A crop disease and insect pest diagnosis system based on knowledge graph

Abstract

The present invention discloses a crop disease and insect pest diagnosis system based on a knowledge graph, which relates to the technical field of disease and insect pest diagnosis. The data acquisition module obtains description information, image data and environmental data of crops; the data processing module extracts symptom keywords, disease and insect pest expansion rate fluctuation characteristics and soil EC value stress characteristics; the knowledge graph construction module establishes an association network including disease and insect pest symptom nodes, environmental condition nodes and dynamic feature nodes; the disease and insect pest differentiation module combines dynamic features and environmental features to evaluate the differentiation accuracy of similar disease and insect pest symptoms; the diagnosis result classification module divides the diagnosis results into accuracy categories and takes corresponding treatment measures; the optimization management module dynamically optimizes the knowledge graph and classification model based on fixed time period analysis of incomplete accuracy diagnosis results, significantly improving the accuracy of disease and insect pest diagnosis, avoiding prevention and control errors caused by misdiagnosis, and providing intelligent technical support for agricultural production.

Description

Technical Field

[0001] The present invention relates to the technical field of disease and insect pest diagnosis, and in particular to a crop disease and insect pest diagnosis system based on a knowledge graph. Background Art

[0002] Knowledge graph-based crop disease and pest diagnosis utilizes artificial intelligence (AI) and knowledge engineering to systematize and structure knowledge in the field of crop diseases and pests, constructing a knowledge network that can be understood and reasoned by machines. This network then uses intelligent algorithms to perform diagnosis and prediction. A knowledge graph is a network graph structure that represents knowledge as nodes (entities) and edges (relationships). In crop disease and pest diagnosis, the knowledge graph construction process involves extracting relevant knowledge from agricultural literature, expert knowledge, and real-world data to form a complete semantic network. For example, "rice," as a node, is connected to "rice blast" through the "susceptible to pests and diseases" relationship; "rice blast" is in turn connected to "gray-brown spots on leaves" through the "typical symptoms" relationship. This structured knowledge graph systematically integrates information on pest and disease classification, symptoms, pathogenesis, triggers, environmental factors, and control measures, providing rich background knowledge for diagnosis.

[0003] The knowledge graph-based diagnostic process primarily relies on reasoning algorithms and intelligent analysis techniques. When farmers upload images of crops or describe symptoms of pests and diseases, the system matches the input information with the knowledge in the knowledge graph and, through logical reasoning or pattern recognition techniques, identifies the most likely pests and diseases and their causes. For example, when the system detects an input description of "irregular grayish-white spots appearing on rice leaves," it combines meteorological data with the patterns of pest and disease occurrence and, through inference from the associations in the knowledge graph, concludes that the diagnosis may be "rice blast" or "bacterial leaf spot." The system can then provide targeted prevention and control recommendations, such as recommending the use of appropriate pesticides or adjusting farming practices.

[0004] The existing technology has the following shortcomings:

[0005] Knowledge graph diagnosis primarily relies on matching symptom descriptions with known pest and disease knowledge. However, in real agricultural scenarios, different pests and diseases can exhibit similar symptoms. For example, irregular grayish-white spots on rice leaves could be rice blast, bacterial leaf spot, or even a non-pest and disease phenomenon caused by environmental stress (such as salinity, pesticide damage, or nutrient deficiency). When the system lacks more detailed graph nodes (such as specific pest and disease identification indicators) or actual image data, misdiagnosis can occur, leading farmers to incorrect prevention and control recommendations, such as using the wrong pesticides or measures, further exacerbating crop damage. Summary of the Invention

[0006] The purpose of the present invention is to provide a crop disease and pest diagnosis system based on knowledge graph to address the shortcomings of the background technology.

[0007] To achieve the above objectives, the present invention provides the following technical solutions: a knowledge graph-based crop disease and insect pest diagnosis system, comprising a data acquisition module, a data processing module, a knowledge graph construction module, a disease and insect pest differentiation module, a diagnosis result classification module, and an optimization management module;

[0008] A data acquisition module is used to obtain crop pest and disease description information, pest and disease image data, and environmental data, wherein the environmental data includes meteorological data and soil data;

[0009] The data processing module is used to perform natural language processing on pest and disease description information to extract symptom keywords, process pest and disease image data using computer vision technology to extract the fluctuation characteristics of pest and disease expansion rate, and process environmental data to obtain soil EC value stress characteristics;

[0010] A knowledge graph construction module is used to build a knowledge graph based on pest and disease knowledge, wherein the knowledge graph includes pest and disease symptom nodes, environmental condition nodes, dynamic feature nodes, and the association relationship between each node;

[0011] The pest and disease differentiation module is used to match the extracted symptom keywords with the symptom information stored in the knowledge graph, differentiate pests and diseases based on the matching results, and evaluate the accuracy of distinguishing similar symptoms of pests and diseases based on the fluctuation characteristics of pest and disease expansion rate and soil EC value stress characteristics;

[0012] The diagnosis result classification module classifies the accuracy of the disease and insect pest diagnosis results according to the evaluation results, and divides the disease and insect pest diagnosis results into accurate diagnosis results, incompletely accurate diagnosis results and inaccurate diagnosis results, and performs corresponding processing;

[0013] The optimization management module further analyzes the accuracy of distinguishing similar symptoms of pests and diseases within a fixed time period for incompletely accurate diagnosis results, and dynamically optimizes the knowledge graph and classification model based on the analysis results to improve the accuracy of crop disease and pest diagnosis.

[0014] Preferably, in the data processing module, the extracted pest and disease expansion rate fluctuation characteristics are analyzed to generate a pest and disease expansion rate fluctuation index. The pest and disease expansion rate fluctuation index is obtained by:

[0015] Obtain disease expansion rate data for different time periods and construct two time series: is the reference expansion rate series, representing the standard disease expansion rate pattern; The disease expansion rate sequence currently collected, each rate point represents the rate of change of the disease area in a certain period of time, for any two points , calculate the local distance: Where, for point and The absolute distance;

[0016] Construct the cost matrix D, which is an m×n matrix, where each element D(i,j) represents the minimum cumulative distance of the path from the starting point (1,1) to the point (i,j). The initial cost is: , ; , ;Fill the matrix sequentially through dynamic programming and cumulatively calculate the minimum matching cost between each point;

[0017] Defining the optimal path is a set of point pair sequences that satisfy the following conditions: Indicates the starting point; Indicates the end point, the path is continuous, no jumping is allowed, and the total cost of the path for: ; Calculate the pest expansion rate fluctuation index, the expression is: Where, is the path length, and WDC is the pest and disease expansion rate fluctuation index.

