Multi-source data fusion strong gluten wheat quality intelligent prediction and planting optimization system

Abstract

The invention relates to the technical field of wheat grain quality intelligent prediction and planting optimization systems, in particular to a multi-source data fused strong gluten wheat quality intelligent prediction and planting optimization system, which comprises a multi-source data acquisition module, a data fusion and storage module, a quality prediction module, a planting optimization module and a visualization and decision support module, the multi-source data acquisition module is used for acquiring data of a soil pH value, an EC value, water content, soil nutrients, soil temperature and soil conductivity by deploying a LoRaWAN soil three-parameter sensor; the data fusion and storage module is used for establishing a big data platform and storing and fusing historical data and real-time monitoring data; the quality prediction module is used for constructing a strong gluten wheat quality prediction model based on a deep learning algorithm; the planting optimization module is used for utilizing the prediction result; the visualization and decision support module is used for providing a user-friendly interface; the accuracy of strong gluten wheat quality prediction is remarkably improved, and dynamic optimization suggestions are provided for planting managers.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent prediction and planting optimization systems for wheat grain quality, and particularly to an intelligent prediction and planting optimization system for strong gluten wheat quality with multi-source data fusion. Background Art

[0002] Due to its high gluten content and strong gluten strength, strong gluten wheat is widely used in food processing such as bread, ramen, and dumplings. However, the quality and yield of wheat are often affected by various factors. Therefore, realizing the intelligent prediction of wheat yield and quality is of great significance for guiding wheat planting and improving yield and quality.

[0003] Currently, in existing planting prediction systems, such as the patent with the publication number CN119398269A, the intelligent prediction method for rice phenological period based on environmental data and deep learning belongs to the field of smart agriculture technology. To achieve accurate prediction of the rice phenological period, the present invention includes obtaining environmental data and phenological period data throughout the year of rice growth, and constructing a phenological model dataset; performing data preprocessing based on the phenological model dataset for subsequent model training and prediction.

[0004] However, it is found in the use of this method that the multi-source data fusion effect of the environment is poor, which reduces the accuracy of prediction, and this method is not convenient for comparing the actual results with the predicted results, reducing the inspection and optimization of the method. Summary of the Invention

[0005] To solve the above technical problems, the present invention provides an intelligent prediction and planting optimization system for strong gluten wheat quality with multi-source data fusion, which can significantly improve the accuracy of strong gluten wheat quality prediction, provide dynamic optimization suggestions for planting managers, adapt to different geographical and climatic conditions, has wide popularization value, improves the quality of strong gluten wheat while reducing resource waste and increasing economic benefits.

[0006] The intelligent prediction and planting optimization system for strong gluten wheat quality with multi-source data fusion of the present invention includes a multi-source data acquisition module, a data fusion and storage module, a quality prediction module, a planting optimization module, and a visualization and decision support module;

[0007] Multi-source data acquisition module: By deploying LoRaWAN soil three-parameter sensors, data such as soil pH value, EC value, moisture content, soil nutrients, soil temperature, and soil conductivity are collected. The temperature, humidity, light, and CO2 data of meteorological conditions are collected through a micro meteorological station. Aerial photography is carried out by a drone equipped with a multi-spectral camera, and at the same time, Sentinel-2 satellite data is coordinated to obtain data of vegetation indices, so as to realize the collection of multi-source data of various soil parameters, meteorological conditions, and vegetation indices;

[0008] Data Fusion and Storage Module: Establish a big data platform to store and fuse historical data and real-time monitoring data, including soil data, meteorological data, planting management records, and quality determination data;

[0009] Quality Prediction Module: Based on deep learning algorithms, build a quality prediction model for strong gluten wheat kernels to predict core quality indicators such as protein content and kernel hardness of wheat kernels;

[0010] Planting Optimization Module: Utilize the prediction results and combine with optimization algorithms to dynamically adjust planting management plans such as fertilization, irrigation, and pest control;

[0011] Visualization and Decision Support Module: Provide a user-friendly interface to display prediction results and optimization suggestions in real-time through the PC or mobile terminal;

[0012] 1. Improve the accuracy of quality prediction: Significantly improve the accuracy of strong gluten wheat kernel quality prediction through multi-source data fusion and deep learning technologies;

[0013] 2. Real-time dynamic optimization: The system realizes real-time updates of quality prediction results and provides dynamic optimization suggestions for planting managers;

[0014] 3. Adapt to multi-regional applications: The system adapts to different geographical and climatic conditions through regionalized model optimization strategies and has broad promotion value;

[0015] 4. Linkage between planting and quality: Link quality prediction with planting optimization to provide users with a scientific management plan that takes both quality and yield into account;

[0016] 5. Improve agricultural benefits: The system improves the quality of strong gluten wheat while reducing resource waste and increasing economic benefits through accurate prediction and optimized management.

[0017] Preferably, it also includes an inspection module, a comparison module, and a display module;

[0018] Inspection Module: Use on-site sampling equipment to collect wheat samples at multiple points and detect the protein content and kernel hardness of the wheat samples;

[0019] Comparison Module: Compare the actual detection data of the wheat samples with the predicted data to test the prediction results, and then send the test results to the quality prediction module to continuously optimize the model according to the feedback data in actual applications; By aligning the actual detection data and the predicted data of the wheat samples in time, ensure that the actual detection data and the predicted data are consistent in the time dimension, and at the same time align the data space to ensure that the data comes from the same plot to ensure the accuracy of the comparison, and use statistical methods to quantify the difference between the actual data and the predicted data;

[0020] Display module: Intuitively display the comparison results between actual data and predicted data in the form of a line chart, and mark the significant difference points in the chart for further analysis.

