Data bloodline construction and reasoning method and system for farmland quality monitoring

Abstract

This disclosure provides a method and system for constructing and inferring data lineage for farmland quality monitoring, applicable to the fields of farmland quality monitoring and geographic information data governance. The method includes acquiring multi-source heterogeneous farmland data from a target area, organizing and standardizing it to generate standard farmland quality monitoring indicator data; then, based on the standard farmland quality monitoring indicator data, constructing a farmland quality evaluation system, generating evaluation result data and lineage metadata; further, based on the evaluation result data and lineage metadata, constructing a farmland quality lineage knowledge graph; and visualizing the farmland quality lineage knowledge graph to enable visualized traceability and inference of farmland quality monitoring data lineage. In this way, a fully traceable lineage network can be constructed, and based on this network, visualized queries and bidirectional inference can be achieved, thereby realizing full-link transparency and computable dynamic impact of the evaluation process.

Description

Technical Field

[0001] This disclosure relates to the field of data processing technology, and in particular to the field of farmland quality monitoring and geographic information data governance technology, specifically to a data lineage construction and reasoning method and system for farmland quality monitoring. Background Technology

[0002] Arable land is the cornerstone of food security, and scientific and traceable monitoring and evaluation of arable land quality is one of the important tasks in agricultural modernization. Current arable land quality evaluation relies on the collaborative analysis of multi-source data, including soil physicochemical properties, topographic conditions, and farming systems. The evaluation process is complex and involves a large amount of heterogeneous data.

[0003] Existing open-source and commercial tools for data lineage governance, such as Apache Atlas, OpenLineage, Graph DB solutions (Neo4j / JanusGraph), and several enterprise-level data lineage products, excel at describing structured data lineage and recording batch ETL steps. However, current technologies have significant shortcomings in the governance of arable land quality data.

[0004] On the one hand, there is a lack of end-to-end data traceability and management methods for multi-source heterogeneous farmland quality data. Farmland quality data comes from diverse sources, including on-site sampling and laboratory testing data, as well as remote sensing imagery, drone imagery, IoT data, and historical survey data. Current systems largely rely on relational databases with flat storage structures, making it difficult to model the complex dependencies between data. This results in a fragmented evaluation process. While the system can display data in tables, it cannot reveal the entire chain of dependencies from data collection and indicator calculation to comprehensive evaluation. Users can only see the final numerical results but cannot efficiently trace the data's source path and calculation logic, lacking transparency and credibility. Therefore, the current need is to address how to construct an end-to-end data lineage model that adapts to spatiotemporal heterogeneity. This model should achieve multi-scale spatial association between "sampling points-plots-regions" while making implicit evaluation calculation logic, such as formulas and weights, explicit into a queryable graph structure, thereby solving the problem of fragmented data and result association in the evaluation process.

[0005] On the other hand, existing farmland quality monitoring data lacks an effective reasoning mechanism between the data and the final evaluation results. System management is "one-way" and "static." When a source data point, such as a soil sampling value, is corrected, the system cannot automatically reason (forward) and update all intermediate indicators and final evaluation results that depend on that data. Similarly, when the final evaluation result is abnormal, the system cannot provide an effective reverse tracing path (reverse) to quickly locate the original data or calculation logic that caused the anomaly. This broken data relationship severely reduces the scientific rigor and timeliness of management decisions. Therefore, it is necessary to address how to establish a dynamic, two-way reasoning mechanism based on lineage mapping, breaking through the traditional static management model to achieve forward propagation reasoning from source data correction to the final evaluation result, as well as reverse diagnostic reasoning from result anomalies to source data. Summary of the Invention

[0006] This disclosure provides a method, apparatus, equipment, and storage medium for constructing and inferring data lineage for farmland quality monitoring.

[0007] According to a first aspect of this disclosure, a method for constructing and inferring data lineage for farmland quality monitoring is provided. The method includes:

[0008] Acquire multi-source heterogeneous farmland data for the target area, wherein the multi-source heterogeneous farmland data includes at least two of the following: field sampling record data, laboratory test data, satellite remote sensing image data, UAV data, Internet of Things / sensor data, historical survey data, and statistical data of the target area;

[0009] The multi-source heterogeneous cultivated land data is organized and standardized to generate standard cultivated land quality monitoring index data;

[0010] Based on the standard cultivated land quality monitoring index data, evaluation result data and lineage metadata are generated;

[0011] Based on the evaluation results data and kinship metadata, a kinship knowledge graph of cultivated land quality is constructed.

[0012] The pedigree knowledge graph of cultivated land quality is visualized to enable data pedigree tracking and reasoning in cultivated land quality monitoring.

[0013] In addition to the aspects and any possible implementations described above, a further implementation is provided, wherein the organization and standardization of the multi-source heterogeneous cultivated land data to generate standard cultivated land quality monitoring index data includes:

[0014] Based on preset classification rules and preset unified identification rules, a primary key index system with the unique identifier of the sampling point as the core is constructed according to the multi-source heterogeneous cultivated land data, and the spatiotemporal unit and standardized cultivated land quality index system are determined to generate initial standard cultivated land quality monitoring index data.

[0015] Based on preset individual data anomaly handling rules and preset data relationship anomaly handling rules, the initial standard cultivated land quality monitoring index data is verified to generate standard cultivated land quality monitoring index data.

[0016] As described above and in any possible implementation, a further implementation is provided, wherein the determination of the spatiotemporal unit and the standardized arable land quality index system includes:

[0017] Using the unique identifier of the sampling point as the micro-spatial unit and the map patch number as the macro-spatial unit, the geometric attributes of the map patch and the information of the administrative region to which it belongs are associated to form a spatial description benchmark that combines points and surfaces, and use it as a spatiotemporal unit;

[0018] Based on a pre-set set of indicators, and using the unique identifier of the sampling point as the primary key, a standardized farmland quality indicator system is determined.

[0019] In addition to the aspects described above and any possible implementations, a further implementation is provided, wherein the method further includes:

[0020] Based on the preset naming rules for derived indicators, derived data are generated according to the initial standard farmland quality monitoring indicator data and stored independently.

[0021] As described above and in any possible implementation, a further implementation is provided, wherein generating evaluation result data and lineage metadata based on the standard arable land quality monitoring index data includes:

[0022] Based on preset index classification rules, the standard cultivated land quality monitoring index data are divided, and then each sampled original index is mapped to a single index score using a corresponding preset mapping algorithm. The unique identifier of the sampling point corresponding to the original index, the preset mapping algorithm, and the unique identifier of the sampling point corresponding to the single index score are stored as lineage metadata. The preset mapping algorithm includes a calculation formula, model parameters, and mapping rules.

[0023] A multi-index weighted aggregation algorithm is used to calculate the comprehensive score of each sampling point based on the score of each single index.

[0024] By spatial mapping, the individual index scores and comprehensive scores of each sampling point are mapped to their respective map patches to obtain the evaluation result data.