[0018] Preferably, in the data processing module, the extracted soil EC stress characteristics are analyzed to generate a soil EC stress anomaly index. The soil EC stress anomaly index is obtained by:

[0019] Collecting a sampling dataset of soil EC values: Input data , each Represents the soil EC value of a sampling point. The soil EC value is standardized and the expression is: ; and are the data mean and standard deviation respectively, and output the standardized data set ; The probability density function is: Where, is the weight of the kth Gaussian distribution, satisfying ; is the probability density function of the kth Gaussian distribution, which is in the form of: ; is the mean of the kth Gaussian distribution, indicating the center of a certain state, is the variance of the k-th Gaussian distribution, indicating the distribution range of the state;

[0020] Random initialization The initial value of , calculates the probability that each data point belongs to the k-th Gaussian distribution: Where, is the responsibility weight of the i-th data point belonging to the k-th distribution, according to Update GMM parameters, the expression is: ; Calculate data points The total probability of , the expression is: ; The soil EC stress abnormality index is defined as point The probability of belonging to a non-normal distribution is expressed as: ; YGF is the soil EC stress abnormality index, for point The probability of belonging to the normal distribution is expressed as: Where, are the weight, mean, and standard deviation of the Gaussian distribution under normal conditions set based on historical data.

[0021] Preferably, in the pest and disease differentiation module, the pest and disease expansion rate fluctuation characteristics and soil EC value stress characteristics are combined to evaluate the accuracy of distinguishing similar symptoms of pests and diseases, specifically:

[0022] The pest and disease expansion rate fluctuation index and the soil EC value stress anomaly index are converted into comprehensive feature vectors, which are used as inputs of the machine learning model. The machine learning model uses the discrimination accuracy value labels of similar pest and disease symptoms predicted by each set of comprehensive feature vectors as the prediction target, and minimizes the sum of the prediction errors of the discrimination accuracy value labels of similar pest and disease symptoms as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence, and the model training is stopped. The discrimination accuracy value of similar pest and disease symptoms is determined according to the model output results. The machine learning model is a polynomial regression model.

[0023] Preferably, in the diagnosis result classification module, the accuracy of the disease and insect pest diagnosis results is classified into accurate diagnosis results, incompletely accurate diagnosis results and inaccurate diagnosis results, specifically:

[0024] Comparing the obtained discrimination accuracy value of similar symptoms of pests and diseases with a gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and comparing the discrimination accuracy value of similar symptoms of pests and diseases with the first standard threshold and the second standard threshold respectively;

[0025] If the accuracy value of distinguishing similar symptoms of pests and diseases is greater than the second standard threshold, it means that the accuracy of distinguishing similar symptoms of pests and diseases is high, and a high-accuracy distinction signal is generated, and the diagnosis result is classified as an accurate diagnosis result;

[0026] If the accuracy value of distinguishing similar symptoms of pests and diseases is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the accuracy of distinguishing similar symptoms of pests and diseases is medium, and a medium accuracy distinction signal is generated, and the diagnosis result is classified as an incomplete accuracy diagnosis result;

[0027] If the accuracy value of distinguishing similar symptoms of pests and diseases is less than the first standard threshold, it means that the accuracy of distinguishing similar symptoms of pests and diseases is low. At this time, a low accuracy distinction signal is generated, and the diagnosis result is classified as an inaccurate diagnosis result.

[0028] Preferably, in the optimization management module, for incompletely accurate diagnosis results, further analysis is performed on the accuracy of distinguishing similar symptoms of pests and diseases within a fixed time period, and the knowledge graph and classification model are dynamically optimized based on the analysis results, specifically:

[0029] In a fixed time period T, key data of incomplete accuracy diagnosis results are collected to form a data set D: Where, is the accuracy value for distinguishing similar symptoms of pests and diseases diagnosed for the i-th time, is the fluctuation index of the expansion rate of the i-th diagnosed pest, is the soil EC stress abnormality index of the i-th diagnosis, For actual disease labels, define a comprehensive evaluation index for discrimination accuracy within a fixed time period , the expression is: ; is the diagnostic accuracy function, according to and the actual label Determine whether the diagnosis is correct: ; N is the number of incomplete accuracy diagnosis samples in time period T.

[0030] Preferably, define the nodes in the knowledge graph and diseased nodes Weight , based on the diagnostic data within a fixed time period T, the weight is updated. The expression is: Where, , is the node in the i-th sample Diseased nodes The contribution prediction value of , η is the learning rate, which is used to control the amplitude of weight update;

[0031] Based on comprehensive feature vector (WDC, YGF) and actual label , retrain the classification model to minimize the discrimination accuracy prediction error: Where, is the objective function, The classification accuracy value predicted by the classification model is used to update the model parameters θ using the gradient descent method: ; is the learning rate, The gradient of the objective function with respect to the model parameter θ defines the overall diagnostic accuracy index after dynamic optimization , the expression is: , N is the total number of samples in the data set.

[0032] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0033] 1. The present invention overcomes the shortcomings of existing technologies in distinguishing diseases with similar symptoms by constructing a crop disease and pest diagnosis system based on a knowledge graph. By introducing the disease and pest expansion rate fluctuation index and the soil EC value stress anomaly index, combined with natural language processing and computer vision technology, the system realizes the integration and dynamic analysis of multi-dimensional data. By optimizing the prediction of the distinction accuracy value through a machine learning model, the diagnostic accuracy of diseases with similar symptoms is further improved. In addition, the system dynamically adjusts the node weights in the knowledge graph and continuously iterates and updates through classification model optimization to ensure efficient and accurate disease differentiation capabilities in complex agricultural scenarios, thereby reducing the risk of misdiagnosis and providing farmers with a scientific and reliable diagnostic basis.

