Crop growth prediction method and system based on multispectral unmanned aerial vehicle monitoring

Abstract

The invention belongs to the technical field of crop growth prediction, and particularly relates to a crop growth prediction method and system based on multispectral unmanned aerial vehicle monitoring. The method comprises the following steps: acquiring multi-dimensional data of real crops in different growth stages under different soil water contents and disease and insect pest states, determining the contribution degree of each vegetation index to crop growth through a factor analysis algorithm, carrying out dimension reduction on hyperspectral data by using a principal component analysis algorithm, fusing with the multi-spectral data, constructing a multi-spectral resolution characteristic space, and carrying out multi-spectral analysis on the hyperspectral data. The method comprises the following steps: firstly, obtaining a real plant height growth fitting function of crops through analogue simulation by combining LiDAR data and vegetation indexes, thirdly, calculating a real growth vegetation index space and obtaining a real growth state space of a standard staged growth period, and finally, inputting the calculated space and function into a model constructed by a reinforcement learning algorithm for training, and accurate prediction of the crop growth state is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of crop growth prediction, and in particular relates to a crop growth prediction method and system based on multispectral unmanned aerial vehicle monitoring. Background Art

[0002] Crop growth monitoring and prediction are crucial for precision agriculture management. It enables timely understanding of crop growth status and the adoption of reasonable agricultural measures such as irrigation, fertilization, and pest and disease control, thereby optimizing resource utilization, improving crop yield and quality, and reducing environmental pollution and production costs. In recent years, multispectral remote sensing technology has been widely used in the field of crop monitoring due to its inherent advantages. However, achieving accurate crop growth prediction based on multispectral drone monitoring still faces many challenges. On the one hand, the processing and analysis of multispectral data requires complex algorithms and models to extract effective crop growth information. On the other hand, the spectral response characteristics of different crops under different growth environments vary, and targeted prediction models need to be established.

[0003] For example, the Chinese patent publication number CN114972684A discloses a crop growth prediction method, apparatus, equipment, and medium. The method includes: determining a lossless two-dimensional image dataset of an original three-dimensional point cloud dataset based on an image backtracking algorithm; fusing the original three-dimensional point cloud dataset, the lossless two-dimensional image dataset, and meteorological factor data, and inverting the temporal and spatial specificity of growth parameters of different crops to obtain an optimal crop growth parameter multidimensional deep learning model, so as to evaluate the growth parameter information of different crops and obtain the temporal and spatial distribution of growth parameters of different crops.

[0004] For example, a Chinese patent with authorization announcement number CN111582554B discloses a crop growth prediction method and system, including: obtaining the normalized vegetation index corresponding to the current year and the normalized vegetation index corresponding to each historical year; fitting and reconstructing the normalized vegetation index corresponding to the current year and each historical year to obtain the current NDVI fitting curve and each historical NDVI fitting curve; using the DTW algorithm to regularize the current NDVI fitting curve and each historical NDVI fitting curve to obtain the current NDVI growth curve and each historical NDVI growth curve; calculating the shortest distance between the current NDVI growth curve and each historical NDVI growth curve, taking the year corresponding to the historical NDVI growth curve with the smallest distance as the best NDVI growth matching year, and predicting the growth of the target crop in the current year.

[0005] The above existing technologies have the following problems: the existing algorithms only predict the overall growth status of crops through single NDVI and image data, but do not take into account that the NDVI corresponding to crops at different growth stages is different, and do not take into account the specific impact of moisture and pests and diseases on the growth status, resulting in low prediction accuracy. To this end, the present invention provides a crop growth prediction method and system based on multispectral drone monitoring. Summary of the Invention

[0006] In response to the shortcomings of the existing technology, the present invention proposes a crop growth prediction method and system based on multispectral UAV monitoring. The method first configures the growth status data of crops in different growth stages under a standard growth environment, and collects multidimensional data of real crops under different soil moisture content and pest and disease conditions at different growth stages. The contribution of each vegetation index to crop growth is determined by the factor analysis algorithm, and the principal component analysis algorithm is used to reduce the dimension of the hyperspectral data, which is then fused with the multispectral data to construct a multispectral resolution feature space. Secondly, LiDAR data and vegetation index are combined to obtain the real plant height growth fitting function of the crop through simulation. Thirdly, the real growth vegetation index space is calculated and the real growth status space of the standard staged growth period is obtained. Finally, the calculated space and function are input into the model constructed by the reinforcement learning algorithm for training to achieve accurate prediction of the crop growth status.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] Crop growth prediction methods based on multispectral UAV monitoring include:

[0009] S1. Configure crop growth status data for each growth stage under a standard growth environment, and obtain multispectral data, crop LiDAR data, and hyperspectral data of real crops at different growth stages, different soil moisture content, and pest and disease conditions;

[0010] S2. Obtain a multi-vegetation index space, and calculate the contribution of each vegetation index to the growth simulation of the corresponding type of crop at different growth stages based on a factor analysis algorithm, and obtain the matching vegetation index for each growth stage of the corresponding type of crop;

[0011] S3. Input the hyperspectral data of different soil moisture contents and pest and disease status into the principal component analysis algorithm for dimensionality reduction, and fuse the hyperspectral data after dimensionality reduction with the corresponding multispectral data to obtain the multispectral resolution feature space of different growth stages;

[0012] S4. Utilize the multispectral resolution feature space of different growth stages to construct vegetation index water and pest and disease regulation factors corresponding to different growth stages. At the same time, through crop LiDAR data and water and pest and disease regulation factors, a simulation algorithm is used to obtain the actual plant height growth fitting function of crops under different growth states.

