Open field vegetable intelligent spraying optimization regulation and control system based on real-time crop information
By designing multi-spectral data fusion and in-depth analysis modules in an intelligent agricultural system, combined with adaptive control, the problem that the existing system cannot effectively integrate multi-source data and lack of real-time feedback is solved, and the precise optimization of open-field vegetable spraying strategy is achieved. Improved crop yield and quality.
Patent Information
- Application Number
- CN202411562630.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-05
- Publication Date
- 2024-12-03
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing intelligent agricultural systems are difficult to effectively integrate multi-source data, unable to fully utilize the potential value of the data, and lack real-time feedback and adaptive optimization capabilities, resulting in insufficient optimization of spraying strategies and inability to accurately respond to crop health status and environmental changes.
A smart spray optimization and regulation system for open-field vegetables based on real-time crop information was designed, including crop information collection module, multi-spectral data fusion and health assessment module, data analysis module, spray optimization module and feedback module, through multi-spectral data fusion , in-depth analysis and adaptive control, dynamically adjust spray parameters, generate personalized spray strategies, and achieve continuous optimization.
It has achieved high accuracy of crop health assessment and the accuracy of spraying strategies, improved resource utilization, reduced waste, improved crop yield and quality, and enhanced the intelligence and practicality of intelligent agricultural management.
Smart Images

Figure CN119068301A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of vegetable spraying, and in particular to an open-field vegetable intelligent spraying optimization and control system based on real-time crop information. Background Art
[0002] Open-field vegetable cultivation is an indispensable part of agricultural production. Its growth process is affected by a variety of environmental factors, such as temperature, humidity, light intensity and soil moisture. In order to ensure the healthy growth and high yield of vegetables, accurate crop management and spraying strategies are essential. With the development of agricultural technology, more and more modern agricultural systems have introduced sensors and intelligent algorithms to optimize crop management through real-time monitoring and data analysis. However, how to effectively integrate multi-source data and dynamically adjust spraying strategies according to the actual needs of crops is still a challenge facing current intelligent agricultural systems.
[0003] In the existing technology, traditional spraying management usually relies on farmers' experience or makes decisions based on data from a single sensor. This method is difficult to fully capture the complex changes in crop health and environmental conditions, resulting in unsatisfactory accuracy and effectiveness of spraying decisions. In addition, although some smart agricultural systems have introduced data collection and analysis modules, they are mostly limited to single-dimensional data analysis and lack deep fusion of multimodal data, and are unable to fully tap the potential value of data, resulting in insufficient optimization of spraying strategies and an inability to effectively adapt to the dynamic growth needs of crops and environmental changes. More importantly, most existing systems lack real-time feedback and adaptive optimization capabilities, making it difficult to adjust spraying strategies in a timely manner to cope with changes in crop health.
[0004] The present invention aims to solve the above-mentioned deficiencies in the prior art and provide an open-field vegetable intelligent spraying optimization and control system based on real-time crop information, which can optimize resource utilization and improve the efficiency and reliability of spraying operations according to the actual needs of crops and environmental changes, thereby achieving higher crop yield and quality and promoting the intelligent and precise development of agricultural production. Summary of the invention
[0005] The present invention provides an open-field vegetable intelligent spraying optimization control system based on real-time crop information.
[0006] The open-field vegetable intelligent spraying optimization and control system based on real-time crop information includes a crop information collection module, a multi-spectral data fusion and crop health assessment module, a data analysis module, a spraying optimization module, a control module and a feedback module, among which; The crop information collection module collects the growth status information of open-field vegetables in real time, including temperature, humidity, soil moisture, light intensity, leaf color and root system signals; The multispectral data fusion and crop health assessment module fuses data from different spectra and conducts a comprehensive assessment of the health status of the crops, specifically including: Spectral data acquisition: Acquire crop image data from hyperspectral, thermal infrared, and near-infrared spectral sources; Data preprocessing: preprocess the collected crop image data, including denoising, correction, and normalization; Multimodal data fusion: Fuse the pre-processed crop image data from different spectral sources to generate comprehensive crop health status information; Health status assessment: Based on comprehensive crop health status information, the health status of crops is assessed, including photosynthesis efficiency, transpiration intensity, early signs of pests and diseases, and nutritional status; The data analysis module analyzes the comprehensive evaluation results of the crop health status, combines historical growth status information and environmental conditions, and generates comprehensive evaluation data of the crop status as a basis for spraying decision-making; The spraying optimization module dynamically adjusts the spraying parameters based on the generated crop status comprehensive evaluation data and generates control instructions. The spraying parameters include spraying amount, spraying time, spraying frequency and spraying area; The control module receives the output control instruction, generates a spraying strategy, including a spraying amount control strategy, a spraying time strategy, a spraying frequency strategy, a spraying area strategy, and a spraying mode strategy, and controls the spraying equipment to operate according to the spraying strategy; The feedback module monitors the execution of the spraying operation in real time to correct the spraying strategy and achieve continuous optimization.
[0007] Optionally, the crop information collection module includes: Temperature collection: Detect the temperature changes of the surrounding environment through the temperature sensor; Humidity collection: monitor the humidity level in the air through humidity sensors; Soil moisture measurement: Measure the water content in the soil through a soil moisture sensor; Light intensity detection: Detect light intensity through light sensor; Leaf color acquisition: Capture leaf color changes through a multispectral camera; Root signal monitoring: Monitor the electrochemical signals of crop roots through root electrochemical sensors.