[0034] 2. Through modular design, the present invention enables the system to grade the accuracy of diagnostic results and provide further optimization measures for incomplete and inaccurate results, making the diagnostic results more reliable. The dynamic optimization management module uses diagnostic data within a fixed time period to continuously update the knowledge graph and classification model, significantly improving the system's adaptability to atypical symptoms such as emerging diseases and environmental stress. This comprehensive and efficient pest and disease diagnosis system not only helps farmers accurately identify diseases, but also optimizes prevention and control measures, avoids the incorrect use of pesticides, reduces crop losses and environmental pollution, and provides strong technical support for intelligent agricultural management. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0036] Figure 1 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0037] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0038] For examples, see Figure 1 As shown, the crop disease and insect pest diagnosis system based on knowledge graph described in this embodiment includes a data acquisition module, a data processing module, a knowledge graph construction module, a disease and insect pest differentiation module, a diagnosis result classification module and an optimization management module;

[0039] A data acquisition module is used to obtain crop pest and disease description information, pest and disease image data, and environmental data, wherein the environmental data includes meteorological data and soil data;

[0040] The data processing module is used to perform natural language processing on pest and disease description information to extract symptom keywords, process pest and disease image data using computer vision technology to extract the fluctuation characteristics of pest and disease expansion rate, and process environmental data to obtain soil EC value stress characteristics;

[0041] A knowledge graph construction module is used to build a knowledge graph based on pest and disease knowledge, wherein the knowledge graph includes pest and disease symptom nodes, environmental condition nodes, dynamic feature nodes, and the association relationship between each node;

[0042] The pest and disease differentiation module is used to match the extracted symptom keywords with the symptom information stored in the knowledge graph, differentiate pests and diseases based on the matching results, and evaluate the accuracy of distinguishing similar symptoms of pests and diseases based on the fluctuation characteristics of pest and disease expansion rate and soil EC value stress characteristics;

[0043] The diagnosis result classification module classifies the accuracy of the disease and insect pest diagnosis results according to the evaluation results, and divides the disease and insect pest diagnosis results into accurate diagnosis results, incompletely accurate diagnosis results and inaccurate diagnosis results, and performs corresponding processing;

[0044] The optimization management module further analyzes the accuracy of distinguishing similar symptoms of pests and diseases within a fixed time period for incompletely accurate diagnosis results, and dynamically optimizes the knowledge graph and classification model based on the analysis results to improve the accuracy of crop disease and pest diagnosis.

[0045] In the data collection module, users enter text descriptions, such as crop symptom information uploaded by farmers or technicians via mobile devices or web platforms. This collection includes disease-related symptom keywords and detailed descriptions, such as: "Irregular grayish-white spots on rice leaves," "Pinhead-sized black spots appear on leaves and gradually expand," and "Leaves turn yellow, and plant growth is poor." Users can enter the crop type, growth stage (e.g., tillering, heading), and the specific time and location of the disease. Using natural language processing technology, key information from the description (e.g., "grayish-white spots," "yellowing leaves") is extracted and stored in a structured format for subsequent analysis.

[0046] Users upload images of crop diseases, such as photos taken by smartphones or drones. The collected content includes: images of disease symptoms on leaves, stems or fruits, such as the shape, color, and edge characteristics of spots. The distribution pattern of the disease (such as concentrated at the tip of the leaf or randomly distributed). Changes in the overall texture of the leaves (such as dryness, curling, and deformity). Field environment images: Collect an overall view of the field to understand the growth of crops and the characteristics of field distribution. Collection method: Static image collection: High-definition photos taken and uploaded by users. Dynamic image collection: Continuous video or multi-angle images taken by drones or agricultural robots. Image quality verification: Use computer vision technology to check image clarity, brightness, etc. and filter out low-quality images.

[0047] Environmental conditions play an important role in the occurrence and development of pests and diseases. Collecting environmental data can significantly improve the accuracy of the diagnostic system. Environmental data is divided into the following two parts:

[0048] Meteorological data collection includes: Online meteorological data interface: obtaining real-time meteorological data from national or regional meteorological service platforms. Field sensors: meteorological station equipment installed in farmland, including temperature and humidity sensors, rainfall gauges, etc. Collection content: Temperature: a key inducement for the occurrence of pests and diseases (e.g., high temperature leads to accelerated reproduction of rice planthoppers). Humidity: high humidity is conducive to the spread of fungal diseases (e.g., rice blast). Rainfall: monitoring the impact of rain on soil moisture and disease spread. Wind speed and direction: affecting the spread direction of flying insect diseases (e.g., rice leaf roller). Formatting and standardizing the meteorological data collected in real time. Correlating meteorological data with the environmental inducing conditions recorded in the pest and disease knowledge graph as an important basis for reasoning.

[0049] Soil data collection: Field soil sensors: Directly collect field soil data through IoT devices. Manual measurement and experimental analysis: Manual sampling and laboratory analysis (such as soil composition testing). Collection content: Soil EC value (electrical conductivity): Reflects the degree of saline-alkali stress and helps distinguish between disease and non-disease causes. Soil moisture: Affects plant root health and the occurrence of diseases (such as root rot). Soil pH value: Detects whether there is an overly acidic or alkaline environment, which may cause nutrient deficiency or saline-alkali stress. Soil nutrient composition: Such as nitrogen, phosphorus, and potassium content, to determine whether similar disease symptoms are caused by fertilizer deficiency. Standardization algorithms are used to unify the dimensions of soil data (such as moisture content is unified into percentage form). Correlate with environmental factors related to crop diseases to identify potential stress characteristics.

[0050] In this application, the data acquisition module forms the basic input for system diagnosis by acquiring pest and disease description information, disease image data, and meteorological and soil environmental data. Pest description information provides semantic features of text descriptions, image data provides visual features, and environmental data provides key external conditions for disease reasoning.

[0051] The data processing module is used to perform natural language processing on pest and disease description information to extract symptom keywords, process pest and disease image data through computer vision technology and extract the fluctuation characteristics of pest and disease expansion rate, and process environmental data to obtain soil EC value stress characteristics.