[0013] S5. Multiply the vegetation index moisture and pest regulation factors corresponding to different growth stages with the vegetation index matching each growth stage to obtain a real growth vegetation index space, and multiply the real growth vegetation index space with the growth state data of the growth period in the standard growth environment to obtain the real growth state space of the standard growth period in each stage;

[0014] S6. Input the real growth state space of the standard staged growth period and the real plant height growth fitting function of crops under different growth states into the reinforcement prediction and adjustment model constructed by the reinforcement learning algorithm for training, obtain the trained reinforcement prediction and adjustment model, and predict the crop growth state through the crop growth data obtained in real time.

[0015] Specifically, the steps for obtaining the matching vegetation index for each growth stage include:

[0016] S201. Extracting the band reflectance data required for calculating each vegetation index by acquiring multispectral data of real crops at different growth stages, different soil moisture contents, and pest and disease conditions, calculating different vegetation indices, and constructing a standardized vegetation index matrix containing N vegetation indices;

[0017] S202: Set each category of crops to include m growth stages, and select q crop samples from each stage. The sample space of

[0018] S203, according to The sample space and standardized vegetation index matrix are input into the factor analysis algorithm to obtain N vegetation indices. The covariance matrix C of the sample points;

[0019] S204 , calculating the eigenvector and eigenvalue of the corresponding vegetation index using the characteristic equation according to the obtained covariance matrix C.

[0020] Specifically, the steps for obtaining the matching vegetation index for each growth stage also include:

[0021] S205. Calculate the contribution rate and factor loading matrix of the nth vegetation index to the kth type of crop at the mth growth stage based on the eigenvector and eigenvalue of the corresponding vegetation index;

[0022] S206: Setting a contribution rate threshold. Based on the contribution rates of the N vegetation indices obtained for the crops in the corresponding category at the mth growth stage and the contribution rate threshold, the vegetation index corresponding to the crop with a cumulative contribution rate greater than the contribution rate threshold is retained as the vegetation index corresponding to the current growth stage of the crop.

[0023] S207, according to the factor loading matrix corresponding to the retained vegetation index, calculating the contribution of the corresponding vegetation index by summing the squares of the corresponding elements in the factor loading matrix;

[0024] S208. Calculate the corresponding vegetation index weighting coefficient using the contribution degree and contribution rate of the corresponding vegetation index, and use the calculated vegetation index weighting coefficient and the vegetation index to obtain a comprehensive matching vegetation index for the current growth stage of the corresponding crop.

[0025] Specifically, the steps for constructing water and pest and disease regulating factors include:

[0026] S401, select from the multi-spectral resolution feature space containing m growth stages according to the S3 process The corresponding samples under different moisture conditions are obtained through the evaluation algorithm, and the growth evaluation scores under the corresponding moisture conditions are obtained; where p represents p soil moisture levels;

[0027] S402. Calculate vegetation indices corresponding to different moisture conditions based on the selected samples;

[0028] S403, obtaining a regression function of the effect of different water conditions on crop growth using a multi-source linear regression algorithm based on vegetation indices corresponding to different water conditions and growth assessment scores under the corresponding water conditions;

[0029] S404, using the coefficient of the regression function for crop growth as a regulating factor of water on crop growth to obtain a water regulating factor sequence;

[0030] S405: Input the water regulation factor sequence into a genetic algorithm for optimization search to obtain the optimal water regulation factor sequence corresponding to different growth stages.

[0031] Specifically, the steps for constructing the water and pest and disease regulating factors also include:

[0032] S406, select from the multi-spectral resolution feature space containing m growth stages according to the S3 process Different disease states and The corresponding samples under different pest conditions are obtained through the evaluation algorithm, and the growth evaluation scores under the corresponding disease conditions and pest conditions are obtained; where u represents u types of disease state samples, and v represents v types of pest state samples;

[0033] S407, repeating the processes of S402-S405 to obtain the optimal disease regulating factor sequence and the optimal pest regulating factor sequence under different disease states and pest states corresponding to different growth stages;

[0034] S408. Construct a water and pest regulation factor space based on the optimal water regulation factor sequence, the optimal disease regulation factor sequence, and the optimal pest regulation factor sequence.

