Agricultural Information Sharing and Management Method and System Based on Multi-Source Data

Abstract

This invention relates to the field of agricultural information management technology, specifically to an agricultural information sharing management method and system based on multi-source data. The method includes the following steps: acquiring terminal time synchronization and confirmation signals; calibrating irrigation and fertilization collection times to generate time reference data; calculating water volume fertilization intensity and converting water content and conductivity to generate standardized data; reconstructing and aggregating time granularity chains; determining plot affiliation based on k-nearest neighbors using coordinate boundaries to generate state; and classifying agricultural information sharing data based on growth stages and consistent changes. In this invention, by unifying the time reference and synchronizing irrigation, fertilization, and soil meteorological data, standard data is generated through the conversion of flow mass, volumetric water content, and conductivity dimensions. Time granularity reconstruction and aggregation are performed to construct a continuous chain. Data affiliation is determined by combining sensor coordinates and plot boundaries, and mean indicators are calculated to classify growth stages and states. This improves data association accuracy and spatial mapping accuracy, thereby enhancing agricultural information sharing capabilities.

Description

Technical Field

[0001] This invention relates to the field of agricultural information management technology, and in particular to a method and system for agricultural information sharing and management based on multi-source data. Background Technology

[0002] The field of agricultural information management technology primarily involves the collection, processing, analysis, and sharing of agricultural production-related data through information technology. Core aspects of this technology include intelligent management of agricultural resources, precise decision-making in agricultural production, and the integration and sharing of agricultural data. Through advanced data processing technologies and information systems, agricultural information management enables real-time monitoring and analysis of data on farmland, climate, soil, crop growth, and other aspects, thereby providing effective management support and decision-making basis. Agricultural information management systems typically include modules for data collection, storage, analysis, and transmission, effectively integrating data from different sources to provide comprehensive agricultural management solutions.

[0003] Traditional agricultural information sharing and management methods and systems based on multi-source data refer to the integration and analysis of multiple data sources to share and manage information during the agricultural production process. This method typically uses technologies such as sensors, satellite remote sensing, and drones to collect multi-dimensional data from agricultural production sites, and then processes and manages the data through information systems. Traditional solutions often rely on a single source of static data, with information sharing and management depending on manual operation or simple integration platforms, making it difficult to meet the increasingly complex demands of the growing volume and types of data in agricultural production. This new method, by integrating data from different sources, such as climate data, soil data, and crop growth data, achieves comprehensive monitoring and optimized management of agricultural production.

[0004] In the process of integrating multi-source agricultural data, existing technologies suffer from inconsistent time references for data collection by various types of equipment, and time misalignment between operation records and environmental monitoring data. This makes it difficult to accurately correlate irrigation and fertilization data with soil and meteorological information. There is a lack of unified dimensional and granularity standards among the data. Spatial mapping of data from different sources relies on manual division and simple correlation. There are biases in the determination of the plot to which sensor data belongs. The data is discrete in both time and space dimensions, making it difficult to form a continuous growth process chain. This results in fragmented management decision-making basis, delayed judgment of state change trends, and affects the pertinence and coordination of agricultural production regulation. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing an agricultural information sharing and management method and system based on multi-source data.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: an agricultural information sharing and management method based on multi-source data, comprising the following steps:

[0007] S1: Obtain the time calibration signal and the returned timestamp confirmation signal sent by the field data acquisition terminal, determine the consistency of irrigation operation time, fertilization operation time, soil moisture and meteorological acquisition time and calibrate them, and generate time reference data;

[0008] S2: Based on the time reference data, obtain irrigation flow, fertilizer application quality, soil volumetric moisture content and electrical conductivity, calculate water volume and fertilizer application intensity, and convert the dimensions of soil volumetric moisture content and electrical conductivity to generate standardized data;

[0009] S3: Obtain a unified time granularity to reconstruct the standardized data, aggregate and calculate the water volume, fertilization intensity, soil volumetric moisture content, electrical conductivity and air temperature within the same time period, and generate a time granularity chain;

[0010] S4: Based on the time granularity chain, obtain the sensor deployment coordinates and plot boundaries, use the spatial inclusion relationship judgment algorithm to determine the plot ownership of sensor data, calculate the average volumetric water content, normalized water volume and fertilization intensity of the same plot, and generate plot status data.

[0011] S5: Based on the plot status data, obtain the crop growth stage time, determine the consistency of the change direction of water volume, fertilizer intensity, volumetric water content and electrical conductivity within adjacent time granularities, divide the growth stage, and generate agricultural management data.

[0012] As a further aspect of the present invention, the time reference data includes irrigation operation time offset, fertilization operation time offset, and soil moisture collection time offset; the standardized data includes water volume, area fertilization intensity, and electrical conductivity standard dimensions; the time granularity chain includes a time index sequence, a water volume aggregate value sequence, a fertilization intensity aggregate value sequence, and a volumetric moisture content aggregate value sequence; the plot status data includes plot number identifier, average plot volumetric moisture content, normalized plot water volume, and normalized plot fertilization intensity; and the agricultural management data includes a growth stage number, the start and end time periods of the growth stage, and the stage consistency ratio.

[0013] As a further aspect of the present invention, the step of obtaining the time reference data specifically includes:

[0014] S101: Obtain the time calibration signal sent by the field acquisition terminal and the timestamp confirmation signal returned by the multi-source agricultural equipment, split the time field and extract the second-level and millisecond-level counts, calculate the difference between irrigation operation time, fertilization operation time, soil moisture acquisition time and meteorological acquisition time, and generate a time deviation vector set;

[0015] S102: Based on the time deviation vector set, the values ​​of multiple deviation vectors are compared with the preset time tolerance threshold item by item, the index positions corresponding to the time tolerance threshold are marked, the arithmetic mean is calculated on the unmarked deviation vectors, the mean result is extracted and bound to the corresponding device identifier to obtain the time calibration parameter sequence.

[0016] S103: Based on the time calibration parameter sequence, perform numerical translation operations on irrigation operation time, fertilization operation time, soil moisture collection time and meteorological collection time respectively, perform addition and subtraction mapping between multiple time values ​​and corresponding calibration parameter sequence elements, reconstruct the time series structure, and generate time reference data.

[0017] As a further aspect of the present invention, the step of obtaining the standardized data specifically includes:

[0018] S201: Based on the time reference data, obtain irrigation flow and fertilization quality records and align them in time sequence. Perform integral calculation on the instantaneous flow value and corresponding time interval in the irrigation flow record to form a cumulative water volume value sequence. At the same time, calculate the fertilization intensity and generate a set of operation quantification sequences.

[0019] S202: Based on the operation quantification sequence set, obtain the soil volumetric water content and soil electrical conductivity, extract the time index corresponding to the two types of values ​​and match the index with the operation quantification sequence set, and perform conversion operations on the soil volumetric water content value and soil electrical conductivity value respectively to obtain the soil dimension conversion sequence.

[0020] S203: Call the operation quantization sequence set and the soil dimension conversion sequence, splice the sequences according to the time scale corresponding to the time base data, extract the minimum and maximum values ​​of the multiple sequence values ​​and perform interval normalization calculation, map the multiple values ​​to a unified scale interval, and generate standardized data.

[0021] As a further aspect of the present invention, the calculation of fertilization intensity is performed using the following formula:

[0022] ;

[0023] in, Represents fertilization intensity. This represents the recorded value of fertilizer application quality. This represents the fertilizer quality compensation constant. Represents the area characterized by irrigation features. This represents the dispersion of fertilizer distribution.

[0024] As a further aspect of the present invention, the step of obtaining the time granularity chain specifically includes:

[0025] S301: Based on the standardized data, obtain a unified time granularity, perform difference calculation on the timestamps of multiple data frames of the standardized data, extract the interval between adjacent timestamps and compare it with the preset time granularity benchmark value, map the time interval values ​​that satisfy the integer multiple relationship to a unified time granularity, and generate a time granularity parameter set.

[0026] S302: Based on the time granularity parameter set, call up the water volume, fertilizer intensity, soil volumetric moisture content, electrical conductivity and air temperature in the standardized data, divide the time window according to the unified time granularity parameter, calculate the arithmetic mean of the values ​​in multiple time windows and aggregate and rearrange them to obtain the granularity aggregated sequence group.

