Crop full life cycle management method and system based on digital twinning

Abstract

The invention provides a crop full-life-cycle management method and system based on digital twinning, and relates to the technical field of intelligent agricultural management.The crop full-life-cycle management method based on digital twinning comprises the steps that sampling point growth environment indexes and crop growth indexes are calculated by obtaining multi-mode sensing data; establishing a plurality of parallel logic judgment paths corresponding to state labels such as a growth stage, a nutrition state and a water shortage degree on a digital twin platform, inputting related indexes and indexes into the platform, matching step by step according to branch logic, dynamically correcting judgment weights, and determining the state label of each path; summarizing the labels and carrying out spatial aggregation, generating a land parcel health degree grading map according to spatial distribution, and carrying out coloring display in a virtual twinborn body; and then a harvesting route map and a storage allocation plan are generated, the influence of a harvesting sequence on earnings is simulated, and an optimal harvesting scheme is output, so that accurate perception, intelligent research and judgment and visual management of crops in the whole life cycle can be realized, and the planting earnings and management efficiency are improved.

Description

Technical Field

[0001] This application relates to the field of smart agricultural management technology, and in particular to a method and system for managing the entire life cycle of crops based on digital twins. Background Technology

[0002] Crop life-cycle management is a core element in ensuring yield and quality, and it is transforming from traditional experience-based management to digital and precision management, with broad application prospects in large-scale planting and smart agriculture.

[0003] In existing technologies, crop management methods based on digital twins mainly acquire multi-source data through satellite remote sensing, field sensors, etc., and combine them with digital twin models to monitor and assess crop growth status. However, these methods often use fixed logic or a single model when determining the status, making it difficult to adapt to complex field scenarios with multiple stresses throughout the entire growth period, and the analysis of the correlation between environment and growth is not in-depth enough.

[0004] Existing methods have shortcomings in plot-level health assessment and harvest optimization. They lack spatial aggregation analysis of status labels from multiple sampling points, making it difficult to generate continuous hierarchical maps and visualizations. Furthermore, the formulation of harvesting routes and storage plans relies heavily on static experience, failing to fully simulate the impact of different harvesting sequences on overall yields, thus limiting decision-making accuracy and efficiency. Therefore, existing technologies suffer from insufficient accuracy and efficiency in crop lifecycle management decision-making. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for crop life cycle management based on digital twins, so as to solve the problem of insufficient accuracy and efficiency in crop life cycle management decision-making in the existing technology.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a method for crop lifecycle management based on digital twins, comprising:

[0007] Acquire multimodal sensing data of crops, including satellite remote sensing images of the planting area and field environmental parameters, and calculate the growth environment index and crop growth index of the corresponding sampling points based on the multimodal sensing data;

[0008] In the digital twin platform, multiple parallel logical judgment paths are established, and the end of each path corresponds to a status label among reproductive stage, nutritional status, or water shortage degree.

[0009] The growth environment index and crop growth index of the current sampling point are input into the digital twin platform, and the matching is performed step by step according to the preset branch logic of each logical judgment path. In the same logical judgment path, the judgment weight of the subsequent matching steps is dynamically adjusted according to the completed matching results until the step-by-step matching of all paths is completed, and the status label corresponding to the end of each path is determined.

[0010] The status labels output by all logical judgment paths are summarized, and the status labels of multiple sampling points within the same plot are spatially aggregated. Based on the degree of aggregation and boundary continuity of each status label in spatial distribution, a health rating map of the corresponding plot is generated and displayed in a virtual twin with hierarchical coloring.

[0011] Based on the health grading map of each sampling point, a harvesting route map and a storage allocation plan are generated in the virtual twin of the digital twin platform, and the impact of different harvesting sequences on overall revenue is simulated to output the optimal harvesting plan.

[0012] Optionally, the growth environment index and crop growth index of the current sampling point are input into the digital twin platform, and matched step by step according to the preset branch logic of each logical judgment path. Within the same logical judgment path, the judgment weight of subsequent matching steps is dynamically adjusted based on the completed matching results until the step-by-step matching of all paths is completed, and the status label corresponding to the end of each path is determined, including:

[0013] Combine the growth environment index and crop growth index of the current sampling point into a data list and input it into each logical judgment path;

[0014] In a single logical judgment path, starting from the first-level judgment node, the values ​​in the data list are compared with the branch logic of the node to obtain the matching result and select the subsequent branch;

[0015] The matching result of the previous node is converted into a correction coefficient and calculated with the preset initial threshold of the next node to obtain the new judgment threshold of the corresponding node.

[0016] The new judgment threshold is used to match the data list values, and the correction coefficient is updated. The operation of entering the next node, calculating the new threshold and matching is iteratively executed until the end of the path is reached and the corresponding associated status label is obtained.

[0017] Complete the matching process for all logical judgment paths to obtain the status label corresponding to the current sampling point.

[0018] Optionally, the matching result of the previous node is converted into a correction coefficient and calculated with the preset initial threshold of the next node to obtain a new judgment threshold for the corresponding node, including:

[0019] The output value corresponding to the matching result of the previous node is input into a preset coefficient conversion function to calculate the correction coefficient.

[0020] Read the initial threshold value that has been pre-set for the next decision node, where the initial threshold value is a baseline value;

[0021] The calculated correction coefficient is multiplied by the read initial threshold, and the corresponding product is set as the new judgment threshold actually used by the next judgment node in this matching process.

[0022] Optionally, multiple parallel logical decision paths are established in the digital twin platform, with each path ending at a status label corresponding to one of the following: reproductive stage, nutritional status, or water scarcity level.

[0023] In the digital twin platform, three parallel logical judgment paths are established, and the judgment targets of the logical judgment paths are respectively set as reproductive stage, nutritional status, and degree of water shortage.

[0024] For each logical judgment path, a hierarchical structure consisting of multiple judgment nodes connected sequentially is constructed, wherein each judgment node contains a judgment condition, which is used to compare a specific parameter in the growth environment index and crop growth index with a preset threshold.

[0025] Based on the comparison results between the parameters and thresholds in the judgment conditions, determine the path branch connecting the current judgment node to the next judgment node;

[0026] At the final decision node of each logical decision path, a status label is associated, which corresponds to one of the following: reproductive stage, nutritional status, or degree of water shortage.

[0027] Optionally, for each logical decision path, a hierarchical structure consisting of multiple decision nodes connected sequentially is constructed, including:

[0028] For each of the aforementioned logical decision paths, the number of levels in the hierarchical structure is set;

[0029] Based on the set number of levels, the decision nodes contained in each level are defined in order from top to bottom;

[0030] Define the connection relationship between decision nodes in the hierarchical structure, so that each output branch of the decision node in the upper level is connected to a specific decision node in the lower level;

[0031] For each decision node in the hierarchical structure, a decision condition is assigned, in which one or more specific parameters selected from the growth environment index and the crop growth index are specified, and one or more thresholds for comparison are set for each specific parameter.

