Regional agricultural planting planning recommendation method and system based on knowledge graph

Abstract

The application relates to the field of wisdom agriculture and agricultural informatization technology, in particular to a regional agricultural planting planning recommendation method and system based on a knowledge graph, which comprises the following steps: acquiring land basic parameters, market order data, daily rainfall, soil volume water content and daily average temperature through an edge gateway; automatically adjusting a sampling time period and extracting time sequence features; calculating a data quality score according to a missing measurement proportion, a sampling frequency ratio and a historical cache coverage mark; constructing an agricultural knowledge graph on a central platform, and screening candidate crops according to a planting previous crop, a continuous cropping avoidance and a water demand matching relationship; combining a static condition matching score, a market order demand matching degree evaluation value, a time sequence condition score and the data quality score to calculate a planting suitability degree; and then combining a regional total irrigation water carrying capacity upper limit and a food crop bottom area lower limit to gradually adjust a planting area, and outputting a regional agricultural planting planning recommendation scheme; and the application improves the rationality of a candidate crop set.

Description

Technical Field

[0001] This invention relates to the fields of smart agriculture and agricultural information technology, specifically to a method and system for recommending regional agricultural planting plans based on knowledge graphs. Background Technology

[0002] With the development of agricultural digitalization and smart agriculture technologies, planting planning recommendations at the county, township, or regional scale have become an important direction in agricultural production management. In order to achieve coordination between crop layout, resource utilization, and market demand within a region, it is particularly important to conduct a comprehensive analysis of land conditions, meteorological moisture, crop rotation relationships, and regional water resource constraints.

[0003] In existing agricultural planting planning recommendations, common methods rely heavily on static plot attributes or historical experience to evaluate crop suitability, making it difficult to reflect the impact of recent rainfall, soil moisture content, and temperature changes on planting decisions in a timely manner. At the same time, field monitoring data is easily affected by factors such as missing measurements, low-frequency sampling, or historical cache coverage. If the reliability of the data is not assessed, the recommendation results may deviate from actual production conditions. In addition, regional planting arrangements need to take into account multiple constraints such as previous crop taboos, avoidance of continuous cropping, crop water requirements, market order quantitative demands, and irrigation water carrying capacity. A single scoring method is insufficient to meet the needs of regional integrated planning.

[0004] Therefore, it is crucial to improve the scientific nature, reliability, and resource utilization efficiency of regional agricultural planting arrangements by combining multi-source agricultural data with dynamic monitoring parameters for time-series feature extraction and data quality evaluation, using agricultural knowledge graphs to constrain and screen candidate crops, and then combining suitability scores and regional total quantity limits to form an executable planting plan recommendation scheme. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for regional agricultural planting planning recommendation based on knowledge graphs. This system can simultaneously process and output planting arrangements based on land parcel conditions, recent meteorological and soil moisture conditions, data reliability, crop rotation relationships, and regional water resource constraints. Furthermore, it effectively prevents the calculation process from falling into infinite loops, improves the stability of system operation, and ensures that water resource constraints converge robustly through multiple iterations. Specifically, the technical solution of this invention is as follows:

[0006] Knowledge graph-based regional agricultural planting planning recommendation methods include:

[0007] Step S1: Obtain multi-source agricultural data through the edge gateway. The multi-source agricultural data includes basic plot parameters, market order data, and dynamic monitoring parameters, including daily rainfall, soil volumetric moisture content, and daily average temperature.

[0008] Step S2: Dynamically adjust the length of the data sampling time period and extract the temporal features within the current time period;

[0009] Step S3: Determine the data quality score based on the data status within the current time period, and upload the data quality score and time series characteristics to the central platform;

[0010] Step S4: Construct an agricultural knowledge graph on the central platform, including land parcel entities, candidate crop entities, and relation edges; use the agricultural knowledge graph to filter candidate crops in the current region based on constraints, and eliminate candidate crops that do not meet the constraints.

[0011] Step S5: For the selected candidate crops, obtain the static condition matching score based on the basic parameters of the plot, obtain the market order demand matching degree evaluation value based on the market order data, and obtain the time series condition score based on the time series characteristics; combine the scores of each and the data quality score to determine the planting suitability score of the candidate crops for the plot.

[0012] Step S6: Based on the planting suitability score and combined with the regional total limit, adjust the planting area and output the recommended regional agricultural planting plan.