[0027] S303: For the granularity aggregated sequence group, search for null value index positions in multiple time windows, extract the difference between two adjacent values ​​and their corresponding time indices, perform linear scaling operations to fill the null value index positions and reconstruct the time series structure to establish a time granularity chain.

[0028] As a further aspect of the present invention, the step of obtaining the land parcel status data specifically includes:

[0029] S401: Based on the time granularity chain, obtain the sensor deployment coordinates and the land parcel boundary, construct a spatial position vector for the sensor deployment coordinates, construct a closed polygon area in combination with the land parcel boundary, determine the topological inclusion relationship between the multi-coordinate vector and the closed polygon area, perform the attribution determination and mapping, and generate a land parcel mapping sequence.

[0030] S402: Based on the plot mapping sequence, call the soil volumetric moisture content in the time granularity chain, aggregate the data by plot number and extract the corresponding values ​​of multiple sensors in the same plot, use summation operation to calculate the ratio with the sensor count value, and generate a plot moisture content mean sequence.

[0031] S403: For the average moisture content sequence of the plot, call the water volume and fertilization intensity in the time granularity chain and obtain the corresponding plot area parameter sequence, normalize the water volume and fertilization intensity respectively, and re-encode them in the same dimension with the average moisture content sequence of the plot to establish plot status data.

[0032] As a further aspect of the present invention, the step of acquiring agricultural management data specifically includes:

[0033] S501: Based on the land parcel status data, obtain the crop growth stage time, extract the crop growth stage identifier and perform time series matching operation with the time index in the land parcel status data, calculate the time series matching distance, and merge the continuous time index range to generate a growth stage time series.

[0034] S502: Based on the time series of the reproductive stage, analyze the sequence of changes in water volume, fertilization intensity, volumetric water content and conductivity within adjacent time granularities and make a time-by-time consistency judgment. Count and summarize the time periods that are judged to be consistent, calculate the ratio based on the time length of the reproductive stage, and generate a consistency ratio sequence.

[0035] S503: For the consistency ratio sequence, obtain a preset ratio threshold parameter set, compare the consistency ratio corresponding to multiple growth stages with the threshold interval, and map the judgment result into a stage state level coding sequence, and recombin it with the crop growth stage time series to establish agricultural management data.

[0036] As a further aspect of the present invention, the step of comparing and judging the consistency ratio and threshold interval corresponding to multiple reproductive stages specifically involves judging the inclusion relationship between the consistency ratio value corresponding to each reproductive stage in the consistency ratio sequence and the multiple ratio intervals in the preset ratio threshold parameter set, and determining the corresponding stage status level coding sequence based on the judgment result.

[0037] The agricultural information sharing and management system based on multi-source data is used to perform the above-mentioned agricultural information sharing and management method based on multi-source data. The system includes:

[0038] The time calibration module acquires the time calibration signal and the returned timestamp confirmation signal sent by the field acquisition terminal, determines the consistency between irrigation operation time, fertilization operation time, soil moisture and meteorological acquisition time and calibrates them, generates time reference data and transmits it to the data standard module;

[0039] The data standard module acquires irrigation flow, fertilizer quality, soil volumetric water content and electrical conductivity based on the time reference data, calculates water volume and fertilizer intensity, converts the dimensions of soil volumetric water content and electrical conductivity, generates standardized data, and transmits it to the time series reconstruction module.

[0040] The time-series reconstruction module reconstructs the standardized data by obtaining a unified time granularity, aggregates and calculates water volume, fertilization intensity, soil volumetric moisture content, electrical conductivity and air temperature within the same time period, generates a time granularity chain and transmits it to the plot analysis module.

[0041] The plot analysis module obtains the sensor deployment coordinates and plot boundaries based on the time granularity chain, uses a spatial inclusion relationship judgment algorithm to determine the plot affiliation of the sensor data, calculates the average volumetric water content, normalized water volume and fertilization intensity of the same plot, generates plot status data and transmits it to the agricultural management module.

[0042] The agricultural management module obtains the crop growth stage time based on the plot status data, determines the consistency of the change direction of water volume, fertilizer intensity, volumetric water content and electrical conductivity within adjacent time granularities, divides the growth stage status, and generates agricultural management data.

[0043] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0044] In this invention, a unified time reference is constructed to achieve synchronous calibration of irrigation and fertilization signals and soil meteorological acquisition signals, enabling multi-source operational data to be matched under the same time coordinate. By combining the dimensional conversion of flow rate, mass, volumetric water content, and conductivity, comparable standard data are formed. After unified time granularity reconstruction and aggregation, a continuous time chain is generated. Then, by combining sensor coordinates and plot boundaries, the data attribution is determined through proximity relationships, and plot mean and normalized index are calculated to form plot status information with stage continuity. This enables consistent judgment of the direction of water and fertilizer environment changes and division of growth stages, improves the accuracy of data association and spatial mapping, and enhances the overall collaborative ability of agricultural information sharing and regulation decision-making. Attached Figure Description

[0045] Figure 1 This is a schematic diagram of the workflow of the present invention;

[0046] Figure 2 This is a flowchart of the steps for obtaining time reference data in this invention;

[0047] Figure 3 This is a flowchart of the standardized data acquisition steps for this invention;

[0048] Figure 4 This is a flowchart illustrating the steps for obtaining the time granularity chain in this invention.

[0049] Figure 5 This is a flowchart illustrating the steps for obtaining land parcel status data according to the present invention.

[0050] Figure 6 This is a flowchart illustrating the steps for acquiring agricultural management data according to the present invention. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0052] Please see Figure 1 This invention provides a technical solution: an agricultural information sharing and management method based on multi-source data, comprising the following steps:

[0053] S1: Obtain the time calibration signal and the returned timestamp confirmation signal sent by the field data acquisition terminal, determine the consistency of irrigation operation time, fertilization operation time, soil moisture and meteorological acquisition time and calibrate them, and generate time reference data;

[0054] S2: Based on time-referenced data, obtain irrigation flow, fertilizer quality, soil volumetric moisture content and electrical conductivity, calculate water volume and fertilizer intensity, convert the dimensions of soil volumetric moisture content and electrical conductivity, and generate standardized data;

[0055] S3: Obtain a unified time granularity to reconstruct standardized data, aggregate and calculate water volume, fertilizer intensity, soil volumetric moisture content, electrical conductivity and air temperature within the same time period, and generate a time granularity chain;

[0056] S4: Based on the time granularity chain, obtain the sensor deployment coordinates and plot boundaries, use the spatial inclusion relationship judgment algorithm to determine the plot ownership of sensor data, calculate the average volumetric water content, normalized water volume and fertilization intensity of the same plot, and generate plot status data.

[0057] S5: Based on the plot status data, obtain the crop growth stage time, determine the consistency of the change direction of water volume, fertilizer intensity, volumetric water content and electrical conductivity within adjacent time granularities, divide the growth stage, and generate agricultural management data.

[0058] The time reference data includes irrigation operation time offset, fertilization operation time offset, and soil moisture collection time offset. The standardized data includes water volume, area fertilization intensity, and electrical conductivity standard dimensions. The time granularity chain includes time index sequence, water volume aggregate value sequence, fertilization intensity aggregate value sequence, and volumetric moisture content aggregate value sequence. The plot status data includes plot number identifier, average plot volumetric moisture content, plot normalized water volume, and plot normalized fertilization intensity. The agricultural management data includes growth stage number, growth stage start and end time period, and stage consistency ratio.

