability score of the agricultural system at spatial location x, y and time t; Indicates the tensor fluctuation amplitude (standard deviation approximation) of the specified area in the past K2 time units; K2 represents the length of the time window used to calculate historical fluctuations; It represents the multi-source perception semantic tensor at the spatial position (x, y) at the time point tk, which contains information features such as crop growth, physiology, meteorology, and soil; represents the average tensor state within the time window [t-K2, t-1]; Represents the state drift potential energy function value, which measures the degree of damage to the system stability caused by the current feedback difference; Represents the stability adjustment coefficient, which is used to balance the weight of feedback energy and system stability; S53. Use the potential energy calculated in the previous step to adjust the next round of strategy output tensor, and let the strategy parameter be expressed as a tensor , the following response gradient descent is used: ; in, represents the policy parameter tensor, i.e., the control model weights or feature map parameters used to generate recommendations at the t-th round of decision making; represents the control model weights or feature map parameters used to generate recommendations in the t+1th round of decision making; Represents the learning rate, which is used to control the update step size of the policy weight to avoid oscillation; It represents the gradient of the drift potential energy with respect to the strategy parameters, reflecting which direction can maximize the reduction of the deviation potential energy; S54. Introduce a first-order memory averaging mechanism to avoid oscillation caused by response updates: ; At the same time, combined with the time window sliding trend prediction module , make regression corrections to the long-term trend deviation and adjust the system strategy target point: ; in, Represents the strategy memory decay factor, which controls the fusion ratio of historical weights and new weights; represents the time trend prediction function, which is used to predict the reasonable interval of future target state based on historical observations; Represents the observation sequence within the time window, which is used by the trend prediction module to speculate on state evolution; Indicates the target agricultural status of the new round, and is dynamically updated through trend forecasting; It should be noted that the time trend prediction function It is a state evolution modeling method based on sliding time series. Its core idea is to use historical observation tensor sequences of a certain time length to extract the potential temporal evolution trend of the agricultural system through nonlinear state regression or embedded tensor decoder models, and then predict the optimal expected state at the next moment or within several steps in the future. This module not only considers the smooth trend of numerical changes, but also combines the interactive coupling characteristics of multi-dimensional channels (such as soil moisture, temperature, pest and disease signals, etc.), and uses a tensor dynamic decoding mechanism to model multimodal evolution paths to ensure that the trend prediction results are robust and practically adaptable, thereby providing a sustainable target reference direction for dynamic strategy adjustment, playing a dual role of system foresight and evolutionary coordination. S55, encapsulate the above feedback, energy, gradient update, trend prediction and other modules into a self-evolution controller unit , and add the time iteration formula: ; in, Represents the state of the agricultural regulation self-evolution controller at the current time and space point (x, y); It represents the iterative function of the controller evolving from t to t+1, using the strategy network evolution function in reinforcement learning; represents the gradient of the potential energy function with respect to the policy parameters.

[0023] Example 2, as Figure 2 As shown, the present invention proposes a multi-source perception fusion agricultural monitoring system, which is applied to the multi-source perception fusion agricultural monitoring method proposed in Example 1, including: a multi-source perception semantic tensor construction module, a multimodal aggregation module, an agricultural state emergence detection module, an agricultural regulation suggestion generation module, and an agricultural strategy dynamic feedback and self-evolution module.