[0025] As described above and in any possible implementation, a further implementation is provided, wherein constructing a farmland quality kinship knowledge graph based on the evaluation result data and kinship metadata includes:

[0026] Using the unique identifier of each sampling point as the root node, the corresponding original indicators and multi-source heterogeneous cultivated land data are connected to form the first layer of kinship;

[0027] The calculation formulas for each sample are mapped to the scores of individual indicators to form attribute nodes, thus constituting the second layer of lineage.

[0028] The root node is associated with the comprehensive score of each sample to form the third layer of bloodline relationship;

[0029] The fourth layer of bloodline relationships is formed by associating the nodes of the sampled patches with their comprehensive scores.

[0030] Based on the preset graph instantiation mapping rules, a pedestal knowledge graph of arable land quality is constructed according to the first layer of kinship, the second layer of kinship, the third layer of kinship, and the fourth layer of kinship.

[0031] In addition to the aspects described above and any possible implementations, a further implementation is provided, wherein the method further includes:

[0032] Respond to user commands such as load, insert, build, and / or query to display the corresponding data.

[0033] According to a second aspect of this disclosure, a data lineage construction and inference system for arable land quality monitoring is provided. The system includes:

[0034] The acquisition module is used to acquire multi-source heterogeneous cultivated land data of the target area. The multi-source heterogeneous cultivated land data includes at least two of the following: field sampling record data, laboratory test data, satellite remote sensing image data, UAV data, Internet of Things / sensor data, historical survey data, and statistical data of the target area.

[0035] The standardization module is used to organize and standardize the multi-source heterogeneous cultivated land data to generate standard cultivated land quality monitoring index data.

[0036] The generation module is also used to generate evaluation result data and lineage metadata based on the standard cultivated land quality monitoring index data;

[0037] The module is used to construct a knowledge graph of pedigree for arable land quality based on the evaluation results data and pedigree metadata.

[0038] The visualization module is used to visualize the pedigree knowledge graph of cultivated land quality, so as to enable the data pedigree of cultivated land quality monitoring to be visualized, traced, and inferred.

[0039] According to a third aspect of this disclosure, an electronic device is provided. The electronic device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the program to implement the method described above.

[0040] According to a fourth aspect of this disclosure, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the method described above.

[0041] This application provides a data lineage construction and reasoning method for farmland quality monitoring. It acquires multi-source heterogeneous farmland data from a target area, organizes and standardizes it to generate standard farmland quality monitoring indicator data. Based on this standard farmland quality monitoring indicator data, a farmland quality evaluation system is constructed, generating evaluation result data and lineage metadata. Then, based on the evaluation result data and lineage metadata, a farmland quality lineage knowledge graph is constructed. The farmland quality lineage knowledge graph is visualized to enable visual traceability and reasoning of farmland quality monitoring data lineage. Based on this, a unified acquisition and preprocessing mechanism for multi-source heterogeneous spatiotemporal data can be established, laying the foundation for building a data governance model based on lineage relationships. During indicator calculation and evaluation, source data entities, calculation logic, evaluation results, and their dependencies are instantiated as nodes and edges in the graph, thereby constructing a fully traceable lineage network. Based on this lineage network, visual querying and bidirectional reasoning are achieved, thus realizing full-link transparency and computable dynamic impact of the evaluation process.

[0042] It should be understood that the description in the Summary of the Invention is not intended to limit the key or essential features of the embodiments of this disclosure, nor is it intended to restrict the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0043] The above and other features, advantages, and aspects of the embodiments of this disclosure will become more apparent from the accompanying drawings and the following detailed description. The drawings are provided for a better understanding of the invention and are not intended to limit the scope of this disclosure. In the drawings, the same or similar reference numerals denote the same or similar elements, wherein:

[0044] Figure 1 A flowchart is shown of a data lineage construction and inference method for farmland quality monitoring according to an embodiment of the present disclosure;

[0045] Figure 2 A flowchart illustrating a data individual anomaly handling strategy according to an embodiment of the present disclosure is shown;

[0046] Figure 3A flowchart illustrating a data relationship anomaly handling strategy according to an embodiment of the present disclosure is shown;

[0047] Figure 4 A diagram illustrating blood relations according to an embodiment of the present disclosure is shown;

[0048] Figure 5 A schematic diagram of a blood relationship establishment system according to an embodiment of the present disclosure is shown;

[0049] Figure 6 A block diagram of a data lineage construction and inference system for farmland quality monitoring according to an embodiment of the present disclosure is shown;

[0050] Figure 7 A block diagram of an exemplary electronic device capable of implementing embodiments of the present disclosure is shown. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0052] Furthermore, the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0053] This disclosure establishes a unified acquisition and preprocessing mechanism for multi-source heterogeneous spatiotemporal data, laying the foundation for building a data governance model based on lineage relationships. During the indicator calculation and evaluation process, the source data entities, calculation logic, evaluation results, and their dependencies are instantiated as nodes and edges in a graph, thereby constructing a lineage network with full-link traceability. Based on this lineage network, visual querying and bidirectional reasoning are realized, thereby achieving full-link transparency of the evaluation process and computable dynamic impact.

[0054] Figure 1 A flowchart of a data lineage construction and inference method 100 for farmland quality monitoring according to an embodiment of the present disclosure is shown.

[0055] In box 110, acquire multi-source heterogeneous farmland data for the target area. The multi-source heterogeneous farmland data includes at least two of the following: field sampling records, laboratory test data, satellite remote sensing image data, drone data, IoT / sensor data, historical survey data, and statistical data for the target area.

[0056] In some embodiments, the target area and the number and location of sampling points within the target area can be set according to the user's actual needs.

[0057] In box 120, multi-source heterogeneous farmland data are organized and standardized to generate standard farmland quality monitoring index data.

[0058] Because the data sources involved in farmland quality assessment are diverse, have different formats, and lack unified spatiotemporal benchmarks, a unified organization and standardization processing mechanism for multi-source heterogeneous spatiotemporal data will be constructed. This mechanism will transform the aforementioned multi-source heterogeneous farmland data into a unified structured form, providing standard input for the subsequent construction of kinship relationships. That is, it will organize and standardize multi-source heterogeneous farmland data to generate standard farmland quality monitoring indicator data.

[0059] In some embodiments, the above-mentioned organization and standardization of multi-source heterogeneous cultivated land data to generate standard cultivated land quality monitoring index data specifically includes:

[0060] Based on preset classification rules and preset unified identification rules, a primary key index system with the unique identifier of sampling points as the core is constructed according to multi-source heterogeneous cultivated land data, and the spatiotemporal units and standardized cultivated land quality index system are determined to generate initial standard cultivated land quality monitoring index data.

[0061] Based on preset rules for handling individual data anomalies and preset rules for handling anomalies in relationships between data, the initial standard cultivated land quality monitoring index data is verified, and standard cultivated land quality monitoring index data is generated.

