Method and system for dynamic acquisition and processing of agricultural information based on management partition

Abstract

The application discloses a management partition-based agricultural information dynamic acquisition and processing method and system, and belongs to the technical field of agricultural information acquisition and processing, and comprises dynamic management partition construction, partition self-adaptive linkage regulation and control, multi-source agricultural data standardized space-time fusion processing and window linkage regulation and control. The application acquires target farmland space-time dynamic factor data, initially delimits a dynamic management partition graph; determines a differentiated collection strategy according to partition attributes, including multi-terminal synchronous time difference self-adaptive correction and collection frequency adjustment, and synchronously collects multi-dimensional agricultural information through multi-modal terminals; standardizes and corrects multi-source data and performs space-time fusion to generate a standardized data set; extracts internal and external correlation laws and evolution characteristics in the partition, generates differentiated management and control decisions based on overlap proportion dynamic regulation, and outputs the decisions to corresponding terminals for execution.

Description

Technical Field

[0001] This invention relates to the field of agricultural information acquisition and processing technology, and in particular to a method and system for dynamic acquisition and processing of agricultural information based on management zones. Background Technology

[0002] Agricultural information acquisition and processing technology based on management zoning is now widely used in precision agriculture monitoring, farmland production management, and agricultural resource allocation. Existing technologies primarily use fixed administrative divisions, pre-defined plot boundaries, and static soil fertility zoning as basic units. They rely on remote sensing imagery, field sensors, and manual inspections to collect agricultural data such as farmland environment, crop growth, soil moisture, and pests and diseases. Conventional data models are then used to complete zoning data statistics and preliminary analysis, providing fundamental support for farmland zoning management.

[0003] In large-scale, dynamic modern agricultural production scenarios, it is difficult to meet the needs of real-time and precise zoning management. The zoning system is static and fixed, with zoning boundaries, scales, and functional attributes remaining fixed for a long time. This makes it unable to adapt to dynamic scenarios such as farmland topography changes, soil fertility changes, crop rotation, and planting structure adjustments. The lack of adaptive updates and temporal iteration mechanisms causes a disconnect between zoning spatial attributes and actual field production conditions, significantly reducing the matching degree and effectiveness of data collection. Information collection models have defects. Traditional fixed-frequency collection methods cannot achieve differentiated scheduling based on zoning characteristics, crop growth stages, and disaster and pest risks. High-risk areas are prone to missed or delayed information collection, while conventional areas suffer from data redundancy and resource waste. Furthermore, the collection dimensions are singular, failing to comprehensively and multidimensionally depict the overall state of farmland. Multi-source data fusion capabilities are weak. Data acquired from different devices and channels differ in format, accuracy, coordinate systems, and update frequency, lacking a unified correction and spatiotemporal registration mechanism, making deep linkage and fusion of heterogeneous data difficult. Traditional models can only perform simple statistical comparisons, failing to uncover data correlation patterns and spatiotemporal evolution characteristics, and lacking intelligent analysis capabilities. The lack of spatiotemporal linkage analysis means that existing systems mostly process static data at single points in time, failing to achieve time-series tracking and trend prediction. This results in a lag in assessing fluctuations in farmland growth, changes in soil moisture, and the spread of disasters, hindering early warning and dynamic control. Furthermore, poor decision-making adaptability is another issue. A uniform data processing and output standard across the entire region cannot adapt to the soil conditions, planting types, and management needs of different zones, leading to coarse-grained decision-making results and preventing precise zoned policy implementation. This severely restricts the intelligent and refined development of agricultural zoned management. Summary of the Invention

[0004] In view of this, embodiments of the present invention provide a method and system for dynamic acquisition and processing of agricultural information based on management zoning. This system can initially delineate and generate a dynamic management zoning map by acquiring spatiotemporal dynamic factor data of the target farmland; determine differentiated acquisition strategies based on zoning attributes, including multi-terminal synchronous time difference adaptive correction and acquisition frequency adjustment, and synchronously acquire multi-dimensional agricultural information through multi-modal terminals; perform standardization correction and spatiotemporal fusion on multi-source data to generate a standardized dataset; extract the internal and external correlation patterns and evolution characteristics of the zoning, dynamically adjust based on the overlap ratio to generate differentiated management and control decisions, and output them to the corresponding terminals for execution.

[0005] On the one hand, a method for dynamic acquisition and processing of agricultural information based on management zoning is provided. This method includes: acquiring spatiotemporal dynamic factor data of the target farmland area; based on the spatiotemporal dynamic factor data, initially delineating management zoning of the target farmland area to generate a dynamic management zoning map; determining differentiated acquisition strategies for each management zoning based on the zoning attribute information of each management zoning in the dynamic management zoning map; and synchronously acquiring multi-dimensional agricultural information of each management zoning through multi-modal acquisition terminals according to the differentiated acquisition strategies. The differentiated acquisition strategies include determining whether to perform adaptive correction of the dynamic threshold for time difference synchronization between multiple terminals based on the zoning update cycle; if so, determining whether to perform adaptive adjustment of the time difference offset linkage acquisition frequency after adaptive correction; if not… The system directly determines whether to implement adaptive adjustment of the time difference offset linkage acquisition frequency; based on the multi-dimensional agricultural information of each management zone, it performs data standardization correction processing on the multi-dimensional agricultural information of each management zone, and performs spatiotemporal fusion processing on the standardized and corrected multi-source heterogeneous data to generate a standardized spatiotemporal fusion dataset for each management zone; based on the standardized spatiotemporal fusion dataset, it extracts the correlation patterns and spatiotemporal evolution characteristics of agricultural data within and between each management zone, and determines whether to implement dynamic adjustment of linkage overlap ratio. If so, it generates differentiated production control decision instructions for each management zone after dynamic adjustment; if not, it directly generates differentiated production control decision instructions for each management zone and outputs them to the terminal execution devices of the corresponding management zones.

[0006] On the other hand, a dynamic agricultural information acquisition and processing system based on management zoning is provided. This system includes: a dynamic management zoning construction module, a zoning adaptive linkage control module, a multi-source agricultural data standardization spatiotemporal fusion processing module, and a window linkage control module. The dynamic management zoning construction module acquires spatiotemporal dynamic factor data of the target farmland area, and based on this data, initially delineates management zoning for the target farmland area, generating a dynamic management zoning map. The zoning adaptive linkage control module determines differentiated acquisition strategies for each management zoning based on the zoning attribute information of each management zoning in the dynamic management zoning map. According to these differentiated acquisition strategies, it synchronously acquires multi-dimensional agricultural information from each management zoning through multi-modal acquisition terminals. The differentiated acquisition strategies include determining whether to perform adaptive correction of the dynamic threshold for synchronous time difference between multiple terminals based on the zoning update cycle; if so, the system will then perform adaptive correction... The system determines whether to perform adaptive adjustment of the time difference offset linkage acquisition frequency. If not, it directly determines whether to perform adaptive adjustment of the time difference offset linkage acquisition frequency. The multi-source agricultural data standardization spatiotemporal fusion processing module is used to perform data standardization and correction processing on the multi-dimensional agricultural information of each management zone based on the multi-dimensional agricultural information of each management zone. Based on the standardized and corrected multi-source heterogeneous data, it performs spatiotemporal fusion processing to generate a standardized spatiotemporal fusion dataset for each management zone. The window linkage control module is used to extract the correlation patterns and spatiotemporal evolution characteristics of agricultural data within and between each management zone based on the standardized spatiotemporal fusion dataset. It determines whether to perform dynamic control of linkage overlap ratio. If yes, it generates differentiated production control decision instructions for each management zone after dynamic control. If no, it directly generates differentiated production control decision instructions for each management zone and outputs them to the terminal execution devices of the corresponding management zones.

[0007] Beneficial effects The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following: 1. This invention overcomes the limitations of traditional static zoning by constructing a dynamic management zoning system based on spatiotemporal dynamic factor weighted clustering. This system enables annual periodic updates and event-driven trigger updates to management zoning based on dynamic factors such as changes in farmland topography, soil fertility, crop rotation, and minor adjustments to ownership. This ensures that zoning boundaries maintain spatial consistency with actual field production conditions, avoiding the low spatial matching problem caused by static zoning. Through differentiated acquisition strategies driven by zoning attribute information (plot characteristics, growth period, disaster risk, and pest and disease probability), and a linkage-based multi-terminal synchronous time difference dynamic threshold adaptive correction and time difference offset linkage acquisition frequency adaptive adjustment mechanism, differentiated scheduling of encrypted acquisition for high-risk zoning and moderate acquisition for regular zoning is achieved. Multimodal acquisition terminals simultaneously acquire five dimensions of information: soil, crops, meteorology, agricultural operations, and agricultural machinery maintenance. This avoids the risk of missed or delayed acquisition of sudden abnormal information and effectively reduces data redundancy and resource waste caused by high-frequency acquisition of regular zoning, improving the adaptability and real-time performance of multi-dimensional zoning information acquisition.