[0021] Preferably, the quality prediction module includes a data collection and preprocessing module, a model selection and construction module, a model training and optimization module, and a model deployment and application module;

[0022] Data collection and preprocessing module: Used to collect sample data of core quality indicators such as protein content and grain hardness of strong gluten wheat, as well as environmental factors that may affect these indicators. Then, process the collected data for missing values and outliers to ensure data quality, scale the data to the same scale to accelerate model convergence, and then extract the morphological characteristics of wheat grains through image processing technology;

[0023] Model selection and construction module: Use a convolutional neural network to extract the morphological characteristics of wheat grains and map the extracted features to the output. Use a recurrent neural network to capture the time series data of environmental changes during the wheat growth process and map the time series features to the output;

[0024] Model training and optimization module: Select an appropriate loss function to measure the difference between the predicted value and the true value. Adjust hyperparameters such as the learning rate and batch size through grid search or random search methods. Use cross-validation to evaluate the generalization ability of the model, and use mean square error and mean absolute error metrics to evaluate the model performance;

[0025] Model deployment and application module: Save the trained model as a file, deploy the model to actual applications, predict the quality indicators of wheat grains in real time, and continuously optimize the model according to the feedback data in actual applications.

[0026] Preferably, the planting optimization module uses the wheat quality indicators predicted by the quality prediction module and environmental data to evaluate the effect of current planting management. According to the feedback data, continuously improve the planting management plan, specifically including the following steps:

[0027] S1. Combine the prediction results with real-time monitoring data to form complete planting status information;

[0028] S2. Identify the problems existing in current planting management by analyzing the prediction results;

[0029] S3. Dynamically adjust the planting management plan using optimization linear programming, genetic algorithms, reinforcement learning, and Bayesian optimization algorithms; adjust the application rates of fertilizers such as nitrogen, phosphorus, and potassium according to soil nutrient levels and prediction results, adjust the irrigation amount and irrigation time according to soil moisture and meteorological data, and adjust the pesticide application amount and prevention and control strategies according to the pest and disease prediction results;

[0030] S4. Deploy the optimized planting management plan to actual planting, monitor the planting effect in real time, collect new data, and re-run the optimization algorithm based on the feedback data to continuously improve the planting management plan; the planting optimization module can dynamically adjust planting management plans such as fertilization, irrigation, and pest and disease control by combining prediction results and optimization algorithms, thereby increasing the yield and quality of wheat. The core of this module lies in data integration, the selection of optimization algorithms, as well as real-time feedback and iterative optimization. Through intelligent planting management, efficient utilization of resources and sustainable development of agricultural production can be achieved.

[0031] Preferably, the data fusion and storage module includes a data storage layer, a data processing layer, a data application layer, and a data fusion layer;

[0032] The data storage layer is used to store long-term accumulated data on soil, meteorology, planting management, and quality determination, and at the same time store real-time data generated by sensors and monitoring devices;

[0033] The data processing layer is used to handle missing values, outliers, and duplicate data, and integrate data from different sources into a unified format and standard;

[0034] The data application layer displays data through visualization tools and provides data access services through APIs;

[0035] The data fusion layer is used to integrate multi-source data into a consistent and usable data set, align data from different sources by timestamp, align data by geographical location, associate data through predefined rules, discover potential relationships between data through machine learning models, and convert data into a unified format; by establishing a big data platform, historical data and real-time monitoring data can be effectively stored and fused, providing data support for planting optimization and quality prediction. The core of the platform lies in the full-process management of data collection, storage, processing, and application, and at the same time, advanced technologies and tools need to be combined to achieve efficient data fusion and analysis.

[0036] Preferably, the on-site sampling device includes a support device, a processing device, a cylinder body, a closing door, a connecting plate, a first rotating shaft, multiple groups of L-shaped plates, blades and a first motor. The cylinder body is installed on the support device, and the support device is used to drive the cylinder body to move. Openings are provided at the bottom and side of the cylinder body respectively. The closing door is slidably arranged at the side opening of the cylinder body. The connecting plate is installed on the outer side wall of the closing door. The first rotating shaft is rotatably installed on the inner side wall of the cylinder body. Multiple groups of L-shaped plates are circumferentially installed on the outer side wall of the first rotating shaft. The blades are installed at the bottom opening of the cylinder body. The first motor is installed on the support device, and the output end of the first motor is connected to the first rotating shaft. The processing device is arranged below the cylinder body and is used to dehull the wheat. The support device drives the cylinder body to move downward, so that the upper part of the planted wheat extends into the cylinder body through the bottom opening of the cylinder body. Then, the first motor drives the first rotating shaft to rotate, so that the first rotating shaft drives multiple groups of L-shaped plates to move circumferentially, so that one of the L-shaped plates cooperates with the blades to cut and store the wheat ears into the cylinder body. Subsequently, the cylinder body is moved to other planting positions. By repeating the above operations, the wheat planted in different areas is respectively stored inside the cylinder body between multiple groups of L-shaped plates, improving the convenience of wheat sampling in different areas. Then, by opening the closing door, it is convenient to put the wheat into the processing device through the side opening of the cylinder body for dehulling, improving the convenience of later wheat detection.