[0032] Optionally, the status labels output by all logical judgment paths are aggregated, and the status labels of multiple sampling points within the same plot are spatially aggregated. Based on the degree of clustering and boundary continuity of each status label in spatial distribution, a health grading map of the corresponding plot is generated, and the grading is displayed in a virtual twin with color coding, including:

[0033] Collect the status labels output by each sampling point within the same plot of land for all logical judgment paths to obtain a set of status labels corresponding to each sampling point;

[0034] Sampling points that are adjacent in location within a plot and have the same set of state labels are grouped into a single independent unit, and all independent units are labeled.

[0035] Calculate the shared boundary length between each independent unit, and merge independent units whose shared boundary length is greater than a preset continuity threshold to form multiple final aggregated regions;

[0036] Based on the number of independent units and sampling point density contained in each of the aggregated regions, a health level is assigned to each aggregated region to generate a health grading map of the corresponding plot.

[0037] On the three-dimensional surface model corresponding to the virtual twin in the digital twin platform, based on the health level map, different colors are used to render and fill the aggregated areas of different health levels.

[0038] Optionally, based on the health grading map of each sampling point, a harvesting route map and a storage and allocation plan are generated in the virtual twin of the digital twin platform, and the impact of different harvesting sequences on overall revenue is simulated to output the optimal harvesting plan, including:

[0039] Based on the health level of each aggregated region in the health grading diagram, each aggregated region is marked in the virtual twin, and the marked aggregated regions are connected to generate a closed-loop harvesting route map without intersections.

[0040] Based on the assumed harvesting sequence in the closed-loop harvesting route map, combined with the estimated yield per unit area corresponding to different health levels, the daily harvest volume is calculated, and a detailed storage allocation plan is generated based on the regional division of storage space.

[0041] In the digital twin platform, the access order of the aggregation area is changed, the daily harvest volume is recalculated for each changed order, and a corresponding warehousing and allocation plan is generated. The total operation time and estimated total output from the start of harvesting to the completion of warehousing are simulated and calculated.

[0042] By comparing the total operation time with the estimated total output obtained from different access sequences, the result with the highest estimated total output and the shortest total operation time is selected as the optimal harvesting scheme.

[0043] Secondly, this application provides a crop lifecycle management system based on digital twins, including:

[0044] The acquisition module is used to acquire multimodal sensing data of crops, including satellite remote sensing images of the planting area and field environmental parameters, and to calculate the growth environment index and crop growth index of the corresponding sampling point based on the multimodal sensing data.

[0045] The module is used to establish multiple parallel logical decision paths in the digital twin platform. The end of each path corresponds to a status label among reproductive stage, nutritional status, or water shortage level.

[0046] The matching module is used to input the growth environment index and crop growth index of the current sampling point into the digital twin platform, and perform step-by-step matching according to the preset branch logic of each logical judgment path. In the same logical judgment path, the judgment weight of the subsequent matching steps is dynamically adjusted according to the completed matching results until the step-by-step matching of all paths is completed, and the status label corresponding to the end of each path is determined.

[0047] The generation module is used to summarize the status labels output by all logical judgment paths, spatially aggregate the status labels of multiple sampling points within the same plot, generate a health rating map of the corresponding plot based on the degree of aggregation and boundary continuity of each status label in spatial distribution, and display the rating in a virtual twin.

[0048] The output module is used to generate a harvesting route map and a storage and allocation plan in the virtual twin of the digital twin platform based on the health grading map of each sampling point, and to simulate the impact of different harvesting sequences on the overall revenue, and output the optimal harvesting plan.

[0049] Thirdly, this application provides an electronic device, comprising:

[0050] Memory, used to store computer programs;

[0051] A processor is used to execute computer programs to implement the steps of the digital twin-based crop lifecycle management method described in the first aspect above.

[0052] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the crop lifecycle management method based on digital twins as described in the first aspect above.

[0053] The crop lifecycle management method based on digital twins provided in this application acquires multimodal sensing data such as satellite remote sensing images and field environmental parameters of the planting area, and calculates the growth environment index and crop growth indicators of sampling points, providing accurate and comprehensive data support for judging crop growth status. By establishing multiple parallel logical judgment paths on the digital twin platform, each path corresponding to a state label, it can realize the synchronous judgment of multi-dimensional crop growth status. By inputting relevant indices and indicators into the platform, matching them step by step according to branch logic and dynamically correcting the judgment weights, it can improve the accuracy and adaptability of state label judgment. By summarizing state labels and performing spatial aggregation, generating a plot health grading map and displaying it with color, it can intuitively present the overall growth status of the plot, making it easy to quickly grasp the distribution of crop health. By generating harvesting routes and storage plans based on the health grading map, simulating the impact of harvesting sequence on yield, and outputting the optimal solution, it can optimize the harvesting and storage process and improve planting yield.

[0054] Furthermore, the growth environment index and crop growth index of the sampling points are combined into a data list and input into each logical judgment path. Within a single path, numerical matching is performed starting from the first-level judgment node, and subsequent branches are selected. The matching result of the previous node is converted into a correction coefficient, and the new judgment threshold for the next node is calculated and iteratively matched until the state label at the end of each path is obtained. This step-by-step matching and iterative approach of dynamically correcting judgment thresholds further improves the accuracy of crop growth status label determination, avoids judgment bias caused by fixed thresholds, and ensures that the status labels of each dimension highly match the actual crop growth situation, providing a reliable basis for subsequent plot health assessment and harvesting plan optimization. Attached Figure Description

[0055] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 A flowchart illustrating a crop lifecycle management method based on digital twins, provided for an embodiment of this application;

[0057] Figure 2A flowchart illustrating the specific implementation of a crop lifecycle management method based on digital twins, as provided in this application embodiment;

[0058] Figure 3 This is a schematic diagram of the structure of a crop life cycle management system based on digital twins, provided as an embodiment of this application. Detailed Implementation

[0059] In the process of transforming crop life-cycle management towards digitalization and precision, existing management methods based on digital twins still have significant shortcomings. These methods often employ fixed logic or single models when determining crop growth status, making it difficult to adapt to the complex field environment throughout the crop's growth cycle and unable to deeply analyze the correlation between the environment and crop growth. Furthermore, in assessing plot health, they lack integrated analysis of the status of multiple sampling points, failing to intuitively present the distribution of crop health. Moreover, harvesting and storage planning relies on experience, failing to fully consider the impact of harvesting sequence on yield, ultimately leading to inaccurate management decisions and difficulty in improving planting profits.