[0024] The multi-source perception semantic tensor construction module is responsible for receiving and fusing heterogeneous data from multiple sources, such as satellite remote sensing, ground sensors, and weather stations. It then processes these heterogeneous data through spatial and temporal fusion and resonance field initialization to construct a high-dimensional semantic tensor. This module provides deep semantic expression of agricultural environment and crop growth information, ensuring the richness and consistency of perception information. The multimodal aggregation module is used to analyze the coupling of multimodal semantic tensors. By constructing a gravitational tensor model between devices, it organically aggregates different modal and time series data, reveals the mutual influence and dynamic coupling relationship between multiple factors in the agricultural system, and provides accurate comprehensive status assessment. The agricultural status emergence detection module is used to dynamically analyze aggregated agricultural status data to identify potential anomalies, emergent behaviors, and nonlinear evolution characteristics, providing a scientific basis for risk warning and status diagnosis, and improving the sensitivity and accuracy of monitoring; The agricultural regulation suggestion generation module is used to automatically generate targeted agricultural regulation strategies and operational suggestions based on detected agricultural status changes and emergent features. These strategies cover irrigation, fertilization, pest and disease control, and other aspects, effectively guiding agricultural production practices, promoting healthy crop growth, and optimizing resource utilization. The agricultural strategy dynamic feedback and self-evolution module is used to combine multi-source perception data after the execution of control measures to construct feedback difference tensors and state drift potential energy, and combine strategy gradient updates and trend predictions to dynamically adjust and optimize agricultural strategies, achieve closed-loop adaptation and continuous evolution of the system, and enhance the intelligence level and long-term stability of the agricultural monitoring and control system.

[0025] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A multi-source perception fusion agricultural monitoring method, characterized in that: The specific implementation steps include the following: S1. By normalizing and mapping multi-source heterogeneous agricultural perception signals into spatiotemporal semantic fourth-order tensors, an agricultural semantic embedding field is constructed to identify potential event excitation sources, and a semantic resonance field is defined to simulate nonlinear interactions between modalities, driving the generation of coupling strength tensors. S2. Construct a dynamic gravitational tensor field based on modal semantic differences and resonance strength, calculate the net force acting on each modality, and iteratively update the tensor state of the semantic particles in the direction of multimodal consensus. Define a convergence judgment indicator and output the fusion result when the average change in the norm is lower than the convergence threshold. S3. By introducing structured perturbations into the semantic tensor field to simulate external stimuli, the normalized response intensity is calculated to measure the degree of change after the perturbation, which is converted into an agricultural state emergence index with enhanced spatial consistency. Based on the threshold, a risk binary map is constructed to identify potential mutation areas. S4. After detecting the risk area, the target steady-state semantic tensor is constructed, the difference is calculated to form a semantic disturbance tensor, which is mapped into a variable disturbance vector through the modal decoupling matrix. Sparse optimization is used to generate the minimum intervention operation vector, and consistency verification is performed through the agricultural knowledge graph to ensure feasibility and agronomic consistency; S5. After executing the control suggestions, the feedback difference tensor is collected, and the drift potential function is constructed to drive the gradient descent to update the strategy parameter tensor. The target state is dynamically corrected by combining the memory averaging mechanism and trend prediction, and the self-evolution controller is encapsulated to realize closed-loop iterative optimization.

2. The agricultural monitoring method based on multi-source perception fusion according to claim 1 is characterized in that: The coupling strength tensor construction process is: The multi-source heterogeneous agricultural sensing data is standardized and mapped to generate a fourth-order modal tensor containing the dimensions of channel number, time length, spatial height and width; By weighted fusion of semantic feature vectors of each modal tensor slice, a three-dimensional agricultural semantic vector field is constructed. The L2 gradient norm of the semantic state vector is calculated and its extreme value is found to locate the excitation source of agricultural events. The full-field semantic resonance intensity field is defined by weighting the Frobenius norm deviation of each modal tensor from its mean through the modal response coefficient. The absolute value of the mixed partial derivative of its function with respect to the dual-modal tensor measures the semantic resonance coupling strength between the modalities, that is, constructing the coupling strength tensor.

