Crop yield prediction method and system based on multi-data fusion

Abstract

The invention provides a crop yield prediction method and system based on multi-data fusion, and relates to the technical field of agricultural information processing, and the method comprises the steps: configuring multi-stage farmland monitoring standards for the designated multi-stage growth stages of crops; regularly acquiring multi-source monitoring equipment data for monitoring the farmland environment; performing feature extraction and dimension reduction processing on the obtained multi-source monitoring equipment data to obtain dimension-reduced multi-source monitoring equipment data, and updating a crop growth element set; obtaining a multi-source deviation data set; arranging the multi-source deviation data in each window according to a sliding step time sequence to form a time sequence multi-source deviation data set; and inputting the time sequence multi-source deviation data set into a crop growth prediction model to obtain a crop yield prediction value. The technical problem that the prior art depends on a single data source and a simple model is solved, and the prediction accuracy of the crop yield is remarkably improved.

Description

Technical Field

[0001] The present application relates to the field of agricultural information processing technology, and in particular to a crop yield prediction method and system based on multi-data fusion. Background Art

[0002] In the field of modern agriculture, accurate prediction of crop yields is of vital importance. With the continuous growth of the global population, people's demand for crop products is also increasing, and agricultural production requires more scientific and efficient management methods. Crop yield forecasting can help farmers, agricultural enterprises and relevant government departments plan resource allocation in advance, formulate reasonable market strategies and ensure a stable supply of agricultural products. Traditional crop yield prediction methods often rely on a single data source or a simple empirical model, ignoring the many complex factors and their interactions in the crop growth process, and cannot accurately consider the different responses of crops to environmental changes at different growth stages. It also lacks the ability to comprehensively consider multiple factors such as soil quality and pesticide residues. Therefore, its prediction accuracy and reliability are low, and there is no way to provide accurate yield forecast data for enterprises. This type of model has a relatively simple structure, fewer factors to consider, and the model accuracy needs to be improved. Traditional crop yield estimation methods are mainly agronomic forecasting methods, statistical forecasting methods, meteorological forecasting methods, etc. Summary of the invention

[0003] This application provides a method and system for predicting crop yields based on multi-data fusion, aiming to solve the technical problem that the existing technology relies on a single data source and a simple empirical model, resulting in low accuracy and reliability of prediction results. In order to further improve the prediction effect of crop yields, the method of the present invention has the characteristics of being easy to implement and having good effect. According to a first aspect of the present disclosure, a method for predicting crop yield based on multi-data fusion is provided, comprising: According to the multi-stage growth stages of crops, multi-stage farmland monitoring standards are configured, and the multi-stage farmland monitoring standards are configured in the intelligent control system, wherein the multi-stage growth stages include germination stage, seedling stage, growth stage and maturity stage; based on the multi-stage growth stages of crops, multi-source monitoring equipment data for monitoring the farmland environment are regularly acquired, wherein the multi-source monitoring equipment is communicatively connected with the intelligent control system; the PCA algorithm is used to perform feature extraction and dimensionality reduction processing on the acquired multi-source monitoring equipment data, to obtain the reduced dimensionality multi-source monitoring equipment data, and to update the crop growth element set, wherein the crop growth element set is constructed based on the content of the multi-stage farmland monitoring standards, and the crop growth element set includes the soil quality of the crop growth. a quantity element set, a weather condition element set and a pesticide residue element set; a sliding window is set, and a sliding step is set according to the multi-level growth stage of the crop, and the average value of the data of each monitoring device updated by each element set in the crop growth element set in the window is calculated, and the average value of the data of each monitoring device in each element feature set is compared with the corresponding monitoring standard in the multi-level farmland monitoring standard, and the deviation value is calculated to obtain a multi-source deviation data set; based on the sliding step, the multi-source deviation data in each window are arranged in the order of the sliding step time to form a time series multi-source deviation data set; the time series multi-source deviation data set is input into the crop growth prediction model to obtain the crop yield prediction value, wherein the prediction model is constructed with ARIMAX as the architecture.