[0008] Optionally, the spectral data collection includes: Hyperspectral image acquisition: Hyperspectral imaging sensors are used to capture hyperspectral images of crop surfaces in the visible and near-infrared ranges, obtain spectral information, and reflect crop pigment distribution, chlorophyll content, nitrogen content, and signs of pests and diseases; Thermal infrared image data acquisition: Thermal infrared imaging sensors are used to detect thermal radiation on the crop surface and generate thermal infrared image data for assessing crop transpiration and moisture conditions; Near-infrared imaging data acquisition: Near-infrared imaging sensors are used to obtain the reflectance characteristics of crops in the near-infrared spectral range, providing additional information about the internal structure and moisture content of crops, including leaf moisture content, tissue moisture content, leaf thickness, cell structure density, vascular bundle structure, light transmittance, and light scattering characteristics.
[0009] Optionally, the data preprocessing includes: De-noising: Gaussian filtering algorithm is used to denoise crop image data to reduce random noise generated during sensor acquisition; Radiometric correction: Radiometric correction of hyperspectral data is performed using the Normalized Difference Vegetation Index (NDVI) algorithm to correct image deviations caused by different observation conditions (such as light intensity, sensor characteristics, etc.); Data normalization: Normalize crop image data.
[0010] Optionally, the multimodal data fusion includes: Feature extraction: Extract feature vectors from pre-processed hyperspectral, thermal infrared and near-infrared crop image data using the principal component analysis (PCA) algorithm; Feature fusion: Use a multimodal neural network (MMNN) to fuse feature vectors extracted from different spectral sources to generate a comprehensive feature vector ; Health status information generation: based on the fused comprehensive feature vector , the health status information of crops is generated through the health status assessment model.
[0011] Optionally, the health status assessment model adopts a convolutional neural network (CNN) model, and the convolutional neural network (CNN) model includes: Multimodal input layer: In the input layer, the fused comprehensive feature vector The spectral data are grouped according to different sources and input into different convolution branches respectively. Each branch performs convolution operation on different spectral data to extract spectral features, which can be expressed as: ; in, is the input feature of the kth spectral branch, is the convolution kernel of the corresponding branch, * represents the convolution operation, is the bias, is the ReLU activation function, is the convolutional layer output feature map of the kth spectral branch; Feature sharing mechanism between convolutional layers: A feature sharing mechanism is introduced between multiple convolutional layers to enable the features of different spectral data to interact and fuse between layers, which can be expressed as: ; in, is the output feature map of the lth layer, is the output feature map of the kth branch in the previous layer, is the convolution kernel weight, is the bias term of the lth layer, n is the number of spectral branches; Multi-scale convolution kernel design: Multi-scale convolution kernels are used to capture different scale features of crops, expressed as: ; in, is the output feature map of the multi-scale convolutional layer at layer l, Represents convolution kernels of different scales, s represents the scale of the convolution kernel, (such as , wait), For the The input feature map of the layer; Feature fusion layer: In the feature fusion layer of the network, feature maps from different spectral branches are fused to form the final comprehensive feature vector for health status assessment, which is expressed as: ; in, is the final comprehensive feature vector, is the weight matrix of the feature fusion layer, is the convolutional layer output feature map of the k-th branch of the l-th layer, is the bias term of the feature fusion layer; Output layer (health status assessment): The output layer is based on the final fused feature vector , the health status information of crops is generated through the fully connected layer, and the corresponding scores or classification results are generated according to different health indicators (such as photosynthesis efficiency, signs of pests and diseases, etc.), which are expressed as: ; in, is the predicted value of the jth health state, is the weight of the output layer, is the bias term.
[0012] Optionally, the health status assessment includes: Photosynthetic efficiency assessment: using a combination of normalized difference vegetation index (NDVI) and chlorophyll fluorescence measurements, based on a comprehensive eigenvector The spectral information in the calculation of photosynthesis efficiency is expressed as: ; Among them, NDVI is the normalized difference vegetation index, are chlorophyll-related characteristics, is the photosynthetically active radiation related feature, is the estimated value of the photosynthesis efficiency of the crop; Estimation of transpiration intensity: based on thermal infrared feature vector and the moisture feature vector , the energy balance model is combined with thermal infrared data to evaluate the transpiration intensity, which is expressed as: ; in, is the thermal infrared characteristics of the crop surface, is the ambient air temperature characteristic, is the water-related characteristic of the crop, It is the evaluation value of crop transpiration intensity; Early signs of pests and diseases detection: Using algorithms based on image processing and pattern recognition, combined with multispectral features and health status features, early signs of pests and diseases are detected, expressed as: ; in, is the kth pest feature (such as disease spots, discoloration, etc.), is the weight of each feature, n is the number of features, scoring detection of early signs of crop pests and diseases; Nutritional status assessment: By analyzing the features related to chlorophyll, nitrogen and water in the comprehensive feature vector, a linear regression model is used to estimate the overall nutritional status of the crop, expressed as: ; in, is the kth nutrition-related feature, is the regression coefficient of the corresponding feature, m is the number of nutrition-related features, is the intercept term of the linear regression model, It is the nutritional status assessment value of the crop.