[0013] Preferably, step S2 includes:

[0014] Obtain the changes in soil volumetric moisture content and air temperature at adjacent time points;

[0015] The length of the data sampling period is dynamically adjusted based on whether the daily rainfall, soil volumetric moisture content change, and temperature change reach the preset rainfall change threshold, moisture content change threshold, and temperature change threshold, respectively.

[0016] Temporal features are extracted within the adjusted time period, including the average soil volumetric moisture content within the time period, the extent of soil volumetric moisture content recovery within a preset time after rainfall, and the number of consecutive days with the daily average temperature greater than or equal to the preset high temperature threshold.

[0017] Preferably, the process of calculating the data quality score in step S3 includes:

[0018] Obtain the data missing rate, the ratio of actual sampling frequency to nominal sampling frequency, and historical cache overwrite flags for the current time period;

[0019] The data quality score is calculated and determined based on the preset basic quality score, the proportion of missing data, the ratio, and the historical cache coverage mark.

[0020] Preferably, the constraint screening in step S4 includes:

[0021] Traverse the agricultural knowledge graph; if the previous crop of the plot in the current region has a non-continuous cropping relationship with the candidate crop, or if the irrigation level of the plot is lower than the minimum water requirement of the candidate crop, then remove the candidate crop from the candidate crop set of the plot.

[0022] Preferably, step S6 includes:

[0023] Recommended crops are assigned to each plot based on the highest planting suitability score to form an initial set of schemes, and the total water demand for the current region is calculated;

[0024] If the total water demand is greater than or equal to the upper limit of the region's total irrigation water carrying capacity, obtain the unit water consumption of the currently recommended crops. In the plots where the unit water consumption exceeds the standard, replace the recommended crops according to the principle of minimizing the difference in planting suitability scores, until the current region's total water demand does not exceed the standard and meets the minimum area requirement for grain crops, and output the recommended agricultural planting plan for the region.

[0025] A knowledge graph-based regional agricultural planting planning and recommendation system includes:

[0026] The data acquisition module is used to acquire multi-source agricultural data, including basic land parcel parameters, market order data, and dynamic monitoring parameters, through the edge gateway;

[0027] The feature extraction module is used to dynamically adjust the length of the data sampling time period and extract the temporal features within the current time period;

[0028] The quality calculation module is used to determine the data quality score based on the data status within the current time period, and upload the data quality score and time series characteristics to the central platform.

[0029] The constraint filtering module is used to construct an agricultural knowledge graph on the central platform and use the agricultural knowledge graph to filter candidate crops in the current region based on constraints.

[0030] The scoring calculation module is used to obtain the static condition matching score, the market order demand matching degree assessment value, and the time series condition score, and to determine the planting suitability score by combining the scores of each item and the data quality score.

[0031] The planning recommendation module is used to adjust the planting area based on the planting suitability score and the regional total limit, and output the regional agricultural planting planning recommendation scheme.

[0032] Preferably, the feature extraction module is specifically used for:

[0033] The data sampling time period is dynamically adjusted based on preset thresholds for rainfall mutation, moisture content mutation, and temperature mutation, and then temporal features are extracted within the adjusted time period.

[0034] Preferably, the quality calculation module is specifically used for:

[0035] Obtain the data missing rate, the ratio of actual sampling frequency to nominal sampling frequency, and the historical cache coverage flag for the current time period, and calculate and determine the data quality score based on the above parameters.

[0036] Preferably, the constraint filtering module is specifically used for:

[0037] If the previous crop of the plot in the current area has a non-continuous cropping relationship with the candidate crop, or if the irrigation level of the plot is lower than the minimum water requirement of the candidate crop, then the candidate crop will be removed from the candidate crop set.

[0038] Preferably, the planning recommendation module is specifically used for:

[0039] Based on the scores, the recommended crops are allocated to calculate the total water demand. When the total water demand exceeds the standard, the recommended crops are replaced in the plots where the unit water consumption exceeds the standard according to the principle of minimizing the score difference, until the regional water demand and guaranteed area requirements are met.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] 1. By adopting the above technical solution, this invention acquires basic plot parameters, market order data, and dynamic monitoring parameters such as daily rainfall, soil volumetric moisture content, and daily average temperature at the edge gateway. It also automatically adjusts the length of the current data sampling period according to changes in rainfall, soil moisture, and temperature, thus solving the problem that traditional planting planning mainly relies on static attributes or historical experience and is difficult to reflect recent changes in meteorological and soil moisture conditions in a timely manner.

