Big data-based pest control information management system

Abstract

The invention discloses a disease and pest control information management system based on big data, and the system comprises a multi-source data collection module which is used for obtaining multi-source data, carrying out the anomaly detection of all data, and processing the abnormal data; the distributed data storage module is used for constructing a plurality of databases and respectively storing the detected and processed data by utilizing the plurality of databases; the big data analysis and scheme generation module is used for generating a prevention and treatment scheme and outputting the prevention and treatment scheme; and the business application module is used for managing user information and presenting a prevention and treatment scheme. According to the method, data sources are increased, data comprehensiveness and accuracy are improved, multi-class databases are constructed, data are classified and managed, management efficiency is improved, during retrieval, the databases are positioned through keywords requested by a user firstly, then the positioned databases are retrieved, retrieval complexity and information processing amount are reduced, and retrieval efficiency is improved. The retrieval efficiency is improved, so that the prevention and treatment scheme generation efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of pest control information management, and in particular to a pest control information management system based on big data. Background Art

[0002] Current agricultural pest control efforts face challenges such as data fragmentation, homogeneous control plans, and insufficient environmental adaptability. Traditional pest information systems primarily rely on manual reporting or single-sensor data (e.g., temperature alone). They lack the ability to integrate multi-dimensional data such as plant physiological status (e.g., leaf spectral characteristics), field microecology (e.g., soil microbial data), and pesticide characteristics (e.g., resistance dynamics). This results in one-sided pest control plans. Furthermore, all information is stored in the same database, requiring the entire database to be read for information retrieval. This increases read time, reduces information retrieval efficiency, and leads to inefficient control plan generation. Summary of the Invention

[0003] The purpose of the present invention is to provide a pest control information management system based on big data to address the above-mentioned problems, increase data sources, improve the accuracy of control plans, classify and manage data by constructing multiple databases, locate the database by keywords during retrieval, and only search the located database, thereby reducing the complexity of retrieval and the amount of information processing, improving retrieval efficiency, and improving the efficiency of generating control plans.

[0004] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is as follows:

[0005] According to one aspect of the present invention, a pest control information management system based on big data is provided, comprising:

[0006] Multi-source data acquisition module, used to obtain plant information data, field environment data, pest and disease characteristic data, pesticide data and historical control data, and perform anomaly detection on all data and process abnormal data;

[0007] A distributed data storage module is used to construct multiple databases, and use the multiple databases to respectively store the detected and processed plant information data, field environment data, pest and disease characteristic data, pesticide data and historical control data;

[0008] A big data analysis and solution generation module is used to receive user requests, extract keywords from the user requests, locate the database based on the keywords, search the located database, integrate the retrieved information corresponding to the keywords to generate a prevention and control plan, and output the prevention and control plan;

[0009] Business application module, used to manage user information and present prevention and control plans.

[0010] Preferably, the multi-source data acquisition module includes a data input unit and a data anomaly detection unit;

[0011] The data input unit includes a manual input port, a sensor data access port, an image data access port and an external system access port;

[0012] The data anomaly detection unit is used to perform anomaly detection on the data collected by the data input unit and process the abnormal data.

[0013] Preferably, the data anomaly detection unit includes a parameterized anomaly detection subunit and a non-parametric anomaly detection subunit;

[0014] The parameterized anomaly detection subunit includes the following detection steps:

[0015] Data grouping: group continuous data;

[0016] Parameter calculation: calculate the sample mean μ and sample standard deviation σ of the sample data of the current group;

[0017] Anomaly detection: Use the calculated sample mean and sample standard deviation to calculate the Atman Z-score model and obtain the calculation threshold for each data point in the current group;

[0018] Threshold judgment: judge whether the calculated threshold exceeds the standard threshold range. When the calculated threshold exceeds the standard threshold range, the data points that exceed the range are marked as abnormal data points.