[0052] Using natural language processing (NLP) technology, the text description information entered by the user is parsed, symptom keywords are extracted, and the unstructured text is converted into structured data that can be used for reasoning. The processing steps include: Text preprocessing: Removing irrelevant characters such as stop words and punctuation, and standardizing the text format. Example input: "Irregular grayish-white spots appear on rice leaves." Keyword extraction: Using lexical analysis and word segmentation techniques, symptom-related keywords are extracted. Example output: {Crop type: rice, Disease site: leaf, Symptom characteristics: irregular grayish-white spots}. Symptom association matching: Using semantic matching technology, the extracted keywords are matched with the standardized symptom vocabulary in the knowledge graph to identify potential disease nodes. For example, the keyword "grayish-white spots" matches the symptom nodes of "rice blast" and "bacterial leaf spot." The output: Structured symptom keyword data, which provides basic input for disease reasoning in the knowledge graph.

[0053] Computer vision technology is used to process user-uploaded disease images to extract visual features of pests and diseases, including dynamic information such as spread rate fluctuations. Processing steps include: Image preprocessing: Denoising, brightness adjustment, and cropping the input image to ensure image quality suitable for analysis. Disease feature extraction: A convolutional neural network (CNN) model is used to extract key visual features from the image: spot color: off-white, brown, or black; spot shape: regularity or irregularity; spot edges: blurred or sharp. Example output: {color: off-white, shape: irregular, edge: blurred}. Fluctuation characteristics of disease spread rate: When a user uploads multiple disease images of the same plant, the system compares images from different time periods and calculates the disease spread rate. Image segmentation techniques (such as Mask R-CNN) are used to identify lesion areas and calculate changes in lesion area: Spread rate = (current lesion area - previous lesion area) / time interval. Example output: Spread rate = 5 cm² / day. Output results: The extracted disease visual feature data and expansion rate fluctuation characteristics provide a temporal dynamic basis for distinguishing highly similar symptoms.

[0054] Collected environmental data (such as soil data) is standardized and feature extracted, with a focus on identifying soil EC stress characteristics as a key distinguishing indicator between disease and non-disease conditions (such as salinity-alkalinity stress). Processing steps include: Data normalization: De-noises and normalizes raw soil data collected by sensors (such as electrical conductivity EC values). Example: A raw EC value of 2.5 mS / cm is normalized to convert it into a comparable indicator. Stress feature extraction: A threshold for the normal range of soil EC values ​​is set (e.g., 0.5 mS / cm to 2.0 mS / cm). When EC values ​​exceed this range, they are flagged as salinity-alkalinity stress characteristics. Example output: Soil EC value = 2.8 mS / cm → Flagged as "Salt-alkalinity Stress." Soil moisture = 30% → Assists in determining whether symptoms are caused by water stress. Environmental data and visual data correlation: Soil EC stress characteristics are fused with disease image features to determine whether symptoms are caused by salinity-alkalinity stress rather than disease infection. Output: Standardized soil EC value data and stress characteristics, providing an environmental basis for systematically eliminating non-disease factors.

[0055] After analyzing the extracted pest and disease expansion rate fluctuation characteristics, the pest and disease expansion rate fluctuation index is generated. The pest and disease expansion rate fluctuation index is obtained as follows:

[0056] Obtain disease expansion rate data for different time periods and construct two time series: : Reference expansion rate series, representing the standard disease expansion rate pattern (such as the typical expansion trend recorded in the knowledge graph). : The currently collected disease expansion rate sequence (such as the rate fluctuation obtained from actual data). Each rate point represents the rate of change of the disease area in a certain time period. The lengths of the two sequences may be different, that is, m≠n, and the time points may not be aligned. For any two points , calculate the local distance: Where, for point and absolute distance.

[0057] Construct the cost matrix D, which is an m×n matrix, where each element D(i,j) represents the minimum cumulative distance of the path from the starting point (1,1) to the point (i,j). The initial cost is: , ; , ; Fill the matrix in sequence through dynamic programming, and cumulatively calculate the minimum matching cost between each point.

[0058] Defining the optimal path is a set of point pair sequences that satisfy the following conditions: Indicates the starting point; Indicates the end point, the path is continuous, no jumping is allowed, and the total cost of the path for: ; Calculate the pest expansion rate fluctuation index, the expression is: Where, is the path length (number of matching points), and WDC is the pest expansion rate fluctuation index.

[0059] The larger the pest and disease expansion rate fluctuation index, the more significant the difference between the current disease expansion rate and the standard expansion rate pattern, that is, the dynamic characteristics of the time series match less well. This larger fluctuation index may indicate:

[0060] The current disease may have significantly different expansion trends from the reference standard diseases (such as rice blast and bacterial leaf spot), and the system has a high accuracy in distinguishing highly similar symptoms.

[0061] However, if the system's knowledge graph is insufficiently covered, such large discrepancies could also be due to expansion anomalies caused by emerging diseases or non-disease factors (such as environmental stress). Therefore, cases with large fluctuations in the index require comprehensive analysis in conjunction with other data (such as soil EC values ​​and image features) to ensure the reliability of the diagnostic results.

[0062] When the pest and disease expansion rate fluctuation index is smaller, it indicates that the current disease expansion rate is highly similar to the dynamic characteristics of the standard expansion model, and the time series matching degree is high. This usually means:

[0063] The current disease is likely to belong to the same category as the standard disease, and the system may have lower accuracy in distinguishing similar symptoms because the similarity in expansion rate may mask the differences between diseases.

[0064] In this case, it is necessary to further rely on other features (such as edge clarity and color characteristics of the symptom image) to improve the differentiation ability, otherwise it is easy to misdiagnose. For example, rice blast and bacterial leaf spot may show the same expansion rate, but they may be easier to distinguish based on lesion morphology or soil characteristics.

[0065] The soil EC stress anomaly index is generated by analyzing the extracted soil EC stress characteristics. The soil EC stress anomaly index is obtained as follows:

[0066] Collecting a sampling dataset of soil EC values: Input data , each Indicates the soil EC value of a sampling point. The data may contain different states (such as normal, mild stress, and severe stress).

[0067] Normalize or standardize the soil EC value, the expression is: ; and are the data mean and standard deviation respectively, and output the standardized data set ; GMM assumes that the data obeys a mixture model of K Gaussian distributions, and the probability density function is: Where, is the weight of the kth Gaussian distribution, satisfying ; is the probability density function of the kth Gaussian distribution, which is in the form of: ; is the mean of the kth Gaussian distribution, indicating the center of a certain state, is the variance of the kth Gaussian distribution, which represents the distribution range of the state.