[0060] To address the aforementioned issues, this application proposes a digital twin-based method for full life-cycle management of crops. This method first acquires satellite remote sensing imagery and field environmental data of the planting area, calculating crop growth-related indicators. Then, it establishes multiple parallel logical judgment paths through a digital twin platform, dynamically adjusting judgment weights to accurately determine the crop growth status. Subsequently, it integrates the status of multiple sampling points to generate an intuitive plot health grading map. Finally, based on the grading map, it formulates harvesting and storage plans, simulates the impact of different harvesting sequences on profitability, and outputs the optimal solution. This solution directly solves the shortcomings of existing technologies, such as poor adaptability of fixed logic, discontinuous health assessment, and reliance on experience for harvesting decisions. It achieves precise perception and visual management throughout the entire crop lifecycle, significantly improving decision-making accuracy and planting returns.

[0061] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0062] The core of this application is to provide a method for managing the entire life cycle of crops based on digital twins. A flowchart of one specific implementation method is shown below. Figure 1 As shown, the method includes:

[0063] S101. Acquire multimodal sensing data of crops, and calculate the growth environment index and crop growth index of the corresponding sampling points based on the multimodal sensing data.

[0064] Multimodal sensing data refers to various types of data used to monitor crop growth, including satellite remote sensing images of the planting area and field environmental parameters. Satellite remote sensing images can intuitively present the macroscopic situation of the entire planting area, while field environmental parameters include environmental data that affect crop growth, such as temperature, humidity, and soil nutrients. The growth environment index is used to measure the environmental suitability of the sampling point, and crop growth indicators are used to reflect the growth status of crops, such as growth rate and robustness.

[0065] In one specific implementation, firstly, satellite remote sensing technology is used to capture overall images of the planting area, clearly showing the distribution of plots and the overall growth status of the crops. Simultaneously, various sensors are strategically deployed in the field to collect environmental parameters such as temperature, humidity, and soil nutrients in real time. These two types of data are then integrated to form complete multimodal sensing data. Next, the acquired multimodal sensing data is processed and analyzed. Based on preset calculation standards, the growth environment index and crop growth index for each sampling point are calculated. The specific calculation process is as follows:

[0066] The growth environment index is obtained by weighted summation of parameters such as temperature, humidity, and soil nutrients after standardization. The specific formula is as follows:

[0067]

[0068] in, This is an index of the growth environment. The parameter weights are calibrated according to crop growth requirements, with values ​​ranging from 0.2 to 0.4, and satisfying the following conditions: ; These are standardized values ​​for temperature, humidity, and soil nutrients, calculated by comparing actual collected values ​​with the appropriate ranges for the corresponding parameters. The value ranges are all... .

[0069] The crop growth index is calculated by combining satellite remote sensing image features and field environmental parameters. The specific formula is as follows:

[0070]

[0071] in, As an indicator of crop growth, The preset weighting coefficients are calibrated according to crop varieties, with values ​​ranging from 0.3 to 0.5, and satisfying the following conditions: ; Canopy coverage, derived from satellite remote sensing image analysis; The vegetation normalization index is calculated from the reflectance of satellite remote sensing images. This is the soil nutrient coefficient, calculated from field soil nutrient parameters.

[0072] S102. Establish multiple parallel logical judgment paths in the digital twin platform. The end of each logical judgment path corresponds to a status label among reproductive stage, nutritional status, or water shortage degree.

[0073] The logical judgment path refers to the process link built in the digital twin platform to determine a specific state of crop growth. Each path runs independently and the judgment process is advanced synchronously. The status label is an identifier used to intuitively represent a certain dimension of crop growth, corresponding to three core growth dimensions: growth stage, nutritional status, and water shortage degree.

[0074] The digital twin platform is built based on the actual geographical information of the planting area, crop growth patterns and field management rules. It can fully map the real planting scene and carry data processing and logical judgment functions. The method of building the digital twin is a conventional solution, which will not be elaborated on in this application.

[0075] Furthermore, the digital twin platform incorporates multiple logical decision paths to synchronously and independently analyze crop growth states across different dimensions, avoiding interference and deviations caused by a single judgment process. Each path employs a multi-level decision node structure, progressively refining judgment conditions from macro to micro perspectives and narrowing the state range layer by layer. Each decision node carries corresponding judgment conditions for accurate comparison of growth parameters with preset standards. Status tags are associated at the end of each path, transforming the final judgment result into an intuitive and readable identifier.

[0076] Optionally, step S102 may specifically include the following steps:

[0077] S1021. In the digital twin platform, establish three parallel logical judgment paths. The judgment targets of the logical judgment paths are set as reproductive stage, nutritional status, and water shortage degree, respectively.

[0078] Among them, the parallel logical judgment path refers to three paths running independently at the same time without interfering with each other, and can synchronously complete the logical judgment of different judgment targets, which can significantly improve the efficiency of state judgment compared with the serial path.

[0079] S1022. For each logical judgment path, construct a hierarchical structure consisting of multiple judgment nodes connected in sequence. Each judgment node contains a judgment condition, which is used to compare specific parameters in the growth environment index and crop growth index with a preset threshold.

[0080] Specifically, step S1022 may include the following processes: setting the number of levels in the hierarchical structure for each logical judgment path; defining the judgment nodes contained in each level in order from top to bottom according to the set number of levels; defining the connection relationship between judgment nodes in the hierarchical structure so that each output branch of the judgment node in the upper level is connected to a specific judgment node in the lower level; assigning judgment conditions to each judgment node in the hierarchical structure, specifying one or more specific parameters selected from the growth environment index and crop growth index in the judgment conditions, and setting one or more thresholds for comparison for each specific parameter.

[0081] In the above steps, the judgment node is the core judgment unit in the logical judgment path, and each node is responsible for completing a specific parameter comparison task. The hierarchical structure refers to an ordered structure formed by connecting multiple judgment nodes sequentially from top to bottom according to the order of the judgment logic; the judgment result of the upper-level node directly determines the judgment direction of the lower-level node. The judgment condition is the core content of the judgment node. By retrieving relevant parameters from the previously calculated growth environment index and crop growth index, it is compared with preset standard values, and the comparison result is output to guide subsequent judgments.

[0082] S1023. Based on the comparison results between the parameters and thresholds in the judgment conditions, determine the path branch connecting the current judgment node to the next judgment node.

[0083] S1024. At the final judgment node of each logical judgment path, associate a status label, which corresponds to one of the following: reproductive stage, nutritional status, or degree of water shortage.

[0084] In this embodiment, three parallel paths are first built on a digital twin platform, and the judgment target corresponding to each path is defined. Then, a hierarchical judgment node structure is constructed for each path, and corresponding judgment conditions are configured. Next, the subsequent path branches are determined based on the parameter comparison results of each judgment node. Finally, the corresponding status label is associated at the end of each path, thereby realizing the hierarchical judgment of the multi-dimensional growth status of crops.

[0085] As an example, firstly, three parallel logical judgment paths are established in the digital twin platform using S1021. The judgment targets are set as growth stage, nutritional status, and water shortage level, respectively. The three paths start simultaneously without interfering with each other. For example, in the wheat planting scenario, the three paths are used to determine whether the wheat is currently in the jointing stage, heading stage, or maturity stage, to determine whether the wheat is nutritious enough, and to determine whether the wheat is water-deficient.