[0062] In some embodiments, preset classification rules and preset unified identifier rules can be set according to the user's actual needs in order to define the classification access specifications and unique identifier generation rules for multi-source heterogeneous cultivated land data, thereby solving the problem of association of heterogeneous data.

[0063] In some embodiments, the preset classification rules can be set to classify and integrate multi-source heterogeneous cultivated land data into the following data organizations:

[0064] Instantaneous sampling data: including field sampling records or laboratory test data; the format can be plain text, such as CSV (Comma-Separated Values) format, or Excel spreadsheet format, i.e., XLSX format; used to characterize the physical and chemical indicators of cultivated land at a specific point in time;

[0065] Spatial monitoring data includes sampling point data, remote sensing / GIS image extraction results (such as administrative division or land parcel data in Shp format), UAV data, and IoT / sensor data; used to characterize the spatial boundaries of cultivated land, land use properties, meteorological conditions, and spatial coordinates of sampling points.

[0066] Historical and time-series data: including historical survey archives of arable land quality, statistical data (such as crop rotation records), etc., to provide benchmark references and historical background constraints for evaluation.

[0067] In some embodiments, the pre-defined unified identification rule mainly establishes a primary key index system centered on the "unique identifier of the sampling point," namely the File Identifier Descriptor (FID). All accessed data adopts standardized naming, with the prefix FID and the suffix being an abbreviation of the data and number format, connected by underscores. For example, UAV image data is named 001_wrj001.

[0068] In some embodiments, a standardized spatiotemporal characteristic description dimension and core indicator system can be constructed to determine the spatiotemporal unit and the standardized cultivated land quality indicator system, thereby generating initial standard cultivated land quality monitoring indicator data.

[0069] In some embodiments, the above-mentioned system for determining spatiotemporal units and standardized arable land quality indicators includes:

[0070] Using the unique identifier of the sampling point as the micro-spatial unit and the map patch number as the macro-spatial unit, the geometric attributes of the map patch and the information of the administrative region to which it belongs are associated to form a spatial description benchmark that combines points and surfaces, and use it as a spatiotemporal unit;

[0071] Based on a pre-set set of indicators, and using the unique identifier of the sampling point as the primary key, a standardized farmland quality indicator system is determined.

[0072] In some embodiments, the preset set of indicators can be set according to the user's actual needs.

[0073] In some embodiments, spatial dimension standardization may include: using “unique identifier of sampling point (FID)” as a micro-spatial unit and “map patch number (such as TB1, TB2, etc.)” as a macro-spatial unit, and automatically associating the geometric attributes of the map patch (such as area SHAPE_AREA, perimeter SHAPE_LENGTH) and the information of the administrative region to which it belongs, to form a spatial description benchmark that combines points and surfaces.

[0074] In some embodiments, the standardized farmland quality indicator system defines a set of indicators covering all elements, i.e., a pre-defined set of indicators, which may include:

[0075] a. Infrastructure and site conditions: drainage capacity (PSNL), irrigation capacity (GGNL), agroforestry grid degree (NTLWHCD), field road accessibility (TDDBZD).

[0076] b. Soil physical properties: topsoil texture (GCZD), topsoil thickness (GCZHD), soil bulk density (TRZDR), soil aggregate structure (TJTJG);

[0077] c. Soil chemical properties: Soil organic matter (YJZ), pH value (PHZ), total nitrogen (QN), available nitrogen (JJD), available phosphorus (YXL), available potassium (SXJ), cation exchange capacity (YLZJHL);

[0078] d. Topographic conditions: ground slope (DMPD), elevation (HB), aspect (PX);

[0079] e. Environmental quality (optional): Heavy metal pollution level (ZJSRCD), pesticide residue (NYCL);

[0080] f. Attribute indicator set: covers farming information such as past crops and current crops.

[0081] In some embodiments, to ensure the accuracy and logical consistency of farmland quality monitoring indicator data files, a hierarchical data anomaly handling mechanism can be set up, which can be a knowledge base-based intelligent data cleaning and anomaly verification mechanism.

[0082] In some embodiments, preset rules for handling individual data anomalies and preset rules for handling anomalies in relationships between data can be set according to the user's actual needs.

[0083] like Figure 2 As shown, the data individual anomaly handling strategy, i.e., the preset data individual anomaly handling rules, may include:

[0084] First, determine the data type;

[0085] If the data type is numerical, then perform a neighborhood consistency check, and then calculate the local mean. and local standard deviation ;judge Where x is the sample value and k is the standard deviation multiple; if it is greater than, it is marked as an anomaly; otherwise, the data is entered.

[0086] If the data is text, a procedure dictionary match is performed; if the match fails, an exception is marked; if the match succeeds, a domain consistency check is performed, and if the check succeeds, the entered data is verified; if the check fails, an exception is marked.

[0087] In some embodiments, the data individual anomaly handling strategy can also be understood as a null and missing value handling and data completion strategy. Specifically, for required fields, such as FID, strong validation is performed; if missing, import is rejected and abnormal data is recorded; for optional fields, such as the storage path for multi-source heterogeneous data, a missing value marker (NULL) is used. To avoid errors caused by inconsistent units, a standard unit is defined for each field, such as total nitrogen g / kg; alkaline nitrogen mg / kg. During import, mismatched units are prompted and the conversion is completed according to the configuration.

[0088] like Figure 3 As shown, the data relationship anomaly handling strategy, i.e., the preset data relationship anomaly handling rules, may include:

[0089] Load the Rule Knowledge Base (RKB);

[0090] Iterate through each rule;

[0091] Calculate the actual deviation from the expected value;

[0092] Calculate whether it deviates from the threshold;

[0093] If yes, mark it as an error; otherwise, enter the data.

[0094] In some embodiments, the RKB model integrates standard constraints, soil science mechanisms, and historical experience models, and consists of a series of rule triples <subject index, relation type, object index>, as shown in Table 1, along with attribute information, which can be used for logical consistency verification based on RKB.

[0095] Table 1: Soil Relationship Rule Base

[0096]

[0097] In some embodiments, during logical consistency verification based on RKB, examples of rules include, but are not limited to, "soil acidity is positively correlated with aluminum ion activity," "pH value is positively correlated with exchangeable calcium / magnesium content," "the ratio of organic carbon to total nitrogen (C / N ratio) must be within a reasonable threshold range," and "electrical conductivity is positively correlated with total soluble salts." The verification process involves the system traversing the multi-indicator combinations at each sampling point and matching them with the rule constraints in RKB. If a combination of indicators (such as a single indicator contradicting other indicators, or the overall combination deviating from the normal confidence interval defined by the rule) violates the constraints, the system determines it as "abnormal relationship" and marks it. The correction mechanism involves triggering a manual review process for the data marked as abnormal, verifying the accuracy of the data, and performing corrections or re-supplementation, ultimately outputting a fully compliant standardized dataset.