[0008] 2. This invention addresses the differences in coordinate systems, data formats, acquisition accuracy, and update frequencies among various acquisition devices and channels. Through coordinate system unification correction, format standardization conversion, accuracy normalization correction, and time alignment processing, combined with a multi-source data fusion method using a spatiotemporal cube, it maps heterogeneous multi-source data to a consistent spatiotemporal benchmark, generating a standardized spatiotemporal fusion dataset. This achieves deep fusion and effective linkage of data across zones, time periods, and terminals, solving the pain points of inconsistent data standards and weak fusion processing. This invention introduces a time series prediction model based on Long Short-Term Memory networks, combined with sliding window dynamic sampling technology (adaptive adjustment of window duration, dynamic control of window overlap ratio coefficient with relative position index and rate of change), to perform linked analysis of historical time series data and real-time fusion datasets, deeply mining the correlation patterns and spatiotemporal evolution characteristics of agricultural data within and between management zones.

[0009] 3. Based on the personalized parameters such as soil endowment, planting type, production shortcomings, and control priorities of different management zones, the output results of the general analysis model are offset corrected and threshold recalibrated to generate differentiated production control decision instructions and accurately output them to the terminal execution devices in each zone. This ensures that the analysis results closely match the actual production management needs of each zone, solving the problem of insufficient accuracy in zone control decisions caused by homogeneous output, and truly realizing the precision agriculture goal of zoned policy implementation and precise control. This invention constructs a closed-loop dynamic optimization link from zoned data collection, fusion analysis, decision output to execution effect tracking and feedback. Execution effect data is fed back to the dynamic zoned algorithm and differentiated collection strategy, driving continuous iterative optimization of the zoned system and collection scheme, thereby improving the intelligence of dynamic management of agricultural zones. Attached Figure Description

[0010] Figure 1 A flowchart of a method for dynamic acquisition and processing of agricultural information based on management partitions provided in an embodiment of this application; Figure 2 A flowchart illustrating the adaptive correction of dynamic threshold for synchronization time difference of agricultural information based on management partitions in a linked multi-terminal method for dynamic acquisition and processing of agricultural information provided in this application embodiment; Figure 3 This is a schematic diagram of the structure of the agricultural information dynamic acquisition and processing system based on management partitions provided in the embodiments of this application. Detailed Implementation

[0011] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0012] like Figure 1 The diagram shown is a flowchart of a method for dynamic acquisition and processing of agricultural information based on management partitions, provided in an embodiment of this application. The method includes the following steps: As the first step of this invention, the spatiotemporal dynamic factor data of the target farmland area is obtained, and based on the spatiotemporal dynamic factor data, the management zoning of the target farmland area is initially delineated to generate a dynamic management zoning map.

[0013] It needs to be explained that the specific process for the initial delineation of management zones for the target farmland area is as follows: Spatiotemporal dynamic factor data of the target farmland area are acquired and normalized to generate a standardized dynamic factor matrix. Based on the standardized dynamic factor matrix, a weighted clustering segmentation algorithm is used to initially delineate the management zones of the target farmland area. The spatiotemporal dynamic factor data includes quantitative factors and qualitative factors. Quantitative factors include spatial distribution data of soil nutrient content, topographic slope data, and plot area data. Qualitative factors include planting type data, crop rotation pattern data, and ownership category data. Each type of dynamic factor is assigned an initial weight, and the weighted Euclidean distance between each spatial unit and the cluster center is calculated. The optimization objective is to minimize the sum of squared weighted distances. The target farmland area is divided into several initial management zones through iterative clustering. During the clustering process, the elbow rule is used to determine the optimal number of partitions, and the clustering results are processed by region merging and boundary smoothing to generate an initial administrative partition map. The initial administrative partition map contains the spatial boundaries and initial partition attributes of each initial partition.

[0014] The initial delineation of management zones for the target farmland area also includes: Preset dynamic update trigger conditions, which include timed trigger conditions and event-driven trigger conditions; The timed trigger condition is triggered periodically every year to adapt to the annual administrative division adjustment needs; The event-driven trigger condition is triggered when the change of any dynamic factor exceeds a preset threshold. The preset threshold is set according to the coefficient of variation or production management sensitivity of each factor. The change of the dynamic factor is calculated by comparing the current measured data with the historical data used in the last partition update. When the dynamic update triggering condition is met, the real-time spatiotemporal dynamic factor data of the current moment is obtained, the normalization process and the initial management partition delineation are repeatedly performed, and all management partitions are re-delineated. Each dynamically updated management zone is assigned a unique zone code, and its spatial boundary vector, zone scale parameters, and functional attribute labels are recorded. The functional attribute labels include the dominant planting type, soil type, and control priority level.

[0015] In this embodiment, when initially delineating and dynamically updating the refined management zoning of the target farmland area, multi-source spatiotemporal dynamic factor data is first collected across the entire target farmland area. This type of spatiotemporal dynamic factor data includes both quantitative factors that can be quantified numerically and qualitative factors that can be represented by discrete categories. Specifically, the quantitative factors include spatial distribution data of soil nutrient content, terrain slope data, and plot area data, which can objectively reflect the basic endowment and topographic conditions of the plots. The qualitative factors specifically include planting type data, crop rotation pattern data, and farmland ownership category data, which can reflect the differences in farmland production patterns and management attributes. After completing the collection of original data of multi-category spatiotemporal dynamic factors, all collected data are processed... The quantitative and qualitative factors, which are inconsistent in terms of dimensions, magnitude, and representation, are uniformly standardized and normalized to eliminate computational interference caused by differences in physical meaning, numerical range, and data dimensionality. This results in a standardized dynamic factor matrix with unified dimensions, consistent scale, and collaborative computation capabilities, which serves as the foundational dataset for farmland zoning clustering analysis. Subsequently, based on the constructed standardized dynamic factor matrix, a weighted clustering algorithm is used to perform refined initial management zoning of the overall target farmland area. Specifically, for each type of spatiotemporal dynamic factor participating in the zoning operation, a unique initial weight is assigned according to its contribution to the differentiation of farmland production and the homogenization of zoning, accurately defining the initial management zoning. To reflect the primary and secondary influences of different factors in farmland zoning, this method further traverses all spatial grid cells of the farmland, calculating the weighted Euclidean distance between each spatial cell and the preset cluster center. Minimizing the sum of squared weighted distances of all spatial cells in the entire region is the core optimization objective. Through multiple iterative clustering operations, the location of cluster centers and cell affiliation are continuously corrected, gradually aggregating spatial cells with similar attributes and homogeneous production conditions into the same plot unit. Ultimately, the overall target farmland area is autonomously divided into several spatially continuous, attribute-homogeneous, and boundary-regular initial management zones. To effectively avoid the subjective errors, overly detailed, redundant, or mixed zones caused by manually setting the number of zones, the iterative clustering operation... Throughout the process, the elbow rule is introduced to scientifically solve and determine the optimal number of partitions, ensuring that the number of partitions is highly matched with the differentiated characteristics of farmland plots. At the same time, regional merging is carried out for the scattered and fragmented patches and adjacent homogeneous fragmented partitions after the initial clustering, and boundary smoothing optimization is performed on the jagged and irregular boundaries of the partitions to effectively eliminate invalid fragmented partitions and correct boundary distortion problems. Finally, an initial management partition map with a regular structure, clear hierarchy, and close fit to the actual field plot morphology is generated. This initial management partition map fully carries the spatial boundary vector information and basic initial partition attribute information of each initial management partition, providing a precise spatial basis for subsequent differentiated collection strategy configuration, accurate matching of agricultural data, and partition situation analysis.To address the technical shortcomings of traditional static farmland management zoning, which cannot adapt to the temporal evolution of farmland production conditions, this embodiment further configures a dual-mode dynamic update triggering mechanism for the initially delineated management zones. Specifically, it presets two types of dynamic update triggering conditions: timed triggering conditions and event-driven triggering conditions. The timed triggering conditions employ an annual periodic automatic triggering mode, capable of adapting to the routine zoning update needs of annual adjustments to farmland planting structures, annual changes in soil fertility, and annual switching of crop rotation patterns, achieving an annual update of the entire zoning baseline. The event-driven triggering conditions are a real-time dynamic monitoring triggering mode, targeting each... The system employs a spatiotemporal dynamic factor approach, setting independent preset thresholds for changes based on each factor's data variation coefficient and its sensitivity to farmland production management. Different factors correspond to different judgment threshold standards, effectively balancing the fluctuation characteristics of different factors with their production impact weight. During daily farmland production, the system compares the real-time data of each spatiotemporal dynamic factor obtained from current measurements with the historical benchmark data used in the previous partition update, calculating the real-time changes of each dynamic factor. When the real-time change of any spatiotemporal dynamic factor exceeds its corresponding preset threshold, the system immediately determines that the event-driven update condition is met and automatically triggers the update. The system employs a comprehensive management zoning iterative update process. When the system detects that any update condition—either a timed trigger or an event-driven trigger—is met, it immediately collects real-time spatiotemporal dynamic factor data for the target farmland area. It then repeatedly executes the complete initial zoning process, including factor normalization, standardized dynamic factor matrix construction, weighted clustering, optimal zoning quantity calculation, region merging, and boundary smoothing. This process comprehensively iterates and updates the boundary positions, zoning scales, and internal attributes of all management zoning areas within the target farmland. After completing the dynamic zoning update, each updated management zoning area is then individually... Each zone is assigned a unique partition code, enabling unique identification of a single zone and unified management of all zones. Simultaneously, it accurately records and stores the spatial boundary vector parameters, zone scale quantization parameters, and corresponding multi-functional attribute tags for each updated management zone. These attribute tags specifically include core attributes such as the zone's dominant planting type, soil type, and control priority level. This ensures that each updated management zone possesses independent, complete, and traceable spatial and production attribute parameters, continuously guaranteeing a high degree of compatibility between dynamic management zones and real-time farmland topography, soil endowment, planting structure, crop rotation status, ownership information, and production control level.