[0086] Secondly, a hierarchical structure is constructed for each path and judgment conditions are configured through S1022. For example, for the "water shortage degree" judgment path, the number of layers is set to 3, and 3 judgment nodes are defined sequentially from top to bottom. The connection relationship is defined as follows: the comparison result of the first layer node is divided into two branches, "meets the standard" and "does not meet the standard", which are respectively connected to two different judgment nodes in the second layer. The comparison result of the second layer node is also connected to the judgment node in the third layer. Judgment conditions are assigned to each node. The first layer node selects the humidity standardized value in the growth environment index as a specific parameter and sets the threshold to 0.4. The second layer node selects the actual field humidity collection value and sets the threshold to 50%. The third layer node selects the soil moisture content parameter and sets the threshold to 15%.

[0087] Next, path branches are determined by S1023 based on the comparison results of the parameters of each node with the threshold. For example, the first-level node compares the standardized humidity value with 0.4. If the value is ≥0.4, it is judged as "compliant" and enters the corresponding node in the second level; if the value is <0.4, it is judged as "non-compliant" and enters another node in the second level; the second-level node compares the actual field humidity value with 50%, and so on, until the comparison of all levels of nodes is completed.

[0088] Finally, status labels are associated at the final node of each path via S1024. For example, at the end of the "Water Deficiency" path, based on the final comparison results, four status labels are associated: "No Water Deficiency," "Slight Water Deficiency," "Moderate Water Deficiency," and "Severe Water Deficiency." At the end of the "Growth Stage" path, status labels such as "Jointing Stage," "Heading Stage," and "Maturity Stage" are associated. At the end of the "Nutritional Status" path, status labels such as "Sufficient Nutrition," "Moderate Nutrition," and "Insufficient Nutrition" are associated. The above example is only one example of this application. In practical applications, the number of levels, judgment parameters, and thresholds can be adjusted according to the needs of crop varieties and planting scenarios. This application does not limit this.

[0089] This application enables the simultaneous and accurate determination of crop growth stage, nutritional status, and water shortage level, clarifying the growth status of crops in various dimensions, providing a reliable status basis for plot health aggregation analysis and harvesting plan optimization, and improving the efficiency and accuracy of crop growth status determination.

[0090] S103. Input the growth environment index and crop growth index of the current sampling point into the digital twin platform, and perform step-by-step matching according to the preset branch logic of each logical judgment path. In the same logical judgment path, dynamically adjust the judgment weight of the subsequent matching steps according to the completed matching results until the step-by-step matching of all paths is completed, and determine the status label corresponding to the end of each path.

[0091] Among them, hierarchical matching refers to comparing the parameters of each judgment node sequentially from top to bottom according to the hierarchical structure in each logical judgment path. The matching result of the previous node directly determines the matching direction of the next node. Dynamic adjustment of judgment weight refers to adjusting the judgment criteria of subsequent nodes based on the matching results of completed nodes, so that the matching process is more in line with the actual growth state of the plant and improves the accuracy of judgment.

[0092] Step S103 combines the growth environment index and crop growth index calculated above with the established logical judgment path, and through step-by-step matching and dynamic weight correction, to accurately determine the crop growth status labels in various dimensions, connecting the data above with the subsequent plot health analysis.

[0093] Optionally, such as Figure 2 As shown, step S103 may specifically include the following steps:

[0094] S1031. Combine the growth environment index and crop growth index of the current sampling point into a data list and input each logical judgment path.

[0095] The data list is an ordered set of data that integrates the growth environment index, crop growth indicators and related parameters of a single sampling point. This makes it easy for the digital twin platform to quickly retrieve and read each parameter, while ensuring that each logical judgment path can synchronously obtain unified and complete judgment data, thus guaranteeing the consistency of multi-path matching.

[0096] Furthermore, the logical judgment path within the digital twin platform is constructed based on crop growth patterns, with parameter adaptation and logical presets. The judgment nodes and conditions in the path are designed around growth environment indices and crop growth indicators, allowing direct reading, comparison, and analysis of input parameter values, thereby ensuring that input data can smoothly enter the hierarchical judgment process.

[0097] S1032. In a single logical judgment path, starting from the first-level judgment node, the values ​​in the data list are compared with the branch logic of the node to obtain the matching result and select the subsequent branch.

[0098] Among them, the branch logic is the pre-set comparison rule for each judgment node, which is used to clarify the comparison method between parameters and thresholds and the corresponding output results. Different comparison results correspond to different subsequent node branches, providing clear guidance for step-by-step matching.

[0099] In actual execution, the comparison of parameter values ​​in the data list with the branch logic is carried out using a unified unit and a unified standard. The system first extracts the parameter values ​​required for the current judgment node from the data list, then directly compares these values ​​with the preset thresholds in the branch logic. Based on the comparison results, the matching status of the current node is determined, and then the path branch to the next node is selected according to preset rules, thereby achieving orderly and stable step-by-step judgment.

[0100] S1033. Convert the matching result of the previous node into a correction coefficient and calculate it with the preset initial threshold of the next node to obtain the new judgment threshold of the corresponding node.

[0101] Specifically, step S1033 may include the following process: inputting the output value corresponding to the matching result of the previous node into a preset coefficient conversion function to calculate the correction coefficient; reading the initial threshold set in advance for the next judgment node, the initial threshold being the base value; multiplying the calculated correction coefficient by the read initial threshold, and setting the corresponding product as the new judgment threshold actually used by the next judgment node in this matching process.

[0102] In the above steps, the correction coefficient is an adjustment parameter calculated based on the matching result of the previous node. It is used to dynamically optimize the judgment criteria of the next node, making subsequent matching more consistent with the current growth state of the organism. The initial threshold is a baseline comparison value pre-set for each judgment node.

[0103] S1034. Match the new judgment threshold with the data list values ​​and update the correction coefficient. Iterate through the operations of entering the next node, calculating the new threshold and matching, until the end of the path is reached and the corresponding associated status label is obtained.

[0104] S1035. Complete the matching process of all logical judgment paths and obtain the status label corresponding to the current sampling point.

[0105] In this embodiment, the growth environment index and crop growth index of the current sampling point are first integrated into a data list. All parallel logic judgment paths are input. Then, starting from the first-level node in each path, the data list values ​​are compared with the node branch logic to obtain the matching result and the subsequent branch is selected. Then, the matching result of the previous node is converted into a correction coefficient, and a new threshold is calculated with the initial threshold of the next node. Then, the new threshold is used to continue matching and updating the correction coefficient. The process is iterated until the end of the path is reached and the status label is obtained. Finally, all path matching is completed and all status labels corresponding to the sampling point are obtained.

[0106] As an example, continuing with the wheat planting scenario mentioned earlier, we assume that the current sampling point has calculated the growth environment index EI=0.5, the crop growth index VI=0.61, and the relevant parameters include temperature standardized value 0.5, humidity standardized value 0.5, soil nutrient standardized value 0.5, canopy coverage 0.6, NDVI=0.7, and soil nutrient coefficient 0.5.