[0098] In some embodiments, the above method further includes:

[0099] Based on the preset naming rules for derived indicators, derived data are generated according to the initial standard farmland quality monitoring indicator data and stored independently.

[0100] In some embodiments, the pre-defined naming rules for derived indicators can be set according to the user's actual needs, thereby enabling derived data management and multi-source complementary optimization.

[0101] On the one hand, the results data is stored in isolation, and the naming rules for derived indicators can be predefined to generate score fields with the suffix "_S" (such as PSNL_S, GGNL_S) and the comprehensive mean score field (score). The above-mentioned derived data is stored in a separate score file and does not write back to overwrite the original source file, thereby ensuring the originality and traceability of the source data.

[0102] On the other hand, the standardized data table generation, after the aforementioned processing and manual verification, produces a standardized farmland quality monitoring indicator data file. This file is stored in tabular form, with the sampling point number (FID) as the primary key, integrating corrected spatiotemporal characteristics and complete quality indicators. In particular, for unstructured heterogeneous data, such as image files, their storage path is recorded in the attribute column corresponding to the sampling point number, realizing a unified index for heterogeneous data in the logical table structure.

[0103] As can be seen, the above data organization and standardization process can be summarized as follows: Organizing multi-source heterogeneous farmland quality data using FID as a unified prefix ensures the logically accurate mounting of different modalities; clarifying the core spatiotemporal units and standardization system required for farmland evaluation defines key spatiotemporal characteristics and standardizes the indicator system; constructing independent storage and standardized tables achieves source data protection and unified indexing of heterogeneous information, completing derivative data management and multi-source complementary optimization; ensuring the integrity and consistency of source data through manual verification and rule detection; based on this, by uniformly organizing, standardizing, and verifying the quality of multi-source heterogeneous farmland data through the above process, a standardized data table with a consistent structure containing a correct farmland quality evaluation system and directly usable for lineage construction can be formed, namely, standard farmland quality monitoring indicator data. This table serves as the input to the farmland quality lineage construction system to establish the lineage logic chain between data.

[0104] In box 130, evaluation result data and lineage metadata are generated based on standard farmland quality monitoring index data.

[0105] In some embodiments, to address the issues of opacity and difficulty in traceability in the current evaluation process, the above-mentioned standard farmland quality monitoring indicator data files can be used to calculate and spatially map farmland quality evaluation indicator scores, thereby achieving full-chain traceability of farmland quality evaluation.

[0106] For example, a hierarchical and progressive evaluation system of "indicator quantification - comprehensive evaluation - spatial mapping" was established. Based on the quality evaluation standard document, it realizes the complete quantitative calculation from the original indicators to the scores of sampling points and patch scores, and records all calculation basis to support lineage traceability. It can be seen that this system not only produces evaluation result data, but more importantly, it simultaneously generates lineage metadata describing the calculation logic, laying the foundation for building a traceable data lineage network.

[0107] In some embodiments, the generation of evaluation result data and lineage metadata based on standard cultivated land quality monitoring index data specifically includes:

[0108] Based on the preset index classification rules, the standard cultivated land quality monitoring index data are divided, and then the corresponding preset mapping algorithms are used to map each sampled original index into a single index score. The unique identifier of the sampling point corresponding to the original index, the preset mapping algorithm, and the unique identifier of the sampling point corresponding to the single index score are stored as lineage metadata. The preset mapping algorithm includes a calculation formula, model parameters, and mapping rules.

[0109] By using a pre-set multi-index weighted aggregation algorithm, the comprehensive score of each sampling point is calculated based on the score of each single index.

[0110] By spatial mapping, the individual index scores and comprehensive scores of each sampling point are mapped to their respective map patches to obtain the evaluation result data.

[0111] In some embodiments, the preset indicator classification rules, preset mapping algorithms, and preset multi-indicator weighted aggregation algorithms can be specifically set according to the user's actual needs.

[0112] For example, in the preset indicator classification rules, indicators can be divided into three categories based on their data characteristics: hierarchical indicators, nonlinear continuous indicators, and linear continuous indicators. The preset mapping algorithm can include predefined mapping tables, nonlinear mapping functions, and linear mapping functions.

[0113] In some embodiments, in "Indicator Quantification", the scores of single monitoring indicators at sampling points can be classified and quantified.

[0114] In some embodiments, after standardizing the multi-source heterogeneous data, the standard cultivated land quality monitoring index data are divided into three categories based on the data characteristics of the indicators: graded indicators, nonlinear continuous indicators, and linear continuous indicators. Different algorithm models are then used to map the original attribute values ​​into standardized single-indicator scores, identified by the suffix "_S". During this process, the calculation formulas, model parameters, and mapping rules can be stored in a structured manner as lineage metadata.

[0115] Specifically, the original indicators are first systematically classified into three categories: graded indicators, nonlinear continuous indicators, and linear continuous indicators. Graded indicators mainly include soil quality grades, soil types, and other qualitative or graded evaluation items. A predefined mapping table is used to convert different grades into single-indicator scores, thereby achieving a quantitative expression of qualitative information. Graded indicators are applicable to qualitative indicators or graded attributes, such as soil quality grades (fully satisfied, satisfied, basically satisfied, not satisfied), etc. Each category corresponds to a standard score, and the mapping rules are implemented through a lookup table. For example, the mapping table for each sampled original indicator to single-indicator scores is shown below:

[0116]

[0117] The soil quality grades are as follows: "fully satisfied" corresponds to a score of 1, "satisfied" corresponds to 0.8, "basically satisfied" corresponds to 0.6, and "not satisfied" corresponds to 0.3. This rule ensures that discrete data from different sources can be uniformly mapped to a standardized numerical range.

[0118] Secondly, nonlinear continuous indices cover soil physicochemical properties, trace element content, and environmental factors. Through nonlinear mapping functions, such as Gaussian functions or reciprocal quadratic functions, the index values ​​are scientifically mapped to ideal reference values. This ensures sensitivity near the optimal values ​​while reasonably penalizing deviations, reflecting the scientific laws of soil and environmental constraints. Nonlinear continuous indices are suitable for continuous indices, especially physicochemical properties exhibiting nonlinear responses, such as organic matter content, total nitrogen content, pH value, salinity, and electrical conductivity. By designing non-isotropic Gaussian or reciprocal quadratic functions, the original values ​​are mapped to scores, as shown below:

[0119]

[0120] Where u is the original value (such as organic matter content, total nitrogen content, pH value, etc.), c is the ideal reference value or optimal value of the indicator, and a>0 is an empirical coefficient or sensitivity adjustment parameter. Each indicator corresponds to different a and c values, which can be obtained from the standard evaluation documents of various regions, as shown in Table 2, thereby realizing dynamic penalties for deviations from the optimal value. This rule can preserve the sensitivity and scientific interpretability of the indicator under different values.