[0037] Preferably, the processing device includes a vibration device, a box body, multiple feed hoppers, multiple screen plates, a crushing rod, multiple vibrating plates, a fan, a blowing air pipe, multiple discharge channels, and multiple ejector rods. A plurality of chambers are provided inside the box body. The multiple feed hoppers are respectively communicated and arranged at the tops of the multiple chambers. The multiple screen plates are respectively installed in the multiple chambers. The tops of the multiple screen plates are respectively communicated with the multiple feed hoppers. The crushing rod rotatably passes through the multiple screen plates. The multiple vibrating plates are installed in the multiple chambers through the vibration device, and the vibration device is used to drive the crushing rod to rotate. The fan is installed on the outer side wall of the box body. The power input end of the fan is connected to the crushing rod. The output end of the fan is communicated with the blowing air pipe. The output end of the blowing air pipe communicates with the inside of the multiple chambers. The multiple discharge channels are respectively communicated and arranged outside the multiple chambers. The multiple ejector rods are all installed at the top of the box body. Move the cylinder above the feed hopper. When the cylinder moves downward, support the connecting plate through the ejector rod, so that the closing door rotates upward and opens during the downward movement of the cylinder, so as to facilitate the wheat between the multiple L-shaped plates to be respectively put into the multiple feed hoppers. The multiple feed hoppers convey the wheat into the multiple screen plates. Drive the crushing rod to rotate through the vibration device, so that the crushing rod hits the wheat in the screen plates, so that the wheat is peeled. The peeled wheat and wheat bran fall to the top of the vibrating plate. Drive the multiple vibrating plates to vibrate up and down through the vibration device, so as to improve the separation effect of wheat grains and wheat bran. While the crushing rod rotates, drive the fan, and the fan blows air into the multiple chambers through the blowing air pipe, so that the air flows to blow the wheat bran, and the blown wheat bran is discharged outward through the multiple discharge channels, improving the convenience of automatic wheat peeling treatment.

[0038] Preferably, the vibration device includes a second motor, a second rotating shaft, multiple eccentric wheels, multiple tension springs, two belt pulleys, and a belt. The second motor is installed on the outer side wall of the box body. The second rotating shaft is rotatably installed on the inner side wall of the box body. The multiple eccentric wheels are all installed on the outer side wall of the second rotating shaft. The outer side walls of the multiple eccentric wheels are respectively in contact with the bottoms of the multiple vibrating plates. The multiple tension springs are all installed on the inner side wall of the box body. The tops of the multiple tension springs are respectively connected to the bottoms of the multiple vibrating plates. The first set of belt pulleys is installed at the end of the second rotating shaft. The second set of belt pulleys is installed at the end of the crushing rod. The belt is sleeved on the outer side walls of the two belt pulleys. The second motor drives the second rotating shaft to rotate, so that the second rotating shaft drives the multiple eccentric wheels to rotate, so that the multiple eccentric wheels cooperate with the multiple tension springs to drive the multiple vibrating plates to vibrate up and down. While the second rotating shaft rotates, drive the crushing rod to rotate through the two belt pulleys and the belt, so that the crushing rod hits the wheat, and at the same time improves the convenience of wind separation of wheat grains and wheat bran.

[0039] Preferably, the support device includes a guiding frame, a hydraulic telescopic arm, a bracket, a guide rail, and an electric rotating table. The guiding frame is installed on the rotating end of the electric rotating table. The hydraulic telescopic arm is slidably installed up and down on the guiding frame. The bracket is installed on the moving end of the hydraulic telescopic arm. The first motor and the cylinder are both installed on the outer side wall of the bracket. The electric rotating table is slidably installed horizontally on the guide rail. The electric rotating table drives the guiding frame to rotate, so that the guiding frame drives the cylinder to adjust its orientation. The moving end of the hydraulic telescopic arm drives the bracket to move horizontally, and at the same time, the hydraulic telescopic arm moves downward, thereby improving the convenience of moving the cylinder to sample wheat. By horizontally moving the electric rotating table, the convenience of the cylinder to separately put the wheat collected from different areas into multiple groups of feed hoppers is improved.

[0040] Preferably, it further includes a vehicle body. The guide rail and the box body are respectively installed on the top of the vehicle body. By moving the position of the vehicle body, the convenience of moving the sampling device to different positions for sampling is improved.

[0041] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0042] 1. Improve the accuracy of quality prediction: Through multi-source data fusion and deep learning technology, significantly improve the accuracy of strong gluten wheat quality prediction; Real-time dynamic optimization: The system realizes real-time update of the quality prediction results and provides dynamic optimization suggestions for planting managers; Adapt to multi-region applications: The system adapts to different geographical and climatic conditions through regional model optimization strategies and has broad promotion value; Linkage between planting and quality: Link quality prediction with planting optimization to provide users with a scientific management plan that takes both quality and yield into account; Improve agricultural benefits: The system improves the quality of strong gluten wheat while reducing resource waste and increasing economic benefits through accurate prediction and optimized management;

[0043] 2. Facilitate wheat sampling in different planting areas;

[0044] 3. Facilitate automatic wheat peeling treatment in different planting areas and improve the convenience of later wheat detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a schematic diagram of the system structure of the present invention;

[0046] Figure 2 is a schematic diagram of the system such as the data fusion and storage module and the quality prediction module;

[0047] Figure 3 is a schematic diagram of the system of the planting optimization module;