[0004] According to a second aspect of the present disclosure, a crop yield prediction system based on multi-data fusion is provided, comprising: A multi-level farmland monitoring standard configuration module, the multi-level farmland monitoring standard configuration module is used to configure multi-level farmland monitoring standards for the designated multi-level growth stages of crops, and configure the multi-level farmland monitoring standards in the intelligent control system, wherein the multi-level growth stages include germination stage, seedling stage, growth stage and maturity stage; a multi-source monitoring equipment data acquisition module, the multi-source monitoring equipment data acquisition module is used to regularly acquire multi-source monitoring equipment data for monitoring the farmland environment based on the multi-level growth stages of crops, wherein the multi-source monitoring equipment is communicatively connected with the intelligent control system; a multi-source monitoring equipment data feature extraction module, the multi-source monitoring equipment data feature extraction module is used to perform feature extraction and dimensionality reduction processing on the acquired multi-source monitoring equipment data using the PCA algorithm, obtain the reduced dimensionality multi-source monitoring equipment data, and update the crop growth element set, wherein the crop growth element set is constructed based on the content of the multi-level farmland monitoring standard, and the crop growth element set includes the soil for crop growth. a soil quality element set, a weather condition element set and a pesticide residue element set; a multi-source deviation data set acquisition module, the multi-source deviation data set acquisition module is used to set a sliding window, the sliding step is set according to the multi-level growth stage of the crop, the average value of each monitoring device data updated by each element set in the crop growth element set in the window is calculated, the average value of each monitoring device data in each element feature set is compared with the corresponding monitoring standard in the multi-level farmland monitoring standard, the deviation value is calculated, and the multi-source deviation data set is obtained; a time series multi-source deviation data set construction module, the time series multi-source deviation data set construction module is used to arrange the multi-source deviation data in each window in the order of the sliding step time based on the sliding step, so as to form a time series multi-source deviation data set; a crop yield prediction value acquisition module, the crop yield prediction value acquisition module is used to input the time series multi-source deviation data set into a crop growth prediction model to obtain a crop yield prediction value, wherein the prediction model is constructed with ARIMAX as the architecture.

[0005] One or more technical solutions provided in the present disclosure have at least the following technical effects or advantages: configuring multi-level farmland monitoring standards for the defined multi-level growth stages of crops, and configuring the multi-level farmland monitoring standards in the intelligent control system, wherein the multi-level growth stages include germination stage, seedling stage, growth stage and maturity stage; regularly acquiring multi-source monitoring equipment data for monitoring the farmland environment based on the multi-level growth stages of crops, wherein the multi-source monitoring equipment is communicatively connected with the intelligent control system; performing feature extraction and dimensionality reduction processing on the acquired multi-source monitoring equipment data using the PCA algorithm to obtain dimensionality-reduced multi-source monitoring equipment data, and updating the crop growth element set, wherein the crop growth element set is constructed based on the content of the multi-level farmland monitoring standards, and the crop The growth element set includes the soil quality element set, weather condition element set and pesticide residue element set for crop growth; a sliding window is set, and the sliding step length is set according to the multi-level growth stage of the crop, and the average value of each monitoring device data updated by each element set in the crop growth element set in the window is calculated, and the average value of each monitoring device data in each element feature set is compared with the corresponding monitoring standard in the multi-level farmland monitoring standard, and the deviation value is calculated to obtain a multi-source deviation data set; based on the sliding step length, the multi-source deviation data in each window is arranged in the order of the sliding step time to form a time series multi-source deviation data set; the time series multi-source deviation data set is input into the crop growth prediction model to obtain the crop yield prediction value, wherein the prediction model is constructed with ARIMAX as the architecture. The technical problem of relying on a single data source or a simple empirical model in the prior art and ignoring many complex factors and their interactions in the crop growth process is solved, and the technical effect of improving the prediction accuracy is achieved.

[0006] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in the present disclosure or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0008] Figure 1 A schematic diagram of a process flow of a crop yield prediction method based on multi-data fusion provided in an embodiment of the present application; Figure 2A schematic diagram of the structure of a crop yield prediction system based on multi-data fusion provided in an embodiment of the present application.

[0009] Explanation of the accompanying drawings: multi-level farmland monitoring standard configuration module 11, multi-source monitoring equipment data acquisition module 12, multi-source monitoring equipment data feature extraction module 13, multi-source deviation data set acquisition module 14, time series multi-source deviation data set construction module 15, crop yield prediction value acquisition module 16. DETAILED DESCRIPTION

[0010] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0011] Embodiment 1 The crop yield prediction method based on multi-data fusion provided in the embodiment of the present disclosure is referred to Figure 1 For illustration, the methods include: S100: configuring a multi-level farmland monitoring standard for the designated multi-level growth stages of crops, and configuring the multi-level farmland monitoring standard in an intelligent control system, wherein the multi-level growth stages include a germination stage, a seedling stage, a growth stage, and a maturity stage.