[0013] Optionally, the data analysis module includes: Historical data matching and weighted analysis: Based on the comprehensive evaluation results of crop health status and historical growth status data, weighted matching analysis is performed, and the weighted Euclidean distance is used to measure the similarity between the current status and the historical status, which is expressed as: ; in, represents the weighted Euclidean distance, is the i-th indicator of the current evaluation result, is the corresponding historical growth status data, is the weight of each indicator, and m is the total number of indicators; Environmental Condition Correction: The current crop health status assessment results are corrected for environmental conditions. The correction factor is used to adjust the comprehensive evaluation results by combining the current environmental data and historical environmental data. It is expressed as: ; in, is the corrected crop health status assessment result, is a correction factor, which indicates the degree of influence of environmental conditions on crop status. is the current environment data, For historical environmental data; Generation of comprehensive crop status evaluation: Using the linear regression model, combined with the corrected evaluation results and weighted historical matching analysis data, the comprehensive crop status evaluation data is generated, expressed as: ; in, It is the final comprehensive evaluation data of crop status. , and are the coefficients of the linear regression model.
[0014] Optionally, the spraying optimization module includes: Optimization of spraying amount: based on comprehensive evaluation data of crop status and target health threshold , dynamically adjust the spraying amount through the proportional integral derivative control (PID) algorithm , expressed as: ; in, is the error at the current time t, , , are the proportional, integral and differential coefficients of the PID controller; Spraying time optimization: By combining comprehensive crop status evaluation data and environmental conditions, a dynamic time warping algorithm is used to determine the optimal spraying time , expressed as: ; Where d is the distance metric for time series matching, is the optimal time matching path, is the time series data of crop status, is the time series data of environmental conditions, and n is the number of data points in the time series; Spraying frequency optimization: Based on the real-time needs of the crop and historical spraying results, the spraying frequency is determined through the exponentially weighted moving average (EWMA) model , expressed as: ; in, is the smoothing factor, is the previous spraying frequency, It is the current comprehensive evaluation data; Optimization of spraying areas: Combine crop status comprehensive evaluation data and geographic information to divide farmland into different spraying areas based on the k-means clustering algorithm , expressed as: ; in, is the location point in the farmland, is the center point of the ith region, is the center point of the mth spraying area; Control instruction generation: Comprehensively optimize the spraying amount, spraying time, spraying frequency and spraying area to generate comprehensive control instructions , expressed as: .
[0015] Optionally, the feedback module includes: Spraying effect monitoring: Using sensor networks to collect real-time data on crop health status after spraying and environmental data and compared with the comprehensive evaluation data before spraying Compare and calculate the difference in spraying effect , expressed as: ; in, Indicates the change value of crop status before and after spraying; Error analysis and correction: By comparing the difference between the spraying effect and the expected effect Perform error analysis and use adaptive control algorithms to adjust spraying strategy parameters in real time based on the errors , expressed as: ; ; in, is the error value, is the current spraying strategy parameter, is the learning rate, is the gradient of the spraying effect on the strategy parameter; Continuous optimization and strategy update: based on revised spraying strategy parameters , dynamically update the spraying strategy for the next cycle , to achieve continuous optimization, expressed as: ; in, is the optimized spraying strategy for the next cycle, is a function that combines the revised policy parameters with the current environmental conditions.
[0016] Beneficial effects of the present invention: The present invention provides a comprehensive and accurate crop health assessment by real-time monitoring of the growth status of open-field vegetables. It utilizes multispectral data fusion and a crop health assessment module, combines data from multiple spectral sources such as hyperspectral, thermal infrared, and near infrared, and generates comprehensive health status information, thereby ensuring a high-precision assessment of the crop health status. At the same time, the data analysis module performs in-depth analysis based on real-time collected information and historical data to generate comprehensive crop status evaluation data, thereby providing a scientific basis for spraying optimization and optimizing the accuracy and effectiveness of spraying strategies.
[0017] The present invention generates a personalized spraying strategy by dynamically adjusting spraying parameters, including spraying amount, spraying time, spraying frequency and spraying area, and combines proportional integral differential control, dynamic time warping, exponential weighted moving average and k-means clustering algorithm to efficiently respond to the actual needs of crops, accurately control spraying operations, maximize resource utilization, reduce waste, reduce production costs, and effectively improve crop yield and quality.
[0018] The present invention monitors the execution effect of the spraying operation in real time, adjusts the spraying strategy in time based on difference analysis and adaptive control algorithm to achieve continuous optimization, and continuously optimizes the spraying strategy through a closed-loop feedback mechanism to ensure that crops can obtain the best growth management under various environmental conditions. It provides reliable technical support for smart agricultural management and improves the intelligence level and practicality of the system. BRIEF DESCRIPTION OF THE DRAWINGS
[0019] In order to more clearly illustrate the technical solutions in the present invention 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 for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0020] Figure 1 A schematic diagram of system function modules according to an embodiment of the present invention; Figure 2Schematic diagram of a multispectral data fusion and crop health assessment module according to an embodiment of the present invention. DETAILED DESCRIPTION
[0021] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.
[0022] It should be noted that the references to "one embodiment", "embodiment", "exemplary embodiments", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).
[0023] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.