[0042] 2. This invention extracts time-series features such as average soil volumetric moisture content, the rate of increase in soil volumetric moisture content after rainfall, and the number of consecutive high-temperature days within the adjusted current time period. At the same time, it calculates data quality scores by combining the proportion of missing data, the ratio of actual sampling frequency to nominal sampling frequency, and historical cache coverage markers. This can reduce the corresponding recommended score when there are missing dynamic data, low-frequency sampling, or historical cache coverage, thus avoiding the unreasonable impact of unreliable monitoring data on planting decisions.

[0043] 3. This invention constructs an agricultural knowledge graph on a central platform that includes plot entities, candidate crop entities, and relationships such as previous crops, non-continuous cropping, and water requirement matching. It also filters candidate crops based on previous crop restrictions and irrigation level constraints, thereby eliminating crops that do not meet the crop rotation relationship and water requirement conditions in advance and improving the rationality of the candidate crop set. Attached Figure Description

[0044] The above and other objects, features and advantages of the present invention will become clearer from the following description of embodiments of the invention with reference to the accompanying drawings, in which:

[0045] Figure 1 A flowchart illustrating the knowledge graph-based regional agricultural planting planning and recommendation method provided in this embodiment of the invention;

[0046] Figure 2 This is a schematic diagram of a module of a knowledge graph-based regional agricultural planting planning and recommendation system provided in an embodiment of this application. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0048] Knowledge graph-based regional agricultural planting planning recommendation methods include:

[0049] Step S1: Obtain multi-source agricultural data through the edge gateway. The multi-source agricultural data includes basic plot parameters, market order data, and dynamic monitoring parameters, including daily rainfall, soil volumetric moisture content, and daily average temperature.

[0050] Step S2: Dynamically adjust the length of the data sampling time period and extract the temporal features within the current time period;

[0051] Step S3: Determine the data quality score based on the data status within the current time period, and upload the data quality score and time series characteristics to the central platform;

[0052] Step S4: Construct an agricultural knowledge graph on the central platform, including land parcel entities, candidate crop entities, and relation edges; use the agricultural knowledge graph to filter candidate crops in the current region based on constraints, and eliminate candidate crops that do not meet the constraints.

[0053] Step S5: For the selected candidate crops, obtain the static condition matching score based on the basic parameters of the plot, obtain the market order demand matching degree evaluation value based on the market order data, and obtain the time series condition score based on the time series characteristics; combine the scores of each and the data quality score to determine the planting suitability score of the candidate crops for the plot.

[0054] Step S6: Based on the planting suitability score and combined with the regional total limit, adjust the planting area and output the recommended regional agricultural planting plan;

[0055] Step S2 includes:

[0056] Obtain the changes in soil volumetric moisture content and air temperature at adjacent time points;

[0057] The length of the data sampling period is dynamically adjusted based on whether the daily rainfall, soil volumetric moisture content change, and temperature change reach the preset rainfall change threshold, moisture content change threshold, and temperature change threshold, respectively.

[0058] Temporal features are extracted within the adjusted time period. These features include the average soil volumetric moisture content within the time period, the extent of soil volumetric moisture content recovery within a preset time after rainfall, and the number of consecutive days with the daily average temperature greater than or equal to a preset high temperature threshold.

[0059] The process of calculating the data quality score in step S3 includes:

[0060] Obtain the data missing rate, the ratio of actual sampling frequency to nominal sampling frequency, and historical cache overwrite flags for the current time period;

[0061] Based on the preset baseline quality score, the proportion of missing data, the ratio, and the historical cache coverage marker, the data quality score is calculated and determined.

[0062] The constraint filtering in step S4 includes:

[0063] Traverse the agricultural knowledge graph; if the previous crop of the plot in the current area has a non-continuous cropping relationship with the candidate crop, or if the irrigation level of the plot is lower than the minimum water requirement irrigation level of the candidate crop, then remove the candidate crop from the candidate crop set of the plot.

[0064] Step S6 includes:

[0065] Recommended crops are assigned to each plot based on the highest planting suitability score to form an initial set of schemes, and the total water demand for the current region is calculated;

[0066] If the total water demand is greater than or equal to the upper limit of the region's total irrigation water carrying capacity, obtain the unit water consumption of the currently recommended crops. In the plots where the unit water consumption exceeds the standard, replace the recommended crops according to the principle of minimizing the difference in planting suitability scores, until the current region's total water demand does not exceed the standard and meets the minimum area requirement for grain crops, and output the recommended agricultural planting plan for the region.