[0012] Further, the multi-level farmland monitoring standard includes: configuring a first monitoring standard, wherein the first monitoring standard is soil quality based on the multi-level growth stages of crops; configuring a second monitoring standard, wherein the second monitoring standard is based on weather conditions at multiple growth stages of the crop; Configuring a third monitoring standard, wherein the third monitoring standard is based on pesticide residues at multiple growth stages of crops; The first monitoring standard, the second monitoring standard and the third monitoring standard are correlated with each other to determine a multi-level farmland monitoring standard.

[0013] Specifically, the multi-stage growth stages of plants are divided according to the comprehensive results of plant physiology, agricultural science and crop cultivation practice, including germination, seedling, growth and maturity. The multi-stage farmland monitoring standards for various crops at various stages are defined based on the historical data of long-term observation records in the region, the research results of agricultural scientists, the experience summary of agricultural technology extension stations and agricultural technicians, and international and domestic agricultural standards, including soil quality, weather conditions and pesticide residues. The soil quality is reflected by the soil pH value, and the weather conditions are reflected by temperature, humidity and light intensity. The specific values ​​of the multi-stage farmland monitoring standards at each growth stage are as follows.

[0014] The multi-level farmland monitoring standards during the germination period include: the appropriate soil pH value is between 6.0 and 7.0, which is conducive to the activity of enzymes in seeds and promotes seed germination; the appropriate temperature is 12-20°C, which helps seeds break dormancy and start the germination process; the appropriate humidity is 50%-60%, at this time, too little water makes it difficult for seeds to germinate, and too much water may cause lack of oxygen and seed rot; the light intensity of crops that require light can be 1000-1500 lux, because seeds mainly rely on their own stored nutrients in the early stages of germination, and light is not a key factor. Excessive light may also inhibit germination; pesticide residues should be as close to zero as possible. Since seeds and young sprouts in the germination period are very sensitive to pesticides, even trace amounts of pesticide residues may have a serious impact on their growth and development, so before sowing, ensure that there are no pesticide residues in the soil that may affect seed germination.

[0015] The multi-level farmland monitoring standards for the seedling stage include: the appropriate soil pH value is 6.0-7.5, which is conducive to the absorption of soil nutrients by the seedling roots; the appropriate temperature is 15-22°C, which is conducive to the growth of the seedling roots and the development of the aboveground parts; the humidity range is 55%-65%, which can meet the water needs of the seedling growth and ensure good soil aeration, which is conducive to root respiration; the light intensity range is 1500-3000 lux, and the seedling stage needs to appropriately enhance the light for photosynthesis; the pesticide residue standard is as low as possible and meets strict safety standards. The seedlings' tolerance to pesticides is still weak, and high concentrations of pesticide residues should be avoided to prevent adverse effects on the growth and development of seedlings.

[0016] The multi-level farmland monitoring standards during the growing period include: the suitable soil pH value is 6.5-7.5, which is conducive to the growth and nitrogen fixation of rhizobia, and is also conducive to the absorption of other nutrients by crops; the temperature range is 18-25°C, which is conducive to the metabolism of crops during the vigorous growth period, including nutrient absorption, photosynthesis and other physiological processes. The humidity range is 60%-70%. Sufficient soil moisture provides water guarantee for crop growth, but water accumulation should also be avoided to maintain good physical structure and aeration of the soil; the light intensity range is 2000-5000 lux. As crops grow, the demand for light intensity increases. At this time, it is necessary to meet the energy demand of photosynthesis and increase the accumulation of photosynthetic products; pesticide residues should comply with the safety production standards of agricultural products. There may be pest and disease control measures during the growing period. It is necessary to strictly control the amount of pesticide used and the safe interval to ensure that the pesticide residue is within the safe range at harvest.