[0059] Please see Figure 2 The specific steps for obtaining time base data are as follows:

[0060] S101: Obtain the time calibration signal sent by the field acquisition terminal and the timestamp confirmation signal returned by the multi-source agricultural equipment, split the time field and extract the second-level and millisecond-level counts, calculate the difference between irrigation operation time, fertilization operation time, soil moisture acquisition time and meteorological acquisition time, and generate a time deviation vector set;

[0061] The system receives time calibration signals transmitted via fiber optic networks from field data acquisition terminals deployed in the core areas of farmland, and records the absolute reference timestamp of the transmitted signals. Simultaneously, it captures timestamp confirmation signals returned along the original communication link from multiple agricultural devices, including irrigation solenoid valve controllers, integrated water and fertilizer machines, embedded soil moisture sensors, and miniature weather stations. For the 40 received data packets containing time attributes, the system performs packet parsing to extract the standard time string from the packet header. The standard time string is then sliced ​​using colons and periods as delimiters, precisely extracting the year, month, date, and hour data from the 19-character string. The system focuses on extracting the second count (positions 15-16) and the millisecond count (positions 18-20) from the end of the string. The extracted second counts are multiplied by 1000 and added to the millisecond counts to convert the result into an 8-digit absolute millisecond timestamp format. The baseline timestamp, recorded by the underlying clock at the moment the calibration signal is emitted, is retrieved. Using the same string slicing and multiplication / addition conversion logic, the corresponding baseline millisecond timestamp is obtained. For each device millisecond timestamp corresponding to irrigation operation time, fertilization operation time, soil moisture collection time, and meteorological collection time, a subtraction operation is performed with the baseline millisecond timestamp. Specifically, the absolute difference between the device timestamp and the baseline millisecond timestamp is obtained by subtracting the baseline millisecond timestamp value from the device timestamp value. The differences generated by one field acquisition terminal and four types of agricultural equipment in 10 consecutive high-frequency calibration communications are sequentially appended to a linear array in local memory, constructing an initial time deviation vector set containing four independent subsets. If the baseline millisecond timestamp is 55950 milliseconds and the timestamp of a certain irrigation device is 56000 milliseconds, then the calculated absolute difference is 50 milliseconds. The final deviation vector set for irrigation equipment is [45, 50, 48, 52, 47, 49, 51, 46, 50, 48], the deviation vector set for fertilization equipment is [60, 62, 59, 65, 61, 63, 58, 64, 60, 61], the deviation vector set for soil moisture sensors is [110, 115, 108, 112, 109, 111, 114, 113, 107, 110], and the deviation vector set for weather stations is [150, 155, 152, 280, 149, 151, 154, 148, 153, 150].

[0062] S102: Based on the time deviation vector set, the values ​​of multiple deviation vectors are compared with the preset time tolerance threshold item by item, the index positions corresponding to the time tolerance threshold are marked, the arithmetic mean of the unmarked deviation vectors is calculated, the mean result is extracted and bound to the corresponding device identifier to obtain the time calibration parameter sequence.

[0063] The system reads all integer elements from the initial time deviation vector set and retrieves the preset time tolerance threshold stored in the local register. This preset time tolerance threshold is precisely set to 200 milliseconds by extracting the upper limit of the 99% confidence interval from the normal distribution characteristics of the round-trip delay of statistical data packets through 200 high-concurrency network stress tests on similar agricultural IoT devices. A loop iterative structure of depth 10 is constructed to read the values ​​in each deviation vector set item by item and compare them with 200 milliseconds. If and only if the read deviation value is strictly greater than 200 milliseconds, it is determined that the data transmission encountered abnormal delay due to network jitter. For the fourth element 280 in the weather station deviation vector set, the comparison shows it is significantly greater than 200 milliseconds. Therefore, in the pre-initialized Boolean type marker array, the value at index position 3 is changed from the default 0 to 1, accurately locking the physical location of the abnormal data. For the valid index positions with a value of 0 in the marker array, the corresponding deviation vector values ​​are stored one by one in the accumulator variable and added together. The final sum obtained from the accumulator is divided by the total number of valid values, and a strict arithmetic mean is calculated and rounded down. For the weather station equipment, after removing the outlier value of 280, the sum of the remaining 9 valid values ​​is 1362, 1362 / 9 = 151 milliseconds. Similarly, the arithmetic mean for irrigation equipment is calculated to be 48 milliseconds, for fertilization equipment to be 61 milliseconds, and for soil moisture sensors to be 110 milliseconds. These four mean results are extracted and packaged into key-value pair structures, using the unique MAC address string fixed at the physical layer of each device as the key name and the calculated mean as the key value for forced memory binding. These four sets of key-value pairs are concatenated into a linear list according to the device registration priority to generate a complete time calibration parameter sequence containing specific compensation features.

[0064] S103: Based on the time calibration parameter sequence, numerical translation operations are performed on irrigation operation time, fertilization operation time, soil moisture collection time and meteorological collection time respectively. Multiple time values ​​are added and subtracted to the corresponding calibration parameter sequence elements to reconstruct the time series structure and generate time reference data.

[0065] Extract the raw time string log data reported in real time by each agricultural device during actual operation, and call the previously generated time calibration parameter sequence in memory. For the irrigation operation time, fertilization operation time, soil moisture collection time, and meteorological collection time recorded in the log, parse the MAC address in the data packet header, and match the corresponding mean calibration parameter according to the address mapping relationship. Extract the valve opening execution time of 14:30:15:200 milliseconds uploaded by the irrigation solenoid valve, and convert its time format to absolute millisecond count, obtaining the corresponding value 52215200. Match the corresponding irrigation equipment calibration parameter of 48 milliseconds. Determine that the device clock is faster than the reference clock, and then perform a subtraction and shift operation between the absolute millisecond count and the calibration parameter, that is, use 52215200-48=52215152 to obtain the accurate real timestamp 52215152 after shifting. If it is determined that the internal clock of the device is configured in time lag compensation mode, then call the addition operation to perform the mapping operation. After completing the translation, addition, and subtraction operations for all message values, the corrected millisecond integers are re-operated using modulo and division operations. Specifically, the modulo operation is taken to obtain the corrected millisecond value, and the integer division by 1000 is used to obtain the second-level cumulative value. This process is then reversed layer by layer to convert the data into a structured standard time series containing year, month, day, hour, minute, second, and millisecond. The reconstructed time series is used as the time reference data, covering the time fields with errors in the original data stream, thus achieving absolute precision alignment of multi-source agricultural equipment data on the standard time axis. The cleaned and aligned irrigation operation start nodes, fertilizer valve response nodes, soil moisture decrease abrupt change nodes, and meteorological temperature extreme value acquisition nodes are synchronously written into the time series database to establish a three-dimensional agricultural monitoring map with absolute time series correlation.

[0066] Please see Figure 3 The specific steps for obtaining standardized data are as follows:

[0067] S201: Based on time reference data, obtain irrigation flow and fertilization quality records and align them in time sequence. Perform integral calculation on the instantaneous flow value and corresponding time interval in the irrigation flow record to form a cumulative water volume value sequence. At the same time, calculate the fertilization intensity and generate a set of operation quantification sequences.

[0068] Based on the generated time-base data, irrigation flow logs and fertilization logs from the time-series database are retrieved. For the irrigation logs, data collected by the pipeline flow meter is retrieved, and the instantaneous flow value of 3.5 cubic meters per hour is extracted with a period of 50 milliseconds. According to the continuous time slice identifier in the time-base data, the trapezoidal rule numerical integration operation is performed on the instantaneous flow value, that is, the instantaneous flow value is multiplied by the time step and accumulated to obtain the cumulative water volume sequence value of 7.0 cubic meters within a 2-hour period. Simultaneously, the relevant parameters of the water and fertilizer integrated machine for the current time period are extracted and substituted into the following formula to calculate the fertilization intensity: The numerator part records the fertilizer quality value through addition. Compensation constant for fertilizer quality The addition aims to compensate for the physical losses intercepted at the blind ends of the pipeline network into the total consumption; the denominator reflects the dispersion of fertilizer distribution. Perform the square root operation and compare it with the irrigation feature area. Multiplication is performed to penalize the reduction of the ideal physical coverage area; finally, division is used to distribute the absolute fertilizer mass evenly across the actual effective area, generating the fertilization intensity. Retrieve the flow data from the weighing sensor, extract the initial mass of 100.0 kg and the final mass of 84.5 kg, and calculate the difference to obtain the fertilizer application quality record value. It is 15.5 kg. For the fertilizer quality compensation constant... The specific settings are based on the reference content. The residual mass is calculated by flushing the pipeline network. In three measurement experiments, the residual mass was 0.5 kg, 0.6 kg, and 0.4 kg respectively. The total is calculated to be 1.5. 1.5 / 3 = 0.5, therefore 0.5 kg is set as the fertilizer mass compensation constant. 300 valid pixels were extracted from the image and multiplied by the area of ​​a single pixel (0.04 square meters) to obtain the irrigation feature area. The area is 12.0 square meters. Fertilizer distribution dispersion. This is a non-numerical spatial state. The quantification process involves deploying 16 rain gauges in the field, calculating the coefficient of variation of rainfall volume (0.9), performing a square operation to convert it into a dimensionless index, and obtaining the fertilization distribution divergence. The value is 0.81. Substituting the above parameters strictly into the fertilizer intensity calculation formula, the specific calculation process is explained below: First, for the numerator of the formula: Secondly, the calculation is performed on the denominator of the formula. The specific calculation process is as follows: Finally, the absolute total fertilizer mass obtained from the numerator (16.0) is divided by the actual effective area obtained from the denominator (10.8), and a division mapping operation is performed. Determine the fertilization intensity The final value is 1.48. This value accurately quantifies the actual fertilizer quality received by crops per unit area, and directly serves as the core quantitative indicator in the operational quantification sequence set, providing an absolute calculation benchmark for the subsequent generation of standardized data within a unified scale range.