[0024] like Figure 1-Figure 2 As shown, the open-field vegetable intelligent spraying optimization and control system based on real-time crop information includes a crop information collection module, a multispectral data fusion and crop health assessment module, a data analysis module, a spraying optimization module, a control module and a feedback module, wherein; The crop information collection module collects real-time growth status information of open-field vegetables, including temperature, humidity, soil moisture, light intensity, leaf color, and root signals; The multispectral data fusion and crop health assessment module integrates data from different spectra and conducts a comprehensive assessment of the health status of crops, including: Spectral data acquisition: Acquire crop image data from hyperspectral, thermal infrared, and near-infrared spectral sources; Data preprocessing: preprocess the collected crop image data, including denoising, correction, and normalization; Multimodal data fusion: Fuse the pre-processed crop image data from different spectral sources to generate comprehensive crop health status information; Health status assessment: Based on comprehensive crop health status information, assess the health status of crops, including photosynthesis efficiency, transpiration intensity, early signs of pests and diseases, and nutritional status; The data analysis module analyzes the comprehensive evaluation results of crop health status, combines historical growth status information and environmental conditions, and generates comprehensive evaluation data of crop status as the basis for spraying decision-making; The spraying optimization module dynamically adjusts the spraying parameters and generates control instructions based on the generated comprehensive evaluation data of crop status. The spraying parameters include spraying amount, spraying time, spraying frequency and spraying area to ensure that the spraying operation meets the actual needs of crops and optimizes resource utilization. The control module receives the output control instructions, generates a spraying strategy, including a spraying amount control strategy, a spraying time strategy, a spraying frequency strategy, a spraying area strategy, and a spraying mode strategy, and controls the spraying equipment to operate according to the spraying strategy; The feedback module monitors the execution of the spraying operation in real time to correct the spraying strategy and achieve continuous optimization; Through the above content, accurate and efficient spraying strategies are provided, which effectively improve the accuracy of crop health monitoring and resource utilization, ensure the optimal growth state of crops under different growth conditions, and at the same time have the ability of continuous optimization and adaptive adjustment, making the system more intelligent and reliable.
[0025] The crop information collection module includes: Temperature collection: Detect the temperature changes of the surrounding environment through the temperature sensor; Humidity collection: monitor the humidity level in the air through humidity sensors; Soil moisture measurement: Measure the water content in the soil through a soil moisture sensor; Light intensity detection: Detect light intensity through light sensor; Leaf color collection: Capture leaf color changes through multispectral cameras to reflect the health status of crops; Root signal monitoring: monitor the electrochemical signals of crop roots through root electrochemical sensors to evaluate the nutrient absorption and health status of the roots; Through the above content, the health status and growth environment of crops can be reflected in a timely manner, rich information support can be provided, and the accuracy of spraying optimization and regulation can be improved, so that crops are always in the best growth state in a dynamic environment and achieve higher yield and quality.
[0026] Spectral data collection includes: Hyperspectral image acquisition: Hyperspectral imaging sensors are used to capture hyperspectral images of crop surfaces in the visible and near-infrared ranges, obtain spectral information, and reflect crop pigment distribution, chlorophyll content, nitrogen content, and signs of pests and diseases; Thermal infrared image data acquisition: Thermal infrared imaging sensors are used to detect thermal radiation from the crop surface and generate thermal infrared image data for assessing crop transpiration and moisture conditions; Near-infrared imaging data acquisition: The near-infrared imaging sensor obtains the reflectance characteristics of crops in the near-infrared spectrum, providing additional information about the internal structure and moisture content of crops, including leaf moisture content, tissue moisture content, leaf thickness, cell structure density, vascular bundle structure, light transmittance, and light scattering characteristics; Through the above content, the comprehensiveness and accuracy of data collection are improved, providing a more reliable basis for subsequent crop health assessment and spraying optimization, thereby improving the intelligence level and practicality of the system.
[0027] Data preprocessing includes: Denoising: Gaussian filtering algorithm is used to denoise crop image data to reduce the random noise generated during sensor acquisition, which is expressed as: ; in, is the pixel value after filtering, is the standard deviation of the Gaussian function, x and y are the coordinates of the pixel; Radiometric correction: The hyperspectral data is radiometrically corrected using the Normalized Difference Vegetation Index (NDVI) algorithm to correct image deviations caused by different observation conditions (such as light intensity, sensor characteristics, etc.), expressed as: ; Among them, NIR represents the reflectance value of the near-infrared band, and R represents the reflectance value of the red light band; Data normalization: Normalize crop image data to standardize the data range to , expressed as: ; in, is the normalized value, X is the original pixel value, and are the minimum and maximum pixel values in the image dataset, respectively; Through the above preprocessing steps, data preprocessing provides high-quality and standardized input data for multispectral data fusion and crop health assessment, ensuring that the system can obtain consistent and reliable crop health assessment results under different spectral conditions.