[0017] The multi-level farmland monitoring standards during the maturity period include: soil pH value of 6.0-7.0 is directly conducive to the improvement of fruit quality; the temperature range is 20-28°C, too high or too low temperature may affect fruit quality or cause crops to mature early or late. The humidity range is 50%-60%, and appropriately reducing humidity during the maturity period will help fruit ripening and sugar accumulation, while preventing diseases caused by excessive humidity; the light intensity range is 2000-4000 lux, sufficient but not too strong light is conducive to photosynthesis and quality formation during fruit ripening; pesticide residues must strictly comply with food safety standards, because the crops in the mature period are about to be harvested, and excessive pesticide residues will directly affect the quality and safety of agricultural products.

[0018] Furthermore, data from multiple monitoring equipment for monitoring farmland environment is obtained regularly, including: Based on the content of the first monitoring standard, configuring a monitoring device that meets the first monitoring standard, the monitoring device comprising a pH sensor; Based on the content of the second monitoring standard, configuring a monitoring device that meets the second monitoring standard, the monitoring device comprising a temperature sensor, a humidity sensor, and a light sensor; Based on the content of the third monitoring standard, configuring monitoring equipment that meets the third monitoring standard, the monitoring equipment includes a pesticide residue detector; Based on the multi-stage growth stages of the crops, monitoring data of the pH sensor, temperature sensor, humidity sensor, light sensor and pesticide residue detector are regularly acquired.

[0019] Specifically, pH sensors are evenly distributed in the farmland. For large areas of farmland, a grid layout is adopted to ensure that the changes in the soil pH value of the entire farmland can be accurately reflected. At the same time, the installation depth of the sensor is determined according to the root distribution of the crop. For shallow-rooted crops, the installation depth is 10-20 cm, and for deep-rooted crops, the installation depth is 20-30 cm. The temperature sensor is installed near the crop canopy to accurately measure the temperature of the crop growth environment. The air humidity sensor is installed in a well-ventilated location that can represent the humidity of the crop growth environment, avoiding interference factors such as direct sunlight and proximity to water sources. The light sensor is installed above the crop canopy to avoid being blocked by crop leaves to accurately measure the light intensity reaching the crop surface. Pesticide residue detection requires regular collection of soil, crop leaves or fruit samples in the farmland and testing according to the operating specifications of the detector.

[0020] S200: Based on the multi-stage growth stages of the crops, regularly obtain data from a multi-source monitoring device for monitoring the farmland environment, wherein the multi-source monitoring device is communicatively connected to the intelligent control system.

[0021] Specifically, the frequency of data collection is determined according to the growth stage of the crops. During the germination and seedling stages, since crops are more sensitive to environmental changes, the collection frequency is set to once every two hours. During the growth and maturity stages, the collection frequency is set to 1-2 times a day. Since pesticide residue detection requires manual collection of test samples, and the detection frequency depends on the use of pesticides, the growth stage of crops, and the safety standards for pesticide use at each stage, the collection frequency is determined to be 14 days after the application of pesticides during the germination and seedling stages, and the detection frequency is increased to once every 7 days during the growth and maturity stages. After collecting various types of monitoring data, the data is transmitted to the intelligent control system by using wireless communication technology.

[0022] Furthermore, before obtaining the reduced-dimensional multi-source monitoring device data, it also includes: The multi-source monitoring equipment data for monitoring the farmland environment is regularly acquired, and the acquired multi-source monitoring equipment data is cleaned, wherein the data cleaning includes missing value processing, outlier detection and standardization processing.

[0023] Specifically, check whether each variable has missing values. For a small number of missing values, use interpolation methods to supplement them. If the missing ratio is large, use statistical methods to estimate the missing values ​​to supplement them. Detect and process outliers, use statistical methods such as Z-score and interquartile range to identify outliers, and correct or remove data. Standardize the format, unify the date and time format, and ensure that all time information is stored in the same format. Standardize numerical precision to ensure that all numerical data has the same representation method; S300: Use the PCA algorithm to perform feature extraction and dimensionality reduction processing on the acquired multi-source monitoring equipment data, obtain the reduced dimensionality multi-source monitoring equipment data, and update the crop growth element set, wherein the crop growth element set is constructed based on the content of the multi-level farmland monitoring standard, and the crop growth element set includes the soil quality element set, weather condition element set and pesticide residue element set for crop growth.