[0067] This embodiment provides a knowledge graph-based method for recommending regional agricultural planting plans, such as... Figure 1 As shown, this method takes the basic parameters of the land plots, market order data and dynamic monitoring parameters collected by the edge gateway as input, completes the adjustment of the current time period length, extraction of time series features and calculation of data quality score on the edge side, and completes the screening of previous crops, continuous crops and water requirements on the central platform using agricultural knowledge graph. After screening, the suitability score of candidate crops is calculated, and finally, the recommended agricultural planting plan for the region is output by combining the upper limit of the total irrigation water carrying capacity of the region and the lower limit of the guaranteed area of ​​grain crops.

[0068] This implementation method simultaneously processes and outputs planting arrangements based on the plot's own conditions, recent meteorological moisture, data reliability, crop rotation relationships, and regional water resource limitations;

[0069] When acquiring multi-source agricultural data, the edge gateway collects basic plot parameters, market order data, and dynamic monitoring parameters according to the initial data sampling time period to generate a plot data set. The basic plot parameters include plot area, slope, soil type, and irrigation level, etc. The market order data includes the guaranteed purchase demand for crops and the target quantitative parameters. The dynamic monitoring parameters include the numerical characteristics of daily rainfall, soil volumetric moisture content, and daily average temperature at the corresponding time.

[0070] During the adjustment of the current time period length, the change in soil volumetric moisture content at adjacent times is calculated, i.e., the absolute value of the difference between the soil volumetric moisture content at the current time and the previous adjacent time, as well as the change in temperature, i.e., the absolute value of the difference between the daily average temperature at the current time and the previous adjacent time. When the daily rainfall exceeds a preset rainfall mutation threshold, or the change in soil volumetric moisture content exceeds a preset moisture content mutation threshold, or the change in temperature exceeds a preset temperature mutation threshold, the current time period length is shortened. In this embodiment, the rainfall mutation threshold is 50 mm / 24h, the moisture content mutation threshold is 10%, and the temperature mutation threshold is 5℃.

[0071] If the above mutation condition is not triggered for a consecutive preset number of days k, the length of the current time period is extended; the value of the consecutive preset number of days k ranges from 3 to 7 days; otherwise, the length of the current time period remains unchanged; the length of the current time period can be adjusted between the preset lower limit and the preset upper limit of the time period length, with the lower limit being 3 days and the upper limit being 15 days, the shortening step being 2 days and the extension step being 1 day;

[0072] Within the current time period after the length adjustment, extract temporal features; the temporal features include at least the average soil volumetric moisture content within the time period, the extent of soil volumetric moisture content recovery within a preset time after rainfall, and the number of consecutive days with the daily average temperature exceeding a preset high temperature threshold.

[0073] The temporal characteristics are composed of three dimensions: the average soil volumetric moisture content within the current time period, the rate of increase in soil volumetric moisture content within a preset time after rainfall, and the number of consecutive days with the daily average temperature exceeding the preset high temperature threshold. The data quality score is calculated based on the proportion of missing data within the current time period, the ratio of the actual sampling frequency to the nominal sampling frequency, and the historical cache coverage flag. The historical cache coverage flag is a Boolean variable, which takes a value of 1 when the data is covered by the historical cache, and a value of 0 otherwise.

[0074] Specifically, by using basic quality points Multiply by the proportion of missing data in sequence and corresponding deduction weights Determined deduction factor The ratio of the actual sampling frequency to the nominal sampling frequency. and corresponding deduction weights Determined deduction factor and the flags overwritten by historical cache and corresponding deduction weights Determined deduction factor The data quality score was calculated. The calculation formula is as follows:

[0075]

[0076] The calculated data quality score and time-series characteristics are uploaded to the central platform together; the ratio of the actual sampling frequency to the nominal sampling frequency... Before being substituted into the calculation, its upper limit is limited to 1 by the minimum value function;

[0077] The central platform constructs an agricultural knowledge graph, in which entities include land plot entities, candidate crop entities, pest and disease entities, and irrigation level entities; relationships include previous crop, avoidance of continuous cropping, and water requirement matching; attributes are used to record information such as plot area, soil type, slope, irrigation level, and crop water requirement level;

[0078] For the target plot and the target candidate crop, traverse the relationship path in the agricultural knowledge graph. If the previous crop of the target plot and the target candidate crop have a non-continuous cropping relationship, or if the irrigation level of the target plot is lower than the minimum water requirement irrigation level of the target candidate crop, then remove the target candidate crop from the candidate crop set.