[0028] Multimodal data fusion includes: Feature extraction: The principal component analysis (PCA) algorithm is used to extract feature vectors from the preprocessed hyperspectral, thermal infrared and near-infrared crop image data, expressed as: ; in, is the characteristic matrix of the original image data, is the eigenvector matrix corresponding to the spectral source, is the feature vector of each spectral source data; Feature fusion: Use a multimodal neural network (MMNN) to fuse feature vectors extracted from different spectral sources to generate a comprehensive feature vector , expressed as: ; in, is the weight of each spectral source eigenvector, is the feature vector extracted from the i-th spectral source, n is the number of spectral sources; Health status information generation: based on the fused comprehensive feature vector , generating crop health status information through a health status assessment model; Through the above content, crop image data from different spectral sources can be effectively fused to generate comprehensive information reflecting the overall health status of the crop, ensuring that the assessment of the crop health status is highly accurate and reliable, and can provide accurate decision support for subsequent spraying optimization.
[0029] The health status assessment model uses a convolutional neural network (CNN) model, which includes: Multimodal input layer: In the input layer, the fused comprehensive feature vector The spectral data are grouped according to different sources and input into different convolution branches respectively. Each branch performs convolution operation on different spectral data to extract spectral specific features, which can be expressed as: ; in, is the input feature of the kth spectral branch, is the convolution kernel of the corresponding branch, * represents the convolution operation, is the bias, is the ReLU activation function, is the convolutional layer output feature map of the kth spectral branch; Feature sharing mechanism between convolutional layers: A feature sharing mechanism is introduced between multiple convolutional layers to enable the features of different spectral data to interact and fuse between layers, thereby enhancing the network's ability to integrate multimodal information, which can be expressed as: ; in, is the output feature map of the lth layer, is the output feature map of the kth branch in the previous layer, is the convolution kernel weight, is the bias term of the lth layer, n is the number of spectral branches; Multi-scale convolution kernel design: Multi-scale convolution kernels are used to capture the different scale features of crops, ensuring that the network can capture both local details and global patterns, expressed as: ; in, is the output feature map of the multi-scale convolutional layer at layer l, Represents convolution kernels of different scales, s represents the scale of the convolution kernel, (such as , etc.), capturing features of different granularities through multi-scale convolution operations. For the The input feature map of the layer; Feature fusion layer: In the feature fusion layer of the network, feature maps from different spectral branches are fused to form the final comprehensive feature vector for health status assessment, which is expressed as: ; in, is the final comprehensive feature vector, is the weight matrix of the feature fusion layer, is the convolutional layer output feature map of the k-th branch of the l-th layer, is the bias term of the feature fusion layer; Output layer (health status assessment): The output layer is based on the final fused feature vector , the health status information of crops is generated through the fully connected layer, and the corresponding scores or classification results are generated according to different health indicators (such as photosynthesis efficiency, signs of pests and diseases, etc.), which are expressed as: ; in, is the predicted value of the jth health state, is the weight of the output layer, is the bias term; Through the above content, the system's ability to identify complex and changeable crop health characteristics has been improved, and the intelligence and practicality of spraying optimization have been enhanced, thereby effectively promoting the precise management and efficient implementation of agricultural production.
[0030] Health status assessment includes: Photosynthetic efficiency assessment: using a combination of normalized difference vegetation index (NDVI) and chlorophyll fluorescence measurements, based on a comprehensive eigenvector The spectral information in the calculation of photosynthesis efficiency is expressed as: ; Among them, NDVI is the normalized difference vegetation index, are chlorophyll-related characteristics, is the photosynthetically active radiation related feature, is the estimated value of the photosynthesis efficiency of the crop; Estimation of transpiration intensity: based on thermal infrared feature vector and the moisture feature vector , the energy balance model is combined with thermal infrared data to evaluate the transpiration intensity, which is expressed as: ; in, is the thermal infrared characteristics of the crop surface, is the ambient air temperature characteristic, is the water-related characteristic of the crop, It is the evaluation value of crop transpiration intensity; Early signs of pests and diseases detection: Using algorithms based on image processing and pattern recognition, combined with multispectral features and health status features, early signs of pests and diseases are detected, expressed as: ; in, is the kth pest feature (such as disease spots, discoloration, etc.), is the weight of each feature, n is the number of features, scoring detection of early signs of crop pests and diseases; Nutritional status assessment: By analyzing the features related to chlorophyll, nitrogen and water in the comprehensive feature vector, a linear regression model is used to estimate the overall nutritional status of the crop, expressed as: ; in, is the kth nutrition-related feature, is the regression coefficient of the corresponding feature, m is the number of nutrition-related features, is the intercept term of the linear regression model, It is the nutritional status assessment value of the crop; Through the above content, more comprehensive and detailed crop health data can be provided, and potential problems in the crop growth process, such as pests and diseases and nutrient deficiencies, can be identified early, so that management measures can be adjusted in time to improve crop yield and quality. In addition, with the help of the fused comprehensive feature vector, the evaluation results are more accurate and reliable, providing strong decision-making support for intelligent agricultural management.
[0031] Data analysis modules include: Historical data matching and weighted analysis: Based on the comprehensive evaluation results of crop health status and historical growth status data, weighted matching analysis is performed, and the weighted Euclidean distance is used to measure the similarity between the current status and the historical status, which is expressed as: ; in, represents the weighted Euclidean distance, is the i-th indicator of the current evaluation result, is the corresponding historical growth status data, is the weight of each indicator, and m is the total number of indicators; Environmental Condition Correction: The current crop health status assessment results are corrected for environmental conditions. The correction factor is used to adjust the comprehensive evaluation results by combining the current environmental data and historical environmental data. It is expressed as: ; in, is the corrected crop health status assessment result, is a correction factor, which indicates the degree of influence of environmental conditions on crop status. is the current environment data, For historical environmental data; Generation of comprehensive crop status evaluation: Using the linear regression model, combined with the corrected evaluation results and weighted historical matching analysis data, the comprehensive crop status evaluation data is generated, expressed as: ; in, It is the final comprehensive evaluation data of crop status. , and is the coefficient of the linear regression model, which is used to adjust the influence weight of different input variables; Through the above content, the accuracy and practicality of the evaluation results are improved, ensuring that spraying decisions are based on comprehensive and accurate crop status information, optimizing resource utilization and crop growth effects.