[0068] Random initialization The initial value of , calculates the probability that each data point belongs to the k-th Gaussian distribution: Where, is the responsibility weight of the i-th data point belonging to the k-th distribution, according to Update GMM parameters, the expression is: ; Calculate data points The total probability of , the expression is: ; The soil EC stress abnormality index is defined as point The probability of belonging to a non-normal distribution is expressed as: ; YGF is the soil EC stress abnormality index, for point The probability of belonging to the normal distribution is expressed as: Where, are the weight, mean, and standard deviation of the Gaussian distribution under normal conditions set based on historical data.

[0069] A higher soil EC stress anomaly index indicates that the current soil salinity and alkali stress level deviates significantly from the normal range. This suggests that plant symptoms similar to those of pests and diseases (such as leaf mottling and yellowing) may be due to non-disease causes (such as salinity and alkali stress or environmental stress). In this case, the system can more accurately distinguish between similar symptoms of pests and diseases because a high anomaly index can identify non-disease causes and eliminate incorrect disease diagnoses, thereby avoiding the misuse of pesticides and other control measures. Therefore, a larger anomaly index helps improve the diagnostic system's ability to distinguish between disease and non-disease causes.

[0070] When the soil EC stress anomaly index decreases, soil salinity levels are closer to normal, and plant symptoms are likely caused by pests and diseases rather than soil stress. In this case, the system needs to rely on other characteristics (such as spread rate and lesion morphology) to distinguish between diseases with similar symptoms. The exclusion of soil stress may reduce the accuracy of pest and disease symptom differentiation, especially between highly similar diseases (such as rice blast and bacterial leaf spot). Therefore, when the anomaly index is low, it is necessary to analyze more disease characteristics to avoid misdiagnosis or ambiguity.

[0071] The knowledge graph construction module is used to establish a knowledge graph based on pest and disease knowledge. The knowledge graph includes pest and disease symptom nodes, environmental condition nodes, dynamic feature nodes, and the association relationships between the nodes.

[0072] The nodes in the knowledge graph are the basic units of knowledge related to pests and diseases. Different node types represent the multi-dimensional characteristics of the occurrence of pests and diseases and their related factors: Pest and disease symptom node definition: describes the typical plant symptoms caused by pests and diseases. Data source: agricultural literature, expert knowledge base, historical disease case data. Main attributes: Symptom description (text): such as "gray-white spots on leaves", "leaf edge withering". Typical characteristics (numerical values): such as lesion area, color, edge characteristics. Disease type: such as rice blast, bacterial leaf spot, rice planthopper, etc. Affected parts: such as leaves, stems, fruits, etc. Example node: Symptom node: "gray-white spots on leaves" Disease type: "Rice blast" Part: "Leaves" Characteristics: "Irregular spots, fast expansion rate"

[0073] Environmental condition node definition: describes external environmental factors related to the occurrence of pests and diseases, including meteorological conditions and soil conditions. Data source: meteorological data interface, sensor monitoring data, experimental data. Main attributes: Meteorological conditions: temperature, humidity, rainfall, wind speed, etc. Soil conditions: EC value, pH value, humidity, nutrient content, etc. Inducing conditions: The triggering effect of the environment on diseases, such as high humidity promoting fungal reproduction. Example nodes: Environmental condition node: "High humidity (>80%)" induces disease: "Rice blast" Environmental condition node: "Excessively high soil EC value (>2.5 mS / cm)" induces symptoms: "Yellowing of leaves due to salt-alkali stress"

[0074] Dynamic feature node definition: Describes the dynamic characteristics of pest and disease symptoms over time. Data source: Time series data, experimental observations of disease expansion. Main attributes: Expansion rate: such as the growth rate of lesion area (unit: cm² / day). Change pattern: such as whether the disease expansion shows linear growth, explosive growth, or a stable stage. Time node: such as "peak expansion period" and "symptom relief period". Example node: Dynamic feature node: "Expansion rate 5 cm² / day" Disease: "Rice blast" Stage: "Mid-term expansion stage"

[0075] The knowledge graph establishes connections through semantic relationships between nodes, reflecting the inherent connections and causal relationships between pest and disease-related factors: Symptom-Disease Relationship: The symptoms caused by a particular pest and disease and their manifestations. Example: Relationship Type: "Causing Symptoms" Rice Blast → Causing Symptoms → Grayish-White Spots; Rice Planthopper Infestation → Causing Symptoms → Leaf Yellowing. Disease-Environment Relationship: The occurrence or spread of a particular pest and disease is influenced by specific environmental conditions. Example: Relationship Type: "Affected by" Rice Blast → Affected by → High Humidity; Salt-Alkali Stress → Affected by → Excessively High Soil EC Value. Dynamic Feature-Symptom Relationship: The dynamic manifestation of a particular disease symptom over time. Example: Relationship Type: "Dynamic Manifestation is" Grayish-White Spots → Dynamic Manifestation is → Expansion Rate of 5 cm² / day.

[0076] Use natural language processing (NLP) technology to extract semantic information about diseases, symptoms, environmental characteristics, and dynamic features from agricultural literature and reports. Extract feature values ​​through image processing and environmental sensor data analysis. Use graph databases (such as Neo4j) or semantic technologies (such as RDF / OWL) to build graph structures. Define node types, attributes, and relationships. Maintain consistency across multiple data sources (literature, sensors, and image analysis results) to eliminate redundant or conflicting nodes. Add new nodes and relationships based on new disease data or user feedback.

[0077] The pest and disease differentiation module is used to match the extracted symptom keywords with the symptom information stored in the knowledge graph, differentiate pests and diseases based on the matching results, and evaluate the accuracy of distinguishing similar symptoms of pests and diseases based on the fluctuation characteristics of the pest and disease expansion rate and the soil EC value stress characteristics.

[0078] Symptom keywords are matched against symptom information in the knowledge graph, extracted from user-uploaded descriptions and image data. Symptom keywords are semantically matched against symptom nodes and their characteristic attributes in the knowledge graph. Matching rules can be based on keyword similarity, semantic similarity (such as Word2Vec or BERT embeddings), graph path weights, and more. Example: Input keywords: "grayish-white spots" and "leaf yellowing." Matching knowledge graph nodes: grayish-white spots → rice blast. Leaf yellowing → rice planthopper. Based on the matching degree, a list of candidate diseases or pests is output, and the matching weight for each disease is calculated. Dynamic features (such as expansion rate) and environmental conditions are combined to re-rank the confidence of candidate diseases and determine the final result.