[0079] For candidate crops that pass the constraint screening, let the target plot be... Any of the candidate crops it retains is The system calculates a static condition matching score, which reflects the degree of matching between basic conditions such as soil type, slope, and irrigation level of the plot and the requirements of candidate crops. The time series condition score reflects the degree to which current soil moisture, the extent of post-rainfall recovery, and the number of consecutive days of high temperatures match the recent growth or sowing needs of candidate crops. ,in, This represents the set of time-series features extracted within the current time period, and the market order matching score reflecting the degree of matching between candidate crop order demand and demand. and multiply by the values ​​that satisfy the normalization condition ( ) Weighting coefficients , , Perform a weighted summation;

[0080] In this embodiment, based on the principle of prioritizing natural conditions in agricultural planning, the specific values ​​of the weighting coefficients are set as follows: =0.4, =0.4, =0.2; then multiply by the data quality score used to reduce the final score when there are missing dynamic data, low-frequency sampling, or historical cache overwriting. This allows for the calculation of the suitability score for planting this crop on the plot. The calculation formula is as follows:

[0081] During the planning and recommendation phase, each plot is first assigned a candidate crop with the highest suitability score to form an initial set of schemes. The total water demand of the current area is then calculated by multiplying the area of ​​each plot by the unit water consumption of the currently recommended crop for that plot and summing the results.

[0082] When the total water demand in the current area exceeds the upper limit of the total irrigation water carrying capacity of the area, the score difference of the plot corresponding to the recommended high water consumption crop is calculated. The score is the difference between the highest planting suitability score of the original recommended crop and the planting suitability score of the second best candidate crop in the candidate crop set of the plot with the second highest score and less water consumption per unit than the original recommended crop.

[0083] The central platform selects the plot with the smallest score difference, replaces its recommended crop with a low water consumption alternative crop, and repeats the water demand judgment and replacement operation until the current total water demand of the region does not exceed the upper limit of the region's total irrigation water carrying capacity and meets the lower limit requirement for the guaranteed area of ​​grain crops, and outputs a recommended agricultural planting plan for the region.

[0084] During the calculation and replacement of scoring differences, if there are no candidate crops with a unit water consumption lower than the currently recommended crop in the candidate crop set of a certain plot, the plot is marked as unreplaceable and removed from the replacement plot pool of the current round to prevent the system from falling into an infinite loop.

[0085] The replaced suboptimal candidate crop automatically becomes the current new recommended crop, and the replaced suboptimal candidate crop automatically becomes the current new recommended crop. This ensures that the highest planting suitability score and the second highest score always refer to the best and second best in the current set of remaining valid candidates during multiple rounds of repeated judgments, ensuring that water resource constraints converge robustly in multiple iterations. The system completes state updates and logical operations based on iterative calculation rules that dynamically reduce water consumption round by round.

[0086] A knowledge graph-based regional agricultural planting planning and recommendation system includes:

[0087] The data acquisition module is used to acquire multi-source agricultural data, including basic land parcel parameters, market order data, and dynamic monitoring parameters, through the edge gateway;

[0088] The feature extraction module is used to dynamically adjust the length of the data sampling time period and extract the temporal features within the current time period;

[0089] The quality calculation module is used to determine the data quality score based on the data status within the current time period, and upload the data quality score and time series characteristics to the central platform.

[0090] The constraint filtering module is used to construct an agricultural knowledge graph on the central platform and use the agricultural knowledge graph to filter candidate crops in the current region based on constraints.

[0091] The scoring calculation module is used to obtain the static condition matching score, the market order demand matching degree assessment value, and the time series condition score, and to determine the planting suitability score by combining the scores of each item and the data quality score.

[0092] The planning recommendation module is used to adjust the planting area based on the planting suitability score and the regional total limit, and output the regional agricultural planting planning recommendation scheme.

[0093] The feature extraction module is specifically used for:

[0094] The changes in soil volumetric moisture content and temperature at adjacent time points are obtained, and the length of the data sampling period is dynamically adjusted according to the preset thresholds for sudden changes in rainfall, moisture content, and temperature. Then, temporal features are extracted within the adjusted time period.

[0095] The quality calculation module is specifically used for:

[0096] Obtain the data missing rate, the ratio of actual sampling frequency to nominal sampling frequency, and historical cache coverage flags for the current time period, and calculate and determine the data quality score based on the above parameters;

[0097] The constraint filtering module is specifically used for:

[0098] If the previous crop of the plot in the current area has a non-continuous cropping relationship with the candidate crop, or if the irrigation level of the plot is lower than the minimum water requirement of the candidate crop, then the candidate crop will be removed from the candidate crop set.