[0107] First, integrate this data into a data list, and simultaneously input three parallel logical judgment paths: reproductive stage, nutritional status, and water shortage level, to ensure that each path can obtain complete judgment data.

[0108] Secondly, the first-level node matching of a single path is carried out through S1032. Taking the path for determining the degree of water shortage as an example, assuming that the initial threshold of the first-level node is 0.4, the preset branch logic of the first-level node is to compare the humidity standardization value. Therefore, the humidity standardization value of 0.5 in the data list is compared with the initial threshold. Since 0.5 > 0.4, the matching result of "meets the standard" is obtained, and the corresponding subsequent branch is selected to enter the second-level node.

[0109] Next, the new judgment threshold for the next node is calculated via S1033. The preset coefficient conversion function is: correction coefficient k = matching result output value × 0.2 + 0.8. Assuming the output value corresponding to the "meeting the standard" matching result of the previous node is 1, substituting it into the function, we get the correction coefficient k = 1 × 0.2 + 0.8 = 1.0. The preset initial threshold for the second layer node is 50%, which serves as the field humidity benchmark. Multiplying the correction coefficient by the initial threshold, we get the new judgment threshold = 1.0 × 50% = 50%.

[0110] Then, using the new judgment threshold in S1034, the matching continues. The actual field moisture values ​​of 60% and 50% in the data list are compared to obtain a "compliant" matching result. At the same time, the correction coefficient is updated. The output value of 1 from this matching is substituted into the transformation function, and the new correction coefficient is still 1.0. The iteration proceeds to the third layer node. The initial threshold of the third layer node is 15%, which serves as the benchmark value for soil moisture content. The new threshold is calculated as 1.0 × 15% = 15%. The actual soil moisture content value of 18% in the data list is compared with the new threshold to obtain a "compliant" result. The iteration continues until the end of the path is reached.

[0111] Finally, the matching process for all three paths is completed via S1035. The water shortage path is associated with the "no water shortage" status label at its end, the growth stage path is associated with the "jointing stage" status label, and the nutrient status path is associated with the "moderate nutrient" status label, thus obtaining all the status labels corresponding to the wheat sampling point. The above example is only one example of this application. In practical applications, the coefficient conversion function, initial threshold, and branch logic can be adjusted according to the needs of crop variety and planting scenario. This application does not limit this.

[0112] In another specific implementation, the coefficient transformation function can employ a linear transformation approach, adjusting the function parameters according to different crop growth stages. For example, during the wheat heading stage, the weight of the correction coefficient is increased, making the threshold correction more closely match the growth characteristics of that stage. Simultaneously, an outlier detection step can be added during the iteration process. If a value in the data list exceeds a reasonable range, a timely alert is issued, and a default threshold is used for matching, ensuring the stability of the matching process.

[0113] This application enables the accurate determination of status labels for each path, avoids judgment bias caused by fixed thresholds, ensures that status labels are highly consistent with the actual growth of crops, and improves the accuracy and adaptability of crop growth status determination.

[0114] S104. Summarize the status labels output by all logical judgment paths, spatially aggregate the status labels of multiple sampling points within the same plot, generate a health rating map of the corresponding plot based on the degree of aggregation and boundary continuity of each status label in spatial distribution, and display the rating in a virtual twin.

[0115] Spatial aggregation refers to classifying and integrating adjacent sampling points with similar growth states within the same plot to form continuous regional units, facilitating overall analysis of crop health status. The health grading map is a graphic representation of crop health levels in different areas within a plot, clearly reflecting differences in crop growth through different grading levels. Virtual twin colorization displays present the health grading results on the 3D surface model of the digital twin platform using different colors, making the health distribution of the plot readily apparent.

[0116] This step combines the status labels of each sampling point obtained above with spatial aggregation and hierarchical display to transform the scattered sampling point data into intuitive plot health information, connecting the status determination in the previous section with the formulation of subsequent harvesting plans, and providing visual support for precise management.

[0117] Optionally, step S104 may specifically include the following steps:

[0118] S1041. Collect the status labels output by each sampling point within the same plot of land for all logical judgment paths, and obtain a set of status labels corresponding to each sampling point.

[0119] Among them, the status label set is a summary of the status labels output by a single sampling point in all logical judgment paths. It includes three core growth information corresponding to the sampling point: reproductive stage, nutritional status, and water shortage level, ensuring that the growth status of each sampling point can be fully included in the aggregate analysis.

[0120] S1042. Group sampling points that are adjacent in location within the plot and have the same set of status labels into the same independent unit, and mark all independent units.

[0121] An independent unit is a small region consisting of sampling points that are adjacent in location and have the same growth state. Each independent unit corresponds to a set of identical state labels.

[0122] S1043. Calculate the shared boundary length between each independent unit, merge independent units whose shared boundary length is greater than the preset continuity threshold, and form the final multiple aggregated regions.

[0123] The shared boundary length refers to the boundary length shared between two adjacent independent units, used to measure the degree of continuity in the spatial distribution of the two units. The preset continuity threshold is a pre-set boundary length standard used to determine whether two adjacent independent units meet the conditions for merging, ensuring that the merged aggregate region has good spatial continuity and avoiding the impact of scattered areas on the accuracy of health analysis.

[0124] S1044. Based on the number of independent units and sampling point density contained in each aggregation region, assign a health level to each aggregation region to generate a health grading map of the corresponding plot.

[0125] Sampling point density refers to the number of sampling points per unit area, used to help determine the stability of the growth status of the aggregated area. Health level is a comprehensive assessment of the growth status of the aggregated area, combining the number of independent units and sampling point density to more comprehensively and accurately reflect the overall health level of crops within the area, ensuring the scientific validity of the health level map.

[0126] S1045. On the three-dimensional surface model corresponding to the virtual twin in the digital twin platform, based on the health level map, different colors are used to render and fill the aggregated areas of different health levels.

[0127] The 3D surface model is a 3D model of the actual planting plot in the digital twin platform, consistent with the spatial location and terrain features of the actual plot. Color rendering and filling associate different health levels with specific colors, intuitively distinguishing areas with different health levels through color differences, allowing managers to quickly identify the distribution of crop growth quality within the plot.

[0128] In this embodiment, all status labels of each sampling point within the same plot are first collected to form a status label set for each sampling point. Then, sampling points that are adjacent in location and have the same label set are merged into independent units and marked. Next, the shared boundary length of each independent unit is calculated, and independent units that meet the conditions are merged to form an aggregated region. Then, the health level is assigned according to the number of independent units and the sampling point density of the aggregated region to generate a health level map. Finally, on the three-dimensional surface model of the virtual twin, the aggregated regions of each level are rendered with different colors to complete the visualization display.