3. The agricultural monitoring method based on multi-source perception fusion according to claim 2 is characterized in that: The iterative update process of tensor state is: A dynamic gravitational field is constructed based on modal semantic differences and resonance strength, which is inversely proportional to the square of the modal tensor difference, driving the energy flow between modalities for adaptive aggregation; By calculating the net force tensor of the gravitational forces exerted on any modality by other modalities, its semantic tensor is driven to iteratively update in the direction of multimodal consensus with an update amplitude, thus achieving cross-modal collaborative fusion and semantic consistency convergence: ； ； in, represents the resultant force of all other modal attractions on modal j at (x, y, t), that is, the sum of all semantic guidance effects from other modalities; Represents the semantic tensor of modality j at this spatiotemporal position; Indicates the magnitude of a semantic tensor update; Represents the semantic gravitational vector exerted by the i-th modality on the j-th modality at the three-dimensional space-time coordinate (x, y, t); (x, y) is the spatial position; t is the timestamp.

4. The agricultural monitoring method based on multi-source perception fusion according to claim 3 is characterized in that: The convergence judgment index is: ； If satisfied: , it is considered that the modal aggregation has reached a stable state and the final fusion tensor is output; in, represents the average change amplitude of mode j during the k-th evolution process; Represents the tensor state of mode j after the kth iteration; Represents the tensor state at the previous iteration; N is the total number of all (x, y, t) spatiotemporal positions involved in the fusion; Indicates the set convergence threshold.

5. The agricultural monitoring method based on multi-source perception fusion according to claim 4 is characterized in that: The process of constructing the risk binary map is as follows: By superimposing a structured perturbation term simulating external stimulation on the fused semantic tensor field, the perturbed modal semantic tensor is constructed: ； in, Represents the semantic tensor of modality j after adding perturbation; Represents the original fused semantic tensor of modality j at position (x, y) and time t; represents the disturbance term; The structural response strength is quantified by calculating the Frobenius norm difference of the semantic tensor before and after perturbation and performing unit perturbation normalization: ； in, represents the response amplitude tensor, that is, the structural response intensity of mode j to the disturbance at point (x, y, t); represents the intensity of the disturbance term itself; represents the stability factor; The modal weights are used to weight the amplitude components of each response, and the spatial Laplace operator is used to enhance the aggregation characteristics to generate the agricultural state emergence index: ； in, represents the agricultural status emergence index; w j represents the modal weight; represents the adjustment coefficient of the Laplace term; represents the spatial Laplacian operator; Construct a binary risk map based on the emergence index map : ； in, Represents the agricultural status risk marker map, marking whether the point (x, y, t) is a potential key emergence area, 1 means an abnormality is detected, and 0 means normal; Represents the emergence index threshold.

6. The agricultural monitoring method of multi-source perception fusion according to claim 5 is characterized in that: The variable disturbance vector generation process is: When a state emergence is detected in a certain area, the expected value of each modal semantic tensor in the normal area with a risk mark of zero is calculated through the conditional expectation operator, and the target steady-state semantic tensor is constructed to define the ideal agricultural state: ； in, The semantic tensor representing the desired agricultural system, i.e., at the spatial location (x, y) and time point t; represents the conditional expectation operator; Indicates the spatial position of the jth mode , the semantic tensor at time point t; A binary map representing the agricultural status risk marker map; The difference between the target state tensor and the current actual state tensor is calculated to construct the modal semantic disturbance tensor, which is mapped into a variable disturbance vector through the modal decoupling matrix trained based on agricultural domain knowledge and historical data: ； ； in, represents the semantic perturbation tensor under modality j; Representing the modal decoupling matrix, mapping the modal semantic tensor space to the specific agricultural control variable space, based on agricultural domain knowledge and historical data training, characterizing the correspondence between each dimension of modal semantics and agricultural variables; represents the disturbance vector of agricultural variables, that is, the agricultural variable value that should be adjusted for mode j at point (x, y, t), that is, the disturbance response value of the p agricultural variables associated with mode j.