[0019] The non-parametric anomaly detection subunit includes the following detection steps:

[0020] Sorting: Arrange the data set in ascending order;

[0021] Quantile calculation: calculate the first quartile Q1 and the third quartile Q3, and calculate the interquartile range IQR from the first quartile Q1 and the third quartile Q3;

[0022] Threshold range calculation: The lower limit and upper limit of the threshold range are calculated by the first quartile Q1, the third quartile Q3 and the interquartile range IQR;

[0023] Threshold judgment: When a data point exceeds the upper or lower limit of the threshold range, the value of the data point is judged as an outlier.

[0024] Preferably, the distributed data storage module includes a database construction unit and a distributed storage unit;

[0025] The database construction unit includes a plant information database, a field environment database, a pest and disease characteristic database, a pesticide database, and a historical control database;

[0026] The distributed storage unit is used to identify the data collected and processed by the multi-source data collection module, identify the data type, and store the data in the corresponding database according to the data type.

[0027] Preferably, the big data analysis and solution generation module includes a user request information extraction unit, a prevention and control solution generation unit, and a prevention and control solution output unit;

[0028] The user request information extraction unit is used to receive a user request and extract keywords of the user request;

[0029] The control plan generating unit locates a database based on keywords, searches the located database, and integrates the retrieved information corresponding to the keywords to generate a control plan;

[0030] The control plan output unit is used to output the control plan.

[0031] Preferably, the user request information extraction unit includes a multimodal data receiving subunit and a natural language processing subunit;

[0032] The multimodal data receiving subunit includes a text data receiving port, an image data receiving port and a voice data receiving port;

[0033] The natural language processing subunit includes the following processing steps:

[0034] Word segmentation: split the sentence requested by the user into words;

[0035] Part-of-speech tagging: Identify the part of speech of the split words and tag them;

[0036] Keyword extraction: extract keywords from split words.

[0037] Keyword output: Output the extracted keywords.

[0038] Preferably, the control plan generation unit includes a database positioning subunit, a database retrieval subunit and an information integration subunit;

[0039] The database positioning subunit is used to locate the database according to the keyword;

[0040] The database search subunit searches the located database according to the keywords and extracts information data related to the keywords;

[0041] The information synthesis subunit integrates the extracted information data to generate a prevention and treatment plan corresponding to the user's request.

[0042] Preferably, the application service module includes a management unit and an information browsing unit;

[0043] The management unit is used to manage user information; the information browsing unit is used to present prevention and treatment plans.

[0044] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0045] The present invention improves the comprehensiveness and accuracy of data by increasing data sources, and improves management efficiency by constructing multiple databases to classify and manage data. When searching, the database is first located by the keyword requested by the user, and then the located database is searched, which reduces the complexity of the search and the amount of information processing, improves the search efficiency, and thus improves the efficiency of generating prevention and control plans. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a functional structure diagram of the present invention. DETAILED DESCRIPTION

[0047] To make the objectives, technical solutions, and advantages of the present invention more clearly understood, the present invention is further described below with reference to the accompanying drawings and by way of preferred embodiments. However, it should be noted that many of the details listed in this specification are merely provided to help the reader gain a thorough understanding of one or more aspects of the present invention, and these aspects of the present invention can be practiced even without these specific details.

[0048] See also Figure 1 The present invention provides a pest control information management system based on big data, which includes a multi-source data acquisition module. The multi-source data acquisition module is used to obtain plant information data, field environment data, pest and disease characteristic data, pesticide data and historical control data, and perform structured processing on all data. Specifically, the multi-source data acquisition module includes a data input unit and a data anomaly detection unit. The data input unit includes a manual input port, a sensor data access port, an image data access port and an external system access port. The manual input port can be used to receive manually typed data. The sensor data access port can be used to receive field environment data such as temperature, humidity, and light sensors. The image data access port can be used to receive plant and pest and disease characteristic image data. The external system access port can be used to receive meteorological data, pesticide data, historical control data, etc. Through the various data receiving ports of the data input unit, multi-channel acquisition of plant information, field environment, pest and disease characteristics, pesticide data and historical control data is achieved.