[0048] Figure 4 is an axonometric structure schematic diagram of the connection between the feed hopper and the box body, etc.;

[0049] Figure 5It is an axonometric partial structure schematic diagram of the connection between the closed door and the connecting plate, etc.;

[0050] Figure 6 It is an axonometric partial structure schematic diagram of the connection between the first rotating shaft and the L-shaped plate, etc.;

[0051] Figure 7 It is an axonometric structure schematic diagram of the connection between the box body and the second motor, etc.;

[0052] Figure 8 It is an axonometric partial structure schematic diagram of the connection between the box body and the screen plate, etc.;

[0053] Figure 9 It is an axonometric structure schematic diagram of the connection between the cylinder body and the support, etc.;

[0054] Figure 10 It is an axonometric partial structure schematic diagram of the connection between the guide frame and the electric rotating table, etc.

[0055] Reference signs in the drawings: 101, cylinder body; 102, closed door; 103, connecting plate; 104, first rotating shaft; 105, L-shaped plate; 106, blade; 107, first motor; 201, box body; 202, feed hopper; 203, screen plate; 204, breaking rod; 205, vibrating plate; 206, blower; 207, air blowing pipe; 208, discharge channel; 209, ejector rod; 301, second motor; 302, second rotating shaft; 303, eccentric wheel; 304, tension spring; 305, pulley; 306, belt; 401, guide frame; 402, hydraulic telescopic arm; 403, support; 404, guide rail; 405, electric rotating table; 501, vehicle body. Detailed implementation manners

[0056] To facilitate the understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. The present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided to make the disclosure of the present invention more thorough and comprehensive.

[0057] Embodiment 1

[0058] The multi-source data fusion-based intelligent prediction and planting optimization system for strong gluten wheat quality of the present invention includes a multi-source data acquisition module, a data fusion and storage module, a quality prediction module, a planting optimization module, and a visualization and decision support module;

[0059] Multi-source data acquisition module: By deploying LoRaWAN soil three-parameter sensors, it collects data on soil pH value, EC value, moisture content, soil nutrients, soil temperature, and soil conductivity. It collects temperature, humidity, light, and CO2 data of meteorological conditions through a micro-meteorological station, conducts aerial photography with a multi-spectral camera carried by a drone, and simultaneously combines Sentinel-2 satellite data to obtain data on vegetation indices, thus realizing the acquisition of multi-source data on various soil parameters, meteorological conditions, and vegetation indices;

[0060] Data fusion and storage module: It establishes a big data platform to store and fuse historical data and real-time monitoring data, including soil data, meteorological data, planting management records, and quality determination data;

[0061] Quality prediction module: Based on deep learning algorithms, it constructs a prediction model for the quality of strong gluten wheat grains to predict core quality indicators such as protein content and grain hardness of wheat grains;

[0062] Planting optimization module: Utilizing the prediction results and combining optimization algorithms, it dynamically adjusts various planting management schemes such as fertilization, irrigation, and pest control;

[0063] Visualization and decision support module: It provides a user-friendly interface to display prediction results and optimization suggestions in real-time through a PC or mobile device;

[0064] In this embodiment, multi-source data fusion: Integrates soil, meteorological, remote sensing, and planting management data to comprehensively improve the accuracy and timeliness of the prediction model;

[0065] Dynamic prediction and optimization: Dynamically adjusts the parameters of the prediction model based on real-time data to achieve the immediacy of planting management suggestions;

[0066] Strong regional adaptability: Optimizes according to the environmental characteristics of different regions, making the system applicable to various climate and soil conditions;

[0067] Intelligent decision support: The system not only provides prediction results but also generates specific planting optimization suggestions to help users achieve efficient management.

[0068] Embodiment 2

[0069] Based on Embodiment 1, the multi-source data fusion intelligent prediction and planting optimization system for strong gluten wheat quality of the present invention further includes an inspection module, a comparison module, and a display module;

[0070] Inspection module: It uses wheat samples collected at multiple points by on-site sampling equipment and detects the protein content and grain hardness of the wheat samples;

[0071] Comparison module: Compare the actual test data of wheat samples with the predicted data to test the prediction results, and then send the test results to the quality prediction module to continuously optimize the model according to the feedback data in actual applications; By aligning the actual test data and the predicted data of wheat samples in time, ensure that the actual test data and the predicted data are consistent in the time dimension, and at the same time align the data space to ensure that the data comes from the same plot to ensure the accuracy of the comparison, and use statistical methods to quantify the difference between the actual data and the predicted data;

[0072] Display module: Intuitively display the comparison results of the actual data and the predicted data in the form of a line chart, and mark the significant difference points in the chart for further analysis;

[0073] The quality prediction module includes a data collection and preprocessing module, a model selection and construction module, a model training and optimization module, and a model deployment and application module;

[0074] Data collection and preprocessing module: Used to collect sample data of core quality indicators such as protein content and kernel hardness of strong gluten wheat, as well as environmental factors that may affect these indicators. Then, process the collected data for missing values and outliers to ensure data quality, scale the data to the same scale to accelerate model convergence, and then extract the morphological characteristics of wheat kernels through image processing technology;

[0075] Model selection and construction module: Use a convolutional neural network to extract the morphological characteristics of wheat kernels and map the extracted features to the output. Use a recurrent neural network to capture the time series data of environmental changes during the wheat growth process and map the time series features to the output;