[0099] The planning recommendation module is specifically used for:

[0100] Based on the scores, the recommended crops are allocated to calculate the total water demand. When the total water demand exceeds the standard, the recommended crops are replaced in the plots where the unit water consumption exceeds the standard according to the principle of minimizing the score difference, until the regional water demand and guaranteed area requirements are met.

[0101] This embodiment provides a knowledge graph-based regional agricultural planting planning and recommendation system, such as... Figure 2As shown; the system includes a data acquisition module, a feature extraction module, a quality calculation module, a constraint screening module, a score calculation module, and a planning recommendation module;

[0102] The data acquisition module is deployed on the edge gateway or used in conjunction with the edge gateway to receive basic plot parameters, market order data, and dynamic monitoring parameters. The edge gateway integrates a programmable logic controller, which serves as the core hardware to perform data acquisition scheduling and preliminary logic judgment for each underlying agricultural sensor.

[0103] The feature extraction module and the quality calculation module can run on the edge side to complete the adjustment of the current time period length, extraction of temporal features, and calculation of data quality scores before the data is uploaded to the central platform; the constraint screening module, the scoring calculation module, and the planning recommendation module can run on the central platform to form the final recommendation scheme by utilizing agricultural knowledge graphs and regional total quantity limits.

[0104] The data acquisition module collects multi-source agricultural data through the edge gateway according to the initial data sampling time period. The multi-source agricultural data includes a combination of data from plot archives, market orders, and field monitoring, specifically including plot area, slope, soil type, irrigation level, crop order information, daily rainfall, soil volumetric moisture content, and daily average temperature. By completing basic access at the edge gateway, dynamic monitoring parameters are directly transmitted to the subsequent feature extraction module.

[0105] The feature extraction module adaptively adjusts the current length of the data sampling period based on daily rainfall, changes in soil volumetric moisture content, and changes in temperature, and generates time-series features. The feature extraction module calculates the change in soil volumetric moisture content at adjacent times by calculating the absolute value of the difference between the current time and the previous adjacent time, and calculates the change in temperature by calculating the absolute value of the difference between the current time and the previous adjacent time's average daily temperature.

[0106] If any of the following conditions are met: daily rainfall exceeds the preset rainfall mutation threshold, soil volumetric moisture content change exceeds the preset moisture content mutation threshold, or temperature change exceeds the preset temperature mutation threshold, the current time period length is shortened; if the above mutation conditions are not triggered for a consecutive preset number of days, the current time period length is extended.

[0107] In other cases, the length of the current time period remains unchanged; the preset time after rainfall is 24 to 72 hours, and the number of consecutive high-temperature days refers to the number of consecutive days when the daily average temperature is greater than the preset high-temperature threshold of 35℃; within the current time period after the length is adjusted, the feature extraction module extracts at least the temporal features of the average soil volumetric moisture content, the increase in soil volumetric moisture content within the preset time after rainfall, and the number of consecutive high-temperature days.

[0108] The quality calculation module obtains the data status within the current time period, performs quality calculation on the data within the current time period based on the data missing ratio, the ratio of actual sampling frequency to nominal sampling frequency, and historical cache overwrite flags, generates a data quality score, and uploads the data quality score and time series characteristics to the central platform.

[0109] The quality calculation module uses a basic quality score, combined with a data loss deduction based on the proportion of missing data, a low-frequency sampling deduction based on the ratio of actual sampling frequency to nominal sampling frequency, and a historical data coverage deduction based on historical cache coverage markers. The corresponding deduction weights can be preset according to sensor type and management requirements. Specifically, for Class I soil sensors with measurement errors within a preset range... Take 0.1, Take 0.15, Take 0.05; for Class II weather stations that are not equipped with this type of sensor, Take 0.2, Take 0.2, Take 0.1;

[0110] When performing low-frequency sampling deduction calculation based on the ratio of actual sampling frequency to nominal sampling frequency, the quality calculation module imposes an upper limit truncation constraint on this ratio, which is recorded as a maximum of 1, to ensure that the data quality score will not exceed the upper limit of the basic quality score when communication anomalies cause repeated reporting.

[0111] The constraint filtering module obtains the candidate crop set for the current region and filters the candidate crops based on the agricultural knowledge graph in the central platform, generating the filtered candidate crop set for each target plot. The agricultural knowledge graph includes plot entities, candidate crop entities, and relationship edges. The relationship edges include previous crop relationships, non-continuous cropping relationships, and water requirement matching relationships.