[0016] As the second step of this invention, based on the partition attribute information of each management partition in the dynamic management partition map, a differentiated acquisition strategy is determined for each management partition. According to the differentiated acquisition strategy, multi-dimensional agricultural information of each management partition is synchronously acquired through a multi-modal acquisition terminal. The differentiated acquisition strategy includes determining whether to perform adaptive correction of the dynamic threshold for synchronized time difference between multiple terminals based on the partition update cycle. If so, it determines whether to perform adaptive adjustment of the time difference offset linkage acquisition frequency after adaptive correction; otherwise, it directly determines whether to perform adaptive adjustment of the time difference offset linkage acquisition frequency. The partition attribute information includes plot characteristic parameters, crop growth stage information, meteorological disaster risk level, and pest and disease outbreak probability level. The differentiated acquisition strategy includes configuration information for acquisition frequency, acquisition time period, and acquisition method combination. The multi-dimensional agricultural information includes soil parameter information, crop growth information, meteorological environment information, agricultural operation information, and agricultural machinery operation and maintenance information.

[0017] It should be understood that Figure 2 The specific process of the linkage multi-terminal synchronization time difference dynamic threshold adaptive correction flowchart of the agricultural information dynamic acquisition and processing method based on management partition provided in this application embodiment is as follows: A correlation mapping function between the partition update cycle and the multi-terminal data synchronization time difference threshold is pre-constructed. The partition update cycle is divided into levels and matched one-to-one with the corresponding level's benchmark synchronization time difference threshold. Simultaneously, it is clarified that the partition update cycle represents the refresh frequency of management partition space and production attribute information, and the multi-terminal synchronization time difference threshold represents the maximum allowable deviation of the collection timestamp of the multi-modal acquisition terminal. Then, the current partition update cycle is read in real time and compared with the open interval formed by the reference lower limit and reference upper limit. If the current cycle is within the interval, the corresponding level's benchmark threshold is directly retrieved as the target synchronization time difference threshold. If the cycle is less than or equal to the reference lower limit, the cycle is substituted into the mapping function to obtain a downward adjustment coefficient, which is multiplied by the benchmark threshold to calculate the target threshold. If the cycle is greater than or equal to the reference upper limit, the cycle is substituted into the function to obtain an upward adjustment coefficient, which is multiplied by the benchmark threshold to calculate the target threshold.

[0018] Furthermore, the specific process for determining whether to perform dynamic threshold adaptive correction of synchronization time difference across multiple terminals is as follows: A mapping function is pre-established to associate the partition update cycle with the multi-terminal data synchronization time difference threshold, and the partition update cycle of different levels is matched one by one with the corresponding benchmark multi-terminal data synchronization time difference threshold. The partition update cycle is used to characterize the refresh frequency of management partition space and production attribute information, and the multi-terminal data synchronization time difference threshold is used to characterize the maximum deviation value between the collection timestamps of each terminal when multi-modal acquisition terminals synchronously collect multi-dimensional agricultural information of the same management partition. The system reads the current partition update cycle in real time, compares the partition update cycle with the preset update cycle reference range, and dynamically adjusts the multi-terminal data synchronization time difference threshold based on the comparison results.

[0019] The specific steps for dynamically adjusting the multi-terminal data synchronization time difference threshold based on the comparison results are as follows: If the current partition update cycle is within the preset update cycle reference range, the benchmark multi-terminal data synchronization time difference threshold matching the corresponding level is retrieved as the target multi-terminal data synchronization time difference threshold. The preset update cycle reference range represents the open interval formed by the preset update cycle reference lower limit and the preset update cycle reference upper limit. If the current partition update cycle is less than or equal to the preset update cycle reference lower limit, the current partition update cycle is input into the associated mapping function, and the time difference threshold reduction coefficient is output. The time difference threshold reduction coefficient is multiplied by the benchmark multi-terminal data synchronization time difference threshold of the corresponding level to obtain the target multi-terminal data synchronization time difference threshold. If the current partition update cycle is greater than or equal to the preset update cycle reference upper limit, the current partition update cycle is input into the associated mapping function, and the time difference threshold adjustment coefficient is output. The time difference threshold adjustment coefficient is multiplied by the benchmark multi-terminal data synchronization time difference threshold of the corresponding level to obtain the target multi-terminal data synchronization time difference threshold.

[0020] In this embodiment, to achieve high-precision time-series collaborative acquisition by multi-modal acquisition terminals under different farmland zoning update conditions, and to avoid the technical problem that a fixed synchronization time difference threshold cannot adapt to the dynamic update rhythm of the management zoning, leading to time-series misalignment of multi-source acquisition data or waste of computing resources, the system pre-constructs a dedicated correlation mapping function between the zoning update cycle and the multi-terminal data synchronization time difference threshold that can represent the dynamic correspondence. Furthermore, through a tiered matching method, different levels and refresh frequencies of zoning update cycles are bound one-to-one with the corresponding accuracy level of the benchmark multi-terminal data synchronization time difference threshold, thus achieving synchronization between the zoning update rhythm and the terminal synchronization accuracy. The quantitative correspondence is defined in this embodiment as follows: the partition update cycle is specifically used to quantify the frequency of iterative refreshes of the spatial boundaries, partition scale, and production attribute information of each dynamic management partition. This directly reflects the drastic dynamic changes in the production status of the target farmland plot. A shorter partition update cycle indicates more frequent changes in farmland partition attributes and a more unstable field production status, while a longer cycle indicates that the farmland partition status tends to be stable in the long term. The multi-terminal data synchronization time difference threshold is used to precisely limit the time difference between multi-modal acquisition terminals and other multi-dimensional agricultural information such as soil parameters, crop growth, meteorological environment, agricultural operations, and agricultural machinery maintenance within the same management partition. The maximum allowable time deviation between data timestamps collected by different acquisition terminals is a core accuracy constraint parameter that ensures the accuracy of subsequent spatiotemporal fusion and the effectiveness of time-series correlation analysis of multi-source heterogeneous agricultural data. During routine field data collection and management, the system reads the latest effective partition update cycle parameter of the current target management partition in real time and compares the real-time acquired partition update cycle value with a preset update cycle reference open interval composed of a preset lower limit and a preset upper limit. Based on the interval position of the partition update cycle, the system achieves adaptive multi-terminal data synchronization time difference threshold. Differentiated dynamic adjustment: Specifically, when the current partition update cycle read in real time is strictly within the open interval formed by the preset update cycle reference lower limit and the preset update cycle reference upper limit, it is determined that the current farmland partition update rhythm is in a normal and stable update condition, the partition attribute change range is moderate and the change rate is stable, and there is no need to offset the benchmark synchronization time difference threshold. The system directly retrieves the benchmark multi-terminal data synchronization time difference threshold that matches the current partition update cycle level as the final effective target multi-terminal data synchronization time difference threshold, ensuring that the multi-terminal data synchronization accuracy meets the time series analysis requirements while maintaining stable terminal scheduling computing power overhead.When the real-time read current partition update cycle is less than or equal to the preset lower limit of the update cycle reference, the current managed partition is determined to be in a high-frequency dynamic update condition. Attributes such as soil endowment, planting structure, disaster risk, and crop growth status change frequently, and field production fluctuates dramatically, requiring higher accuracy in the time-series alignment of multi-source data. In this case, the system inputs the current real-time partition update cycle value into a pre-built association mapping function. Through function calculation, it adaptively outputs a time difference threshold reduction coefficient adapted to the high-frequency update condition. This reduction coefficient is then multiplied by the benchmark multi-terminal data synchronization time difference threshold corresponding to the current partition level to obtain a more accurate and strictly constrained target multi-terminal data synchronization time difference threshold. By actively narrowing the allowable range of terminal acquisition timestamp deviation, it effectively suppresses the time-series data misalignment problem caused by asynchronous acquisition from multiple terminals in high-frequency changing scenarios, significantly improving the spatiotemporal matching consistency of multi-source data in dynamically changing partitions. When the update cycle is greater than or equal to the preset upper limit of the update cycle reference, the current management partition is determined to be in a low-frequency steady-state update condition. The partition's spatial attributes and production attributes remain stable over a long period, and there are no significant fluctuations in field production status. Therefore, there is no need to maintain stringent synchronization constraints requiring high precision and high computing power. In this case, the system also inputs the current real-time partition update cycle value into the correlation mapping function. The function adaptively solves and outputs an adjustment coefficient for the time difference threshold adapted to the steady-state condition. This adjustment coefficient is then multiplied by the corresponding benchmark multi-terminal data synchronization time difference threshold to obtain a moderately relaxed target multi-terminal data synchronization time difference threshold. This reasonably relaxes the multi-terminal synchronization deviation constraint without affecting the accuracy of steady-state partition data analysis, reducing the ineffective computing power consumption and data transmission pressure caused by high-frequency precision synchronization of multi-modal terminals. Ultimately, it achieves adaptive matching of the optimal terminal synchronization time difference accuracy based on the dynamic update frequency of the management partition, completing the intelligent linkage adaptation between the dynamic change characteristics of the partition and the multi-terminal data synchronization accuracy.