[0024] Specifically, extract and organize data obtained from multi-source monitoring equipment, including data collected from multiple pH sensors, temperature sensors, humidity sensors, light sensors, and pesticide residue detectors. Set the data matrix X, where each row represents a sample, such as monitoring data at different time points or different farmland locations, and each column represents a feature, such as pH value, temperature, humidity, etc. Calculate the covariance matrix Cov(X) of the data matrix X to understand the linear relationship between various features. The covariance matrix reflects the correlation between different features. If the covariance between two features is positive, it means they are positively correlated. If it is negative, it means negative correlation. If it is close to 0, it means no correlation. The formula for the covariance matrix is: Where n is the number of samples, Represents the transposed matrix of matrix X; By solving the equation Cov(X)v=λV, we get the eigenvalue λ and eigenvector V, where the eigenvalue represents the variance explained by each principal component, and the eigenvector determines the direction of the principal component. The eigenvalues ​​are arranged in order from large to small, and the corresponding eigenvectors are also sorted accordingly. According to the set retained variance ratio or the number of principal components, the first k principal components are selected. In general, the principal components whose cumulative variance contribution rate reaches a certain proportion can be selected. For example, if the cumulative variance contribution rate of the first three principal components reaches 85%, these three principal components are selected for dimensionality reduction. The selection of principal components can determine which sensor-acquired features best reflect the key environmental conditions at this stage.

[0025] Perform data dimensionality reduction on the principal component data. For each sample Xi in the original data matrix X, the dimension is reduced by projecting it onto the selected principal component. The reduced-dimensional data Yi=Xi*Vk, where Vk is a matrix consisting of the selected k eigenvectors.

[0026] S400: Set a sliding window, set the sliding step according to the multi-level growth stage of the crop, calculate the average value of the data of each monitoring device updated in each element set in the crop growth element set in the window, compare the average value of the data of each monitoring device in each element feature set with the corresponding monitoring standard in the multi-level farmland monitoring standard, calculate the deviation value, and obtain a multi-source deviation data set.

[0027] Specifically, the sliding window size is set according to the time span of the multi-stage growth stage and the frequency of data changes. If the growth stage lasts for a long time and the data changes relatively slowly, a larger sliding window is set. Conversely, if the growth stage changes rapidly, a smaller step size can be set to obtain a more detailed data change situation. Assuming that the germination period is 10 days, the sliding window is set to 3 days and the sliding step size is 1 day. In the longer stage such as the growth period, the sliding window is set to 5-7 days and the sliding step size is 2 days.

[0028] Obtain the data of each monitoring device in the element feature set and calculate the average value in the sliding window. Taking the soil quality element set as an example, if there are multiple data points of pH sensors in a sliding window, add these data points and divide them by the number of data points to get the average pH value in the sliding window. The same method is used to calculate the average value of monitoring equipment data such as temperature, humidity, light intensity, pesticide residue, etc. in their respective element sets.

[0029] Get the multi-level farmland monitoring standards in step S100 and calculate the deviation value. Taking soil pH as an example, if the average pH value calculated in a sliding window is 6.5, and the standard range for this growth stage is 6.0 - 7.0, then the deviation value is 0. If the calculated average value is 5.5, the deviation value is -0.5. In the same way, calculate the deviation values ​​of each element such as temperature, humidity, light intensity, pesticide residue and the corresponding standards, and combine these deviation values ​​to form a multi-source deviation data set. The multi-source deviation data set reflects the deviation of the crop growth environment from the ideal monitoring standard in different sliding windows, providing an important data basis for subsequent analysis.

[0030] S500: Based on the sliding step length, the multi-source deviation data in each window is arranged in order of the sliding step time to form a time series multi-source deviation data set.

[0031] Specifically, a multi-source deviation data set is obtained, and a starting window is determined. The starting time of the window is the starting time of the entire monitoring process. Starting from the first sliding window, the data is arranged in sequence according to the set sliding step. For example, if the sliding step is 1 day, the data of the first window is the multi-source deviation data of the 1st to 3rd day, then the data of the second window is arranged next to the data of the first window. For the multi-source deviation data in each sliding window, the order of the deviation values ​​corresponding to the internal elements, soil quality, weather conditions, pesticide residues and other elements remains unchanged, and the data of different windows are arranged in chronological order. If there are n sliding windows, after such an arrangement, a time-series multi-source deviation data set containing n groups of multi-source deviation data is formed. This data set reflects the changes of multi-source deviation data over time during the entire monitoring period.

[0032] S600: Inputting the time series multi-source deviation data set into a crop growth prediction model to obtain a crop yield prediction value, wherein the prediction model is constructed based on ARIMAX architecture.