[0079] Based on natural language processing (NLP) technology, structured keywords are extracted from the description entered by the user. For example, the input description is: "Irregular gray-white spots appear on rice leaves and gradually expand." Extract keywords: {"gray-white spots", "irregular", "gradually expand"}. Using computer vision technology, features (such as lesion color, shape, and distribution) are extracted from the disease image. For example: color = gray-white, edge = fuzzy, expansion rate = 5 cm² / day. In the symptom nodes of the knowledge graph, nodes that fully or partially match the extracted keywords are retrieved. Example: Symptom keyword: "gray-white spots" knowledge graph matching nodes: {"rice blast", "bacterial leaf spot"}. Use deep learning embedding models (such as BERT, Word2Vec) to calculate the semantic similarity between the input keywords and the symptom description in the knowledge graph: Q is the extracted symptom keyword, and S is the symptom description stored in the knowledge graph. Output a list of possible pest and disease candidates by sorting by similarity.

[0080] The matching process combines dynamic features and environmental conditions: Dynamic feature matching: Compare the dynamic curves of expansion rate or morphological change. Environmental condition matching: Check whether the disease-inducing conditions (such as humidity > 80%, high temperature, etc.) are met.

[0081] Taking into account the similarity of symptoms, similarity of dynamic characteristics, and degree of environmental matching, the matching weight of each pest and disease is calculated: ; Where α, β, and γ are weight factors that adjust the importance of symptoms, dynamic characteristics, and environmental conditions. is the comprehensive matching weight of candidate disease i. Sort the candidate diseases and select the disease with the highest weight as the diagnosis result. Output confidence , the expression is: Confidence is used to explain the reliability of the diagnosis results. If the highest confidence is lower than the set threshold, a list of candidate diseases and their confidence levels is output, prompting the user for further confirmation.

[0082] Candidate diseases are re-ranked based on dynamic features such as expansion rate and morphological changes to improve classification accuracy. For example, if the expansion rate is 5 cm² / day, "rice blast" is prioritized (its dynamic features are highly consistent). Environmental condition data (such as humidity and soil EC values) are used to exclude diseases that do not meet the inducing conditions. For example, if the soil EC value is >2.5 mS / cm, salinity stress is more likely to be diagnosed than disease infection. Based on user feedback on diagnostic results (such as confirmation of disease type), the weights of knowledge graph nodes are dynamically updated to continuously optimize the matching algorithm.

[0083] Application scenarios include: Single disease diagnosis: For example, direct diagnosis of rice blast or rice planthoppers. Differentiation of similar symptoms: For example, distinguishing between rice blast and bacterial leaf spot, combining expansion rate and environmental conditions to provide more accurate diagnostic results. Dynamic change monitoring: In continuous monitoring, dynamic characteristics are combined to distinguish between disease and non-disease stress.

[0084] The pest and disease expansion rate fluctuation index and the soil EC value stress anomaly index are converted into comprehensive feature vectors, which are used as inputs of the machine learning model. The machine learning model uses the discrimination accuracy value labels of similar pest and disease symptoms predicted by each set of comprehensive feature vectors as the prediction target, and minimizes the sum of the prediction errors of the discrimination accuracy value labels of similar pest and disease symptoms as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence, and the model training is stopped. The discrimination accuracy value of similar pest and disease symptoms is determined according to the model output results. The machine learning model is a polynomial regression model.

[0085] The method for obtaining the accuracy value of distinguishing similar symptoms of pests and diseases is to obtain the corresponding function expression from the comprehensive feature vector training data of the trained machine learning model: Where, is the output function of the model, WDC is the fluctuation index of pest and disease expansion rate, YGF is the soil EC value stress abnormality index, and ASD is the accuracy value of distinguishing similar symptoms of pests and diseases.

[0086] The diagnosis result classification module divides the accuracy of the disease and insect pest diagnosis results according to the evaluation results, and divides the disease and insect pest diagnosis results into accurate diagnosis results, incompletely accurate diagnosis results and inaccurate diagnosis results, and performs corresponding processing.

[0087] Comparing the obtained discrimination accuracy value of similar symptoms of pests and diseases with a gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and comparing the discrimination accuracy value of similar symptoms of pests and diseases with the first standard threshold and the second standard threshold respectively;

[0088] If the accuracy value of distinguishing similar symptoms of pests and diseases is greater than the second standard threshold, it means that the accuracy of distinguishing similar symptoms of pests and diseases is high, and a high-accuracy distinction signal is generated, and the diagnosis result is classified as an accurate diagnosis result;

[0089] If the accuracy value of distinguishing similar symptoms of pests and diseases is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the accuracy of distinguishing similar symptoms of pests and diseases is medium, and a medium accuracy distinction signal is generated, and the diagnosis result is classified as an incomplete accuracy diagnosis result;

[0090] If the accuracy value of distinguishing similar symptoms of pests and diseases is less than the first standard threshold, it means that the accuracy of distinguishing similar symptoms of pests and diseases is low. At this time, a low accuracy distinction signal is generated, and the diagnosis result is classified as an inaccurate diagnosis result.

[0091] Processing of Accuracy Diagnosis Results: When the ASD (Accuracy Standard Deviation) for distinguishing similar symptoms between pests and diseases exceeds the second standard threshold, the diagnostic result is highly reliable and the system is able to distinguish highly similar symptoms. Direct Diagnosis Output: The diagnosed disease name, cause, characteristics, and control recommendations are directly displayed to the user. Example: Diagnosis result: "Rice Blast." Control Recommendation: Use of a fungicide (tricyclazole) at a spray concentration of 200-300 times the recommended concentration, or increased field ventilation and dehumidification. Providing a Detailed Reasoning Path: The knowledge graph displays the diagnostic reasoning process, including: the matching degree between symptom keywords and knowledge graph symptom nodes; similarity analysis between dynamic features (such as expansion rate) and standard features; and the weighting of the impact of environmental conditions (such as soil EC value) on the diagnosis. This helps users understand the diagnostic basis and increase confidence. Result Storage and Optimization: Diagnosis results are stored as sample data for further model optimization. Success stories are recorded to enhance the knowledge graph's weight assignment and association analysis capabilities.