[0021] It should be further explained that the specific process for determining whether to perform adaptive adjustment of the time difference offset linkage acquisition frequency is as follows: A one-to-one mapping relationship between the dynamic threshold correction offset for multi-terminal synchronization time difference and the collection interval adjustment ratio is pre-built. The timestamps of multi-terminal synchronization collection in each management partition are collected in real time. The actual collection time difference is calculated. The difference between the actual collection time difference and the currently effective dynamic threshold for multi-terminal synchronization time difference is calculated to obtain the real-time dynamic threshold correction offset for multi-terminal synchronization time difference. The real-time multi-terminal synchronization time difference dynamic threshold correction offset is compared with the threshold correction offset critical interval, and the collection frequency is dynamically adjusted according to the comparison result. The collection frequency is used to characterize the collection interval of multi-dimensional agricultural information in each management area. If the real-time multi-terminal synchronization time difference dynamic threshold correction offset is within the threshold correction offset critical interval, then the sampling frequency benchmark value will be used as the target sampling frequency. The threshold correction offset critical interval represents the open interval formed by the threshold correction offset critical lower limit and the threshold correction offset critical upper limit.

[0022] The specific procedure for dynamically adjusting the collection frequency based on the comparison results is as follows: If the real-time multi-terminal synchronization time difference dynamic threshold correction offset is less than or equal to the threshold correction offset critical lower limit, it is determined that the multi-terminal acquisition synchronization deviation is small. Then, the current dynamic threshold correction offset is input into the corresponding mapping relationship, and the first-level acquisition interval stretching ratio is output. The acquisition interval benchmark value and the first-level acquisition interval stretching ratio are multiplied to obtain the target acquisition interval, so as to extend the acquisition interval duration of agricultural information in each dimension and reduce the overall acquisition frequency of the partition. If the real-time multi-terminal synchronization time difference dynamic threshold correction offset is greater than or equal to the threshold correction offset critical upper limit, it is determined that the multi-terminal acquisition synchronization deviation is large. Then, the current dynamic threshold correction offset is input into the corresponding mapping relationship, and the first-level acquisition interval compression ratio is output. The acquisition interval benchmark value and the first-level acquisition interval compression ratio are multiplied to obtain the target acquisition interval, so as to shorten the acquisition interval time of agricultural information in various dimensions and increase the overall acquisition frequency of the partition.

[0023] In this embodiment, a one-to-one quantitative mapping relationship is pre-constructed between the dynamic threshold correction offset of multi-terminal synchronization time difference and the acquisition interval adjustment ratio. This ensures that different magnitudes of timing synchronization deviation can be accurately matched with the specific acquisition frequency adjustment range. During routine intelligent data acquisition operations in farmland, the system continuously collects the synchronization acquisition timestamp data of multimodal acquisition terminals deployed in each management zone in real time. By performing difference calculations on the same-source acquisition timestamps of multiple terminals, the actual acquisition time difference between terminals is obtained. Furthermore, this real-time calculated actual acquisition time difference is compared with the current zone. Under operating conditions, the adaptively corrected and effective multi-terminal synchronization time difference dynamic threshold is used for precise difference calculation. This quantifies the real-time multi-terminal synchronization time difference dynamic threshold correction offset, which accurately represents the degree of current multi-terminal timing synchronization deviation. The sign and magnitude of this offset value directly reflect the degree of redundancy and excess in the multi-terminal acquisition timing alignment, providing a quantitative basis for the intelligent dynamic adjustment of subsequent acquisition frequency. In this embodiment, the acquisition frequency specifically refers to the information on soil parameters, crop growth, meteorological environment, agricultural operations, and agricultural machinery maintenance in each management zone. The standardized collection interval for multi-dimensional agricultural information is determined by the length of the collection interval; a longer interval results in a lower collection frequency, while a shorter interval results in a higher collection frequency. After obtaining the real-time dynamic threshold correction offset, the system imports this offset value into a preset threshold correction offset critical open interval, defined by the lower and upper threshold correction offset thresholds, to perform interval comparison and judgment. The system adaptively and dynamically adjusts the collection frequency based on the differences in operating conditions within the offset interval. Specifically, when the real-time multi-terminal synchronization time difference dynamic threshold correction offset is strictly within the threshold correction offset range... When the current multi-terminal acquisition synchronization deviation is within the open interval formed by the lower limit of the quantity criticality and the upper limit of the threshold correction offset, it is determined that the timing alignment accuracy of the multi-source acquisition data is moderate and fully meets the accuracy requirements of farmland data analysis, feature extraction and trend prediction. There is no problem of excessive synchronization accuracy or excessive timing deviation. There is no need to dynamically correct the preset acquisition frequency. The system directly determines the pre-configured acquisition frequency benchmark value as the target acquisition frequency of the current partition, maintains the stability and balance of the agricultural information acquisition density of the partition, and ensures the continuity and effectiveness of routine farmland monitoring data.When the real-time multi-terminal synchronization time difference dynamic threshold correction offset is less than or equal to the threshold correction offset critical lower limit, it is determined that the current multi-terminal acquisition synchronization deviation is extremely small, the terminal timing synchronization accuracy has a high margin, and the spatiotemporal alignment consistency of multi-source data is excellent. If the high-frequency dense acquisition mode continues to be used, a large amount of duplicate and redundant data will be generated, resulting in an ineffective waste of terminal acquisition energy consumption, data transmission bandwidth, and cloud storage computing power. At this time, the system substitutes the current real-time dynamic threshold correction offset into the preset mapping relationship, adaptively matches and outputs the corresponding first-level acquisition interval stretching ratio. By multiplying the preset acquisition interval benchmark value with the first-level acquisition interval stretching ratio, the target acquisition interval adapted to the current high-precision synchronization condition is obtained, effectively extending the acquisition interval time of agricultural information in various dimensions, reasonably reducing the overall acquisition frequency of the corresponding management area, and significantly reducing invalid acquisition operations without affecting the accuracy of farmland status monitoring and data analysis results, thus optimizing the energy consumption of terminal equipment and the allocation of system computing resources. When the real-time multi-terminal synchronization time difference dynamic threshold correction offset is greater than or equal to the threshold correction offset critical upper limit, it is determined that the current multi-terminal acquisition synchronization deviation significantly exceeds the standard, and the problem of multi-modal terminal acquisition timing misalignment is prominent. Relying solely on existing acquisition data is insufficient to accurately depict the real-time dynamic changes in farmland, easily leading to problems such as lagging trend analysis and biased management decisions. In this case, the system inputs the current real-time dynamic threshold correction offset into a preset mapping relationship, adaptively outputs the corresponding first-level acquisition interval compression ratio, and multiplies the acquisition interval benchmark value with the first-level acquisition interval compression ratio to obtain a smaller target acquisition interval. This effectively shortens the acquisition interval time for agricultural information in various dimensions, proactively increases the overall acquisition frequency of the corresponding management zone, and compensates for data defects caused by multi-terminal timing synchronization deviation by increasing the spatiotemporal sampling density. It also compensates for the adverse effects of timing alignment errors on subsequent data fusion, feature mining, and trend prediction, ultimately achieving adaptive dynamic matching of the optimal acquisition frequency based on multi-terminal real-time synchronization conditions.