[0032] The application optimization module includes: Optimization of spraying amount: based on comprehensive evaluation data of crop status and target health threshold , dynamically adjust the spraying amount through the proportional integral derivative control (PID) algorithm , expressed as: ; in, is the error at the current time t, , , are the proportional, integral and differential coefficients of the PID controller; Spraying time optimization: By combining comprehensive crop status evaluation data and environmental conditions, a dynamic time warping algorithm is used to determine the optimal spraying time , expressed as: ; Where d is the distance metric for time series matching, is the optimal time matching path, is the time series data of crop status, is the time series data of environmental conditions, and n is the number of data points in the time series; Spraying frequency optimization: Based on the real-time needs of the crop and historical spraying results, the spraying frequency is determined through the exponentially weighted moving average (EWMA) model , expressed as: ; in, is the smoothing factor, is the previous spraying frequency, It is the current comprehensive evaluation data; Optimization of spraying areas: Combine crop status comprehensive evaluation data and geographic information to divide farmland into different spraying areas based on the k-means clustering algorithm , expressed as: ; in, is the location point in the farmland, is the center point of the ith region, is the center point of the mth spraying area; Control instruction generation: Comprehensively optimize the spraying amount, spraying time, spraying frequency and spraying area to generate comprehensive control instructions , expressed as: ; Through the above contents, efficient and accurate crop management can be achieved, waste can be reduced, costs can be lowered, and the accuracy of crop health monitoring can be improved, providing strong support for smart agricultural management.
[0033] Target health threshold Based on historical data settings: ; in, is the historical health status data under similar environmental conditions, m is the number of historical data, is the average value of historical data, is the comprehensive evaluation data of the current crop status, It is a dynamic adjustment factor used to adjust the difference between historical data and current status.
[0034] The feedback module includes: Spraying effect monitoring: Using sensor networks to collect real-time data on crop health status after spraying and environmental data and compared with the comprehensive evaluation data before spraying Compare and calculate the difference in spraying effect , expressed as: ; in, Indicates the change in crop status before and after spraying, used to evaluate the actual effect of spraying; Error analysis and correction: By comparing the difference between the spraying effect and the expected effect Perform error analysis and use adaptive control algorithms to adjust spraying strategy parameters in real time based on the errors , expressed as: ; ; in, is the error value, is the current spraying strategy parameter, is the learning rate, is the gradient of the spraying effect on the strategy parameter; Continuous optimization and strategy update: based on revised spraying strategy parameters , dynamically update the spraying strategy for the next cycle , to achieve continuous optimization, expressed as: ; in, is the optimized spraying strategy for the next cycle, It is a function that combines the revised strategy parameters with the current environmental conditions to generate the final spraying plan; Function that combines the revised strategy parameters with the current environmental conditions It is expressed as: ; in, is the optimized spraying strategy for the next cycle, is the modified spraying strategy parameter, is the environmental condition data at the current time t (such as temperature, humidity, light, etc.); Through the above content, not only the maintenance effect of crop health status is improved, but also resource waste is minimized, the efficiency and reliability of spraying operations are improved, and strong support is provided for the smart agricultural system to ensure that crops can obtain the best growth management under various environmental conditions.
[0035] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.
[0036] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. An open-field vegetable intelligent spraying optimization and control system based on real-time crop information, characterized in that: It includes crop information collection module, multispectral data fusion and crop health assessment module, data analysis module, spraying optimization module, control module and feedback module, among which; The crop information collection module collects the growth status information of open-field vegetables in real time, including temperature, humidity, soil moisture, light intensity, leaf color and root system signals; The multispectral data fusion and crop health assessment module fuses data from different spectra and conducts a comprehensive assessment of the health status of the crops, specifically including: Spectral data acquisition: Acquire crop image data from hyperspectral, thermal infrared, and near-infrared spectral sources; Data preprocessing: preprocess the collected crop image data, including denoising, correction, and normalization; Multimodal data fusion: Fuse the pre-processed crop image data from different spectral sources to generate comprehensive crop health status information; Health status assessment: Based on comprehensive crop health status information, assess the health status of crops, including photosynthesis efficiency, transpiration intensity, early signs of pests and diseases, and nutritional status; The data analysis module analyzes the comprehensive evaluation results of the crop health status, combines historical growth status information and environmental conditions, and generates comprehensive evaluation data of the crop status as a basis for spraying decision-making; The spraying optimization module dynamically adjusts the spraying parameters based on the generated crop status comprehensive evaluation data and generates control instructions. The spraying parameters include spraying amount, spraying time, spraying frequency and spraying area; The control module receives the output control instruction, generates a spraying strategy, including a spraying amount control strategy, a spraying time strategy, a spraying frequency strategy, a spraying area strategy, and a spraying mode strategy, and controls the spraying equipment to operate according to the spraying strategy; The feedback module monitors the execution of the spraying operation in real time to correct the spraying strategy and achieve continuous optimization.