[0076] Model training and optimization module: Select an appropriate loss function to measure the difference between the predicted value and the true value, adjust hyperparameters such as the learning rate and batch size through grid search or random search methods, use cross-validation to evaluate the generalization ability of the model, and use mean square error and mean absolute error indicators to evaluate the model performance;

[0077] Model deployment and application module: Save the trained model as a file, deploy the model to actual applications, predict the quality indicators of wheat kernels in real time, and continuously optimize the model according to the feedback data in actual applications;

[0078] The planting optimization module uses the wheat quality indicators predicted by the quality prediction module and environmental data to evaluate the effect of current planting management, and continuously improves the planting management plan according to the feedback data, specifically including the following steps:

[0079] S1. Combine the prediction results with the real-time monitoring data to form complete planting status information;

[0080] S2. Identify the problems existing in the current planting management by analyzing the prediction results;

[0081] S3. Dynamically adjust the planting management plan by using linear programming, genetic algorithm, reinforcement learning, and Bayesian optimization algorithm; adjust the application rates of fertilizers such as nitrogen, phosphorus, and potassium according to the soil nutrient level and prediction results, adjust the irrigation amount and irrigation time according to the soil humidity and meteorological data, and adjust the pesticide application rate and prevention and control strategy according to the pest and disease prediction results;

[0082] S4. Deploy the optimized planting management plan to actual planting, monitor the planting effect in real time, collect new data, and re-run the optimization algorithm according to the feedback data to continuously improve the planting management plan;

[0083] The data fusion and storage module includes a data storage layer, a data processing layer, a data application layer, and a data fusion layer;

[0084] The data storage layer is used to store the long-term accumulated data of soil, meteorology, planting management, and quality determination, and at the same time store the real-time data generated by sensors and monitoring devices;

[0085] The data processing layer is used to process missing values, outliers, and duplicate data, and integrate the data from different sources into a unified format and standard;

[0086] The data application layer displays the data through visualization tools and provides data access services through APIs;

[0087] The data fusion layer is used to integrate multi-source data into a consistent and available data set, align the data from different sources according to the timestamp, align the data according to the geographical location, associate the data through predefined rules, discover the potential relationships between the data through machine learning models, and convert the data into a unified format;

[0088] In this embodiment, the planting optimization module can dynamically adjust the planting management plans such as fertilization, irrigation, and pest and disease prevention and control by combining the prediction results and optimization algorithms, thereby improving the yield and quality of wheat. The core of this module lies in the integration of data, the selection of optimization algorithms, and real-time feedback and iterative optimization. Through intelligent planting management, the efficient utilization of resources and the sustainable development of agricultural production can be achieved. By establishing a big data platform, historical data and real-time monitoring data can be effectively stored and fused, providing data support for planting optimization and quality prediction. The core of the platform lies in the full-process management of data collection, storage, processing, and application, and at the same time, advanced technologies and tools need to be combined to achieve efficient data fusion and analysis.

[0089] Embodiment 3

[0090] Based on Embodiment 1, for the intelligent prediction and planting optimization system of strong gluten wheat quality with multi-source data fusion of the present invention, the on-site sampling device includes a support device, a processing device, a cylinder body 101, a closing door 102, a connecting plate 103, a first rotating shaft 104, multiple groups of L-shaped plates 105, blades 106 and a first motor 107. The cylinder body 101 is installed on the support device, and the support device is used to drive the cylinder body 101 to move its position. Openings are respectively provided at the bottom and side of the cylinder body 101. The closing door 102 is slidably arranged at the side opening of the cylinder body 101. The connecting plate 103 is installed on the outer side wall of the closing door 102. The first rotating shaft 104 is rotatably installed on the inner side wall of the cylinder body 101. Multiple groups of L-shaped plates 105 are circumferentially installed on the outer side wall of the first rotating shaft 104. The blades 106 are installed at the bottom opening of the cylinder body 101. The first motor 107 is installed on the support device, and the output end of the first motor 107 is connected to the first rotating shaft 104. The processing device is arranged below the cylinder body 101 and is used for peeling the wheat;

[0091] The processing device includes a vibration device, a box body 201, multiple groups of feed hoppers 202, multiple groups of sieve plates 203, crushing rods 204, multiple groups of vibrating plates 205, a blower 206, a blowing air pipe 207, multiple groups of discharge channels 208 and multiple groups of ejector rods 209. Multiple chambers are arranged inside the box body 201. Multiple groups of feed hoppers 202 are respectively communicated and arranged at the tops of the multiple chambers. Multiple groups of sieve plates 203 are respectively installed inside the multiple chambers. The tops of the multiple groups of sieve plates 203 are respectively communicated with the multiple groups of feed hoppers 202. The crushing rods 204 rotatably pass through the multiple groups of sieve plates 203. The multiple groups of vibrating plates 205 are installed inside the multiple chambers through the vibration device, and the vibration device is used to drive the crushing rods 204 to rotate. The blower 206 is installed on the outer side wall of the box body 201. The power input end of the blower 206 is connected to the crushing rods 204. The output end of the blower 206 is communicated with the blowing air pipe 207, and the output end of the blowing air pipe 207 communicates with the inside of the multiple chambers. The multiple groups of discharge channels 208 are respectively communicated and arranged outside the multiple chambers. The multiple groups of ejector rods 209 are all installed at the top of the box body 201;