[0033] Furthermore, the crop growth prediction model includes: Obtaining the historical annual yield of the crop, converting it into multi-stage growth stage data, and obtaining multi-stage growth stage sample yield data; Obtaining a sample soil quality element set, a sample weather condition element set, and a sample pesticide residue element set at multiple growth stages, wherein the sample soil quality element set includes pH sensor sample data, the sample weather condition element set includes temperature sensor sample data, humidity sensor sample data, and light sensor sample data, and the sample pesticide residue element set includes pesticide residue detector sample data; Set a sample sliding window and a sample sliding step size, obtain a sample multi-source deviation data set, and construct a sample time series multi-source deviation data set, wherein the sample multi-source deviation data set includes pH sensor sample deviation data, temperature sensor sample deviation data, humidity sensor sample deviation data, light sensor sample deviation data, and pesticide residue detector sample deviation data; Associating the sample time series multi-source deviation data with the sample yield data of the multi-stage growth stage, extracting the sample time series data in the sample time series multi-source deviation data set, and forming a sample time series data set; The sample time series data set is used as the target variable, and the sample time series multi-source deviation data set is used as the exogenous variable, and input into the ARIMAX model to determine the sample annual output as the training sample; Based on the training samples, supervised training is performed on the ARIMAX model to generate a crop growth prediction model; Based on the training samples, the crop growth prediction model is verified, and a model convergence test and retraining process are performed until the convergence conditions are met, thereby obtaining the constructed crop growth prediction model.

[0034] Furthermore, the crop yield prediction method based on multi-data fusion also includes: Regularly obtain data from multi-source monitoring equipment that monitors farmland environment, and obtain predicted crop yield data through crop growth prediction models; Determine to draw a coordinate axis curve, the horizontal axis is the time variable, the time variable depends on the multi-level growth stage division of the crop, and for the vertical axis, the acquired multi-source monitoring equipment data are respectively drawn in different sub-graphs, and the multi-level farmland monitoring standards and predicted yields corresponding to the data of each monitoring equipment are drawn in the corresponding sub-graphs, and double coordinate axis marking is adopted; Data visualization technology is used to display the drawn sub-graphs on the intelligent control system. Based on data analysis and prediction results, a decision support system is built to provide decision support for increasing crop yields.

[0035] Specifically, historical data from previous years are collected as samples to generate a sample time series data set. Through the ARIMAX model, the sample time series data set is used as the target variable to determine the sample year output and create a prediction model.

[0036] Specifically, a time series multi-source deviation data set is obtained and the data is fed into the prediction model. The obtained multi-source data are plotted in different sub-graphs. For example, a separate sub-graph is drawn for soil moisture data to intuitively show the change of soil moisture over time. The sub-graph is labeled with double coordinate axes, one axis is used to represent multi-source deviation data, such as the left axis represents the soil moisture deviation value, and the other axis is used to represent the yield data corresponding to the deviation value calculated by the prediction model. In this way, different types of data and their relationships can be clearly displayed in the same sub-graph, and the crop yield data in this state can be clearly determined.

[0037] Embodiment 2 Based on the same inventive concept as the crop yield prediction method based on multi-data fusion in the aforementioned embodiment, Figure 2 As shown, the present disclosure also provides a crop yield prediction system based on multi-data fusion, the system comprising: A multi-level farmland monitoring standard configuration module 11, the multi-level farmland monitoring standard configuration module 11 is used to configure multi-level farmland monitoring standards for the multi-level growth stages of the designated crops, and configure the multi-level farmland monitoring standards in the intelligent control system, wherein the multi-level growth stages include germination stage, seedling stage, growth stage and maturity stage; A multi-source monitoring device data acquisition module 12, the multi-source monitoring device data acquisition module 12 is used to regularly acquire multi-source monitoring device data for monitoring the farmland environment based on the multi-stage growth stages of the crops, wherein the multi-source monitoring device is communicatively connected with the intelligent control system; A multi-source monitoring equipment data feature extraction module 13, the multi-source monitoring equipment data feature extraction module 13 is used to use the PCA algorithm to perform feature extraction and dimensionality reduction processing on the acquired multi-source monitoring equipment data, obtain the reduced dimensionality multi-source monitoring equipment data, and update the crop growth element set, wherein the crop growth element set is constructed based on the content of the multi-level farmland monitoring standard, and the crop growth element set includes a soil quality element set, a weather condition element set, and a pesticide residue element set for crop growth; A multi-source deviation data set acquisition module 14, the multi-source deviation data set acquisition module 14 is used to set a sliding window, the sliding step length is set according to the multi-level growth stage of the crop, calculate the average value of each monitoring device data updated by each element set in the crop growth element set in the window, compare the average value of each monitoring device data in each element feature set with the corresponding monitoring standard in the multi-level farmland monitoring standard, calculate the deviation value, and obtain a multi-source deviation data set; A time series multi-source deviation data set construction module 15, wherein the time series multi-source deviation data set construction module 15 is used to arrange the multi-source deviation data in each window in the order of the sliding step time based on the sliding step length to form a time series multi-source deviation data set; The crop yield prediction value acquisition module 16 is used to input the time series multi-source deviation data set into the crop growth prediction model to obtain the crop yield prediction value, wherein the prediction model is constructed based on ARIMAX architecture.