[0049] The data anomaly detection unit is used to perform anomaly detection on the data received by the data input unit. The data anomaly detection unit uses statistical methods to detect outliers on the original data and identify abnormal data points. Specifically, the data anomaly detection unit includes a parametric anomaly detection subunit and a non-parametric anomaly detection subunit. The parametric anomaly detection subunit constructs a normal distribution model for continuous data (such as temperature and humidity data), calculates the mean μ and standard deviation σ, and marks data points that deviate from the range of μ±3σ as outliers. The specific implementation steps are as follows:

[0050] Data grouping: Group continuous data. Grouping can be done by time, seasonal cycle, spatial dimension, environmental data, or biological data. For example, by time, grouping can be done based on plant growth cycle stages (seedling / flowering / fruiting) or seasonal cycles (spring / summer / autumn / winter); by spatial dimension, grouping can be done based on geographical region (southern / northern planting areas) or facility type (open-air / greenhouse / shed); by data type, grouping can be done based on environmental data (temperature / humidity / light) or biological data (pest and disease incidence / crop growth indicators).

[0051] Parameter calculation: Calculate the sample mean μ and sample standard deviation σ of the sample data of the current group.

[0052] The sample mean μ is calculated by the following formula:

[0053]

[0054] Where n is the sample size, x i is the value of the i-th sample data point.

[0055] The sample standard deviation σ is calculated by the following formula:

[0056]

[0057] Where n is the sample size, x i is the value of the i-th sample data point, and μ is the sample mean.

[0058] Anomaly Detection: Using the calculated sample mean μ and sample standard deviation σ, we calculate the Atlanto Z-score model and obtain the threshold for each data point in the current group. The Atlanto Z-score model's results represent the degree to which a data point deviates from the standard range.

[0059] The Z-score model of Artman is:

[0060]

[0061] Among them, Z is the threshold, x is the value of the sample data point, μ is the sample mean; σ is the sample standard deviation.

[0062] Threshold judgment: Determines whether the calculated threshold exceeds the standard threshold range. If the calculated threshold exceeds the standard threshold range, the data point that exceeds the range is marked as an abnormal data point. For example, if the standard threshold range is [-3, 3], data points with |Z| ≤ 3 are classified as normal data; data points with 3 < |Z| ≤ 4 are classified as suspected abnormal data and a yellow warning is set; data points with |Z| > 4 are classified as highly abnormal data and a red warning is set.

[0063] The above process is used to detect anomalies in the input continuous data, and early warnings are used to remind staff to delete or modify the abnormal data.

[0064] Specific application scenarios:

[0065] Scenario: Real-time data from temperature and humidity sensors in a tea garden

[0066] Parameters: Temperature μ = 25°C, σ = 2°C; Humidity μ = 75%, σ = 5%

[0067] Anomaly Detection:

[0068] Temperature = 32°C, 3.5>3, the temperature data is abnormal. The staff can set it to delete the data and then store the rest, or modify the data before storing it.

[0069] Humidity = 40%, -7<-3, the humidity data is abnormal.

[0070] The non-parametric detection model is used to detect discrete data. It calculates the difference between the third quartile and the first quartile of the data set, and uses the difference to calculate the upper and lower limits to obtain the standard range. Data points that deviate from the standard range are marked as outliers. The specific implementation steps are as follows:

[0071] Sorting: Arrange the data set in ascending order;

[0072] Quantile calculation: First calculate the first quartile Q1, which is the maximum value of the smallest 25% of the data (lower quartile); then calculate the third quartile Q3, which is the minimum value of the largest 25% of the data (upper quartile). Specifically: when the sample size n is an odd number, use linear interpolation to determine Q1 and Q3; when n is an even number, directly take the value in the middle position; finally calculate IQR, which is the interquartile range, that is, the difference between Q3 and Q1, and the calculation formula is: IQR = Q3-Q1.