2. The open-field vegetable intelligent spraying optimization control system based on real-time crop information according to claim 1 is characterized in that: The crop information collection module comprises: Temperature collection: Detect the temperature changes of the surrounding environment through the temperature sensor; Humidity collection: monitor the humidity level in the air through humidity sensors; Soil moisture measurement: Measure the water content in the soil through a soil moisture sensor; Light intensity detection: Detect light intensity through light sensor; Leaf color acquisition: Capture leaf color changes through a multispectral camera; Root signal monitoring: Monitor the electrochemical signals of crop roots through root electrochemical sensors.
3. The open-field vegetable intelligent spraying optimization control system based on real-time crop information according to claim 1 is characterized in that: The spectral data collection comprises: Hyperspectral image acquisition: Hyperspectral imaging sensors are used to capture hyperspectral images of crop surfaces in the visible and near-infrared ranges, obtain spectral information, and reflect crop pigment distribution, chlorophyll content, nitrogen content, and signs of pests and diseases; Thermal infrared image data acquisition: Thermal infrared imaging sensors are used to detect thermal radiation on the crop surface and generate thermal infrared image data for assessing crop transpiration and moisture conditions; Near-infrared imaging data acquisition: Near-infrared imaging sensors are used to obtain the reflectance characteristics of crops in the near-infrared spectral range, providing additional information about the internal structure and moisture content of crops, including leaf moisture content, tissue moisture content, leaf thickness, cell structure density, vascular bundle structure, light transmittance, and light scattering characteristics.
4. The open-field vegetable intelligent spraying optimization control system based on real-time crop information according to claim 3 is characterized in that: The data preprocessing includes: De-noising: Gaussian filtering algorithm is used to denoise crop image data to reduce random noise generated during sensor acquisition; Radiometric correction: Radiometric correction of hyperspectral data is performed through the normalized difference vegetation index algorithm to correct image deviations caused by different observation conditions; Data normalization: Normalize crop image data.
5. The open-field vegetable intelligent spraying optimization control system based on real-time crop information according to claim 4 is characterized in that: The multimodal data fusion includes: Feature extraction: Extract feature vectors from pre-processed hyperspectral, thermal infrared and near-infrared crop image data using principal component analysis algorithm; Feature fusion: Use a multimodal neural network to fuse feature vectors extracted from different spectral sources to generate a comprehensive feature vector ; Health status information generation: based on the fused comprehensive feature vector , the health status information of crops is generated through the health status assessment model.
6. The open-field vegetable intelligent spraying optimization control system based on real-time crop information according to claim 5 is characterized in that: The health status assessment model adopts a convolutional neural network model, and the convolutional neural network model includes: Multimodal input layer: In the input layer, the fused comprehensive feature vector The spectral data are grouped according to different sources and input into different convolution branches respectively. Each branch performs convolution operation on different spectral data to extract spectral features, which can be expressed as: ; in, is the input feature of the kth spectral branch, is the convolution kernel of the corresponding branch, * represents the convolution operation, is the bias, is the ReLU activation function, is the convolutional layer output feature map of the kth spectral branch; Feature sharing mechanism between convolutional layers: A feature sharing mechanism is introduced between multiple convolutional layers to enable the features of different spectral data to interact and fuse between layers, which can be expressed as: ; in, is the output feature map of the lth layer, is the output feature map of the kth branch in the previous layer, is the convolution kernel weight, is the bias term of the lth layer, n is the number of spectral branches; Multi-scale convolution kernel design: Multi-scale convolution kernels are used to capture different scale features of crops, expressed as: ; in, is the output feature map of the multi-scale convolutional layer at layer l, Represents convolution kernels of different scales, s represents the scale of the convolution kernel, For the The input feature map of the layer; Feature fusion layer: In the feature fusion layer of the network, feature maps from different spectral branches are fused to form the final comprehensive feature vector for health status assessment, which is expressed as: ; in, is the final comprehensive feature vector, is the weight matrix of the feature fusion layer, is the convolutional layer output feature map of the k-th branch of the l-th layer, is the bias term of the feature fusion layer; Output layer: The output layer is based on the final fused feature vector , the health status information of crops is generated through the fully connected layer, and the corresponding scores or classification results are generated according to different health indicators, which are expressed as: ; in, is the predicted value of the jth health state, is the weight of the output layer, is the bias term.