[0069] S202: Based on the work quantification sequence set, obtain soil volumetric moisture content and soil electrical conductivity, extract the time index corresponding to the two types of values ​​and match the index with the work quantification sequence set, and perform conversion operations on the soil volumetric moisture content value and soil electrical conductivity value respectively to obtain the soil dimension conversion sequence.

[0070] The generated job quantization sequence set is read, and the absolute millisecond timestamps of the start and end timestamps of the jobs in the sequence are parsed. Based on the above time interval, a high-frequency polling command is initiated to the underlying soil sensor network to read the raw dielectric constant data returned by the frequency domain reflectance sensor buried at a depth of 200 mm in the root layer, and simultaneously read the raw conductivity voltage signal collected by the four-terminal electrode sensor at the same physical location. The dielectric constant data is converted and calculated using a polynomial fitting curve to obtain the corresponding raw value of soil volumetric moisture content as 28.5%. Simultaneously, the conductivity voltage signal is converted according to temperature compensation logic, combined with the current soil temperature of 18 degrees Celsius, to extract the soil conductivity baseline value of 1.45 millisiemens per centimeter. The occurrence timestamps carried by the above 28.5% moisture content value and 1.45 millisiemens per centimeter conductivity value are extracted, accurate to the absolute millisecond level, and converted into a time index string. This time index string is compared and matched character by character with the index primary key containing the time baseline data in the job quantization sequence set. When the absolute difference between the two timestamps is less than the preset 10-millisecond alignment threshold, the spatiotemporal association is confirmed to be successful. Subsequently, independent conversion operations are performed on the successfully matched soil volumetric moisture content and soil electrical conductivity. Using preset dimensionless conversion base parameters, the percentage-formatted soil volumetric moisture content of 28.5% is divided by the soil porosity calibration value of 45.0%, yielding a dimensionless relative saturation result of 0.63. Simultaneously, the soil electrical conductivity of 1.45 millisiemens per centimeter is multiplied by the soil solution density constant of 1.05 grams per cubic centimeter, yielding a solute mass fraction index value of 1.52. The converted relative saturation and solute mass fraction are then stored in a two-dimensional array and appended with a successfully matched time index stamp, generating a soil dimensionless conversion sequence with a unified structure and continuous temporal characteristics.

[0071] S203: Call the operation quantification sequence set and the soil dimension conversion sequence, splice the sequence according to the time scale corresponding to the time base data, extract the minimum and maximum values ​​of the multiple sequence values ​​and perform interval normalization calculation, map the multiple values ​​to a unified scale interval, and generate standardized data.

[0072] The generated quantified work sequence set and soil dimension-converted sequence are invoked. The cumulative water volume (7.0 cubic meters) and fertilization intensity (1.48) from the quantified work sequence set are extracted as core input parameters. Simultaneously, the relative saturation (0.63) and solute mass fraction (1.52) from the soil dimension-converted sequence are extracted. Based on the absolute millisecond-level timescale primary key provided by the previously generated time base data, work parameters and soil parameters belonging to the same sampling instant (52215152 milliseconds) are placed in the same horizontal data row, and a matrix-dimensional sequence splicing operation is performed. For the spliced ​​multidimensional sequence column vector, the entire record set is traversed sequentially to locate the absolute boundaries of the values. The minimum value (0.5 cubic meters) and maximum value (15.0 cubic meters) of the cumulative water volume dimension are extracted, and the minimum value (0.2) and maximum value (3.0) of the fertilization intensity dimension are extracted. Based on these boundaries, the min-maximum linear mapping mechanism is invoked to independently perform interval normalization calculations on the spliced ​​values. Taking fertilization intensity data as an example, the current fertilization intensity value of 1.48 is subtracted from the extracted minimum value of 0.2 to obtain a difference of 1.28. The range between the maximum value of 3.0 and the minimum value of 0.2 is calculated to be 2.8. Then, 1.28 / 2.8 = 0.45, yielding the normalized value of 0.45 for this dimension. Using the same normalization calculation logic of dividing the difference by the range, the cumulative water volume of 7.0 cubic meters is mapped to a dimensionless value of 0.44, the relative saturation of 0.63 is mapped to 0.58, and the solute mass fraction of 1.52 is mapped to 0.52. Through this process, various operational and environmental monitoring values ​​that originally had different physical meanings are strictly mapped to a unified scale interval of 0 to 1. The characteristics of various results distributed in the unified scale interval are packaged by time series to generate standardized data that completely eliminates dimensional differences and can be directly used for subsequent algorithms.

[0073] Please see Figure 4 The specific steps for obtaining the time granularity chain are as follows:

[0074] S301: Based on standardized data, obtain a unified time granularity, perform difference calculation on the timestamps of multiple data frames of standardized data, extract the interval between adjacent timestamps and compare it with the preset time granularity benchmark value, map the time interval values ​​that satisfy the integer multiple relationship to a unified time granularity, and generate a time granularity parameter set;

[0075] The system reads standardized data that has been processed and whose values ​​are between 0 and 1, and parses the absolute millisecond-level timestamp primary key attached to each frame record in the data stream. For the timestamps of consecutively input data frames, it performs a term-by-term interpolation operation, extracting the timestamp 1620000000000 milliseconds of the first frame and subtracting it from the timestamp 1620000005000 milliseconds of the second frame to obtain the first adjacent timestamp interval as 5000 milliseconds. It then iterates through the first 1000 frames of data, extracting a set of irregular timestamp intervals containing fluctuations of 5000 milliseconds, 10000 milliseconds, and 15000 milliseconds. A pre-stored time granularity benchmark value is called, which is obtained by averaging the physical response delay of farmland micro-irrigation solenoid valves through hundreds of actual measurements and rounding up, precisely set to 5000 milliseconds. Each value in the extracted set of irregular timestamp intervals is compared modulo this 5000-millisecond benchmark value. When the absolute value of the remainder when a timestamp interval value such as 10000 milliseconds divided by the baseline value of 5000 milliseconds is strictly equal to 0, it is determined that it satisfies an integer multiple relationship. For original timestamp nodes that satisfy this relationship, their absolute time position is retained; for offset nodes with a non-zero remainder, their timestamps are converged to the nearest integer multiple of 5000 milliseconds using rounding logic. In this way, the occurrence time points of the entire time series are forcibly aligned with the physical grid, eliminating millisecond-level time deviations caused by sensor sampling frequency jitter, and generating a time granularity parameter set consisting of an absolutely aligned time node sequence.

[0076] S302: Based on the time granularity parameter set, call up the water volume, fertilizer intensity, soil volumetric moisture content, electrical conductivity and air temperature in the standardized data, divide the time window according to the unified time granularity parameter, calculate the arithmetic mean of the values ​​in multiple time windows and aggregate and rearrange them to obtain the granularity aggregated sequence group.