[0024] As the third step of the present invention, based on the multi-dimensional agricultural information of each management zone, data standardization and correction processing is performed on the multi-dimensional agricultural information of each management zone, and spatiotemporal fusion processing is performed on the standardized and corrected multi-source heterogeneous data to generate a standardized spatiotemporal fusion dataset for each management zone.

[0025] It should be understood that after completing the full-domain collection of multi-dimensional agricultural information for each differentiated management zone, systematic standardization correction and high-precision spatiotemporal fusion processing are carried out on the multi-source heterogeneous agricultural data synchronously acquired by multi-modal collection terminals, covering soil parameter information, crop growth information, meteorological environment information, agricultural operation information, and agricultural machinery operation and maintenance information. This is to construct a unified, standardized, and spatiotemporally aligned standardized spatiotemporal fusion dataset. The specific implementation process is as follows: Due to the inconsistencies in hardware models, collection principles, sampling accuracy, output formats, and collection time intervals of various collection terminals deployed in the field, the original multi-dimensional agricultural information obtained suffers from inconsistent dimensions and numerical values. Numerous issues, including large differences in magnitude, interference from abnormal noise, misaligned timestamps, spatial coordinate offsets, missing data, and heterogeneous field formats, prevent direct application to cross-dimensional correlation analysis, time-series evolution feature mining, and regional trend assessment. Therefore, a comprehensive data standardization and correction process is first required for all original multi-dimensional agricultural information corresponding to each management zone. This includes multiple preprocessing operations such as abnormal data removal, missing data completion, unit normalization, coordinate system registration, timestamp calibration, and unified data accuracy correction. Furthermore, anomalies caused by sensor drift, environmental interference, and momentary equipment malfunctions are eliminated using equipment error calibration algorithms. The temporal interpolation completion algorithm fills in the gaps in short-term data collection. It performs global normalization and scaling for heterogeneous indicators of different dimensions and magnitudes, such as soil nutrients, temperature and humidity, crop height, vegetation index, and operation frequency. Simultaneously, it uniformly transforms the spatial coordinates of all terminal-collected data to the benchmark geographic coordinate system of the target farmland management zone, and uniformly calibrates the timestamps of scattered and asynchronous multi-source data to the system's standard time series benchmark. This completely eliminates the data heterogeneity barriers caused by multi-device, multi-dimensional, and multi-time series collection, resulting in standardized and corrected agricultural data with unified format, consistent scale, unified spatiotemporal benchmark, and reliable and valid values. Based on the completed data standardization and correction processing... Based on this, further refined spatiotemporal fusion processing is carried out on the preprocessed multi-source standardized data within each management zone. Relying on the spatial grid matching and temporal sequence alignment fusion mechanism, the dispersed agricultural data from different collection terminals, different collection dimensions, and different collection times within the same management zone are reconstructed by spatial dimension superposition and temporal dimension association. This achieves deep coupling and matching of point data, time series data, and spatial plot data, so that the agricultural data corresponding to each management zone can achieve accurate spatial location correspondence, continuous temporal sequence connection, and multi-dimensional information complementarity and fusion, effectively solving the problems of fragmentation, one-sided information, and spatiotemporal mismatch of traditional single-dimensional data.Through the aforementioned hierarchical and step-by-step standardized correction and temporal-spatial joint fusion processing, a standardized spatiotemporal fusion dataset is ultimately generated for each management zone. This dataset possesses high spatiotemporal consistency, high data integrity, and high-dimensional comprehensiveness. It comprehensively carries all-dimensional spatiotemporal information on dynamic changes in soil endowment, temporal evolution of crop growth, continuous fluctuations in field weather, agricultural operation status, and agricultural machinery operation trajectories for each zone. This provides high-quality, highly reliable, and spatiotemporally unified basic data support for subsequent time-series dynamic analysis models to accurately extract agricultural data correlation patterns, spatiotemporal evolution characteristics, crop growth and soil moisture fluctuation trends, and disaster spread trends within and between zones. This ensures the accuracy and reliability of subsequent trend predictions and zone-specific management and control decision-making outputs from the data source.

[0026] As the fourth step of this invention, based on the standardized spatiotemporal fusion dataset, the correlation patterns and spatiotemporal evolution characteristics of agricultural data within and between each management zone are extracted, and it is determined whether to perform dynamic adjustment of the linkage overlap ratio. If so, differentiated production control decision instructions for each management zone are generated after dynamic adjustment; otherwise, differentiated production control decision instructions for each management zone are directly generated and output to the terminal execution device of the corresponding management zone.

[0027] It should be noted that the specific process for determining whether to implement dynamic adjustment of the overlap ratio is as follows: The minimum sliding window duration threshold, the maximum sliding window duration threshold, the maximum overlap ratio coefficient, and the minimum overlap ratio coefficient are pre-configured. The minimum overlap ratio coefficient and the maximum overlap ratio coefficient are between 0 and 1, and the minimum overlap ratio coefficient is less than the maximum overlap ratio coefficient. Obtain the partition attribute data of the management partition corresponding to the current prediction task, and obtain the sliding window duration of the current time series analysis based on the partition attribute data. The partition attribute data includes the crop growth stage, agricultural data update frequency, and farmland status change rate. Based on the current sliding window duration, the minimum sliding window duration threshold, and the maximum sliding window duration threshold, the relative position index of the current window duration within the preset threshold range is obtained; If the current sliding window duration is within the sliding window duration threshold range, the corresponding index value is calculated according to the relative position of the range. The sliding window duration threshold range represents the closed interval formed by the minimum sliding window duration threshold and the maximum sliding window duration threshold. If the current sliding window duration is less than the minimum sliding window duration threshold, the relative position index is determined to be zero. If the current sliding window duration exceeds the maximum sliding window duration threshold, the relative position index is determined to be a unit extreme value.

[0028] Determining whether to implement dynamic adjustment of the overlap ratio also includes: Based on the calculated relative position index, the initial adaptive overlap ratio coefficient is obtained by dynamic matching through a preset monotonically increasing nonlinear mapping rule. The monotonically increasing nonlinear mapping rule is used to characterize the corresponding matching limit overlap ratio coefficient when the relative position index reaches an extreme value. The relative position index is divided into three intervals by pre-setting two inflection points. Different intervals correspond to the overlap ratio coefficients, and the real-time farmland status change rate index of the current management zone is obtained. The change rate index includes at least one of soil moisture change rate, crop growth change rate, and meteorological element change rate. The initial adaptive overlap ratio coefficient is dynamically corrected based on the real-time change rate index to obtain the final overlap ratio coefficient. If the real-time rate of change exceeds the preset high change threshold, the initial overlap ratio coefficient is incrementally corrected and the corrected value is limited to not exceeding the preset maximum overlap ratio coefficient. If the real-time rate of change index is lower than the preset low change threshold, the initial overlap ratio coefficient is reduced and the corrected value is limited to not be lower than the preset minimum overlap ratio coefficient. Based on the current sliding window duration and the final overlap ratio coefficient, the sliding step size of the time series window is calculated. The obtained sliding step size is then used to complete the sliding window sampling on the time series agricultural data sequence, thereby completing the extraction of agricultural data correlation patterns and spatiotemporal evolution characteristic analysis for each management zone.