[0092] The vibration device includes a second motor 301, a second rotating shaft 302, multiple groups of eccentric wheels 303, multiple groups of tension springs 304, two groups of belt pulleys 305 and a belt 306. The second motor 301 is installed on the outer side wall of the box body 201. The second rotating shaft 302 is rotatably installed on the inner side wall of the box body 201. The multiple groups of eccentric wheels 303 are all installed on the outer side wall of the second rotating shaft 302. The outer side walls of the multiple groups of eccentric wheels 303 are respectively in contact with the bottoms of the multiple groups of vibrating plates 205. The multiple groups of tension springs 304 are all installed on the inner side wall of the box body 201. The tops of the multiple groups of tension springs 304 are respectively connected to the bottoms of the multiple groups of vibrating plates 205. The first group of belt pulleys 305 is installed at the end of the second rotating shaft 302. The second group of belt pulleys 305 is installed at the end of the crushing rods 204. The belt 306 is sleeved on the outer side walls of the two groups of belt pulleys 305;

[0093] The support device includes a guide frame 401, a hydraulic telescopic arm 402, a bracket 403, a guide rail 404, and an electric rotating table 405. The guide frame 401 is installed on the rotating end of the electric rotating table 405. The hydraulic telescopic arm 402 is slidably installed up and down on the guide frame 401. The bracket 403 is installed on the moving end of the hydraulic telescopic arm 402. The first motor 107 and the cylinder 101 are both installed on the outer side wall of the bracket 403. The electric rotating table 405 is slidably installed horizontally on the guide rail 404;

[0094] It further includes a vehicle body 501, and the guide rail 404 and the box body 201 are respectively installed on the top end of the vehicle body 501;

[0095] In this embodiment, the support device drives the cylinder body 101 to move downward, so that the upper part of the planted wheat extends into the cylinder body 101 through the bottom opening of the cylinder body 101. Then, the first motor 107 drives the first rotating shaft 104 to rotate, so that the first rotating shaft 104 drives multiple groups of L-shaped plates 105 to move circumferentially, so that one group of L-shaped plates 105 cooperates with the blade 106 to cut and store the wheat ears into the cylinder body 101. Subsequently, the cylinder body 101 is moved to other planting positions, and by repeating the above operations, the wheat planted in different areas is respectively stored inside the cylinder body 101 between multiple groups of L-shaped plates 105, improving the convenience of wheat sampling in different areas. Then, by opening the closing door 102, it is convenient to put the wheat into the processing device through the side opening of the cylinder body 101 for peeling, improving the convenience of later wheat detection; the cylinder body 101 is moved above the feed hopper 202. When the cylinder body 101 moves downward, the connecting plate 103 is supported by the ejector rod 209, so that the closing door 102 rotates upward to open during the downward movement of the cylinder body 101, so that the wheat between multiple groups of L-shaped plates 105 is respectively put into multiple groups of feed hoppers 202. Multiple groups of feed hoppers 202 convey the wheat into multiple groups of sieve plates 203. The vibration device drives the crushing rod 204 to rotate, so that the crushing rod 204 hits the wheat in the sieve plate 203, so that the wheat is peeled. The peeled wheat and wheat bran fall to the top of the vibration plate 205. The vibration device drives multiple groups of vibration plates 205 to vibrate up and down, so as to improve the separation effect of wheat grains and wheat bran. While the crushing rod 204 rotates, it drives the fan 206. The fan 206 blows air into multiple chambers through the air blowing pipe 207, so that the air flow blows the wheat bran, and the blown wheat bran is discharged outward through multiple groups of discharge channels 208, improving the convenience of automatic wheat peeling treatment; the second motor 301 drives the second rotating shaft 302 to rotate, so that the second rotating shaft 302 drives multiple groups of eccentric wheels 303 to rotate, so that multiple groups of eccentric wheels 303 cooperate with multiple groups of tension springs 304 to drive multiple groups of vibration plates 205 to vibrate up and down. While the second rotating shaft 302 rotates, it drives the crushing rod 204 to rotate through two groups of belt wheels 305 and a belt 306, so that the crushing rod 204 hits the wheat, and at the same time improves the convenience of wind separation of wheat grains and wheat bran.

[0096] As Figures 1 to 10 shown, when the multi-source data fusion-based strong gluten wheat quality intelligent prediction and planting optimization system of the present invention is working, 1. Data collection: Use sensors and remote sensing equipment to obtain soil, meteorological and growth data of the wheat planting area, and update them to the data storage module in real time;

[0097] 2. Data processing: Clean, denoise and format the collected multi-source data, and generate training data required for the quality prediction model through data fusion algorithms;

[0098] 3. Quality prediction: Input real-time monitoring data into the deep learning model to predict the key quality indicators of strong gluten wheat and output the prediction results;

[0099] 4. Optimization suggestions: Generate an optimized planting management plan based on the quality prediction results and meteorological change trends, and push the plan to users through the platform;

[0100] 5. Dynamic adjustment: Users adjust planting management measures in real time according to the optimization suggestions, such as adjusting the fertilization amount, irrigation frequency or pest control plan. The system records the adjustment data and continuously optimizes the model.

[0101] The main functions achieved by the present invention are as follows: improving the accuracy of quality prediction: significantly improving the accuracy of strong gluten wheat quality prediction through multi-source data fusion and deep learning technology; real-time dynamic optimization: the system realizes real-time update of quality prediction results and provides dynamic optimization suggestions for planting managers; adapting to multi-region applications: the system adapts to different geographical and climatic conditions through regionalized model optimization strategies and has wide popularization value; linkage between planting and quality: linking quality prediction with planting optimization to provide users with a scientific management plan that takes both quality and yield into account.