[0073] Threshold range calculation: The lower limit of the range is calculated as: Q1 - 1.5 × IQR; the upper limit of the range is calculated as: Q3 + 1.5 × IQR.

[0074] Threshold judgment: When a data point exceeds the threshold range, the value of the data point is judged as an outlier.

[0075] Specific application scenarios:

[0076] Scenario 1: Detecting abnormal frequency of pests and diseases

[0077] Monthly aphid occurrence data for a certain orchard: [2, 5, 3, 8, 12, 45, 6, 7, 9, 10]

[0078] Processing flow:

[0079] Sorting: [2,3,5,6,7,8,9,10,12,45]

[0080] Calculation: Q1 (number 2.5): Q3 (number 7.5): IQR:9.5-4=5.5.

[0081] Threshold range calculation: lower limit is: 4-1.5×5.5=-4.25, upper limit is: 9.5+1.5×5.5=17.75.

[0082] Threshold judgment: Data point 45>17.75, the data point 45 is judged as severely abnormal data.

[0083] Scenario 2: Pesticide Usage Compliance Check

[0084] Pesticide usage per mu in a certain area: [0.5, 0.8, 1.2, 1.5, 2.0, 5.0, 0.7]

[0085] Calculation results: Q1 = 0.7, Q3 = 1.75, IQR = 1.05, upper limit = 1.75 + 1.5 × 1.05 = 3.325.

[0086] Threshold judgment: Data point 5.0>3.325 exceeds the threshold range, and data point 5.0 is marked as abnormal data.

[0087] Scenario 3: Processing non-normal sensor data

[0088] Soil pH monitoring data: [4.2, 4.5, 4.8, 5.0, 5.2, 5.5, 6.8, 7.0]

[0089] Processing results: The IQR method was used to detect that 6.8 and 7.0 were outliers (the normal pH in soil acidification areas should be <6.5), and 6.8 and 7.0 were marked as outliers.

[0090] By performing outlier detection on continuous and discrete data, abnormal data in the data can be processed, reducing the waste of storage resources and increasing data accuracy, providing a high-quality data foundation for subsequent precise analysis.

[0091] The system also includes a distributed data storage module, which is used to construct multiple databases and use them to store plant information data, field environment data, pest and disease characteristic data, pesticide data, and historical control data after anomaly detection. The distributed data storage module includes a database construction unit and a distributed storage unit. The database construction unit constructs the plant information database, field environment database, pest and disease characteristic database, pesticide database, and historical control database, respectively, to implement data classification management. The plant information database contains plant information (such as name, plant family, growth cycle, and geographical distribution), biological characteristics (such as light requirements, temperature threshold, soil pH preference, and stress resistance), and planting records (such as variety source, seedling raising method, transplanting time, and historical yield data). The field environment database contains real-time monitoring data (such as temperature and humidity, light intensity, soil moisture, wind speed and direction), meteorological data (such as historical rainfall, accumulated temperature, frost days, extreme weather events, etc.), and environmental control records (such as irrigation time, fertilizer type, and greenhouse ventilation records). The pest and disease characteristic database contains disease information (such as disease name, pathogen, symptom description, disease pattern, etc.), pest information (such as pest name, life cycle, affected part, resistance level, natural enemy type, etc.) and image data (such as high-definition pictures of diseased parts, microscopic photos of pests, etc.). The pesticide database contains basic attributes (such as pesticide name, active ingredient, formulation, toxicity level, etc.), usage specifications (such as recommended dosage, dilution ratio, safe interval, incompatibility taboos, etc.) and environmental information (such as pesticide residue standards, biodegradability, toxicity data to natural enemies, etc.). The historical control database contains control records (such as user ID, plant variety, pest and disease type, control time, use plan, the use plan includes pesticide type, dosage and application method) and expert advice. Expert advice is mainly for control plans for new plant varieties.