[0121] Table 2: Indicator Mapping and Score Calculation Parameters

[0122]

[0123] Finally, the linear continuous index performs linear normalization on environmental factors such as slope and sampling depth that exhibit a linear relationship with numerical values, achieving an interpretable mapping between the index and the score. The linear continuous index is suitable for indicators that have a linear impact with numerical values, such as environmental constraint factors like ground slope, sampling depth, or height. By mapping the raw data to a score interval through a linear function, it ensures a direct linear relationship between the score and the index change, facilitating the calculation and interpretation of comprehensive evaluations, as shown below:

[0124]

[0125] Where u is the original value (such as slope, sampling depth, etc.), a is the proportional coefficient, and b is the offset coefficient. Each indicator corresponds to a different value of a and b.

[0126] Based on the above calculation rules, a score column corresponding to each sampling point is generated, and the formula and input parameters of each calculation step are retained, realizing full-chain traceability from a single indicator to the score node. Each score node is associated with its original indicator value, calculation formula, and mapping function parameters, thereby ensuring that the source of each calculation result can be completely traced in terms of lineage.

[0127] In some embodiments, the "Comprehensive Evaluation" can be a weighted aggregate calculation of the comprehensive evaluation scores of the sampling points.

[0128] In some embodiments, a score for each sampling point is calculated using a multi-index aggregation method to reflect the overall farmland quality level of that sampling point.

[0129] First, during the comprehensive scoring process, the contribution weight of each individual indicator score allows the sub-nodes to be traced back to each individual indicator score node, and further back to the original sampled data. The preset multi-indicator weighted aggregation algorithm is shown below:

[0130]

[0131] in, n Indicates the number of indicator data. s j Indicates indicator data, This indicates the weight of the corresponding indicator.

[0132] Next, blood ties are established, and the comprehensive score nodes generated in the above process are weighted by parameters. By establishing logical connections with the scoring nodes of each individual indicator, a reverse tracing path is achieved from "overall results" to "individual indicators" and then to "raw data".

[0133] In some embodiments, spatial topological mapping of farmland quality evaluation scores can be performed in the "spatial mapping" comprehensive evaluation.

[0134] In some embodiments, to address the spatial scale differences between sampling points (point data) and management patches (area data), spatial topology matching can be used to map evaluation results from micro to macro levels. The mapping process is as follows:

[0135] a. Use spatial analysis tools, such as Spatial Join or Point in Polygon algorithms in GIS, to perform topological inclusion determination on the geographic coordinates of each sampling point and the polygon boundaries of all polygons;

[0136] b. If the sampling point P_i falls within the geometric range of the patch Polygon_k, then establish the spatial attribution relationship and assign the single index score and comprehensive score of P_i to the patch Polygon_k;

[0137] c. If a patch contains multiple sampling points, the final score of the patch layer is calculated according to preset rules, such as area weighting or arithmetic mean.

[0138] After establishing the mapping relationship, this process establishes a logical connection between the data space and the geographic space, ensuring that the quality evaluation results of the map features can be accurately traced back to the specific sampling point data within them.

[0139] As can be seen, the above scoring calculation and spatial mapping process can be summarized as follows: mapping the original indicators into quantifiable scores and preserving the complete calculation chain to support lineage tracing; calculating the comprehensive score of sampling points through multi-indicator weighted aggregation, reflecting the overall farmland quality while ensuring that each score is traceable to the single indicator and the original data; and summarizing the sampling point scores to their respective map patches through spatial mapping to achieve weighted comprehensive evaluation at the map patch level, while maintaining traceability from point to surface. Based on this, the entire process involves more than just numerical calculations; its core outputs include two parts: firstly, the calculation result data, i.e., the evaluation result data, which covers the single indicator scores and comprehensive scores of sampling points and map patches; secondly, the generation of lineage metadata, which structurally records the "input data FID - calculation formula / mapping table - model parameters (a,b,c,w) - output data ID" used in each calculation. These two outputs will serve as direct inputs for the subsequent construction of a farmland quality lineage knowledge graph, enabling the subsequently generated graph nodes to have self-explanatory capabilities and completely solving the "fragmentation" problem of the current evaluation process.

[0140] In box 140, a knowledge graph of pedigree for farmland quality is constructed based on the evaluation results data and pedigree metadata.

[0141] In some embodiments, after the above-mentioned index calculation and metadata generation are completed, it will enter the instantiation construction stage of the data lineage graph. The core of constructing the cultivated land quality lineage knowledge graph lies in using the graph structure characteristics of a graph database, such as Neo4j, to transform the above-mentioned "evaluation result data" and "lineage metadata" into a knowledge graph composed of nodes (Node) and edges (Edge), and constructing a multi-level and full-link traceable cultivated land quality evaluation lineage network. Its specific construction process is divided into the following four levels: the construction of the source data entity layer (root node and attribute association), the construction of the logical calculation layer, the construction of the index aggregation layer, and the construction of the spatial mapping layer. In other words, by constructing multi-level ternary relationships between sampling points, attribute values, single-item scores, comprehensive scores, and patch nodes, the lineage knowledge graph is constructed to achieve full-chain lineage mapping of cultivated land quality evaluation data.

[0142] In some embodiments, constructing the cultivated land quality lineage knowledge graph according to the evaluation result data and lineage metadata specifically includes:

[0143] Taking the unique identifier of each sampling point as the root node, connecting the corresponding original indicators and multi-source heterogeneous cultivated land data to form the first-level lineage relationship;

[0144] Taking the mapping between the calculation formula of each sampling and the single-index score as the attribute node to form the second-level lineage relationship;

[0145] Taking the comprehensive score of each sampling to associate with the root node to form the third-level lineage relationship;

[0146] Taking the comprehensive score of each sampling to associate with the patch node to form the fourth-level lineage relationship;

[0147] Based on the preset graph instantiation mapping rules, construct the cultivated land quality lineage knowledge graph according to the first-level lineage relationship, the second-level lineage relationship, the third-level lineage relationship, and the fourth-level lineage relationship.

[0148] In some embodiments, in the source data entity layer, that is, in the construction process of the first-level lineage relationship, it is necessary to establish the foundation of data traceability. Taking the FID of the sampling point as the root node of the graph, and mounting the original attribute indicators and multi-source heterogeneous data files of the sampling point as attribute nodes.

[0149] First, perform attribute association, that is, traverse the standardized data table. For the existing index fields, construct a triple <FID, index field name, field value>. For example, establish the relationship <001, available potassium, 98.85>, indicating that the available potassium content of sampling point 001 is 98.85.

[0150] Next, perform heterogeneous data indexing. That is, for unstructured multi-source heterogeneous data, construct triples <FID, storage path, heterogeneous data file name>. For example, establish the relationship <001,F:\wrj\001\_wrj,001\_wrj>.