[0129] As an example, continuing the wheat planting scenario described earlier, a wheat field has 20 sampling points. Each sampling point has a corresponding set of status labels. For example, some sampling points have the label set as "jointing stage, adequate nutrition, no water shortage," some as "jointing stage, insufficient nutrition, slight water shortage," and some as "heading stage, sufficient nutrition, no water shortage." First, the status labels of all sampling points are collected through S1041, integrating the three types of status labels of each sampling point into a label set to ensure that the growth information of each sampling point is complete and without omission.

[0130] Secondly, the sampling points are merged and marked using S1042. Sampling points that are adjacent in location and have the same set of labels are found. Three adjacent sampling points with the set of labels "jointing stage, moderate nutrition, no water shortage" are merged into independent unit 1. Two adjacent sampling points with the set of labels "jointing stage, insufficient nutrition, mild water shortage" are merged into independent unit 2. Four adjacent sampling points with the set of labels "heading stage, sufficient nutrition, no water shortage" are merged into independent unit 3. And so on, finally 20 sampling points are merged into 6 independent units, which are marked as unit 1 to unit 6 respectively.

[0131] Next, the shared boundary length of each independent unit is calculated via S1043, with a preset continuity threshold of 5 meters. The calculation shows that the shared boundary length between unit 1 and unit 3 is 6 meters, which is greater than the preset threshold, so the two units are merged into aggregate region A. The shared boundary length between unit 2 and its adjacent unit 4 is 4 meters, which is less than the threshold, so they are not merged and remain independent units. The remaining units do not meet the merging conditions, resulting in 5 aggregate regions.

[0132] Then, health levels were assigned to each aggregated area via S1044. The health level calculation method combined the number of independent units and the sampling point density for comprehensive evaluation. Aggregated area A contains 7 merged independent units with a high sampling point density and is rated as Level 1 healthy; Unit 2 contains 2 independent units with a medium sampling point density and is rated as Level 3 healthy; the remaining aggregated areas are rated as Level 2 and Level 4 healthy respectively based on the corresponding parameters, thereby generating a health level map for the wheat plot.

[0133] Finally, using S1045, the 3D surface model of the digital twin platform is colored and displayed. Level 1 healthy areas are rendered in green, level 2 in light green, level 3 in yellow, and level 4 in orange. Each aggregated area is filled with its corresponding color. Managers can visually observe the distribution of wheat health within the plot through the 3D model, quickly identifying high-quality growing areas and areas requiring improvement. The above example is merely one illustration of this application. In practical applications, the continuity threshold, health level classification standards, and coloring scheme can be adjusted according to the needs of crop variety and planting scale. This application does not impose any limitations on these adjustments.

[0134] In another specific implementation, spatial aggregation can employ spatial clustering algorithms. By calculating the spatial distance between sampling points, adjacent sampling points with similar states are automatically merged, improving the efficiency and accuracy of merging independent units. Simultaneously, the allocation of health levels can incorporate more reference parameters, combined with appropriate standards for crop growth stages, further optimizing the scientific basis of the grading and making the health grading map more closely aligned with actual planting needs.

[0135] This application clearly presents the spatial differences in crop growth, making it easier for managers to quickly grasp the overall health status of the plot, providing accurate spatial references for the formulation of harvesting plans, and improving the intuitiveness and efficiency of crop management.

[0136] S105. Based on the health grading map of each sampling point, generate a harvesting route map and a storage allocation plan in the virtual twin of the digital twin platform, simulate the impact of different harvesting sequences on the overall revenue, and output the optimal harvesting plan.

[0137] The harvesting route map is a path planning diagram used to guide harvesting operations, clearly defining the harvesting sequence and route to ensure that harvesting operations are carried out in an orderly and efficient manner. The storage and allocation plan is a detailed arrangement for the storage and handling of harvested crops after harvest, based on the harvest volume and storage conditions, ensuring the orderly storage of harvested crops and reducing losses. The optimal harvesting plan is the best operational plan selected by comprehensively considering factors such as harvesting efficiency and yield.

[0138] This step S105 combines the health grading map generated earlier with route planning, storage allocation, and multi-scheme simulation to output the optimal harvesting plan. It connects with the previous plot health analysis to complete the final stage of crop life cycle management and achieve precise optimization of harvesting and storage.

[0139] Optionally, step S105 may specifically include the following steps:

[0140] S1051. Based on the health level of each aggregated region in the health grading map, mark each aggregated region in the virtual twin and connect the marked aggregated regions to generate a closed-loop harvesting route map without intersections.

[0141] Among them, the closed-loop harvesting route map refers to a harvesting route where the starting point and the ending point coincide and the path does not intersect. It can avoid repeated travel or missed areas during the harvesting process and improve the efficiency of harvesting operations.

[0142] S1052. Based on the assumed harvesting sequence of the closed-loop harvesting route map, combined with the estimated yield per unit area corresponding to different health levels, calculate the daily harvesting volume, and generate a detailed storage allocation plan based on the regional division of storage space.

[0143] The estimated yield per unit area is calculated based on the crop health level of the aggregated area and the crop growth pattern. The higher the health level, the higher the estimated yield per unit area. The daily harvest volume is the total amount that can be harvested in a single day, calculated by combining the operating efficiency of the harvesting equipment and the harvesting time. The storage space area division divides the storage area according to the crop health level and storage requirements, which facilitates classified storage, precise management, and reduces crop loss.

[0144] S1053. Change the access order of the aggregation area in the digital twin platform, recalculate the daily harvest volume for each changed order and generate the corresponding warehousing and allocation plan, simulate and calculate the total operation time and estimated total output from the start of harvesting to the completion of warehousing.

[0145] The order of visits to aggregation areas refers to the sequence in which each aggregation area is harvested during the harvesting operation. Changing the order of visits can create various different harvesting plans. Total operation time is the entire time from the start of the harvesting operation to the completion of harvesting and storage of all crops, including the time spent on each stage such as harvesting, picking, transportation, and storage. Estimated total yield is the total quantity of crops harvested from all aggregation areas, calculated based on the estimated yield and area of ​​each aggregation area.

[0146] S1054. Compare the total operating time and estimated total output obtained from different access sequences, and select the result with the highest estimated total output and the shortest total operating time as the optimal harvesting plan.

[0147] In this embodiment, each aggregation area is first marked according to the health grading map, and each area is connected to generate a closed-loop harvesting route map without intersection. Then, based on the assumed harvesting order of the route and the estimated yield per unit area, the daily harvesting volume is calculated and a storage and allocation plan is formulated. Next, the harvesting order is changed, and the daily harvesting volume, total operating time and estimated total output are recalculated. Finally, the parameters of multiple schemes are compared, the optimal harvesting scheme is selected, and the optimization planning of harvesting and storage is completed.