7. The agricultural monitoring method of multi-source perception fusion according to claim 6 is characterized in that: The verification process for consistency verification through the agricultural knowledge graph is as follows: By mapping the operation variable vector through the agricultural operation response matrix, the optimization problem of minimizing the sum of squared response errors and the L1 norm regularization of the operation vector is solved to generate the minimum intervention strategy operation vector with sparsity constraints: ； in, represents the agricultural operation vector; A (j) represents the agricultural operation response matrix, which describes the impact intensity of each operation on agricultural variables; O represents the operation variable vector; represents the regularization parameter, which controls the weight of the sparse constraint term; ||O||1 represents the 1-norm of the operation vector; Consistency verification is performed through the agricultural knowledge graph, and judgment functions are used to screen whether the operation plan complies with agricultural regulations: ； in, Represents the agricultural knowledge graph; Indicates final agricultural practice recommendations after consistency verification and revision; Used to determine whether the operation plan complies with agricultural knowledge and standards; Indicates adjustments to unreasonable strategies.

8. The agricultural monitoring method of multi-source perception fusion according to claim 7 is characterized in that: The update process of the policy parameter tensor is: According to regulatory recommendations After execution, during the delay Then collect new perception tensors Compare and build the feedback difference tensor: ; in, Indicates that during policy execution The multi-source perception tensor is then recaptured; represents the feedback difference tensor; Construct the state drift potential energy function: ； ； Among them, S x,y,t represents the stability score of the agricultural system at spatial location x, y and time t; Indicates the tensor fluctuation amplitude of the specified area in the past K2 time units; K2 represents the time window length used to calculate historical fluctuations; T x,y,t-k Represents the multi-source perception semantic tensor at the spatial position (x, y) at the time point tk; represents the average tensor state within the time window [t-K2, t-1]; Represents the state drift potential energy function value; represents the stability adjustment coefficient; The potential energy calculated in the previous step is used to adjust the next round of strategy output tensor, and the strategy parameters are expressed as tensors , the following response gradient descent is used: ； in, represents the policy parameter tensor, i.e., the control model weights or feature map parameters used to generate recommendations at the t-th round of decision making; represents the control model weight used to generate recommendations in the t+1th round of decision making; represents the learning rate; represents the gradient of the drift potential with respect to the policy parameters.

9. The agricultural monitoring method of multi-source perception fusion according to claim 8 is characterized in that: The correction process of dynamically correcting the target state is: The memory averaging mechanism uses a decay factor to fuse the new strategy weights with historical weights to avoid oscillation. At the same time, the trend prediction module uses historical observation tensor sequence modeling to model multi-channel coupling characteristics, dynamically decode and predict future target states to achieve dynamic correction: ； ； in, represents the strategy memory decay factor; represents the time trend prediction function; represents the observation sequence within the time window; Indicates the target agricultural status of a new round.

10. An agricultural monitoring system with multi-source sensing fusion, used to execute the agricultural monitoring method with multi-source sensing fusion according to any one of claims 1 to 9, characterized in that: include: The multi-source perception semantic tensor construction module is responsible for receiving and fusing multi-source heterogeneous data, and then processing it through spatial and temporal fusion and resonance field initialization to construct a high-dimensional semantic tensor, completing the deep semantic expression of agricultural environment and crop growth information. The multimodal aggregation module is used to analyze the coupling of multimodal semantic tensors. By building a gravitational tensor model between devices, it organically aggregates different modal and time series data. The agricultural status emergence detection module is used to dynamically analyze aggregated agricultural status data, identify potential anomalies, emergent behaviors, and nonlinear evolution characteristics, and provide a scientific basis for risk warning and status diagnosis; The agricultural regulation suggestion generation module is used to automatically generate targeted agricultural regulation strategies and operational suggestions based on the detected agricultural status changes and emergent features; The agricultural strategy dynamic feedback and self-evolution module is used to combine multi-source perception data after the execution of control measures to construct feedback difference tensors and state drift potential energy, and dynamically adjust and optimize agricultural strategies by combining policy gradient updates and trend predictions.

Images