[0038] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0039] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A crop yield prediction method based on multi-data fusion, characterized in that: The method comprises: According to the multi-stage growth stages of the crops, multi-stage farmland monitoring standards are configured, and the multi-stage farmland monitoring standards are configured in the intelligent control system, wherein the multi-stage growth stages include germination stage, seedling stage, growth stage and maturity stage; Based on the multi-stage growth stages of the crops, regularly acquiring data from a multi-source monitoring device for monitoring the farmland environment, wherein the multi-source monitoring device is communicatively connected with the intelligent control system; Using the PCA algorithm to perform feature extraction and dimensionality reduction processing on the acquired multi-source monitoring equipment data, obtain the reduced-dimensional multi-source monitoring equipment data, and update the crop growth element set, wherein the crop growth element set is constructed based on the content of the multi-level farmland monitoring standard, and the crop growth element set includes a soil quality element set, a weather condition element set, and a pesticide residue element set for crop growth; A sliding window is set, and a sliding step length is set according to the multi-level growth stage of the crop, and the average value of the data of each monitoring device updated by each element set in the crop growth element set in the window is calculated, and the average value of the data of each monitoring device in each element feature set is compared with the corresponding monitoring standard in the multi-level farmland monitoring standard, and the deviation value is calculated to obtain a multi-source deviation data set; Based on the sliding step length, the multi-source deviation data in each window is arranged in the order of the sliding step time to form a time series multi-source deviation data set; The time series multi-source deviation data set is input into a crop growth prediction model to obtain a crop yield prediction value, wherein the prediction model is constructed based on ARIMAX architecture.

2. The method for predicting crop yield based on multi-data fusion according to claim 1, characterized in that: Before obtaining the reduced-dimensional multi-source monitoring equipment data, including: The multi-source monitoring equipment data for monitoring the farmland environment is regularly acquired, and the acquired multi-source monitoring equipment data is cleaned, wherein the data cleaning includes missing value processing, outlier detection and standardization processing.

3. The crop yield prediction method based on multi-data fusion according to claim 1, characterized in that: The multi-level farmland monitoring standard includes: configuring a first monitoring standard, wherein the first monitoring standard is soil quality based on the multi-level growth stages of crops; configuring a second monitoring standard, wherein the second monitoring standard is based on weather conditions at multiple growth stages of the crop; Configuring a third monitoring standard, wherein the third monitoring standard is based on pesticide residues at multiple growth stages of crops; The first monitoring standard, the second monitoring standard and the third monitoring standard are correlated with each other to determine a multi-level farmland monitoring standard.

4. The crop yield prediction method based on multi-data fusion as claimed in claim 3, characterized in that: Regularly obtain data from multiple sources of monitoring equipment to monitor farmland environment, including: Based on the content of the first monitoring standard, configuring a monitoring device that meets the first monitoring standard, the monitoring device comprising a pH sensor; Based on the content of the second monitoring standard, configuring a monitoring device that meets the second monitoring standard, the monitoring device comprising a temperature sensor, a humidity sensor, and a light sensor; Based on the content of the third monitoring standard, configuring monitoring equipment that meets the third monitoring standard, the monitoring equipment includes a pesticide residue detector; Based on the multi-stage growth stages of the crops, monitoring data of the pH sensor, temperature sensor, humidity sensor, light sensor and pesticide residue detector are regularly acquired.