[0148] As an example, continuing with the wheat planting scenario above, a health grading map has been generated for this wheat plot, containing five aggregated regions: Level 1 Health Region A, Level 2 Health Region B, Level 3 Health Region C, and Level 4 Health Regions D and E, with the estimated yield decreasing sequentially for each region. First, the five aggregated regions are marked in the virtual twin using S1051. Combining the spatial location of each region, a path planning method is used to connect the regions and generate a closed-loop harvesting route map without intersections. Assuming the initial harvesting order is A, B, C, D, E, this ensures that the route is non-repetitive and without omissions.

[0149] Secondly, the daily harvest volume is calculated using S1052, and a storage and allocation plan is formulated. It is known that the harvesting equipment operates for 8 hours daily, harvesting 0.5 mu per hour. The areas of each aggregation zone are: Zone A 10 mu, Zone B 8 mu, Zone C 6 mu, Zone D 5 mu, and Zone E 4 mu. The estimated yields for zones one through four are 500 kg / mu, 450 kg / mu, 400 kg / mu, and 350 kg / mu, respectively. The daily harvest volume is calculated as follows: the daily harvestable area is 8 × 0.5 = 4 mu. Following the initial sequence, on the first day, 4 mu of Zone A is harvested, yielding 4 × 500 = 2000 kg; on the second day, the remaining 6 mu of Zone A is harvested, yielding 6 × 500 = 3000 kg, and so on, completing the harvest calculation for all zones. Meanwhile, based on the division of storage space, wheat from first- and second-level healthy areas is stored in the high-quality storage area, while wheat from third- and fourth-level areas is stored in the ordinary storage area. The daily harvest time and storage location of crops are clearly defined, forming a detailed storage and allocation plan.

[0150] Next, by changing the access order of the aggregation area through S1053, multiple harvesting plans are generated. For example, Plan 2 has the order A, C, B, E, D, and Plan 3 has the order B, A, D, C, E. For each plan, the daily harvest volume and warehousing allocation plan are recalculated, the entire harvesting process is simulated, and the total operating time and estimated total output are calculated. Taking Plan 2 as an example, the daily harvesting area and harvest volume are recalculated, and the total operating time and estimated total output are calculated by combining the transportation and warehousing time, and compared with the initial plan.

[0151] Then, the total operating time and estimated total output of all schemes are compared using S1054. After calculation, the estimated total output of the initial scheme is 10×500+8×450+6×400+5×350+4×350=5000+3600+2400+1750+1400=14150 kg, and the total operating time is (10+8+6+5+4)÷4=33÷4=8.25 days; Scheme 2 has the same estimated total output, with a total operating time of 8.5 days; Scheme 3 has an estimated total output of 14000 kg, with a total operating time of 8 days. In summary, the initial scheme has the highest estimated total output and the shortest total operating time, and is therefore determined as the optimal harvesting scheme. The above example is only one example of this application. In practical applications, the route planning, harvesting sequence, and evaluation criteria can be adjusted according to factors such as harvesting equipment efficiency, storage conditions, and market demand. This application does not limit these adjustments.

[0152] In another specific implementation, route planning can employ a shortest path algorithm, combining the spatial distance and harvesting difficulty of each aggregation area to automatically generate the optimal closed-loop route, further improving harvesting efficiency. Simultaneously, when simulating different harvesting sequences, a revenue calculation step can be added, taking into account the market price corresponding to crop quality, and comprehensively considering total yield, operating time, and revenue, making the selection of the optimal harvesting plan more aligned with actual revenue needs.

[0153] This application generates reasonable harvesting routes and storage allocation plans by combining health grading charts, simulates the operational effects of different harvesting sequences, selects the optimal harvesting scheme, realizes the orderly and efficient development of harvesting operations, optimizes the allocation of storage resources, reduces operational time and crop loss, and improves planting income and the overall efficiency of the entire life cycle management.

[0154] Figure 3 This is a schematic diagram illustrating a specific implementation of a crop lifecycle management system based on digital twins, as provided in this application. (Refer to...) Figure 3 The system may include:

[0155] The acquisition module 31 is used to acquire multimodal sensing data of crops, including satellite remote sensing images of the planting area and field environmental parameters, and to calculate the growth environment index and crop growth index of the corresponding sampling points based on the multimodal sensing data.

[0156] Module 32 is established to create multiple parallel logical judgment paths in the digital twin platform. The end of each path corresponds to a status label among reproductive stage, nutritional status, or water shortage level.

[0157] The matching module 33 is used to input the growth environment index and crop growth index of the current sampling point into the digital twin platform, and perform step-by-step matching according to the preset branch logic of each logical judgment path. In the same logical judgment path, the judgment weight of the subsequent matching steps is dynamically adjusted according to the completed matching results until the step-by-step matching of all paths is completed, and the status label corresponding to the end of each path is determined.

[0158] The generation module 34 is used to summarize the status labels output by all logical judgment paths, spatially aggregate the status labels of multiple sampling points within the same plot, generate a health rating map of the corresponding plot based on the degree of aggregation and boundary continuity of each status label in spatial distribution, and display the rating in a virtual twin.

[0159] Output module 35 is used to generate a harvesting route map and a storage allocation plan in the virtual twin of the digital twin platform based on the health grading map of each sampling point, and to simulate the impact of different harvesting sequences on the overall revenue, and output the optimal harvesting plan.

[0160] The crop life cycle management system based on digital twins in this application is used to implement the aforementioned crop life cycle management method based on digital twins. Therefore, the specific implementation of the crop life cycle management system based on digital twins can be found in the embodiment section of the crop life cycle management method based on digital twins above. The specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0161] This application also provides an electronic device, including: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the above-described digital twin-based crop lifecycle management method.

[0162] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-described digital twin-based crop lifecycle management methods.

[0163] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0164] Embodiments of the present invention also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the embodiments of the crop life cycle management method based on digital twins described above.

[0165] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0166] The foregoing has provided a detailed description of a method and system for crop lifecycle management based on digital twins, as provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for managing the entire life cycle of crops based on digital twins, characterized in that, include: Acquire multimodal sensing data of crops, including satellite remote sensing images of the planting area and field environmental parameters, and calculate the growth environment index and crop growth index of the corresponding sampling points based on the multimodal sensing data; In the digital twin platform, multiple parallel logical judgment paths are established, and the end of each path corresponds to a status label among reproductive stage, nutritional status, or water shortage degree. The growth environment index and crop growth index of the current sampling point are input into the digital twin platform, and the matching is performed step by step according to the preset branch logic of each logical judgment path. In the same logical judgment path, the judgment weight of the subsequent matching steps is dynamically adjusted according to the completed matching results until the step-by-step matching of all paths is completed, and the status label corresponding to the end of each path is determined. The status labels output by all logical judgment paths are summarized, and the status labels of multiple sampling points within the same plot are spatially aggregated. Based on the degree of aggregation and boundary continuity of each status label in spatial distribution, a health rating map of the corresponding plot is generated and displayed in a virtual twin with hierarchical coloring. Based on the health grading map of each sampling point, a harvesting route map and a storage allocation plan are generated in the virtual twin of the digital twin platform, and the impact of different harvesting sequences on overall revenue is simulated to output the optimal harvesting plan.