[0029] In this embodiment, a dynamic adjustment mechanism for the overlap ratio of a linked sliding window is constructed to adapt to the needs of differentiated time-series analysis of farmland zoning. In specific implementation, the system first pre-configures the threshold and coefficient benchmark parameters required for dynamic time-series analysis. Specifically, this includes the minimum sliding window duration threshold and the maximum sliding window duration threshold used to constrain the effective time span of the time-series analysis, as well as the maximum and minimum overlap ratio coefficients of the sliding window used to limit the adjustment range of the window overlap ratio. The values ​​of the minimum and maximum overlap ratio coefficients are strictly limited to between 0 and 1, and the minimum overlap ratio is always guaranteed. The coefficient value is less than the maximum overlap ratio coefficient value, thus providing a legal and effective numerical constraint range for subsequent dynamic adjustment of the overlap ratio. When conducting the task of analyzing the correlation patterns and spatiotemporal evolution characteristics of agricultural time-series data in various management zones, the system first retrieves the partition attribute data of the target management zone corresponding to the current analysis and prediction task. This partition attribute data specifically includes three core parameters that can comprehensively characterize the dynamic evolution characteristics of farmland: crop growth stage, agricultural data update frequency, and farmland state change rate. Based on the above multi-dimensional partition attribute data, the system comprehensively determines and obtains a time-series analysis sliding window that is suitable for the current production conditions of the partition. The duration ensures precise matching between the window's basic analysis scale and the real-time production status, data update rhythm, and crop growth characteristics of the partition. Based on this, the system compares the currently effective sliding window duration with preset minimum and maximum sliding window duration thresholds, quantifying and calculating the relative position index of the current window duration within the preset threshold closed interval. Specifically, the judgment logic is as follows: when the current sliding window duration falls exactly within the threshold closed interval formed by the minimum and maximum sliding window duration thresholds, the system accurately calculates the relative position index according to the interval ratio of the current window duration. The corresponding continuously changing relative position index value is calculated to achieve a refined quantitative representation of the window scale. When the current sliding window duration is less than the preset minimum sliding window duration threshold, it is determined that the current situation belongs to an ultra-short time window analysis scenario, with a small data sample size and strong sensitivity to time series fluctuations. The relative position index is directly fixed to zero to match the control benchmark of the extreme short time window. When the current sliding window duration is greater than the preset maximum sliding window duration threshold, it is determined that the current situation belongs to an ultra-long time window analysis scenario, with a long time series coverage period and significant steady-state characteristics of the data. The relative position index is directly fixed to a unit extreme value to match the control benchmark of the ultra-long time series window.After obtaining the accurate relative position index, the system uses a pre-set monotonically increasing nonlinear mapping rule to match the initial value of the overlap coefficient. This nonlinear mapping rule has the constraint characteristic of a one-to-one correspondence between the exponential extreme value and the limit overlap coefficient. It can adaptively match the relative position differences of different window scales to obtain a smooth transition and continuously changing initial adaptive overlap ratio coefficient, effectively avoiding the feature extraction discontinuity problem caused by fixed parameters. At the same time, the system divides the complete relative position index interval into three continuous control intervals by pre-setting two segmented inflection points. Different overlap ratio adjustment logics are matched for different intervals to achieve layered differentiated adaptation of short, medium and long time series windows. To further eliminate single Relying solely on window scale adjustment leads to limitations in adapting to changing conditions. The system synchronously collects real-time farmland status change rate indicators for the current management zone. These indicators encompass at least one core dynamic parameter from soil moisture change rate, crop growth change rate, and meteorological element change rate, accurately reflecting the instantaneous volatility of current farmland production conditions. Based on these real-time change rate indicators, the initial adaptive overlap ratio coefficient is dynamically corrected, ultimately outputting a final overlap ratio coefficient adapted to real-time farmland conditions. Specifically, the correction logic is as follows: when the real-time farmland change rate indicator exceeds a preset high change threshold, it is determined that the current farmland status in the zone is fluctuating drastically and exhibits frequent abrupt changes, requiring increased time-series sampling. To ensure continuity, the initial overlap ratio coefficient is incrementally corrected, while strictly limiting the corrected coefficient value to no more than the preset maximum overlap ratio coefficient. This guarantees that time series data remains continuous without jumps or omissions of abrupt changes in high-frequency change scenarios. When the real-time change rate of farmland is lower than the preset low change threshold, it is determined that the current farmland production status of the current partition is approaching a steady state, the data fluctuations are mild, and there is no significant abrupt change information. The window overlap can be appropriately reduced to reduce redundant calculations. Therefore, the initial overlap ratio coefficient is reduced, and the corrected coefficient value is strictly limited to no less than the preset minimum overlap ratio coefficient. This reduces the computational cost of model iteration while ensuring the accuracy of basic time series analysis. After completing all dynamic corrections and determining the maximum overlap ratio coefficient, the final step is to adjust the overlap ratio coefficient accordingly. After determining the final overlap ratio, the system accurately calculates the sliding step size for a single slide of the time-series window by combining the currently effective sliding window duration with the final overlap ratio. Based on this adaptively updated sliding step size, the system performs segmented sampling and truncation of the sliding window on the continuous time-series agricultural data sequence. Relying on a scale-adaptive and condition-matched time-series sampling mechanism, it comprehensively extracts the data change patterns within each management zone, the spatiotemporal correlations between zones, and the long- and short-term spatiotemporal evolution characteristics. This comprehensively improves the accuracy, continuity, and computational efficiency of farmland time-series situation analysis, achieving an intelligent upgrade from fixed constant control of time-series analysis window parameters to multi-factor adaptive control of zone attributes, window scale, and farmland dynamic change rate.

[0030] like Figure 3The diagram shown is a structural schematic of the agricultural information dynamic acquisition and processing system based on management partitions provided in this application embodiment. It includes: a dynamic management partition construction module, a partition adaptive linkage control module, a multi-source agricultural data standardization spatiotemporal fusion processing module, and a window linkage control module. The dynamic management partition construction module is used to acquire spatiotemporal dynamic factor data of the target farmland area, and based on this data, to initially delineate management partitions for the target farmland area, generating a dynamic management partition map. The partition adaptive linkage control module is used to determine differentiated acquisition strategies for each management partition according to the partition attribute information of each management partition in the dynamic management partition map. Based on these differentiated acquisition strategies, it synchronously acquires multi-dimensional agricultural information of each management partition through multi-modal acquisition terminals. The differentiated acquisition strategies include determining whether to perform dynamic threshold adaptive correction of the linkage multi-terminal synchronous time difference based on the partition update cycle; if so, it then performs adaptive... After correction, it should be determined whether to perform adaptive adjustment of the time difference offset linkage acquisition frequency. If not, it should be determined directly whether to perform adaptive adjustment of the time difference offset linkage acquisition frequency. The multi-source agricultural data standardization spatiotemporal fusion processing module is used to perform data standardization correction processing on the multi-dimensional agricultural information of each management zone based on the multi-dimensional agricultural information of each management zone. Based on the standardized correction, spatiotemporal fusion processing is performed to generate a standardized spatiotemporal fusion dataset for each management zone. The window linkage control module is used to extract the correlation patterns and spatiotemporal evolution characteristics of agricultural data within and between each management zone based on the standardized spatiotemporal fusion dataset, and determine whether to perform dynamic control of linkage overlap ratio. If yes, it generates differentiated production control decision instructions for each management zone after dynamic control. If no, it directly generates differentiated production control decision instructions for each management zone and outputs them to the terminal execution devices of the corresponding management zones.