[0102] The first motor 107, the fan 206, the second motor 301, the hydraulic telescopic arm 402, the electric rotating platform 405 and the vehicle body 501 of the intelligent prediction and planting optimization system for strong gluten wheat with multi-source data fusion of the present invention are purchased on the market. Those skilled in the industry only need to install and operate according to the attached operation manuals, without the need for creative labor from those skilled in the art.

[0103] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art in the technical field, without departing from the technical principle of the present invention, several improvements and modifications can be made, and these improvements and modifications should also be regarded as the protection scope of the present invention.

Claims

1. A strong gluten wheat quality intelligent prediction and planting optimization system for multi-source data fusion, characterized in that It includes a multi-source data acquisition module, a data fusion and storage module, a quality prediction module, a planting optimization module, and a visualization and decision support module; Multi-source data acquisition module: By deploying LoRaWAN soil three-parameter sensors, it collects data on soil pH value, EC value, moisture content, soil nutrients, soil temperature, and soil conductivity. It collects temperature, humidity, light, and CO2 data of meteorological conditions through a micro-meteorological station, conducts aerial photography with a multi-spectral camera carried by a drone, and cooperates with Sentinel-2 satellite data at the same time, so as to obtain data on vegetation indices, and realizes the acquisition of multi-source data on various soil parameters, meteorological conditions, and vegetation indices; Data fusion and storage module: It establishes a big data platform to store and fuse historical data and real-time monitoring data, including soil data, meteorological data, planting management records, and quality determination data; Quality prediction module: Based on deep learning algorithms, it constructs a prediction model for the quality of strong gluten wheat kernels to predict core quality indicators such as the protein content and kernel hardness of wheat kernels; Planting optimization module: Using the prediction results and combining with optimization algorithms, it dynamically adjusts various planting management plans for fertilization, irrigation, and pest control; Visualization and decision support module: It provides a user-friendly interface to display the prediction results and optimization suggestions in real time through the PC side or the mobile side.

2. The intelligent prediction and planting optimization system for strong gluten wheat quality with multi-source data fusion according to claim 1, characterized in that, It also includes an inspection module, a comparison module, and a display module; Inspection module: It uses on-site sampling equipment to collect wheat samples at multiple points and detects the protein content and kernel hardness of the wheat samples; Comparison module: It compares the actual detection data of the wheat samples with the predicted data to test the prediction results, and then sends the test results to the quality prediction module to continuously optimize the model according to the feedback data in actual applications; By aligning the time of the actual detection data of the wheat samples with the predicted data, it ensures that the actual detection data and the predicted data are consistent in the time dimension, and at the same time aligns the data space to ensure that the data comes from the same plot to ensure the accuracy of the comparison, and uses statistical methods to quantify the difference between the actual data and the predicted data; Display module: It intuitively displays the comparison results of the actual data and the predicted data in the form of a line chart, and marks the significant difference points in the chart for further analysis.

3. The intelligent prediction and planting optimization system for the quality of strong gluten wheat with multi-source data fusion according to claim 1, characterized in that, The quality prediction module includes a data collection and preprocessing module, a model selection and construction module, a model training and optimization module, and a model deployment and application module; Data collection and preprocessing module: It is used to collect sample data on the protein content and kernel hardness of strong gluten wheat, as well as environmental factors that may affect these indicators. Then it processes the collected data for missing values and outliers to ensure data quality, scales the data to the same scale to accelerate model convergence, and then extracts the morphological features of wheat kernels through image processing techniques; Model selection and construction module: It uses a convolutional neural network to extract the morphological features of wheat kernels and maps the extracted features to the output. It uses a recurrent neural network to capture the time series data of environmental changes during the growth process of wheat and maps the time series features to the output; Model Training and Optimization Module: Select an appropriate loss function to measure the difference between the predicted value and the true value. Adjust hyperparameters such as the learning rate and batch size through grid search or random search methods. Use cross-validation to evaluate the generalization ability of the model, and use mean squared error and mean absolute error metrics to evaluate the model performance. Model Deployment and Application Module: Save the trained model as a file, deploy the model to the actual application, predict the wheat grain quality indicators in real time, and continuously optimize the model according to the feedback data in the actual application.

4. The intelligent prediction and planting optimization system for the quality of strong gluten wheat with multi-source data fusion according to claim 1, wherein, The planting optimization module uses the wheat quality indicators predicted by the quality prediction module and environmental data to evaluate the effect of the current planting management. According to the feedback data, continuously improve the planting management plan, which specifically includes the following steps: S1. Combine the prediction results with the real-time monitoring data to form complete planting status information. S2. Identify the problems existing in the current planting management by analyzing the prediction results. S3. Dynamically adjust the planting management plan using optimization methods such as linear programming, genetic algorithms, reinforcement learning, and Bayesian optimization algorithms; adjust the application rates of fertilizers such as nitrogen, phosphorus, and potassium according to the soil nutrient level and prediction results, adjust the irrigation amount and irrigation time according to the soil humidity and meteorological data, and adjust the pesticide application amount and prevention and control strategies according to the pest and disease prediction results. S4. Deploy the optimized planting management plan to the actual planting, monitor the planting effect in real time, collect new data, and re-run the optimization algorithm according to the feedback data to continuously improve the planting management plan.