[0112] For the target plot and the target candidate crop, the constraint filtering module traverses the relationship paths in the agricultural knowledge graph. If there is a non-continuous cropping relationship between the previous crop of the target plot and the target candidate crop, or if the irrigation level of the target plot is lower than the minimum water requirement irrigation level of the target candidate crop, the constraint condition is triggered, and the target candidate crop is removed from the candidate crop set of the target plot.

[0113] The scoring calculation module obtains candidate crops that have passed the constraint conditions and scores them based on the basic parameters of the plot, market order data, time series characteristics and data quality scores to generate a planting suitability score. The scoring calculation module weights and sums the static condition matching score, market order matching score and time series condition score, and multiplies it by the data quality score as a parameter coefficient.

[0114] The planning recommendation module obtains the planting suitability score of each plot and gradually adjusts the initial scheme set according to the upper limit of the region's total irrigation water carrying capacity and the lower limit of the guaranteed area for grain crops, generating a regional agricultural planting planning recommendation scheme; the initial scheme set refers to the data set generated after selecting the candidate crop with the highest planting suitability score for each plot.

[0115] The score difference refers to the difference between the highest planting suitability score and the second highest score of the candidate crop whose unit water consumption is less than that of the original recommended crop. The planning recommendation module repeatedly performs water demand judgment and recommended crop replacement, and prioritizes replacing the plot with the smallest score difference, so that the total water demand of the region gradually decreases and meets the minimum area requirement for grain crops.

[0116] To ensure the robustness of the replacement process, when the planning recommendation module fails to find a candidate crop with a unit water consumption lower than the current recommended crop in a certain plot, the plot is locked and removed from the replacement queue for this round. After each replacement of the recommended crop for a plot is completed, the module dynamically updates the current recommended crop for that plot and, in the next round of water demand judgment, re-finds a new second-best option based on the updated recommended crop status to calculate the score difference.

[0117] This design, which processes business logic through step-by-step state updates, relies on iterative calculation rules that dynamically reduce water consumption round by round. This ensures clear boundaries for system calculation logic, effectively prevents the calculation process from falling into an infinite loop, and improves the stability of system operation.

[0118] As one application example, this system can be deployed on county or township agricultural management platforms; field sensors and weather stations send daily rainfall, soil volumetric moisture content, and daily average temperature to the edge gateway, which receives data according to the initial data sampling time period; the feature extraction module adjusts the length of the current time period based on changes in rainfall, soil moisture, and temperature, and extracts temporal features;

[0119] The quality calculation module calculates the data quality score and uploads both to the central platform; the constraint filtering module in the central platform calls the agricultural knowledge graph to first eliminate candidate crops that do not meet the previous crop taboo and water requirement matching conditions; the scoring calculation module calculates the planting suitability score for the remaining candidate crops;

[0120] The planning recommendation module forms an initial set of schemes based on the highest score. When the total water demand in the current region exceeds the upper limit of the region's total irrigation water carrying capacity, it replaces the crops with low water consumption alternatives according to the principle of minimizing the score difference, until the upper limit of the region's total irrigation water carrying capacity and the lower limit of the guaranteed area for grain crops are met. Finally, the system outputs a regional agricultural planting planning recommendation scheme that includes recommended crops and corresponding planting areas for each plot.

[0121] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A knowledge graph-based method for regional agricultural planting planning and recommendation, characterized in that, include: Step S1: Obtain multi-source agricultural data through the edge gateway. The multi-source agricultural data includes basic plot parameters, market order data, and dynamic monitoring parameters, wherein the dynamic monitoring parameters include daily rainfall, soil volumetric moisture content, and daily average temperature. Step S2: Dynamically adjust the length of the data sampling time period and extract the temporal features within the current time period; Step S3: Determine the data quality score based on the data status within the current time period, and upload the data quality score and the time series features to the central platform; Step S4: Construct an agricultural knowledge graph on the central platform, including land parcel entities, candidate crop entities, and relation edges; use the agricultural knowledge graph to filter candidate crops in the current region based on constraints, and eliminate candidate crops that do not meet the constraints. Step S5: For the selected candidate crops, obtain the static condition matching score based on the basic parameters of the plot, obtain the market order demand matching degree evaluation value based on the market order data, and obtain the time series condition score based on the time series features. By combining the scores of each item and the data quality score, the suitability score for planting the candidate crops in the plot is determined. Step S6: Based on the planting suitability score, adjust the planting area according to the regional total limit, and output the recommended regional agricultural planting plan.