[0031] In this embodiment, the system consists of four core functional modules: a dynamic management partition construction module, a partition adaptive linkage control module, a multi-source agricultural data standardization spatiotemporal fusion processing module, and a window linkage control module. These modules work together in a hierarchical manner to complete the refined intelligent monitoring and differentiated management of farmland across the entire area. The modules are connected sequentially from top to bottom, forming a closed-loop intelligent processing flow from spatial partition construction, differentiated data collection, multi-source data fusion preprocessing to temporal feature mining and intelligent decision output. The specific collaborative implementation process is as follows: The dynamic management partition construction module, as the spatial foundation support module of the entire system, first collects spatiotemporal dynamic factor data covering topography, soil, planting, crop rotation, ownership, and other categories for the target farmland area during operation. Based on the collected multi-dimensional spatiotemporal dynamic factor data, the scientific and refined initial delineation of the target farmland area management zoning is completed. Through multi-factor weighted clustering, optimal zoning quantity solution, boundary smoothing and topology optimization, a dynamic management zoning map with accurate spatial boundaries, homogeneous plot attributes, and differentiated production characteristics is generated. At the same time, the dynamic iterative update of the management zoning is realized by combining a timed periodic update and event-driven update dual trigger mechanism. This provides a precise, dynamic spatial unit base that is highly matched with the time-series farmland production status for subsequent differentiated collection strategy configuration, zoning data processing, zoning time series analysis and differentiated decision output. It completely solves the problems of low level of refined management and poor data matching caused by the homogeneous division of the entire field and the fixed spatial units in traditional farmland monitoring and management.The zonal adaptive linkage control module, as the core intelligent control module of the system's data acquisition end, relies on the real-time dynamic management zoning map output by the dynamic management zoning construction module. It retrieves multi-dimensional zoning attribute information for each independent management zoning, including plot characteristic parameters, crop growth stage information, meteorological disaster risk level, and pest and disease outbreak probability level. Based on the differentiated production endowments and dynamic operating conditions of each zoning, it generates targeted, differentiated acquisition strategies adapted to the specific operating conditions of each zoning. These strategies include personalized configurations of acquisition frequency, acquisition time period, and acquisition method combinations. Subsequently, relying on multi-modal acquisition terminals deployed in the field, it collects information on soil parameters, crop growth, meteorological environment, agricultural operations, and agricultural machinery maintenance for each management zoning. The system synchronously and differentiates the collection of agricultural information, and incorporates a two-level linkage adaptive control logic during the collection strategy execution process. First, it determines whether to perform a linkage-based multi-terminal synchronization time difference dynamic threshold adaptive correction operation based on the current management partition's update cycle. If it determines that synchronization time difference threshold adaptive correction is needed, it first performs a dynamic adaptive update of the multi-terminal data synchronization time difference threshold, and then further determines and executes a time difference offset linkage collection frequency adaptive adjustment operation. If it determines that synchronization time difference threshold correction is not needed, it directly enters the time difference offset linkage collection frequency adaptive adjustment determination process. Through this hierarchical linkage control logic, the system achieves real-time adaptability of multi-terminal synchronization accuracy and partition collection density to changes in farmland partition dynamics and terminal synchronization status. The system should be adapted to effectively address the technical shortcomings of traditional farmland data collection methods, such as fixed collection parameters, poor zoning adaptability, and the inability to dynamically match synchronization accuracy and collection frequency with field conditions. It should achieve highly adaptable, low-redundancy, and high-precision zoning-differentiated data collection. The multi-source agricultural data standardization spatiotemporal fusion processing module, as the core carrier for system data preprocessing and data quality optimization, receives multi-dimensional raw agricultural information from various management zones collected by the zoning adaptive linkage control module. Addressing the multi-source heterogeneity defects in multimodal terminal-collected data, such as format heterogeneity, inconsistent dimensions, inconsistent accuracy, spatiotemporal benchmark misalignment, abnormal noise interference, and local data loss, it uniformly performs outlier removal, missing data interpolation, and dimension normalization on all raw agricultural information. Comprehensive standardized correction processing, including spatiotemporal coordinate registration and time stamp calibration, completely eliminates data benchmark differences caused by multi-device, multi-dimensional, and multi-time-series acquisition. Based on the standardized data obtained through correction, further spatial dimension fusion and temporal dimension correlation fusion processing are carried out on multi-source standardized data to achieve deep coupling and complementary reconstruction of multi-dimensional, multi-time-series, and multi-location agricultural data within the same management area. Finally, standardized spatiotemporal fusion datasets for each management area with unified spatiotemporal benchmarks, high integrity, high consistency, and high comprehensiveness are generated. This provides high-quality, standardized, and directly usable benchmark data for subsequent time-series dynamic analysis models to accurately mine the evolution patterns of farmland, ensuring the accuracy of subsequent feature extraction and trend prediction from the data source.The window linkage control module, as the core computation and execution module for the system's time series analysis and intelligent decision output, relies on the standardized spatiotemporal fusion datasets of each zone generated by the multi-source agricultural data standardization spatiotemporal fusion processing module. Through a time series dynamic analysis model, it deeply mines and extracts the time series evolution patterns, state fluctuation characteristics, spatial correlation patterns, and risk contagion evolution characteristics within each management zone, as well as those between adjacent management zones. Simultaneously, it adds a window linkage overlap ratio dynamic control judgment logic to the feature analysis and trend prediction process. Based on the current zone conditions and time series analysis requirements, it determines whether to execute the sliding window overlap ratio linkage dynamic control operation. If it determines that linkage overlap ratio dynamic control is necessary, it first completes the adaptive optimization adjustment of the window overlap ratio and sliding step size through window threshold interval determination, relative position index solution, nonlinear coefficient mapping, and farmland change rate condition correction. Then, based on... The optimized time-series sampling parameters complete accurate time-series feature analysis and trend inference, generating differentiated production control decision instructions for each management zone. If it is determined that no dynamic adjustment of overlap ratio is required, time-series analysis is directly completed based on conventional fixed window parameters, and corresponding zone-specific production control decision instructions are output. Finally, the differentiated production control decision instructions generated for all zones are accurately distributed and output to the terminal execution devices of the corresponding management zones. This achieves the intelligent agricultural management goal of zoned policy implementation, precise control, and dynamic regulation for farmland plots with different production endowments, risk levels, crop growth states, and zone shortcomings. The entire system of four modules is coupled and coordinated, forming a full-process, closed-loop, and adaptive intelligent farmland management system consisting of dynamic spatial zoning, differentiated intelligent acquisition, multi-source data standardization fusion, adaptive time-series analysis, and zone-specific precise decision implementation.

[0032] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the scope and intent of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application is also intended to include such modifications and variations.

Claims

1. A method for dynamic acquisition and processing of agricultural information based on management zoning, characterized in that, Includes the following steps: Acquire spatiotemporal dynamic factor data of the target farmland area, and based on the spatiotemporal dynamic factor data, initially delineate the management zoning of the target farmland area to generate a dynamic management zoning map; Based on the partition attribute information of each management partition in the dynamic management partition map, a differentiated collection strategy for each management partition is determined. According to the differentiated collection strategy, multi-dimensional agricultural information of each management partition is collected synchronously through a multi-modal collection terminal. The differentiated collection strategy includes determining whether to perform dynamic threshold adaptive correction of the time difference of the linkage multi-terminal synchronization based on the partition update cycle. If yes, then after the adaptive correction, it is determined whether to perform adaptive adjustment of the time difference offset linkage collection frequency. If no, it is directly determined whether to perform adaptive adjustment of the time difference offset linkage collection frequency. Based on the multi-dimensional agricultural information of each management zone, data standardization and correction processing is performed on the multi-dimensional agricultural information of each management zone. Based on the standardized and corrected multi-source heterogeneous data, spatiotemporal fusion processing is performed to generate a standardized spatiotemporal fusion dataset for each management zone. Based on a standardized spatiotemporal fusion dataset, the correlation patterns and spatiotemporal evolution characteristics of agricultural data within and between each management zone are extracted. It is then determined whether to implement dynamic adjustment of the linkage and overlap ratio. If so, differentiated production control decision instructions for each management zone are generated after dynamic adjustment. If not, differentiated production control decision instructions for each management zone are directly generated and output to the terminal execution devices of the corresponding management zones.

2. The method for dynamic acquisition and processing of agricultural information based on management zoning as described in claim 1, characterized in that: The specific process for the initial delineation of management zones for the target farmland area is as follows: Spatiotemporal dynamic factor data of the target farmland area are acquired and normalized to generate a standardized dynamic factor matrix. Based on the standardized dynamic factor matrix, a weighted clustering segmentation algorithm is used to initially delineate the management zones of the target farmland area. The spatiotemporal dynamic factor data includes quantitative factors and qualitative factors. The quantitative factors include spatial distribution data of soil nutrient content, topographic slope data, and plot area data. The qualitative factors include planting type data, crop rotation pattern data, and ownership category data. Each type of dynamic factor is assigned an initial weight, and the weighted Euclidean distance between each spatial unit and the cluster center is calculated. The optimization objective is to minimize the sum of squared weighted distances. The target farmland area is divided into several initial management zones through iterative clustering. During the clustering process, the elbow rule is used to determine the optimal number of partitions, and the clustering results are processed by region merging and boundary smoothing to generate an initial management partition map, which includes the spatial boundaries and initial partition attributes of each initial partition.

3. The method for dynamic acquisition and processing of agricultural information based on management zoning as described in claim 2, characterized in that: The initial delineation of management zones for the target farmland area also includes: Preset dynamic update trigger conditions, including timed trigger conditions and event-driven trigger conditions; The timed triggering condition is triggered periodically every year to adapt to the annual administrative division adjustment needs; The event-driven trigger condition is triggered when the change of any dynamic factor exceeds a preset threshold. The preset threshold is set according to the coefficient of variation or production management sensitivity of each factor. The change of the dynamic factor is calculated by comparing the current measured data with the historical data used in the last partition update. When the dynamic update triggering condition is met, the real-time spatiotemporal dynamic factor data at the current moment is obtained, and the normalization process and initial management partition delineation are repeatedly performed to redefine all management partitions. Each dynamically updated management zone is assigned a unique zone code, and its spatial boundary vector, zone scale parameters, and functional attribute labels are recorded. The functional attribute labels include the dominant planting type, soil type, and control priority level.

4. The method for dynamic acquisition and processing of agricultural information based on management zoning as described in claim 1, characterized in that: The specific process for determining whether to perform dynamic threshold adaptive correction of multi-terminal synchronization time difference is as follows: A mapping function is pre-established to associate the partition update cycle with the multi-terminal data synchronization time difference threshold, and the partition update cycle of different levels is matched one by one with the corresponding benchmark multi-terminal data synchronization time difference threshold. The partition update cycle is used to characterize the refresh frequency of management partition space and production attribute information, and the multi-terminal data synchronization time difference threshold is used to characterize the maximum deviation value between the collection timestamps of each terminal when the multimodal acquisition terminal synchronously collects multidimensional agricultural information of the same management partition. The system reads the current partition update cycle in real time, compares the partition update cycle with the preset update cycle reference range, and dynamically adjusts the multi-terminal data synchronization time difference threshold based on the comparison results.