[0092] Handling inaccurate diagnostic results: ASD < the first standard threshold indicates low confidence in the diagnostic result, possibly due to: atypical symptoms or insufficient data. Insufficient system knowledge graph coverage prevents new diseases or non-disease factors (such as stress). Solutions include: Marking as failed diagnosis or unknown disease: This clearly informs the user that the diagnostic result is inaccurate, possibly due to an unknown disease or complex stress. Example: Prompt message: "The system cannot accurately determine the disease type. Please take preliminary measures according to the following." Providing general control measures: Based on symptom characteristics, users are advised to take common disease prevention and environmental control measures: Strengthen field management, such as improving ventilation and reducing humidity. Improve soil conditions, such as reducing soil EC through irrigation and flushing under saline-alkali stress. Utilize broad-spectrum control measures, such as selecting a broad-spectrum fungicide or biological control solution. Unmatched symptom and environmental data are stored as "outliers," triggering the following mechanisms: Knowledge graph expansion: Adding new disease nodes or stress feature nodes to enhance knowledge coverage. Model retraining: Using outlier data to update the machine learning model and optimize its diagnostic capabilities. If the diagnosis result is an unknown disease, the relevant data (such as images, environmental characteristics, and dynamic characteristics) will be uploaded to the agricultural diagnosis platform for manual review by experts.

[0093] The optimization management module further analyzes the accuracy of distinguishing similar symptoms of pests and diseases within a fixed time period for incompletely accurate diagnosis results, and dynamically optimizes the knowledge graph and classification model based on the analysis results to improve the accuracy of crop disease and pest diagnosis.

[0094] In a fixed time period T, key data of incomplete accuracy diagnosis results are collected to form a data set D: Where, is the accuracy value for distinguishing similar symptoms of pests and diseases diagnosed for the i-th time, is the fluctuation index of the expansion rate of the i-th diagnosed pest and disease. is the soil EC stress abnormality index of the i-th diagnosis, It is the actual disease label, which is used to verify the accuracy of the diagnosis results (obtained through subsequent follow-up visits or expert feedback).

[0095] Define a comprehensive evaluation metric for discrimination accuracy within a fixed time period , the expression is: ; is the diagnostic accuracy function, according to and the actual label Determine whether the diagnosis is correct: ; N is the number of incomplete accuracy diagnosis samples in time period T.

[0096] Analyze the key features that lead to misdiagnosis in incomplete accuracy diagnosis samples (deficiencies in symptom nodes, environmental condition nodes, and dynamic feature nodes) and dynamically adjust the structure and weights of the knowledge graph: Based on expert feedback or system review results, supplement the missing disease, symptom, or environmental condition nodes in the knowledge graph. Reweight node relationships with high misdiagnosis rates to improve matching accuracy.

[0097] Defining nodes in the knowledge graph and diseased nodes Weight , based on the diagnostic data within a fixed time period T, the weight is updated. The expression is: Where, , is the node in the i-th sample Diseased nodes The contribution prediction value of , η is the learning rate, which is used to control the amplitude of weight update.

[0098] Based on comprehensive feature vector (WDC, YGF) and actual label , retrain the classification model to minimize the discrimination accuracy prediction error: Where, is the objective function (loss function), is the discrimination accuracy value predicted by the classification model. The model parameters θ are updated using the gradient descent method: ; is the learning rate, is the gradient of the objective function with respect to the model parameter θ. Define the overall diagnostic accuracy index after dynamic optimization , the expression is: Where, The closer it is to 1, the higher the system's accuracy in diagnosing pests and diseases with similar symptoms.

[0099] In this embodiment, accurate diagnosis and dynamic optimization of crop pests and diseases are achieved through the collaborative work of multiple modules. The data acquisition module obtains pest and disease description information, image data, and environmental data such as meteorological and soil data; the data processing module extracts symptom keywords, pest and disease expansion rate fluctuation characteristics, and soil EC value stress characteristics through natural language processing and computer vision technology; the knowledge graph construction module establishes a knowledge graph based on pest and disease knowledge, including symptom nodes, environmental condition nodes, dynamic feature nodes, and their associated relationships; the pest and disease differentiation module matches the extracted symptom keywords with the knowledge graph and evaluates the accuracy of distinguishing similar symptoms of pests and diseases based on expansion rate and soil stress characteristics; the diagnosis result classification module divides the diagnosis results into accurate diagnosis results, incomplete accuracy diagnosis results, and inaccurate diagnosis results based on the differentiation accuracy, and takes targeted measures; the optimization management module further analyzes the incomplete accuracy diagnosis results, dynamically optimizes the knowledge graph and classification model, and continuously improves the accuracy of the diagnosis system.