[0077] The generated time granularity parameter set is used to access a time node array aligned with a fixed step size of 5000 milliseconds. Using this node array as a physical scale, continuous and non-overlapping time windows are segmented on the time axis. For example, a first closed time window is established with a start date of 1620000000000 milliseconds and an end date of 1620000005000 milliseconds. Simultaneously, multi-dimensional environmental and control features from the standardized data are retrieved, including normalized water volume mapping values ​​(0.44), fertilizer intensity mapping values ​​(0.45), relative saturation derived from soil volumetric water content (0.58), and solute mass fraction derived from soil electrical conductivity (0.52). Additionally, air temperature mapping values ​​(0.65) are retrieved from a meteorological station at the same spatial location and normalized by dividing by the highest extreme value of 50 degrees Celsius. The data points with their original timestamps are then placed into the corresponding 5000-millisecond time windows based on their occurrence time. An arithmetic mean is calculated for multiple high-frequency sampled values ​​falling within a single time window. Within the first time window, three fertilization intensity mapping values ​​(0.45, 0.46, and 0.47) are extracted. These three values ​​are summed to obtain a total of 1.38, which is then divided by the total sample size of 3 to calculate the unique representative value of fertilization intensity within that window: 0.46. The same summation and division logic is then performed in parallel for water volume, relative saturation, solute mass fraction, and air temperature within the same window. The average values ​​of the five dimensions calculated within a single window are horizontally concatenated according to a fixed feature sequence to form a single-row multidimensional vector. The vector outputs of all time windows are sequentially arranged along the time axis, and then aggregated and rearranged to generate a granular aggregated sequence group with strictly equidistant temporal features.

[0078] S303: For granular aggregated sequence groups, search for null index positions in multiple time windows, extract the difference between two adjacent values ​​and their corresponding time indices, perform linear scaling operations to fill the null index positions and reconstruct the time series structure to establish a time granular chain.

[0079] The newly generated granular aggregated sequence group is traversed, and the multi-dimensional vector content corresponding to each 5000-millisecond time window is checked row by row. During the traversal process, when a feature value of any dimension within a certain time window is detected as NULL (the memory control character), it is determined that there is a null index position. For example, if the 4th time window timestamp is found to be 1620000015000 milliseconds and there is a gap due to packet loss in the wireless sensor network, then for this null index position, a reverse and forward search is performed along both ends of the time axis to extract the fertilizer intensity value of 0.46 for the 3rd window (timestamp 1620000010000 milliseconds) and the fertilizer intensity value of 0.50 for the 5th window (timestamp 1620000020000 milliseconds). The time index difference operation is then performed on the two valid time nodes, i.e., 1620000020000 - 1620000010000 = 10000 milliseconds. The time span between the null node and the preceding node is 5000 milliseconds. Based on the aforementioned difference parameter, a linear scaling operation is performed. The following equation is used: 0.50 - 0.46 = 0.04. Dividing 0.04 by the total time span of 10000 gives the slope, which is then multiplied by the null node's time span of 5000: 0.04 / 10000*5000 = 0.02. Adding this increment 0.02 to the preceding value of 0.46 yields the precise interpolated value of 0.48 for the null index position. Using the exact same difference operation and slope multiplication-addition logic, independent operations are performed and filled for the remaining four data dimensions. After completing the operations and filling for all missing dimensions, physical breakpoints in the sequence are eliminated, and the time series structure is reconstructed, generating a time-granularity chain.

[0080] Please see Figure 5 The specific steps for obtaining land parcel status data are as follows:

[0081] S401: Based on the time granularity chain, obtain the sensor deployment coordinates and plot boundaries, construct a spatial location vector for the sensor deployment coordinates, construct a closed polygon area by combining the plot boundaries, determine the topological inclusion relationship between the multi-coordinate vector and the closed polygon area, perform the attribution determination and mapping, and generate a plot mapping sequence.

[0082] The system reads the physical media access control addresses of 15 sensor nodes from the time-granularity chain record. By accessing the locally deployed spatial geographic information database, it extracts the two-dimensional planar coordinates of each node in the Cartesian coordinate system. For the sensor node with address code A1, its x-coordinate (500.0 meters) and y-coordinate (600.0 meters) are obtained, constructing a spatial location vector as a combination of x-coordinate 500.0 and y-coordinate 600.0. Simultaneously, it retrieves the spatial boundary information of three target planting plots pre-stored in the geographic information database, extracting the coordinate sequences of the four spatial boundary vertices of plot 1: x-coordinate 400.0 and y-coordinate 500.0, x-coordinate 600.0 and y-coordinate 500.0, x-coordinate 600.0 and y-coordinate 700.0, and x-coordinate 400.0 and y-coordinate 700.0. The extracted four boundary vertex coordinates of plot 1 are then concatenated end-to-end to construct a two-dimensional closed polygon region for that plot. The topology operation logic of the ray crossing method is initiated. Starting from the sensor spatial position vector (horizontal coordinate 500.0, vertical coordinate 600.0) with address code A1, a virtual ray is emitted horizontally in the positive direction of the horizontal coordinate (i.e., the positive direction of the X-axis). The equation of the line is that the vertical coordinate is always 600.0 and the horizontal coordinate is greater than or equal to 500.0. Four boundary segments of the closed polygon are extracted sequentially for intersection determination calculation: For the first boundary (from x-coordinate 400.0 and y-coordinate 500.0 to x-coordinate 600.0 and y-coordinate 500.0), the y-coordinate of this segment is always 500.0, and it does not coincide with the ray's y-coordinate of 600.0, so the number of intersection points is 0; For the second boundary (from x-coordinate 600.0 and y-coordinate 500.0 to x-coordinate 600.0 and y-coordinate 700.0), its y-coordinate interval [500.0, 700.0] covers the ray's y-coordinate of 600.0. The x-coordinate of the intersection point is calculated to be 600.0 through linear interpolation. Since the x-coordinate of the sensor's starting point 500.0 is less than the x-coordinate of the intersection point 600.0, it is determined that the virtual ray and this boundary are in a positive direction. One valid intersection point is generated in the direction of the ray. For the third boundary (from x-coordinate 600.0 and y-coordinate 700.0 to x-coordinate 400.0 and y-coordinate 700.0), the y-coordinate of this line segment is always 700.0, and it does not coincide with the ray's y-coordinate 600.0, so the number of intersection points is 0. For the fourth boundary (from x-coordinate 400.0 and y-coordinate 700.0 to x-coordinate 400.0 and y-coordinate 500.0), its y-coordinate interval [500.0, 700.0] covers the ray's y-coordinate 600.0, and the calculated x-coordinate of the intersection point is 400.0. However, since the sensor's starting point x-coordinate 500.0 is greater than the intersection point x-coordinate 400.0, this intersection point is located in the opposite direction of the ray's emission, so it is determined that no valid intersection point is generated, and the number of intersection points is 0. Summing the determination results of the above four boundaries algebraically, the total number of valid intersection points between the ray and the closed polygon of plot 1 is 1.Based on topological geometry theorems, parity checking logic is triggered. Since the total number of intersections is 1 (odd), it is definitively determined that the sensor node with address code A1 belongs entirely to plot 1 in terms of spatial topology. Similarly, topological operations are performed using this virtual ray and the boundary polygons of plots 2 and 3, resulting in 0 intersections (even), indicating that the node is not within any of the other plots. The plot identifier corresponding to the node determined to be internal is extracted as plot 1, and the sensor node with address code A1 is directly associated with plot 1 in the underlying dictionary structure. The same coordinate extraction and ray intersection topological judgment operations are performed on the remaining 14 sensor nodes, completing the spatial attribution of all node coordinates and generating a set of key-value pairs containing the unique identifier of each node and the corresponding plot identifier—the plot mapping sequence.

[0083] S402: Based on the plot mapping sequence, call the soil volumetric moisture content in the time granularity chain, aggregate the data by plot number and extract the corresponding values ​​of multiple sensors in the same plot, use summation operation to calculate the ratio with the sensor count value, and generate the plot moisture content mean sequence.