[0151] Thus, establish a direct association between the sampling points and their attribute nodes, and save the field name or path as the relationship text. In this way, each sampling point is connected to the corresponding attribute node through its unique number, forming the first layer of relationship in the lineage.

[0152] In some embodiments, in the logical calculation layer, that is, during the construction of the second layer of lineage, adopt the "logical entity-based on lineage dependence" modeling method to solve the problem of opaque evaluation process.

[0153] First, perform logical mapping. That is, for each single-index score column ending with "_S", extract the "original attribute value" it depends on as the head entity, extract the "score value" as the tail entity, and instantiate the "calculation formula and parameters" recorded above as the edge connecting the two.

[0154] Next, perform triple construction, that is, construct the relationship of <original attribute value, calculation formula description, index score value>. For example, establish the relationship <98.85, "y=1 / (1+0.000081*(u-181.622535)^2)", 0.64>.

[0155] The relationship edge here is no longer a simple connection line, but carries the specific calculation logic (formula). This realizes the explicit mapping from the original data to the index score, enabling users to directly consult the derivation process of the score in the graph, which forms the second level of the lineage.

[0156] In some embodiments, in the index aggregation layer, that is, during the construction of the third layer of lineage, this layer aims to express the contribution relationship of local indexes to the overall quality, and embed the index scores into the comprehensive evaluation system through two types of paths.

[0157] First, build the weight contribution path. That is, establish a directed edge from each single score node to the comprehensive score (Score) node, and use "weight" as the attribute of the edge. Construct the triple <single score, weight attribute, Score>, for example <0.64,weight:0.25,0.85>, which intuitively reflects the weighted logic of "local scores converging into the overall evaluation".

[0158] Next, build the overall evaluation path, that is, establish an association from the sampling point FID directly to the Score node, and construct the triple <FID, "cultivated land quality score", Score>, for example <001, cultivated land quality score, 0.85>.

[0159] The Score node itself, as an important lineage node, has numerical attributes and composition information, and can support subsequent visual display and query. By establishing the relationship between the Score node and its corresponding index scores, the reliability of the data source for comprehensive evaluation quality is ensured, facilitating tracing the lineage relationship between the overall quality and its components. At the same time, a direct association with the root node is established, facilitating tracing the overall quality evaluation of the sampling points. In summary, the relationships established among the comprehensive score node, its index score nodes, and the root node form the third layer of the lineage relationship.

[0160] In some embodiments, during the construction of the spatial mapping layer, i.e., the fourth layer of the lineage relationship, the evaluation results at the statistical level are mapped to geographical spatial units (map patches), achieving the unity of macro and micro.

[0161] First, perform spatial attribution association, that is, based on the above spatial matching results, establish the attribution relationship between the sampling point FID and the map patch number node, and construct a triple <FID, "belongs to map patch", map patch number>, such as <001, TB1, 1TB01>.

[0162] Next, perform geometric attribute extension, that is, attach the geometric information of the map patch to the map patch node, and construct a triple <map patch number, geometric attribute name, attribute value>, such as <1TB01, "SHAPE_AREA", 45514.61>.

[0163] Subsequently, perform quality evaluation projection, that is, establish the evaluation relationship between the Score node and the map patch node, and construct a triple <Score, "quality", map patch number>, such as <0.85, "quality", 1TB01>.

[0164] This level reflects the projection of the cultivated land quality results in the geographical space, enabling the atlas to not only reflect the calculation logic at the data level but also be associated with actual geographical space entities. Through the establishment of the above three relationships, the quality evaluation is mapped from the data space to the real geographical space. Through the association of the root node and the score node, a quality evaluation system lineage chain from the sampling point to the map patch is realized, thus forming the fourth layer of the lineage relationship.

[0165] As can be seen, the above-mentioned process of constructing kinship source data can be summarized as follows: taking the sampling point FID as the root node, connecting its indicators and heterogeneous data to form the first layer of kinship; establishing a clear mapping between attribute nodes and their indicator score nodes through corresponding calculation formulas to form the second layer of kinship; aggregating the scores of each individual item through weights in the comprehensive score node, while directly associating it with the root node of the sampling point, to form the third layer of kinship; mapping the sampling point scores to the patch nodes, and establishing the association between the patch and its geometric attributes and comprehensive score, to form the fourth layer of kinship, thereby realizing the full-chain kinship modeling of cultivated land quality evaluation from data space to geographic space. Based on this, a clearly hierarchical kinship map is constructed through the above four steps. Figure 4 As shown, this map fully connects the entire chain of "original sampling points - single index scores - comprehensive scores - map spatial units", ensuring the traceability of data sources, the interpretability of calculation logic, and the scientific nature of evaluation results in the process of monitoring and evaluating farmland quality.

[0166] In some embodiments, the preset map instantiation mapping rules can be set according to the user's actual needs.

[0167] In some embodiments, after the underlying construction of the kinship map is completed, the map can be automatically instantiated based on mapping rules.

[0168] Specifically, an automated mapping engine for "relational data to graph data" can be built-in. This engine first reads the latest standardized source data file and calls the aforementioned evaluation algorithm module to generate intermediate score files. Subsequently, based on predefined mapping logic, the two-dimensional tabular data is converted into a topological structure in a graph database (such as Neo4j).

[0169] Node instantiation in the topology includes:

[0170] Map the FID column in the table to the root node, the Score column to the aggregation node, and the patch number to the spatial entity node.

[0171] Relation instantiation in a topology includes:

[0172] Following the principle of "data is logic", the original indicator field names are instantiated as "attribute association" edges; the generated calculation formula metadata is instantiated as "calculation logic" edges; the weight parameters are instantiated as "contribution factors" edges; and the map patch attribution is instantiated as "spatial projection" edges.

[0173] Through the above process, the entire lineage structure from micro-source data to macro-land parcel quality can be fully reconstructed in the graph database, providing a data foundation for subsequent inference.

[0174] In box 150, the pedigree knowledge graph of arable land quality is visualized to enable the data pedigree of arable land quality monitoring to be visually traced and inferred.

[0175] In some embodiments, visualizing the pedigree knowledge graph of arable land quality is primarily aimed at addressing the issues of dynamic updates and usability of complex relationships within the graph. Therefore, an interactive reasoning mechanism integrating "dynamic data updates and progressive visualization exploration" is proposed.

[0176] In some embodiments, the above method further includes:

[0177] Respond to user commands such as load, insert, build, and / or query to display the corresponding data.

[0178] like Figure 5 As shown, this mechanism changes the current static report management model, realizing dynamic two-way reasoning in the farmland quality evaluation process, while also possessing good usability. In other words, through a graphical interactive interface, it enables the supplementation of farmland quality evaluation lineage and dynamic visual traceability, allowing users to intuitively explore the data source and calculation logic from any node along the entire chain.