7. The open-field vegetable intelligent spraying optimization control system based on real-time crop information according to claim 6 is characterized in that: The health status assessment includes: Photosynthetic efficiency assessment: using a combination of normalized difference vegetation index and chlorophyll fluorescence measurements, based on an integrated eigenvector The spectral information in the calculation of photosynthesis efficiency is expressed as: ; Among them, NDVI is the normalized difference vegetation index, are chlorophyll-related characteristics, is the photosynthetically active radiation related feature, is the estimated value of the photosynthesis efficiency of the crop; Estimation of transpiration intensity: based on thermal infrared feature vector and the moisture feature vector , the energy balance model is combined with thermal infrared data to evaluate the transpiration intensity, which is expressed as: ; in, is the thermal infrared characteristics of the crop surface, is the ambient air temperature characteristic, is the water-related characteristic of the crop, It is the evaluation value of crop transpiration intensity; Early signs of pests and diseases detection: Using algorithms based on image processing and pattern recognition, combined with multispectral features and health status features, early signs of pests and diseases are detected, expressed as: ; in, is the kth pest feature, is the weight of each feature, n is the number of features, scoring detection of early signs of crop pests and diseases; Nutritional status assessment: By analyzing the features related to chlorophyll, nitrogen and water in the comprehensive feature vector, a linear regression model is used to estimate the overall nutritional status of the crop, expressed as: ; in, is the kth nutrition-related feature, is the regression coefficient of the corresponding feature, m is the number of nutrition-related features, is the intercept term of the linear regression model, It is the nutritional status assessment value of the crop.
8. The open-field vegetable intelligent spraying optimization control system based on real-time crop information according to claim 7 is characterized in that: The data analysis module includes: Historical data matching and weighted analysis: Based on the comprehensive evaluation results of crop health status and historical growth status data, weighted matching analysis is performed, and the weighted Euclidean distance is used to measure the similarity between the current status and the historical status, which is expressed as: ; in, represents the weighted Euclidean distance, is the i-th indicator of the current evaluation result, is the corresponding historical growth status data, is the weight of each indicator, and m is the total number of indicators; Environmental Condition Correction: The current crop health status assessment results are corrected for environmental conditions. The correction factor is used to adjust the comprehensive evaluation results by combining the current environmental data and historical environmental data. It is expressed as: ; in, is the corrected crop health status assessment result, is a correction factor, which indicates the degree of influence of environmental conditions on crop status. is the current environment data, For historical environmental data; Generation of comprehensive crop status evaluation: Using the linear regression model, combined with the corrected evaluation results and weighted historical matching analysis data, the comprehensive crop status evaluation data is generated, expressed as: ; in, It is the final comprehensive evaluation data of crop status. , and are the coefficients of the linear regression model.
9. The open-field vegetable intelligent spraying optimization control system based on real-time crop information according to claim 8 is characterized in that: The spraying optimization module includes: Optimization of spraying amount: based on comprehensive evaluation data of crop status and target health threshold , dynamically adjust the spraying amount through the proportional integral differential control algorithm , expressed as: ; in, is the error at the current time t, , , are the proportional, integral and differential coefficients of the PID controller; Spraying time optimization: By combining comprehensive crop status evaluation data and environmental conditions, a dynamic time warping algorithm is used to determine the optimal spraying time , expressed as: ; Where d is the distance metric for time series matching, is the optimal time matching path, is the time series data of crop status, is the time series data of environmental conditions, and n is the number of data points in the time series; Spraying frequency optimization: According to the real-time needs of crops and historical spraying effects, the spraying frequency is determined through the exponentially weighted moving average model , expressed as: ; in, is the smoothing factor, is the previous spraying frequency, It is the current comprehensive evaluation data; Optimization of spraying areas: Combine crop status comprehensive evaluation data and geographic information to divide farmland into different spraying areas based on the k-means clustering algorithm , expressed as: ; in, is the location point in the farmland, is the center point of the ith region, is the center point of the mth spraying area; Control instruction generation: Comprehensively optimize the spraying amount, spraying time, spraying frequency and spraying area to generate comprehensive control instructions , expressed as: 。 10. The open-field vegetable intelligent spraying optimization control system based on real-time crop information according to claim 1 is characterized in that: The feedback module comprises: Spraying effect monitoring: Using sensor networks to collect real-time data on crop health status after spraying and environmental data and compared with the comprehensive evaluation data before spraying Compare and calculate the difference in spraying effect , expressed as: ; in, Indicates the change value of crop status before and after spraying; Error analysis and correction: By comparing the difference between the spraying effect and the expected effect Perform error analysis and use adaptive control algorithms to adjust spraying strategy parameters in real time based on the errors , expressed as: ; ; in, is the error value, is the current spraying strategy parameter, is the learning rate, is the gradient of the spraying effect on the strategy parameter; Continuous optimization and strategy update: based on revised spraying strategy parameters , dynamically update the spraying strategy for the next cycle , to achieve continuous optimization, expressed as: ; in, is the optimized spraying strategy for the next cycle, is a function that combines the revised policy parameters with the current environmental conditions.
Citation Information
Patent Citations
Plant condition quantitative inversion method, device and equipment based on multi-modal deep learning
CN118247558A
Crop disease and pest detection method and system based on multi-modal deep learning
CN118258769A
Beef carcass quality rating method based on image analysis in combination with machine learning
CN118537664A
Open field vegetable fertilizer and pesticide spraying multi-parameter matching regulation and control method
CN118642374A
System for monitoring the health status of plants in precision agriculture using image processing and convolutional neural network
DE202022102591U1
Cited By
Intelligent greenhouse crop growth state monitoring and management system based on Internet of Things
CN119721508A
Agricultural seedling raising control method and platform for deep learning teaching
CN120259013A
Intelligent detection method for agricultural diseases and insect pests
CN120314329A
Traditional Chinese medicinal material standard base intelligent irrigation and fertilization method and decision making system based on Internet of Things
CN120787596A