[0084] The mapping sequence of the generated plots is retrieved, and the independent memory partition identified as plot number one is locked. Using this as the retrieval key, the current data slice with a timestamp of 1620000005000 milliseconds is traversed in the time granularity chain, filtering and separating the data records of the four target sensor nodes belonging to plot number one. Within the current time slice, the raw monitoring values ​​of soil volumetric moisture content independently collected by these four target sensors are extracted. The moisture content percentage values ​​of these four nodes obtained by reading the underlying serial communication interface registers are 28.5, 29.0, 30.5, and 28.0, respectively. A dedicated floating-point accumulator is allocated in the memory arithmetic area to sum the moisture content of multiple sensors within the same plot. 28.5 + 29.0 + 30.5 + 28.0 = 116.0, indicating that the total soil moisture content of plot number one at this time scale is 116.0. The count value of the number of sensors bound to plot number one, recorded in the mapping sequence, is read synchronously as 4. Subsequently, an arithmetic division operation was triggered, using the summed total moisture content value of 116.0 as the dividend and the sensor count value of 4 as the divisor, to perform a ratio calculation, precisely yielding a value of 29.0. This value of 29.0 represents the overall areal soil moisture saturation state of plot 1 under the current timestamp of 1620000005000 milliseconds. Following the same retrieval, data extraction, iterative summation, and division arithmetic logic, the moisture content values ​​of the 6 sensors in plot 2 and the 5 sensors in plot 3 were independently averaged, yielding a comprehensive average of 26.5 for plot 2 and 31.2 for plot 3. Continuing along the positive time axis of the time granularity chain, the above plot-level data aggregation calculation operation was strictly performed for each subsequent fixed 5000-millisecond time window. The macroscopic values ​​of plot 1 (29.0), plot 2 (26.5), and plot 3 (31.2) generated at various time scales are arranged in a one-dimensional sequence according to their absolute occurrence time and plot number identifier. This effectively eliminates extreme value interference caused by local soil cavities from single-point sensors, ultimately generating a sequence of average plot moisture content values ​​that are representative of real space.

[0085] S403: For the average moisture content sequence of the plot, call the water volume and fertilization intensity in the time granularity chain and obtain the corresponding plot area parameter sequence. Normalize the water volume and fertilization intensity respectively, and re-encode them in the same dimension with the average moisture content sequence of the plot to establish plot status data.

[0086] The stable value of 29.0 corresponding to plot 1 at the timestamp of 1620000005000 milliseconds is locked in the generated average water content sequence of plots. Based on this absolute timestamp index, the cumulative water volume value of plot 1 within this time period and the fertilization intensity characteristics that have been compensated in the previous steps are retrieved synchronously from the time granularity chain. The water volume record value for this time period is extracted as 120.0 liters, and the fertilization intensity result value of 1.48 kg / m² is obtained by substituting the aforementioned parameters. The static attribute configuration relationship table of plots is accessed to obtain the spatial area parameter sequence of plot 1, and its actual physical effective area is extracted as 12.0 m². The data homomorphic normalization preprocessing process is started to completely eliminate the difference in absolute scale of values ​​caused by the physical dimensions of different sensor data sources. For the extracted water volume of 120.0 liters, the maximum safe water volume benchmark of 200.0 liters and the minimum intervention benchmark of 0.0 liters set in the irrigation control records of this plot within the past 30 days are retrieved. Performing a pure arithmetic operation to normalize the range, subtracting the minimum value of 0.0 from the current value of 120.0 (120.0 - 0.0 = 120.0), and then subtracting the minimum value of 0.0 from the maximum value of 200.0 to obtain the range of 200.0, finally dividing the difference 120.0 / 200.0 by 0.60, yields a normalized dimensionless floating-point value of 0.60 for the water volume characteristic. For the fertilization intensity of 1.48 kg / m², the maximum threshold benchmark of 2.00 and the baseline of 0.00 (no fertilization background) mandated in the Fertilization Operation Safety Standard White Paper are used. Using the same subtraction and division process, 1.48 - 0.00 = 1.48 is divided by the range of 2.00, resulting in a normalized dimensionless value of 0.74 for the fertilization intensity. Similarly, for a mean moisture content of 29.0, the extreme value of total water saturation (40.0) and the lower limit of extreme drought (10.0) are used. The result is (29.0-10.0) / (40.0-10.0) = 0.63, yielding a dimensionless value of 0.63 for the moisture content. A floating-point array structure with three elements is created in a contiguous memory address space. The water volume value (0.60), fertilizer intensity value (0.74), and mean moisture content value (0.63) of the same time slice and belonging to the same plot dimension are directly written into the array according to a preset index order. This operation preserves the temporal and spatial mapping of the original data, transforming it into a pure numerical feature set at the underlying level. This series of range operations and vector recombination is performed on all plots and time slices in the time granularity chain, unifying the heterogeneous parameters from multiple sources to a standard range of 0 to 1, ultimately establishing plot state data containing the full multidimensional homomorphic matrix.

[0087] Please see Figure 6 The specific steps for obtaining agricultural management data are as follows:

[0088] S501: Obtain the crop growth stage time based on the plot status data, extract the crop growth stage identifier and perform time series matching operation with the time index in the plot status data, calculate the time series matching distance, and merge the continuous time index range to generate the growth stage time series.

[0089] The system reads the multidimensional feature vector containing soil volumetric moisture content and fertilization intensity encapsulated in the plot status data and extracts the absolute timestamp index bound to this data structure. It accesses a pre-defined crop phenological database and, based on the currently planted maize variety, retrieves its full life-cycle growth stage identifier set and corresponding standard reference times. For the time matching stage, a time-series distance evaluation model is constructed. (The formula is missing from the original text.) middle, Represents the temporal matching distance. This represents the total number of nodes involved in the computation of the land parcel status data. The subscript indicates the specific reproductive stage model being calculated. The subscript represents the current permutation number of the node participating in the calculation; The first in the land parcel status data The time index corresponding to each node. Refers to the actual data source of observation; Representing the The standard time for identifying the crop growth stage corresponding to each node; The scaling factor representing the growth cycle of a crop variety. Refers to a specific variety category; Representing the The weights for determining the temporal state of each node are determined. Regarding the overall formula's operational logic and purpose, firstly, the absolute time deviation between the actual time index and the standard fertility time of each node is calculated using subtraction and absolute value operations. This aims to eliminate the error cancellation effect caused by time drift in both positive and negative directions, accurately reflecting the deviation magnitude at a single point. Then, the absolute deviation is multiplied by a scaling factor to globally scale and calibrate the inherent time shift caused by the varying rates of accumulated temperature absorption for different crop varieties. Next, a division operation is used to divide the calibrated deviation by the temporal state determination weight corresponding to each node. This aims to assign higher confidence to nodes in the core observation area and reduce their distance deviation penalty value. Finally, a summation method is used to accumulate the weighted adjusted deviation values ​​of all nodes as the evaluation standard. Parameter acquisition. The value is set to 3. The total accumulated temperature parameter of corn under standard conditions is extracted as 2800.0 degrees Celsius, while the actual historical total accumulated temperature is 2940.0 degrees Celsius. 2940.0 / 2800.0 = 1.05, and the coefficient is calculated. The value is 1.05. The text of the physical region of the first node is extracted as the core observation area, mapped to a baseline value of 1.00, with a communication arrival rate of 85.0%. The weight is calculated by multiplying these two values. The value is 0.85; the second node region is the core observation area, with a baseline value of 1.00 and a communication rate of 90.0%, and the weights are calculated accordingly. The first node has a base value of 0.90; the third node is an edge region with a base value of 0.80 and a communication rate of 95.0%. The weights are calculated as follows. The value is 0.76. The relative growth time of each node is extracted, and the actual time of the first node is... It takes 10.0 days, standard time. It takes 8.0 days; the second node is actually It is 11.0 days, standard. It takes 9.0 days; the third node is actually It is 12.0 days, standard. The time is 11.0 days. Substitute the above parameters into the formula for calculation. The specific calculation process is as follows: The results indicate that the three consecutive extracted time index nodes are highly fitted to the standard temporal characteristics of the target crop growth stage. The advantage of the formula lies in eliminating the error in accumulated temperature absorption rate caused by microclimate differences by introducing a growth cycle scaling factor calculated based on accumulated temperature comparison. A preset maximum tolerance benchmark value of 10.00 was obtained, and comparing 6.18 with 10.00 determined that no abnormal offset was triggered. Using this determination as the trigger condition, left and right boundary exploration and interval merging operations were performed on the continuous time index range to generate the growth stage time series.