5. The method for predicting crop yield based on multi-data fusion according to claim 4, characterized in that: Crop growth prediction models, including: Obtaining the historical annual yield of the crop, converting it into multi-stage growth stage data, and obtaining multi-stage growth stage sample yield data; Obtaining a sample soil quality element set, a sample weather condition element set, and a sample pesticide residue element set at multiple growth stages, wherein the sample soil quality element set includes pH sensor sample data, the sample weather condition element set includes temperature sensor sample data, humidity sensor sample data, and light sensor sample data, and the sample pesticide residue element set includes pesticide residue detector sample data; Set a sample sliding window and a sample sliding step size, obtain a sample multi-source deviation data set, and construct a sample time series multi-source deviation data set, wherein the sample multi-source deviation data set includes pH sensor sample deviation data, temperature sensor sample deviation data, humidity sensor sample deviation data, light sensor sample deviation data, and pesticide residue detector sample deviation data; Associating the sample time series multi-source deviation data with the sample yield data of the multi-stage growth stage, extracting the sample time series data in the sample time series multi-source deviation data set, and forming a sample time series data set; The sample time series data set is used as the target variable, and the sample time series multi-source deviation data set is used as the exogenous variable, and input into the ARIMAX model to determine the sample annual output as the training sample; Based on the training samples, supervised training is performed on the ARIMAX model to generate a crop growth prediction model; Based on the training samples, the crop growth prediction model is verified, and a model convergence test and retraining process are performed until the convergence conditions are met, thereby obtaining the constructed crop growth prediction model.

6. The method for predicting crop yield based on multi-data fusion according to claim 1, characterized in that: Also includes: Regularly obtain data from multi-source monitoring equipment that monitors farmland environment, and obtain predicted crop yield data through crop growth prediction models; Determine to draw a coordinate axis curve, the horizontal axis is the time variable, the time variable depends on the multi-level growth stage division of the crop, and for the vertical axis, the acquired multi-source monitoring equipment data are respectively drawn in different sub-graphs, and the multi-level farmland monitoring standards and predicted yields corresponding to the data of each monitoring equipment are drawn in the corresponding sub-graphs, and double coordinate axis marking is adopted; Data visualization technology is used to display the drawn sub-graphs on the intelligent control system. Based on data analysis and prediction results, a decision support system is built to provide decision support for increasing crop yields.

7. The crop yield prediction system based on multi-data fusion is characterized by: The system comprises: A multi-level farmland monitoring standard configuration module, which is used to configure multi-level farmland monitoring standards for the defined multi-level growth stages of crops, and configure the multi-level farmland monitoring standards in the intelligent control system, wherein the multi-level growth stages include germination stage, seedling stage, growth stage and maturity stage; A multi-source monitoring device data acquisition module, the multi-source monitoring device data acquisition module is used to regularly acquire multi-source monitoring device data for monitoring the farmland environment based on the multi-stage growth stages of the crops, wherein the multi-source monitoring device is communicatively connected with the intelligent control system; A multi-source monitoring equipment data feature extraction module, the multi-source monitoring equipment data feature extraction module is used to use the PCA algorithm to perform feature extraction and dimensionality reduction processing on the acquired multi-source monitoring equipment data, obtain the reduced dimensionality multi-source monitoring equipment data, and update the crop growth element set, wherein the crop growth element set is constructed based on the content of the multi-level farmland monitoring standard, and the crop growth element set includes a soil quality element set, a weather condition element set, and a pesticide residue element set for crop growth; A multi-source deviation data set acquisition module, the multi-source deviation data set acquisition module is used to set a sliding window, the sliding step length is set according to the multi-level growth stage of the crop, calculate the average value of each monitoring device data updated by each element set in the crop growth element set in the window, compare the average value of each monitoring device data in each element feature set with the corresponding monitoring standard in the multi-level farmland monitoring standard, calculate the deviation value, and obtain a multi-source deviation data set; A time series multi-source deviation data set construction module, wherein the time series multi-source deviation data set construction module is used to arrange the multi-source deviation data in each window in the order of the sliding step time based on the sliding step length to form a time series multi-source deviation data set; The crop yield prediction value acquisition module is used to input the time series multi-source deviation data set into the crop growth prediction model to obtain the crop yield prediction value, wherein the prediction model is constructed based on ARIMAX architecture.

Images