[0092] The distributed storage unit is used to identify the data collected and processed by the multi-source data acquisition module, identify the data type, and store the data in the corresponding database according to the data type, providing convenience for subsequent data reading.

[0093] The system also includes a big data analysis and solution generation module. This module receives user requests, extracts keywords from the user requests, locates a database based on the keywords, searches the database, integrates the retrieved information corresponding to the keywords, generates a prevention and control plan, and outputs the prevention and control plan. Specifically, the big data analysis and solution generation module includes a user request information extraction unit, a prevention and control plan generation unit, and a prevention and control plan output unit.

[0094] The user request information extraction unit includes a multimodal data receiving subunit and a natural language processing subunit. The multimodal data receiving subunit supports the reception of text, image and voice data. The text data is the text description manually input by the user (such as "yellow spots appear on the leaves of Panax notoginseng"); the image data is the photos of pest and disease parts uploaded by the user (such as leaves and fruit spots); and the voice data is converted into text through voice recognition technology (such as "Please inquire about the prevention and control methods of rice blast disease"). The natural language processing subunit identifies and splits the received data to form search keywords. Specifically, the natural language processing process is as follows:

[0095] Word segmentation: Use Jieba word segmentation tool to split the sentence into words (e.g. split “yellow spots appear on Panax notoginseng leaves” into “panax notoginseng / leaves / appear / yellow spots”).

[0096] Part-of-speech tagging: Identify and tag the parts of speech of the split words, such as nouns (variety names), verbs (symptom descriptions), and adjectives (degree words).

[0097] Keyword extraction: Extract keywords, such as "Panax notoginseng", "leaf", and "yellow spot".

[0098] Keyword output: Output the extracted keywords.

[0099] The prevention and control plan generation unit includes a database positioning subunit, a database retrieval subunit and an information integration subunit. The database positioning subunit is used to locate the database according to keywords, for example, the plant information database is located according to the keyword "Sanqi", and the historical prevention and control database is located according to the keyword "yellow spot". The database retrieval subunit searches the located database according to the keywords and extracts information data related to the keywords. The information synthesis subunit integrates the extracted information data and generates a prevention and control plan corresponding to the user request. In order to remove useless data, in this embodiment, it also includes an information filtering subunit. The information filtering subunit extracts geographic location data or environmental data based on keywords or user information, and removes prevention and control plans that are useless to the current user or contrary to the current environment based on the geographic location and environmental data, thereby removing useless information and improving the accuracy of prevention and control plan generation.

[0100] Case: Small-scale green growers

[0101] User input: "I am growing Panax notoginseng on a small scale in Yunnan. The leaves are showing yellow-brown spots with yellow halos around the edges."

[0102] Keyword extraction: "Yunnan", "small scale", "Panax notoginseng", "leaves", "yellow-brown", "lesions".

[0103] The information retrieved was that the "Bacillus subtilis + yellow board trapping" scheme successfully prevented and controlled Panax notoginseng black spot disease and that Yunnan is currently in the rainy season.

[0104] Generation plan: Root irrigation with 500 times diluted Bacillus subtilis wettable powder (the humidity in Yunnan is high during the rainy season, so the foliage spraying will be automatically changed to root irrigation), once every 7 days, for 2 consecutive times.

[0105] The control plan output unit is used to output the generated control plan.

[0106] The system also includes a business application module, which is used to manage user information and present prevention and treatment plans. Specifically, the business application module includes an administration unit and an information browsing unit. The administration unit is used to implement user information management (such as distinguishing between ordinary users, expert users, and administrator roles), database maintenance (such as data updates and permission management), and system log auditing. The information browsing unit is used to provide user queries, prevention and treatment plans, symptom image libraries, disease cycle maps, etc.