[0090] S502: Based on the time series of the reproductive stage, analyze the sequence of changes in water volume, fertilization intensity, volumetric water content and conductivity within adjacent time granularities and make consistency judgments at each time step. Count and summarize the time points that are judged to be consistent, calculate the ratio based on the time length of the reproductive stage, and generate a consistency ratio sequence.

[0091] The generated time series of reproductive stages is read and located to a specific jointing stage time window. Within this window, with a fixed sampling period of 5000 milliseconds, the feature sequence streams of four core dimensions—water volume, fertilization intensity, soil volumetric moisture content, and soil electrical conductivity—are synchronously extracted from the plot status data at the corresponding time nodes. A first-order difference operation is performed on each extracted dimension data sequence to accurately determine its dynamic evolution trend between adjacent sampling times. Specifically, the soil volumetric moisture content value of the current sampling point (29.5%) is extracted and compared with the value of the previous adjacent sampling point (28.2%). The difference between 29.5% and 28.2% is 1.3%, resulting in a positive change of 1.3%. The change direction identifier for the moisture content is then set to 1, indicating that the value is in an increasing phase. Using the same difference comparison rule, if the calculated difference is less than 0, the change direction identifier is set to -1, indicating that the value is in a decreasing phase; if the difference is equal to 0, the identifier is set to 0, indicating that the value remains stable. For a specific observation point within the jointing stage, the direction identifiers for water volume change (1), fertilization intensity change (1), volumetric water content change (1), and electrical conductivity change (1) are sequentially extracted and calculated. For the sequence of change directions across these four dimensions, a time-by-time consistency logic judgment process is initiated. The four direction identifiers at the same sampling point are input to the logic gate operation unit for homomorphic comparison. A consistent state is determined if and only if water volume increases, fertilization intensity increases, and soil water content and electrical conductivity simultaneously show an upward trend (i.e., all four direction identifier values ​​are 1). If a physical paradox occurs where the water volume identifier is 1 but the water content identifier is -1, it is directly determined as an inconsistent state. A matching counter with a value of 0 is initialized in memory. Each time a consistent state is detected on the time axis, the counter is incremented by 1. After traversing a total of 17280 independent sampling points within the jointing stage, a total count of 15206 consistent states is obtained. The total number of sampling points for this reproductive stage, 17280, was simultaneously acquired. An arithmetic division operation was performed, using the consistency state count value of 15206 as the dividend and the total number of sampling points of 17280 as the divisor, to calculate the consistency percentage for this reproductive stage as 0.88. Following the same data extraction, difference calculation, logical homomorphism determination, and percentage division logic, the consistency percentage for the seedling stage was extracted as 0.75, and the consistency percentage for the emasculation stage as 0.82. The calculated percentages for each reproductive stage were then linearly arranged and encapsulated in absolute chronological order to generate a complete consistency percentage sequence.

[0092] S503: For the consistency ratio sequence, obtain the preset ratio threshold parameter set, compare the consistency ratio corresponding to multiple growth stages with the threshold interval, and map the judgment result into a stage state level coding sequence. Combine and reorganize it with the crop growth stage time series to establish agricultural management data.

[0093] The system retrieves the specific numerical results corresponding to each growth stage from the generated consistency ratio sequence and simultaneously accesses the local strategy database to obtain the preset agricultural management ratio threshold parameter set. This parameter set is obtained by discretizing the intervals based on the growth response characteristics of different crop varieties under standard fertilizer effect control experiments over the past three years. For the consistency ratio value of 0.88 calculated at the jointing stage, the threshold interval comparison and mapping assignment procedure is initiated. The preset high-efficiency management interval range is set to 0.85 to 1.00, with a corresponding state level code of 3; the general management interval range is set to 0.65 to 0.85, with a corresponding state level code of 2; and the risk warning interval range is set to 0.00 to 0.65, with a corresponding state level code of 1. The consistency ratio value of 0.88 at the jointing stage is compared with each of the above-mentioned threshold intervals. Logical operations identify that the value 0.88 is strictly greater than 0.85 and less than 1.00, accurately falling within the high-efficiency management interval. Therefore, the determination result for this jointing stage is directly mapped to level code 3. For the obtained seedling consistency ratio of 0.75, using the same interval comparison logic, it was identified as falling within the general management interval of 0.65 to 0.85, and its level code was mapped to 2. This process is repeated iteratively to complete the interval comparison and coding transformation of the ratio values ​​for all stages throughout the entire growth period, generating a stage status level coding sequence strictly aligned with the time of each growth stage. The underlying growth stage time series is obtained, and structured data containing start and end timestamps and semantic labels is extracted. A joint recombination operation is performed, tightly coupling and encapsulating the stage name, start absolute timestamp, end absolute timestamp, and the newly generated corresponding level code. In a specific jointing stage data packet, the start time 1715308800000 milliseconds, the end time 1715827200000 milliseconds, and the mapped status level 3 are merged and stored across tables, packaged into a control instruction reference frame. This data recombination process is repeated, transforming scattered time-series features and evaluation levels into a multi-dimensional data set with direct control guidance significance, ultimately establishing complete agricultural management data.

[0094] Table 1. Mapping Configuration of Agricultural Management Decision-Making Status

[0095] ;

[0096] To further clarify the specific mapping relationship, this embodiment provides an agricultural management decision-making status mapping configuration table, see Table 1 for details.

[0097] The time calibration module acquires the time calibration signal and the returned timestamp confirmation signal sent by the field acquisition terminal, determines the consistency between irrigation operation time, fertilization operation time, soil moisture and meteorological acquisition time and calibrates them, generates time reference data and transmits it to the data standard module;

[0098] The data standard module acquires irrigation flow, fertilizer quality, soil volumetric moisture content and electrical conductivity based on time reference data, calculates water volume and fertilizer intensity, converts the dimensions of soil volumetric moisture content and electrical conductivity, generates standardized data and transmits it to the time series reconstruction module.

[0099] The time-series reconstruction module reconstructs standardized data with a unified time granularity, aggregates and calculates water volume, fertilization intensity, soil volumetric moisture content, electrical conductivity and air temperature within the same time period, generates a time granularity chain and transmits it to the plot analysis module.

[0100] The plot analysis module obtains the sensor deployment coordinates and plot boundaries based on the time granularity chain, uses a spatial inclusion relationship judgment algorithm to determine the plot ownership of the sensor data, calculates the average volumetric water content, normalized water volume and fertilization intensity of the same plot, generates plot status data and transmits it to the agricultural management module.

[0101] The agricultural management module obtains the crop growth stage time based on plot status data, judges the consistency of the change direction of water volume, fertilizer intensity, volumetric water content and electrical conductivity within adjacent time granularities, divides the growth stage status, and generates agricultural management data.