2. The method according to claim 1, characterized in that, The growth environment index and crop growth index of the current sampling point are input into the digital twin platform, and matched step by step according to the preset branch logic of each logical judgment path. Within the same logical judgment path, the judgment weight of subsequent matching steps is dynamically adjusted based on the completed matching results until the step-by-step matching of all paths is completed, and the status label corresponding to the end of each path is determined, including: Combine the growth environment index and crop growth index of the current sampling point into a data list and input it into each logical judgment path; In a single logical judgment path, starting from the first-level judgment node, the values ​​in the data list are compared with the branch logic of the node to obtain the matching result and select the subsequent branch; The matching result of the previous node is converted into a correction coefficient and calculated with the preset initial threshold of the next node to obtain the new judgment threshold of the corresponding node. The new judgment threshold is used to match the data list values, and the correction coefficient is updated. The operation of entering the next node, calculating the new threshold and matching is iteratively executed until the end of the path is reached and the corresponding associated status label is obtained. Complete the matching process for all logical judgment paths to obtain the status label corresponding to the current sampling point.

3. The method according to claim 2, characterized in that, The matching result of the previous node is converted into a correction coefficient, and then calculated with the preset initial threshold of the next node to obtain a new judgment threshold for the corresponding node, including: The output value corresponding to the matching result of the previous node is input into a preset coefficient conversion function to calculate the correction coefficient. Read the initial threshold value that has been pre-set for the next decision node, where the initial threshold value is a baseline value; The calculated correction coefficient is multiplied by the read initial threshold, and the corresponding product is set as the new judgment threshold actually used by the next judgment node in this matching process.

4. The method according to claim 1, characterized in that, In the digital twin platform, multiple parallel logical decision paths are established. The end of each path corresponds to a status label from one of the following: reproductive stage, nutritional status, or water scarcity level: In the digital twin platform, three parallel logical judgment paths are established, and the judgment targets of the logical judgment paths are respectively set as reproductive stage, nutritional status, and degree of water shortage. For each logical judgment path, a hierarchical structure consisting of multiple judgment nodes connected sequentially is constructed, wherein each judgment node contains a judgment condition, which is used to compare a specific parameter in the growth environment index and crop growth index with a preset threshold. Based on the comparison results between the parameters and thresholds in the judgment conditions, determine the path branch connecting the current judgment node to the next judgment node; At the final decision node of each logical decision path, a status label is associated, which corresponds to one of the following: reproductive stage, nutritional status, or degree of water shortage.

5. The method according to claim 4, characterized in that, For each logical decision path, a hierarchical structure is constructed consisting of multiple decision nodes connected sequentially, including: For each of the aforementioned logical decision paths, the number of levels in the hierarchical structure is set; Based on the set number of levels, the decision nodes contained in each level are defined in order from top to bottom; Define the connection relationship between decision nodes in the hierarchical structure, so that each output branch of the decision node in the upper level is connected to a specific decision node in the lower level; For each decision node in the hierarchical structure, a decision condition is assigned, in which one or more specific parameters selected from the growth environment index and the crop growth index are specified, and one or more thresholds for comparison are set for each specific parameter.

6. The method according to claim 1, characterized in that, The status labels output from all logical decision paths are aggregated. The status labels from multiple sampling points within the same plot are spatially aggregated. Based on the clustering degree and boundary continuity of each status label in spatial distribution, a health grading map for the corresponding plot is generated and displayed in a virtual twin with grading and coloring, including: Collect the status labels output by each sampling point within the same plot of land for all logical judgment paths to obtain a set of status labels corresponding to each sampling point; Sampling points that are adjacent in location within a plot and have the same set of state labels are grouped into a single independent unit, and all independent units are labeled. Calculate the shared boundary length between each independent unit, and merge independent units whose shared boundary length is greater than a preset continuity threshold to form multiple final aggregated regions; Based on the number of independent units and sampling point density contained in each of the aggregated regions, a health level is assigned to each aggregated region to generate a health grading map of the corresponding plot. On the three-dimensional surface model corresponding to the virtual twin in the digital twin platform, based on the health level map, different colors are used to render and fill the aggregated areas of different health levels.

7. The method according to claim 1, characterized in that, Based on the health grading map of each sampling point, a harvesting route map and warehousing allocation plan are generated in the virtual twin of the digital twin platform. The impact of different harvesting sequences on overall revenue is simulated, and the optimal harvesting plan is output, including: Based on the health level of each aggregated region in the health grading diagram, each aggregated region is marked in the virtual twin, and the marked aggregated regions are connected to generate a closed-loop harvesting route map without intersections. Based on the assumed harvesting sequence in the closed-loop harvesting route map, combined with the estimated yield per unit area corresponding to different health levels, the daily harvest volume is calculated, and a detailed storage allocation plan is generated based on the regional division of storage space. In the digital twin platform, the access order of the aggregation area is changed, the daily harvest volume is recalculated for each changed order, and a corresponding warehousing and allocation plan is generated. The total operation time and estimated total output from the start of harvesting to the completion of warehousing are simulated and calculated. By comparing the total operation time and the estimated total output obtained from different access sequences, the result with the highest estimated total output and the shortest total operation time is selected as the optimal harvesting scheme.

8. A crop lifecycle management system based on digital twins, characterized in that, include: The acquisition module is used to acquire multimodal sensing data of crops, including satellite remote sensing images of the planting area and field environmental parameters, and to calculate the growth environment index and crop growth index of the corresponding sampling point based on the multimodal sensing data. The module is used to establish multiple parallel logical decision paths in the digital twin platform. The end of each path corresponds to a status label among reproductive stage, nutritional status, or water shortage level. The matching module is used to input the growth environment index and crop growth index of the current sampling point into the digital twin platform, and perform step-by-step matching according to the preset branch logic of each logical judgment path. In the same logical judgment path, the judgment weight of the subsequent matching steps is dynamically adjusted according to the completed matching results until the step-by-step matching of all paths is completed, and the status label corresponding to the end of each path is determined. The generation module is used to summarize the status labels output by all logical judgment paths, spatially aggregate the status labels of multiple sampling points within the same plot, generate a health rating map of the corresponding plot based on the degree of aggregation and boundary continuity of each status label in spatial distribution, and display the rating in a virtual twin. The output module is used to generate a harvesting route map and a storage and allocation plan in the virtual twin of the digital twin platform based on the health grading map of each sampling point, and to simulate the impact of different harvesting sequences on the overall revenue, and output the optimal harvesting plan.

9. An electronic device, characterized in that, include: Memory, used to store computer programs; A processor, configured to execute the computer program to implement the steps of the crop lifecycle management method based on digital twins as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, enables the implementation of the crop lifecycle management method based on digital twins as described in any one of claims 1 to 7.

Images