5. The method for dynamic acquisition and processing of agricultural information based on management zoning as described in claim 4, characterized in that: The specific steps for dynamically adjusting the multi-terminal data synchronization time difference threshold based on the comparison results are as follows: If the current partition update cycle is within the preset update cycle reference range, then the benchmark multi-terminal data synchronization time difference threshold matching the corresponding level is retrieved as the target multi-terminal data synchronization time difference threshold. The preset update cycle reference range represents the open interval formed by the preset update cycle reference lower limit and the preset update cycle reference upper limit. If the current partition update cycle is less than or equal to the preset update cycle reference lower limit, the current partition update cycle is input into the associated mapping function, and the time difference threshold reduction coefficient is output. The time difference threshold reduction coefficient is multiplied by the benchmark multi-terminal data synchronization time difference threshold of the corresponding level to obtain the target multi-terminal data synchronization time difference threshold. If the current partition update cycle is greater than or equal to the preset update cycle reference upper limit, the current partition update cycle is input into the associated mapping function, and the time difference threshold adjustment coefficient is output. The time difference threshold adjustment coefficient is multiplied by the benchmark multi-terminal data synchronization time difference threshold of the corresponding level to obtain the target multi-terminal data synchronization time difference threshold.

6. The method for dynamic acquisition and processing of agricultural information based on management zoning as described in claim 1, characterized in that: The specific process for determining whether to perform adaptive adjustment of the time difference offset linkage acquisition frequency is as follows: A one-to-one mapping relationship between the dynamic threshold correction offset for multi-terminal synchronization time difference and the collection interval adjustment ratio is pre-built. The timestamps of multi-terminal synchronization collection in each management partition are collected in real time. The actual collection time difference is calculated. The difference between the actual collection time difference and the currently effective dynamic threshold for multi-terminal synchronization time difference is calculated to obtain the real-time dynamic threshold correction offset for multi-terminal synchronization time difference. The real-time multi-terminal synchronization time difference dynamic threshold correction offset is compared with the threshold correction offset critical interval, and the collection frequency is dynamically adjusted according to the comparison result. The collection frequency is used to characterize the collection interval of multi-dimensional agricultural information in each management zone. If the real-time multi-terminal synchronization time difference dynamic threshold correction offset is within the threshold correction offset critical interval, then the sampling frequency benchmark value will be used as the target sampling frequency. The threshold correction offset critical interval represents the open interval formed by the threshold correction offset critical lower limit and the threshold correction offset critical upper limit.

7. The method for dynamic acquisition and processing of agricultural information based on management zoning as described in claim 6, characterized in that: The specific process for dynamically adjusting the acquisition frequency based on the comparison results is as follows: If the real-time multi-terminal synchronization time difference dynamic threshold correction offset is less than or equal to the threshold correction offset critical lower limit, it is determined that the multi-terminal acquisition synchronization deviation is small. Then, the current dynamic threshold correction offset is input into the corresponding mapping relationship, and the first-level acquisition interval stretching ratio is output. The acquisition interval benchmark value and the first-level acquisition interval stretching ratio are multiplied to obtain the target acquisition interval, so as to extend the acquisition interval duration of agricultural information in each dimension and reduce the overall acquisition frequency of the partition. If the real-time multi-terminal synchronization time difference dynamic threshold correction offset is greater than or equal to the threshold correction offset critical upper limit, it is determined that the multi-terminal acquisition synchronization deviation is large. Then, the current dynamic threshold correction offset is input into the corresponding mapping relationship, and the first-level acquisition interval compression ratio is output. The acquisition interval benchmark value and the first-level acquisition interval compression ratio are multiplied to obtain the target acquisition interval, so as to shorten the acquisition interval time of agricultural information in various dimensions and increase the overall acquisition frequency of the partition.

8. The method for dynamic acquisition and processing of agricultural information based on management zoning as described in claim 1, characterized in that: The specific process for determining whether to perform dynamic adjustment of the linkage overlap ratio is as follows: The minimum sliding window duration threshold, the maximum sliding window duration threshold, the maximum overlap ratio coefficient, and the minimum overlap ratio coefficient are pre-configured. The minimum overlap ratio coefficient and the maximum overlap ratio coefficient are between 0 and 1, and the minimum overlap ratio coefficient is less than the maximum overlap ratio coefficient. Obtain the partition attribute data of the management partition corresponding to the current prediction task, and obtain the sliding window duration of the current time series analysis based on the partition attribute data. The partition attribute data includes the crop growth stage, agricultural data update frequency, and farmland status change rate. Based on the current sliding window duration, the minimum sliding window duration threshold, and the maximum sliding window duration threshold, the relative position index of the current window duration within the preset threshold range is obtained; If the current sliding window duration is within the sliding window duration threshold range, the corresponding index value is calculated according to the relative position of the range. The sliding window duration threshold range represents the closed interval formed by the minimum sliding window duration threshold and the maximum sliding window duration threshold. If the current sliding window duration is less than the minimum sliding window duration threshold, the relative position index is determined to be zero. If the current sliding window duration exceeds the maximum sliding window duration threshold, the relative position index is determined to be a unit extreme value.

9. The method for dynamic acquisition and processing of agricultural information based on management zoning as described in claim 8, characterized in that: The determination of whether to perform dynamic adjustment of the linkage overlap ratio also includes: Based on the calculated relative position index, an initial adaptive overlap ratio coefficient is obtained by dynamic matching through a preset monotonically increasing nonlinear mapping rule. The monotonically increasing nonlinear mapping rule is used to characterize the corresponding limit overlap ratio coefficient when the relative position index takes an extreme value. The relative position index is divided into three intervals by pre-setting two inflection points. Different intervals correspond to the overlap ratio coefficients, and the real-time farmland status change rate index of the current management zone is obtained. The change rate index includes at least one of soil moisture change rate, crop growth change rate, and meteorological element change rate. The initial adaptive overlap ratio coefficient is dynamically corrected based on the real-time change rate index to obtain the final overlap ratio coefficient. If the real-time rate of change exceeds the preset high change threshold, the initial overlap ratio coefficient is incrementally corrected and the corrected value is limited to not exceeding the preset maximum overlap ratio coefficient. If the real-time rate of change index is lower than the preset low change threshold, the initial overlap ratio coefficient is reduced and the corrected value is limited to not be lower than the preset minimum overlap ratio coefficient. Based on the current sliding window duration and the final overlap ratio coefficient, the sliding step size of the time series window is calculated. The obtained sliding step size is then used to complete the sliding window sampling on the time series agricultural data sequence, thereby completing the extraction of agricultural data correlation patterns and spatiotemporal evolution characteristic analysis for each management zone.

10. A system applying the method for dynamic acquisition and processing of agricultural information based on management zoning as described in any one of claims 1-9, characterized in that, include: The system includes a dynamic management partitioning construction module, a partitioning adaptive linkage control module, a multi-source agricultural data standardization spatiotemporal fusion processing module, and a window linkage control module. The dynamic management partition construction module is used to acquire spatiotemporal dynamic factor data of the target farmland area, and based on the spatiotemporal dynamic factor data, to initially delineate the management partitions of the target farmland area and generate a dynamic management partition map. The partition adaptive linkage control module is used to determine the differentiated acquisition strategy for each management partition based on the partition attribute information of each management partition in the dynamic management partition map. According to the differentiated acquisition strategy, multi-dimensional agricultural information of each management partition is synchronously acquired through multi-modal acquisition terminals. The differentiated acquisition strategy includes determining whether to perform dynamic threshold adaptive correction of linkage multi-terminal synchronization time difference based on the partition update cycle. If yes, it determines whether to perform adaptive adjustment of time difference offset linkage acquisition frequency after adaptive correction. If no, it directly determines whether to perform adaptive adjustment of time difference offset linkage acquisition frequency. The multi-source agricultural data standardization spatiotemporal fusion processing module is used to perform data standardization correction processing on the multi-dimensional agricultural information of each management zone according to the multi-dimensional agricultural information of each management zone, and perform spatiotemporal fusion processing on the standardized and corrected multi-source heterogeneous data to generate a standardized spatiotemporal fusion dataset for each management zone. The window linkage control module is used to extract the correlation patterns and spatiotemporal evolution characteristics of agricultural data within and between management zones based on a standardized spatiotemporal fusion dataset, and to determine whether to perform dynamic control of linkage overlap ratio. If so, it generates differentiated production control decision instructions for each management zone after dynamic control. If not, it directly generates differentiated production control decision instructions for each management zone and outputs them to the terminal execution devices of the corresponding management zones.

Images