[0107] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A pest control information management system based on big data, characterized in that: include: Multi-source data acquisition module, used to obtain plant information data, field environment data, pest and disease characteristic data, pesticide data and historical control data, and perform anomaly detection on all data and process abnormal data; A distributed data storage module is used to construct multiple databases, and use the multiple databases to respectively store the detected and processed plant information data, field environment data, pest and disease characteristic data, pesticide data and historical control data; A big data analysis and solution generation module is used to receive user requests, extract keywords from the user requests, locate the database based on the keywords, search the database, integrate the retrieved information corresponding to the keywords to generate a prevention and control plan, and output the prevention and control plan; Business application module, used to manage user information and present prevention and control plans.

2. The pest control information management system based on big data according to claim 1, characterized in that: The multi-source data acquisition module includes a data input unit and a data anomaly detection unit; The data input unit includes a manual input port, a sensor data access port, an image data access port and an external system access port; The data anomaly detection unit is used to perform anomaly detection on the data collected by the data input unit and process the abnormal data.

3. The pest control information management system based on big data according to claim 2, characterized in that: The data anomaly detection unit includes a parameterized anomaly detection subunit and a non-parametric anomaly detection subunit; The parameterized anomaly detection subunit includes the following detection steps: Data grouping: group continuous data; Parameter calculation: calculate the sample mean μ and sample standard deviation σ of the sample data of the current group; Anomaly detection: Use the calculated sample mean and sample standard deviation to calculate the Atman Z-score model and obtain the calculation threshold for each data point in the current group; Threshold judgment: judge whether the calculated threshold exceeds the standard threshold range. When the calculated threshold exceeds the standard threshold range, the data points that exceed the range are marked as abnormal data points. The non-parametric anomaly detection subunit includes the following detection steps: Sorting: Arrange the data set in ascending order; Quantile calculation: calculate the first quartile Q1 and the third quartile Q3, and calculate the interquartile range IQR from the first quartile Q1 and the third quartile Q3; Threshold range calculation: The lower limit and upper limit of the threshold range are calculated by the first quartile Q1, the third quartile Q3 and the interquartile range IQR; Threshold judgment: When a data point exceeds the upper or lower limit of the threshold range, the value of the data point is judged as an outlier.

4. The pest control information management system based on big data according to claim 1, characterized in that: The distributed data storage module includes a database construction unit and a distributed storage unit; The database construction unit includes a plant information database, a field environment database, a pest and disease characteristic database, a pesticide database, and a historical control database; The distributed storage unit is used to identify the data collected and processed by the multi-source data collection module, identify the data type, and store the data in the corresponding database according to the data type.

5. The pest control information management system based on big data according to claim 1, characterized in that: The big data analysis and solution generation module includes a user request information extraction unit, a prevention and control solution generation unit, and a prevention and control solution output unit; The user request information extraction unit is used to receive a user request and extract keywords of the user request; The control plan generating unit locates a database based on keywords, searches the located database, and integrates the retrieved information corresponding to the keywords to generate a control plan; The control plan output unit is used to output the control plan.

6. The pest control information management system based on big data according to claim 4, characterized in that: The user request information extraction unit includes a multimodal data receiving subunit and a natural language processing subunit; The multimodal data receiving subunit includes a text data receiving port, an image data receiving port and a voice data receiving port; The natural language processing subunit includes the following processing steps: Word segmentation: split the sentence requested by the user into words; Part-of-speech tagging: Identify the part of speech of the split words and tag them; Keyword extraction: extract keywords from split words; Keyword output: Output the extracted keywords.

7. The pest control information management system based on big data according to claim 6, characterized in that: The control plan generation unit includes a database positioning subunit, a database retrieval subunit and an information integration subunit; The database positioning subunit is used to locate the database according to the keyword; The database search subunit searches the located database according to the keywords and extracts information data related to the keywords; The information synthesis subunit integrates the extracted information data to generate a prevention and treatment plan corresponding to the user's request.

8. The pest control information management system based on big data according to claim 1, characterized in that: The application service module includes a management unit and an information browsing unit; The management unit is used to manage user information; the information browsing unit is used to present prevention and treatment plans.

Images