[0179] Through the graphical interface, users can select the Load command, Insert command, Build command, and / or Query command. The Load command loads data; the Insert command inserts data and allows users to choose between "Insert Sampling Points" or "Insert Attributes"; the Build command constructs a hierarchical relationship graph; and the Query command provides a dynamic visual query of kinship relationships.

[0180] In some embodiments, a source-driven update strategy, namely a source data-driven dynamic update mechanism, is designed to address the time-sensitive nature of farmland monitoring data.

[0181] Specifically, a unified insertion format is provided. After reading the existing source data file through the graphical interface, users can select insertion commands such as "Insert Sampling Points" or "Insert Other Attributes". In the scenario of inserting sampling points, a form interface containing all original fields can be automatically generated. Categorical fields are filled in by drop-down selection, and numeric fields support text input. After the user completes the input, a new row is appended to the in-memory dataset, and a new standardized data file is immediately generated, retaining only the original standardized data fields and excluding the score column.

[0182] In some embodiments, to address the problem that full-scale graph display is too complex and makes it difficult to identify critical paths, lineage visualization and bidirectional reasoning are designed, and a progressive exploration method of "on-demand loading and drilling down layer by layer" is proposed, supporting two reasoning modes: forward propagation and reverse diagnosis.

[0183] Specifically, lineage visualization includes: after the user enters any node name in the interface, such as a certain FID, a certain original attribute, a certain score node, or a certain patch number, the system first retrieves all the incoming and outgoing edge information of the node in the knowledge graph database, and displays each relationship item by item as a visual button or card component. Each button represents a lineage path that can be further explored, such as "originating from a certain original attribute value", "mapped to a certain indicator score", "converging to the Score node", "associated with the patch number", etc.

[0184] The positive influence propagation reasoning process includes: when a user clicks on a node, the next level of adjacent nodes is read immediately, and the process continues to expand in the same way, forming a dynamically extendable lineage link. This node-by-node expansion interaction mode avoids the visual congestion caused by loading the entire complex graph structure at once, allowing users to understand the dependencies between data layer by layer from point to chain and from chain to network. In the actual tracing process, users can explore the lineage link from any direction; for example, starting from an original attribute value node, one can expand downwards to view its score mapping formula and the mapped score node; continuing to expand from the score node, one can view its contribution weight and contribution value to the comprehensive score; expanding outwards from the Score node, one can see the spatial unit of the graph corresponding to the comprehensive evaluation result, realizing a complete tracing from "sampling point → attribute → index score → quality evaluation → spatial entity of graph".

[0185] The reverse anomaly diagnosis reasoning process includes: after discovering a potentially abnormal patch, the user can start from the patch node and trace back along the lineage path, sequentially displaying "patch evaluation → comprehensive score → key deduction indicators → original attribute anomalies"; through this process, reverse diagnosis from the result anomaly to the source cause is realized, helping users quickly locate the specific indicators and original data sources that lead to the decline in arable land quality, such as locating that the low score of a certain patch is due to the pH value exceeding the standard at sampling point 001, thereby providing accurate decision support.

[0186] According to the embodiments of this disclosure, the following technical effects are achieved:

[0187] By acquiring multi-source heterogeneous cultivated land data from the target area and organizing and standardizing it, standard cultivated land quality monitoring index data is generated. Based on this data, a cultivated land quality evaluation system is constructed, generating evaluation result data and lineage metadata. Then, based on the evaluation result data and lineage metadata, a cultivated land quality lineage knowledge graph is constructed. This knowledge graph is then visualized to enable visual traceability and reasoning of cultivated land quality monitoring data lineage. Based on this, a unified acquisition and preprocessing mechanism for multi-source heterogeneous spatiotemporal data can be established, laying the foundation for building a data governance model based on lineage relationships. During index calculation and evaluation, source data entities, computational logic, evaluation results, and their dependencies are instantiated as nodes and edges in the graph, thus constructing a fully traceable lineage network. Based on this lineage graph, visual queries and bidirectional reasoning are then implemented, achieving full-link transparency and computable dynamic impact of the evaluation process.

[0188] In summary, this disclosure constructs a farmland quality evaluation system that integrates data standardization, indicator mapping, comprehensive evaluation, and pedigree tracing for multi-source heterogeneous spatiotemporal data, including ground sampling, remote sensing imagery, UAV imagery, IoT / sensor data, and historical survey time-series data. It addresses core issues in current technologies such as the difficulty in unified management of multi-source data, opaque evaluation processes, and lack of result traceability. This significantly improves the scientific rigor, reliability, and interpretability of farmland quality evaluation, making it suitable for agricultural farmland quality grading, data auditing, and decision support scenarios. Furthermore, based on farmland quality data, the system achieves end-to-end transparency and computable dynamic impact through the design of a spatiotemporal primary key index, rule validation library, and graph schema architecture. Meanwhile, a kinship map was constructed using farmland quality data from sampling points. By establishing a multi-level data structure covering sampling points, original attributes, individual scores, comprehensive scores, and map entity data, this disclosure achieves a full-chain expression from raw data to the final evaluation result. This allows farmland quality evaluation to no longer rely on black-box calculations, but clearly presents the contribution of each indicator and each calculation step to the final result.

[0189] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this disclosure is not limited to the described order of actions, because according to this disclosure, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to this disclosure.

[0190] The above is an introduction to the method embodiments. The following describes the solution described in this disclosure further through device embodiments.

[0191] Figure 6 A block diagram of a data lineage construction and inference system 600 for farmland quality monitoring according to an embodiment of the present disclosure is shown. Figure 6 As shown, the system 600 includes:

[0192] The acquisition module 610 is used to acquire multi-source heterogeneous cultivated land data of the target area. The multi-source heterogeneous cultivated land data includes at least two of the following: field sampling record data, laboratory test data, satellite remote sensing image data, UAV data, Internet of Things / sensor data, historical survey data, and statistical data of the target area.

[0193] Standardization module 620 is used to organize and standardize multi-source heterogeneous cultivated land data to generate standard cultivated land quality monitoring index data.

[0194] The generation module 630 is used to generate evaluation result data and lineage metadata based on standard cultivated land quality monitoring index data;

[0195] Module 640 is used to construct a knowledge graph of pedigree for arable land quality based on evaluation results data and pedigree metadata.

[0196] The visualization module 650 is used to visualize the pedigree knowledge graph of arable land quality, so as to enable the data pedigree of arable land quality monitoring to be visualized, traced, and inferred.

[0197] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the described module can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0198] The acquisition, storage, and application of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0199] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0200] Figure 7A block diagram of an exemplary electronic device 700 capable of implementing embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0201] Electronic device 700 includes a computing unit 701, which can perform various appropriate actions and processes according to a computer program stored in ROM 702 or a computer program loaded into RAM 703 from storage unit 708. RAM 703 can also store various programs and data required for the operation of electronic device 700. The computing unit 701, ROM 702, and RAM 703 are interconnected via bus 704. I / O interface 705 is also connected to bus 704.