2. The knowledge graph-based regional agricultural planting planning recommendation method according to claim 1, characterized in that, Step S2 includes: Obtain the changes in soil volumetric moisture content and air temperature at adjacent time points; The length of the data sampling period is dynamically adjusted based on whether the daily rainfall, the change in soil volumetric moisture content, and the change in temperature reach preset thresholds for sudden changes in rainfall, moisture content, and temperature, respectively. The time-series features are extracted within the adjusted time period. These features include the average soil volumetric moisture content within the time period, the rate of increase in soil volumetric moisture content within a preset time after rainfall, and the number of consecutive days with the average daily temperature greater than or equal to a preset high-temperature threshold.

3. The knowledge graph-based regional agricultural planting planning recommendation method according to claim 1, characterized in that, The process of calculating the data quality score in step S3 includes: Obtain the data missing rate, the ratio of actual sampling frequency to nominal sampling frequency, and historical cache overwrite flags for the current time period; The data quality score is calculated and determined based on the preset base quality score, the data missing rate, the ratio, and the historical cache coverage flag.

4. The knowledge graph-based regional agricultural planting planning recommendation method according to claim 1, characterized in that, The constraint filtering in step S4 includes: Traverse the agricultural knowledge graph; if the previous crop of the plot in the current area has a non-continuous cropping relationship with the candidate crop, or if the irrigation level of the plot is lower than the minimum water requirement irrigation level of the candidate crop, then remove the candidate crop from the candidate crop set of the plot.

5. The knowledge graph-based regional agricultural planting planning recommendation method according to claim 1, characterized in that, Step S6 includes: Recommended crops are assigned to each plot based on the highest planting suitability score to form an initial set of schemes, and the total water demand for the current region is calculated; If the total water demand is greater than or equal to the upper limit of the region's total irrigation water carrying capacity, obtain the unit water consumption of the currently recommended crop. In the plots where the unit water consumption exceeds the standard, replace the recommended crop according to the principle of minimizing the difference in planting suitability scores, until the current region's total water demand does not exceed the standard and meets the minimum area requirement for grain crops, and output the recommended agricultural planting plan for the region.

6. A regional agricultural planting planning and recommendation system based on knowledge graphs, characterized in that: include: The data acquisition module is used to acquire multi-source agricultural data, including basic land parcel parameters, market order data, and dynamic monitoring parameters, through the edge gateway; The feature extraction module is used to dynamically adjust the length of the data sampling time period and extract the temporal features within the current time period; The quality calculation module is used to determine the data quality score based on the data status within the current time period, and upload the data quality score and the time series features to the central platform. The constraint filtering module is used to construct an agricultural knowledge graph on the central platform and use the agricultural knowledge graph to filter candidate crops in the current region under constraints. The scoring calculation module is used to obtain the static condition matching score, the market order demand matching degree evaluation value, and the time series condition score, and to determine the planting suitability score by combining the scores of the above items and the data quality score. The planning and recommendation module is used to adjust the planting area based on the planting suitability score and the regional total limit, and output a regional agricultural planting planning recommendation scheme.

7. The knowledge graph-based regional agricultural planting planning and recommendation system according to claim 6, characterized in that, The feature extraction module is specifically used for: The changes in soil volumetric moisture content and temperature at adjacent time points are obtained, and the length of the data sampling time period is dynamically adjusted according to preset thresholds for sudden changes in rainfall, moisture content, and temperature, so as to extract the temporal features within the adjusted time period.

8. The knowledge graph-based regional agricultural planting planning and recommendation system according to claim 6, characterized in that, The quality calculation module is specifically used for: Obtain the data missing rate, the ratio of actual sampling frequency to nominal sampling frequency, and the historical cache coverage flag for the current time period, and calculate and determine the data quality score based on the above parameters.

9. The knowledge graph-based regional agricultural planting planning and recommendation system according to claim 6, characterized in that, The constraint filtering module is specifically used for: If the previous crop of the plot in the current area has a non-continuous cropping relationship with the candidate crop, or if the irrigation level of the plot is lower than the minimum water requirement of the candidate crop, then the candidate crop will be removed from the candidate crop set.

10. The knowledge graph-based regional agricultural planting planning and recommendation system according to claim 6, characterized in that, The planning recommendation module is specifically used for: Based on the scores, the recommended crops are allocated to calculate the total water demand. When the total water demand exceeds the standard, the recommended crops are replaced in the plots where the unit water consumption exceeds the standard according to the principle of minimizing the score difference, until the regional water demand and guaranteed area requirements are met.

Images