5. The intelligent prediction and planting optimization system for the quality of strong gluten wheat with multi-source data fusion according to claim 1, characterized in that, The data fusion and storage module includes a data storage layer, a data processing layer, a data application layer, and a data fusion layer. The data storage layer is used to store the long-term accumulated soil, meteorological, planting management, and quality measurement data, and at the same time store the real-time data generated by sensors and monitoring devices. The data processing layer is used to process missing values, outliers, and duplicate data, and integrate data from different sources into a unified format and standard. The data application layer displays data through visualization tools and provides data access services through APIs. The data fusion layer is used to integrate multi-source data into a consistent and usable dataset, align data from different sources according to timestamps, align data according to geographical locations, associate data through predefined rules, discover potential relationships between data through machine learning models, and convert data into a unified format.

6. The intelligent prediction and planting optimization system for the quality of strong gluten wheat with multi-source data fusion according to claim 2, characterized in that The on-site sampling device includes a support device, a processing device, a cylinder body (101), a closing door (102), a connecting plate (103), a first rotating shaft (104), multiple groups of L-shaped plates (105), blades (106) and a first motor (107). The cylinder body (101) is installed on the support device, and the support device is used to drive the cylinder body (101) to move its position. Openings are respectively provided at the bottom and side of the cylinder body (101). The closing door (102) is slidably arranged at the side opening of the cylinder body (101). The connecting plate (103) is installed on the outer side wall of the closing door (102). The first rotating shaft (104) is rotatably installed on the inner side wall of the cylinder body (101). Multiple groups of L-shaped plates (105) are circumferentially installed on the outer side wall of the first rotating shaft (104). The blades (106) are installed at the bottom opening of the cylinder body (101). The first motor (107) is installed on the support device, and the output end of the first motor (107) is connected to the first rotating shaft (104). The processing device is arranged below the cylinder body (101), and the processing device is used for wheat peeling treatment.

7. The intelligent prediction and planting optimization system for strong gluten wheat quality with multi-source data fusion according to claim 6, characterized in that, The processing device includes a vibration device, a box body (201), multiple groups of feed hoppers (202), multiple groups of sieve plates (203), crushing rods (204), multiple groups of vibrating plates (205), a fan (206), a blowing air pipe (207), multiple groups of discharge channels (208) and multiple groups of ejector rods (209). Multiple chambers are arranged inside the box body (201). Multiple groups of feed hoppers (202) are respectively communicated and arranged at the tops of the multiple chambers. Multiple groups of sieve plates (203) are respectively installed in the multiple chambers. The tops of the multiple groups of sieve plates (203) are respectively communicated with the multiple groups of feed hoppers (202). The crushing rods (204) rotatably pass through the multiple groups of sieve plates (203). The multiple groups of vibrating plates (205) are installed in the multiple chambers through the vibration device, and the vibration device is used to drive the crushing rods (204) to rotate. The fan (206) is installed on the outer side wall of the box body (201). The power input end of the fan (206) is connected to the crushing rods (204). The output end of the fan (206) is communicated with the blowing air pipe (207). The output end of the blowing air pipe (207) communicates with the inside of the multiple chambers. Multiple groups of discharge channels (208) are respectively communicated and arranged outside the multiple chambers. Multiple groups of ejector rods (209) are all installed at the top of the box body (201).

8. The intelligent prediction and planting optimization system for the quality of strong gluten wheat with multi-source data fusion according to claim 7, characterized in that, The vibration device includes a second motor (301), a second rotating shaft (302), multiple sets of eccentric wheels (303), multiple sets of tension springs (304), two sets of belt pulleys (305) and a belt (306). The second motor (301) is installed on the outer side wall of the box body (201), the second rotating shaft (302) is rotatably installed on the inner side wall of the box body (201), multiple sets of eccentric wheels (303) are all installed on the outer side wall of the second rotating shaft (302), the outer side walls of multiple sets of eccentric wheels (303) are respectively in contact with the bottoms of multiple sets of vibrating plates (205), multiple sets of tension springs (304) are all installed on the inner side wall of the box body (201), the tops of multiple sets of tension springs (304) are respectively connected to the bottoms of multiple sets of vibrating plates (205), the first set of belt pulleys (305) is installed at the end of the second rotating shaft (302), the second set of belt pulleys (305) is installed at the end of the crushing rod (204), and the belt (306) is sleeved on the outer side walls of the two sets of belt pulleys (305).

9. The intelligent prediction and planting optimization system for the quality of strong gluten wheat with multi-source data fusion according to claim 6, characterized in that, The support device includes a guide frame (401), a hydraulic telescopic arm (402), a bracket (403), a guide rail (404) and an electric rotating table (405). The guide frame (401) is installed on the rotating end of the electric rotating table (405), the hydraulic telescopic arm (402) is slidably installed up and down on the guide frame (401), the bracket (403) is installed on the moving end of the hydraulic telescopic arm (402), the first motor (107) and the cylinder body (101) are both installed on the outer side wall of the bracket (403), and the electric rotating table (405) is slidably installed horizontally on the guide rail (404).

10. The intelligent prediction and planting optimization system for strong gluten wheat quality with multi-source data fusion according to claim 7 or 9, characterized in that, It further includes a vehicle body (501), and the guide rail (404) and the box body (201) are respectively installed on the top of the vehicle body (501).

Images