[0100] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0101] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer program are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center via a wired (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0102] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A crop disease and insect pest diagnosis system based on knowledge graph, characterized by: It includes data collection module, data processing module, knowledge graph construction module, pest and disease differentiation module, diagnosis result classification module and optimization management module; A data acquisition module is used to obtain crop pest and disease description information, pest and disease image data, and environmental data, wherein the environmental data includes meteorological data and soil data; The data processing module is used to perform natural language processing on pest and disease description information to extract symptom keywords, process pest and disease image data using computer vision technology to extract the fluctuation characteristics of pest and disease expansion rate, and process environmental data to obtain soil EC value stress characteristics; Specifically, the method includes: analyzing the extracted pest and disease expansion rate fluctuation characteristics to generate a pest and disease expansion rate fluctuation index. The pest and disease expansion rate fluctuation index is obtained by obtaining the pest and disease expansion rate data in different time periods and constructing two time series: is the reference expansion rate series, representing the standard disease expansion rate pattern; The disease expansion rate sequence currently collected, each rate point represents the rate of change of the disease area in a certain period of time, for any two points , calculate the local distance: Where, for point and The absolute distance; Construct the cost matrix D, which is an m×n matrix, where each element D(i,j) represents the minimum cumulative distance of the path from the starting point (1,1) to the point (i,j). The initial cost is: , ; , ;Fill the matrix sequentially through dynamic programming and cumulatively calculate the minimum matching cost between each point; Defining the optimal path is a set of point pair sequences that satisfy the following conditions: Indicates the starting point; Indicates the end point, the path is continuous, no jumping is allowed, and the total cost of the path for: ; Calculate the pest expansion rate fluctuation index, the expression is: Where, is the path length, WDC is the pest expansion rate fluctuation index; After analyzing the extracted soil EC stress characteristics, the soil EC stress anomaly index is generated. The soil EC stress anomaly index is obtained by: collecting a sampling data set of soil EC values: input data , each Represents the soil EC value of a sampling point. The soil EC value is standardized and the expression is: ; and are the data mean and standard deviation respectively, and output the standardized data set ; The probability density function is: Where, is the weight of the kth Gaussian distribution, satisfying ; is the probability density function of the kth Gaussian distribution, which is in the form of: ; is the mean of the kth Gaussian distribution, indicating the center of a certain state, is the variance of the k-th Gaussian distribution, indicating the distribution range of the state; Random initialization The initial value of , calculates the probability that each data point belongs to the k-th Gaussian distribution: Where, is the responsibility weight of the i-th data point belonging to the k-th distribution, according to Update GMM parameters, the expression is: ; Calculate data points The total probability of , the expression is: ; The soil EC stress abnormality index is defined as point The probability of belonging to a non-normal distribution is expressed as: ; YGF is the soil EC stress abnormality index, for point The probability of belonging to the normal distribution is expressed as: Where, are the weight, mean, and standard deviation of the Gaussian distribution under normal conditions set based on historical data; A knowledge graph construction module is used to build a knowledge graph based on pest and disease knowledge, wherein the knowledge graph includes pest and disease symptom nodes, environmental condition nodes, dynamic feature nodes, and the association relationship between each node; The pest and disease differentiation module is used to match the extracted symptom keywords with the symptom information stored in the knowledge graph, differentiate pests and diseases based on the matching results, and evaluate the accuracy of distinguishing similar symptoms of pests and diseases based on the fluctuation characteristics of pest and disease expansion rate and soil EC value stress characteristics; The diagnosis result classification module classifies the accuracy of the disease and insect pest diagnosis results according to the evaluation results, and divides the disease and insect pest diagnosis results into accurate diagnosis results, incompletely accurate diagnosis results and inaccurate diagnosis results, and performs corresponding processing; The optimization management module further analyzes the accuracy of distinguishing similar symptoms of pests and diseases within a fixed time period for incompletely accurate diagnosis results, and dynamically optimizes the knowledge graph and classification model based on the analysis results to improve the accuracy of crop disease and pest diagnosis.

2. The crop disease and insect pest diagnosis system based on knowledge graph according to claim 1, characterized in that: In the pest and disease differentiation module, the accuracy of distinguishing similar symptoms of pests and diseases is evaluated by combining the fluctuation characteristics of pest and disease expansion rate and the stress characteristics of soil EC value. Specifically: The pest and disease expansion rate fluctuation index and the soil EC value stress anomaly index are converted into comprehensive feature vectors, which are used as inputs of the machine learning model. The machine learning model uses the discrimination accuracy value labels of similar pest and disease symptoms predicted by each set of comprehensive feature vectors as the prediction target, and minimizes the sum of the prediction errors of the discrimination accuracy value labels of similar pest and disease symptoms as the training target. The machine learning model is trained until the sum of the prediction errors reaches convergence, and the model training is stopped. The discrimination accuracy value of similar pest and disease symptoms is determined according to the model output results. The machine learning model is a polynomial regression model.

3. The crop disease and insect pest diagnosis system based on knowledge graph according to claim 2, characterized in that: In the diagnosis result classification module, the accuracy of the disease and insect pest diagnosis results is divided into accurate diagnosis results, incompletely accurate diagnosis results and inaccurate diagnosis results. Specifically: Comparing the obtained discrimination accuracy value of similar symptoms of pests and diseases with a gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and comparing the discrimination accuracy value of similar symptoms of pests and diseases with the first standard threshold and the second standard threshold respectively; If the accuracy value of distinguishing similar symptoms of pests and diseases is greater than the second standard threshold, it means that the accuracy of distinguishing similar symptoms of pests and diseases is high, and a high-accuracy distinction signal is generated, and the diagnosis result is classified as an accurate diagnosis result; If the accuracy value of distinguishing similar symptoms of pests and diseases is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the accuracy of distinguishing similar symptoms of pests and diseases is medium, and a medium accuracy distinction signal is generated, and the diagnosis result is classified as an incomplete accuracy diagnosis result; If the accuracy value of distinguishing similar symptoms of pests and diseases is less than the first standard threshold, it means that the accuracy of distinguishing similar symptoms of pests and diseases is low. At this time, a low accuracy distinction signal is generated, and the diagnosis result is classified as an inaccurate diagnosis result.

4. The crop disease and insect pest diagnosis system based on knowledge graph according to claim 1, characterized in that: In the optimization management module, for incompletely accurate diagnosis results, further analysis is conducted on the accuracy of distinguishing similar symptoms of pests and diseases within a fixed time period. Based on the analysis results, the knowledge graph and classification model are dynamically optimized. Specifically: In a fixed time period T, key data of incomplete accuracy diagnosis results are collected to form a data set D: Where, is the accuracy value for distinguishing similar symptoms of pests and diseases diagnosed for the i-th time, is the fluctuation index of the expansion rate of the i-th diagnosed pest, is the soil EC stress abnormality index of the i-th diagnosis, For actual disease labels, define a comprehensive evaluation index for discrimination accuracy within a fixed time period , the expression is: ; is the diagnostic accuracy function, according to and the actual label Determine whether the diagnosis is correct: ; N is the number of incomplete accuracy diagnosis samples in time period T.

5. The crop disease and insect pest diagnosis system based on knowledge graph according to claim 4, characterized in that: Defining nodes in the knowledge graph and diseased nodes Weight , based on the diagnostic data within a fixed time period T, the weight is updated. The expression is: Where, , is the node in the i-th sample Diseased nodes The contribution prediction value of , η is the learning rate, which is used to control the amplitude of weight update; Based on comprehensive feature vector (WDC, YGF) and actual label , retrain the classification model to minimize the discrimination accuracy prediction error: Where, is the objective function, The classification accuracy value predicted by the classification model is used to update the model parameters θ using the gradient descent method: ; is the learning rate, The gradient of the objective function with respect to the model parameter θ defines the overall diagnostic accuracy index after dynamic optimization , the expression is: , N is the total number of samples in the data set.

Images