[0202] Multiple components in electronic device 700 are connected to I / O interface 705, including: input unit 706, such as keyboard, mouse, etc.; output unit 707, such as various types of displays, speakers, etc.; storage unit 708, such as disk, optical disk, etc.; and communication unit 707, such as network card, modem, wireless transceiver, etc. Communication unit 707 allows electronic device 700 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0203] The computing unit 701 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 701 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 701 performs the various methods and processes described above, such as method 100. For example, in some embodiments, method 100 may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 708.

[0204] In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 700 via ROM 702 and / or communication unit 707. When the computer program is loaded into RAM 703 and executed by computing unit 701, one or more steps of method 100 described above may be performed. Alternatively, in other embodiments, computing unit 701 may be configured to perform method 100 by any other suitable means (e.g., by means of firmware).

[0205] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0206] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0207] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0208] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including voice input, speech input, or tactile input).

[0209] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0210] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0211] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0212] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for constructing and inferring data lineage for farmland quality monitoring, characterized in that, include: Acquire multi-source heterogeneous farmland data for the target area, wherein the multi-source heterogeneous farmland data includes at least two of the following: field sampling record data, laboratory test data, satellite remote sensing image data, UAV data, Internet of Things / sensor data, historical survey data, and statistical data of the target area; The multi-source heterogeneous cultivated land data is organized and standardized to generate standard cultivated land quality monitoring index data; Based on the standard cultivated land quality monitoring index data, evaluation result data and lineage metadata are generated; The step of generating evaluation result data and lineage metadata based on the standard cultivated land quality monitoring index data includes: dividing the standard cultivated land quality monitoring index data according to preset index classification rules, and then using corresponding preset mapping algorithms to map each sampled original index into a single index score, and storing the unique identifier of the sampling point corresponding to the original index, the preset mapping algorithm, and the unique identifier of the sampling point corresponding to the single index score as lineage metadata; wherein, the preset mapping algorithm includes a calculation formula, model parameters, and mapping rules; calculating the comprehensive score of each sampling point based on the single index score using a preset multi-index weighted aggregation algorithm; and mapping the single index score and comprehensive score of each sampling point to the corresponding map patch through spatial mapping to obtain the evaluation result data; Based on the evaluation results data and kinship metadata, a kinship knowledge graph of arable land quality is constructed. This construction includes: using the unique identifier of each sampling point as the root node, connecting the corresponding original indicators and multi-source heterogeneous arable land data to form a first-layer kinship relationship; establishing a mapping between the calculation formula of each sample and the single-indicator score as attribute nodes to form a second-layer kinship relationship; associating the comprehensive score of each sample with the root node to form a third-layer kinship relationship; associating the comprehensive score of each sample with the patch nodes to form a fourth-layer kinship relationship; and constructing the arable land quality kinship knowledge graph based on preset graph instantiation mapping rules, according to the first-layer kinship relationship, the second-layer kinship relationship, the third-layer kinship relationship, and the fourth-layer kinship relationship. The pedigree knowledge graph of cultivated land quality is visualized to enable data pedigree tracking and reasoning in cultivated land quality monitoring.

2. The method according to claim 1, characterized in that, The process of organizing and standardizing the multi-source heterogeneous cultivated land data to generate standard cultivated land quality monitoring index data includes: Based on preset classification rules and preset unified identification rules, a primary key index system with the unique identifier of the sampling point as the core is constructed according to the multi-source heterogeneous cultivated land data, and the spatiotemporal unit and standardized cultivated land quality index system are determined to generate initial standard cultivated land quality monitoring index data. Based on preset individual data anomaly handling rules and preset data relationship anomaly handling rules, the initial standard cultivated land quality monitoring index data is verified to generate standard cultivated land quality monitoring index data.

3. The method according to claim 2, characterized in that, The system for determining spatiotemporal units and standardized arable land quality indicators includes: Using the unique identifier of the sampling point as the micro-spatial unit and the map patch number as the macro-spatial unit, the geometric attributes of the map patch and the information of the administrative region to which it belongs are associated to form a spatial description benchmark that combines points and surfaces, and use it as a spatiotemporal unit; Based on a pre-set set of indicators, and using the unique identifier of the sampling point as the primary key, a standardized farmland quality indicator system is determined.

4. The method according to claim 3, characterized in that, The method further includes: Based on the preset naming rules for derived indicators, derived data are generated according to the initial standard farmland quality monitoring indicator data and stored independently.

5. The method according to any one of claims 1 to 4 is characterized in that, The method further includes: Respond to user commands such as load, insert, build, and / or query to display the corresponding data.

6. A data lineage construction and inference system for farmland quality monitoring, characterized in that, include: The acquisition module is used to acquire multi-source heterogeneous cultivated land data of the target area. The multi-source heterogeneous cultivated land data includes at least two of the following: field sampling record data, laboratory test data, satellite remote sensing image data, UAV data, Internet of Things / sensor data, historical survey data, and statistical data of the target area. The standardization module is used to organize and standardize the multi-source heterogeneous cultivated land data to generate standard cultivated land quality monitoring index data. The generation module is further configured to generate evaluation result data and lineage metadata based on the standard cultivated land quality monitoring index data. Specifically, the generation module is configured to divide the standard cultivated land quality monitoring index data according to preset index classification rules, and then use corresponding preset mapping algorithms to map each sampled original index into a single index score. The module stores the unique identifier of the original index corresponding to the sampling point, the preset mapping algorithm, and the unique identifier of the single index score corresponding to the sampling point as lineage metadata. The preset mapping algorithm includes a calculation formula, model parameters, and mapping rules. A preset multi-index weighted aggregation algorithm is used to calculate the comprehensive score of each sampling point based on the single index score. Through spatial mapping, the single index score and comprehensive score of each sampling point are mapped to their respective map patches to obtain the evaluation result data. A construction module is used to construct a knowledge graph of farmland quality lineage based on the evaluation result data and lineage metadata. Specifically, the construction module is used to connect the corresponding original indicators and multi-source heterogeneous farmland data with the unique identifier of each sampling point as the root node to form a first-layer lineage relationship; to establish a mapping between the calculation formula of each sample and the single indicator score as attribute nodes to form a second-layer lineage relationship; to associate the root node with the comprehensive score of each sample to form a third-layer lineage relationship; to associate the patch nodes with the comprehensive score of each sample to form a fourth-layer lineage relationship; and to construct the knowledge graph of farmland quality lineage based on the preset graph instantiation mapping rules, according to the first-layer lineage relationship, the second-layer lineage relationship, the third-layer lineage relationship, and the fourth-layer lineage relationship. The visualization module is used to visualize the pedigree knowledge graph of cultivated land quality, so as to enable the data pedigree of cultivated land quality monitoring to be visualized, traced, and inferred.

7. An electronic device, characterized in that, include: At least one processor; as well as A memory that is communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method described in any one of claims 1-5.

8. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-5.

Images