[0102] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for agricultural information sharing and management based on multi-source data, characterized in that, Includes the following steps: S1: Obtain the time calibration signal and the returned timestamp confirmation signal sent by the field data acquisition terminal, determine the consistency of irrigation operation time, fertilization operation time, soil moisture and meteorological acquisition time and calibrate them, and generate time reference data; S2: Based on the time reference data, obtain irrigation flow, fertilizer application quality, soil volumetric moisture content and electrical conductivity, calculate water volume and fertilizer application intensity, and convert the dimensions of soil volumetric moisture content and electrical conductivity to generate standardized data; S3: Obtain a unified time granularity to reconstruct the standardized data, aggregate and calculate the water volume, fertilization intensity, soil volumetric moisture content, electrical conductivity and air temperature within the same time period, and generate a time granularity chain; S4: Based on the time granularity chain, obtain the sensor deployment coordinates and plot boundaries, use the spatial inclusion relationship judgment algorithm to determine the plot ownership of sensor data, calculate the average volumetric water content, normalized water volume and fertilization intensity of the same plot, and generate plot status data. S5: Based on the plot status data, obtain the crop growth stage time, determine the consistency of the change direction of water volume, fertilizer intensity, volumetric water content and electrical conductivity within adjacent time granularities, divide the growth stage, and generate agricultural management data; The specific steps for obtaining the agricultural management data are as follows: S501: Based on the land parcel status data, obtain the crop growth stage time, extract the crop growth stage identifier and perform time series matching operation with the time index in the land parcel status data, calculate the time series matching distance, and merge the continuous time index range to generate a growth stage time series. S502: Based on the time series of the reproductive stage, analyze the sequence of changes in water volume, fertilization intensity, volumetric water content and conductivity within adjacent time granularities and make a time-by-time consistency judgment. Count and summarize the time periods that are judged to be consistent, calculate the ratio based on the time length of the reproductive stage, and generate a consistency ratio sequence. S503: For the consistency ratio sequence, obtain a preset ratio threshold parameter set, compare the consistency ratio corresponding to multiple growth stages with the threshold interval, and map the judgment result into a stage state level coding sequence, and recombin it with the crop growth stage time series to establish agricultural management data. Specifically, the interval comparison judgment of the consistency ratio corresponding to multiple reproductive stages and the threshold interval is performed by comparing the consistency ratio value corresponding to each reproductive stage in the consistency ratio sequence with the multiple ratio intervals in the preset ratio threshold parameter set, and determining the corresponding stage status level coding sequence based on the judgment result. The time reference data includes irrigation operation time offset, fertilization operation time offset, and soil moisture collection time offset. The standardized data includes water volume, area fertilization intensity, and electrical conductivity standard dimensions. The time granularity chain includes a time index sequence, a water volume aggregate value sequence, a fertilization intensity aggregate value sequence, and a volumetric moisture content aggregate value sequence. The plot status data includes plot number identifier, average plot volumetric moisture content, normalized plot water volume, and normalized plot fertilization intensity. The agricultural management data includes growth stage number, growth stage start and end time period, and stage consistency percentage.

2. The agricultural information sharing and management method based on multi-source data according to claim 1, characterized in that, The specific steps for obtaining the time reference data are as follows: S101: Obtain the time calibration signal sent by the field acquisition terminal and the timestamp confirmation signal returned by the multi-source agricultural equipment, split the time field and extract the second-level and millisecond-level counts, calculate the difference between irrigation operation time, fertilization operation time, soil moisture acquisition time and meteorological acquisition time, and generate a time deviation vector set; S102: Based on the time deviation vector set, the values ​​of multiple deviation vectors are compared with the preset time tolerance threshold item by item, the index positions corresponding to the time tolerance threshold are marked, the arithmetic mean is calculated on the unmarked deviation vectors, the mean result is extracted and bound to the corresponding device identifier to obtain the time calibration parameter sequence. S103: Based on the time calibration parameter sequence, perform numerical translation operations on irrigation operation time, fertilization operation time, soil moisture collection time and meteorological collection time respectively, perform addition and subtraction mapping between multiple time values ​​and corresponding calibration parameter sequence elements, reconstruct the time series structure, and generate time reference data.

3. The agricultural information sharing and management method based on multi-source data according to claim 1, characterized in that, The specific steps for obtaining the standardized data are as follows: S201: Based on the time reference data, obtain irrigation flow and fertilization quality records and align them in time sequence. Perform integral calculation on the instantaneous flow value and corresponding time interval in the irrigation flow record to form a cumulative water volume value sequence. At the same time, calculate the fertilization intensity and generate a set of operation quantification sequences. S202: Based on the operation quantification sequence set, obtain the soil volumetric water content and soil electrical conductivity, extract the time index corresponding to the two types of values ​​and match the index with the operation quantification sequence set, and perform conversion operations on the soil volumetric water content value and soil electrical conductivity value respectively to obtain the soil dimension conversion sequence. S203: Call the operation quantization sequence set and the soil dimension conversion sequence, splice the sequences according to the time scale corresponding to the time base data, extract the minimum and maximum values ​​of the multiple sequence values ​​and perform interval normalization calculation, map the multiple values ​​to a unified scale interval, and generate standardized data.

4. The agricultural information sharing and management method based on multi-source data according to claim 3, characterized in that, The calculation of fertilization intensity uses the following formula: ； in, Represents fertilization intensity. This represents the recorded value of fertilizer application quality. This represents the fertilizer quality compensation constant. Represents the area characterized by irrigation features. This represents the dispersion of fertilizer distribution.

5. The agricultural information sharing and management method based on multi-source data according to claim 1, characterized in that, The specific steps for obtaining the time granularity chain are as follows: S301: Based on the standardized data, obtain a unified time granularity, perform difference calculation on the timestamps of multiple data frames of the standardized data, extract the interval between adjacent timestamps and compare it with the preset time granularity benchmark value, map the time interval values ​​that satisfy the integer multiple relationship to a unified time granularity, and generate a time granularity parameter set. S302: Based on the time granularity parameter set, call up the water volume, fertilizer intensity, soil volumetric moisture content, electrical conductivity and air temperature in the standardized data, divide the time window according to the unified time granularity parameter, calculate the arithmetic mean of the values ​​in multiple time windows and aggregate and rearrange them to obtain the granularity aggregated sequence group. S303: For the granularity aggregated sequence group, search for null value index positions in multiple time windows, extract the difference between two adjacent values ​​and their corresponding time indices, perform linear scaling operations to fill the null value index positions and reconstruct the time series structure to establish a time granularity chain.

6. The agricultural information sharing and management method based on multi-source data according to claim 1, characterized in that, The specific steps for obtaining the land parcel status data are as follows: S401: Based on the time granularity chain, obtain the sensor deployment coordinates and the land parcel boundary, construct a spatial position vector for the sensor deployment coordinates, construct a closed polygon area in combination with the land parcel boundary, determine the topological inclusion relationship between the multi-coordinate vector and the closed polygon area, perform the attribution determination and mapping, and generate a land parcel mapping sequence. S402: Based on the plot mapping sequence, call the soil volumetric moisture content in the time granularity chain, aggregate the data by plot number and extract the corresponding values ​​of multiple sensors in the same plot, use summation operation to calculate the ratio with the sensor count value, and generate a plot moisture content mean sequence. S403: For the average moisture content sequence of the plot, call the water volume and fertilization intensity in the time granularity chain and obtain the corresponding plot area parameter sequence, normalize the water volume and fertilization intensity respectively, and re-encode them in the same dimension with the average moisture content sequence of the plot to establish plot status data.

7. An agricultural information sharing and management system based on multi-source data, characterized in that, The agricultural information sharing and management method based on multi-source data according to any one of claims 1-6, wherein the system comprises: The time calibration module acquires the time calibration signal and the returned timestamp confirmation signal sent by the field acquisition terminal, determines the consistency between irrigation operation time, fertilization operation time, soil moisture and meteorological acquisition time and calibrates them, generates time reference data and transmits it to the data standard module; The data standard module acquires irrigation flow, fertilizer quality, soil volumetric water content and electrical conductivity based on the time reference data, calculates water volume and fertilizer intensity, converts the dimensions of soil volumetric water content and electrical conductivity, generates standardized data, and transmits it to the time series reconstruction module. The time-series reconstruction module reconstructs the standardized data by obtaining a unified time granularity, aggregates and calculates water volume, fertilization intensity, soil volumetric moisture content, electrical conductivity and air temperature within the same time period, generates a time granularity chain and transmits it to the plot analysis module. The plot analysis module obtains the sensor deployment coordinates and plot boundaries based on the time granularity chain, uses a spatial inclusion relationship judgment algorithm to determine the plot affiliation of the sensor data, calculates the average volumetric water content, normalized water volume and fertilization intensity of the same plot, generates plot status data and transmits it to the agricultural management module. The agricultural management module obtains the crop growth stage time based on the plot status data, determines the consistency of the change direction of water volume, fertilizer intensity, volumetric water content and electrical conductivity within adjacent time granularities, divides the growth stage status, and generates agricultural management data.

Images