[0035] Specifically, the steps for constructing the true plant height growth fitting function include:

[0036] S409, constructing a three-dimensional model of standard heights corresponding to different growth stages of the crop using a three-dimensional algorithm based on LiDAR data of different growth stages of the crop and growth status data of different growth stages of the crop under a standard growth environment;

[0037] S410, based on the three-dimensional model of standard heights corresponding to different growth stages of crops and the space of water and pest and disease regulation factors corresponding to different growth stages, obtain a fitting function of the actual plant height growth corresponding to different growth stages of the corresponding type of crops through a simulation algorithm and a nonlinear fitting function;

[0038] S411. Obtain a prediction error based on the actual plant height data and the predicted data of the actual plant height growth fitting function, and feed the prediction error back to the simulation algorithm for adjustment to obtain a actual plant height growth fitting function with a prediction error less than a configured prediction error threshold, as the actual plant height growth fitting function for the corresponding growth stage of the current type of crop.

[0039] Specifically, the steps for building and training the enhanced prediction and adjustment model include:

[0040] S601, constructing the current input state based on the acquired water and pest and disease regulation factor space, the growth state of the staged growth period under the standard growth environment, and the actual plant height growth fitting function;

[0041] S602: Set the current moment trigger information according to the current moment input state, specifically:

[0042] If the current crop is the kth type of crop in the mth growth stage, the action execution information is triggered. That is, the water and pest regulation factors of the kth type of crop in the mth growth stage are called from the water and pest regulation factor space, and the plant height growth fitting function and the corresponding comprehensive matching vegetation index are adjusted and corrected. The corrected plant height growth fitting function and the corresponding comprehensive matching vegetation index are built into the corresponding time series prediction sub-model to predict the current crop growth status;

[0043] S603: Constructing an action to be executed at the current moment based on the trigger information at the current moment, and constructing a reward function based on the error between the predicted result and the actual growth state, the plant height growth fitting function, and the correction loss of the corresponding comprehensive matching vegetation index;

[0044] S604: Input the current input state, current trigger information, current execution action, and reward function into the reinforcement prediction and adjustment model for training, obtain a trained reinforcement prediction and adjustment model, and output the crop growth state corresponding to the current moment.

[0045] The crop growth prediction system based on multispectral UAV monitoring includes: data acquisition module, index matching module and simulation module;

[0046] The data acquisition module is used to configure the growth status data of crops in different growth stages under a standard growth environment, and obtain multispectral data, crop LiDAR data and hyperspectral data of real crops in different growth stages, different soil moisture content and pest and disease conditions;

[0047] The index matching module includes an index space construction unit and a matching analysis unit;

[0048] Index space construction unit, used to obtain different types of vegetation indices and construct multiple vegetation index spaces;

[0049] The matching analysis unit calculates the contribution of each vegetation index to the growth simulation of the corresponding type of crops at different growth stages through a factor analysis algorithm, and obtains the matching vegetation index of each growth stage of the corresponding type of crops;

[0050] The simulation module includes a feature fusion unit, a factor construction unit, and a function simulation unit;

[0051] The feature fusion unit is used to input the hyperspectral data of different soil moisture contents and pest and disease status into the principal component analysis algorithm for dimensionality reduction, and fuse the hyperspectral data after dimensionality reduction with the corresponding multispectral data to obtain the multispectral resolution feature space of different growth stages;

[0052] The factor construction unit uses the multi-spectral resolution feature space of different growth stages to construct the vegetation index moisture and pest and disease regulation factors corresponding to different growth stages.

[0053] Specifically, the system also includes a growth state space module and a prediction module;

[0054] The growth state space module is used to multiply the vegetation index moisture and pest regulation factors corresponding to the same growth stage with the matching vegetation index of each growth stage to obtain the real growth vegetation index space, and multiply the real growth vegetation index space with the growth state data of the staged growth period under the standard growth environment to obtain the real growth state space of the standard staged growth period;

[0055] The prediction module is used to input the real growth state space of the standard staged growth period and the real plant height growth fitting function of crops under different growth states into the reinforcement prediction adjustment model constructed by the reinforcement learning algorithm for training, obtain the trained reinforcement prediction adjustment model, and predict the crop growth state through the crop growth data obtained in real time.

[0056] An electronic device includes a memory and a processor. The memory stores a computer program. When the processor executes the computer program, a crop growth prediction method based on multispectral unmanned aerial vehicle monitoring is implemented.

[0057] A computer-readable storage medium stores computer instructions, which, when executed, execute a crop growth prediction method based on multispectral unmanned aerial vehicle monitoring.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] In response to the shortcomings of the existing technology, the present invention constructs a high-precision crop growth status prediction system through multi-source data fusion and intelligent algorithms, which significantly improves the accuracy and practicality of the prediction. Specifically, first, the contribution of each vegetation index to different growth stages is calculated through a factor analysis algorithm, and the vegetation index matching each growth stage is obtained, which solves the problem in the existing technology that a single NDVI cannot adapt to different growth stages. Secondly, principal component analysis (PCA) is used to reduce the dimension of hyperspectral data and fuse it with multispectral data to construct a multispectral resolution feature space, which enhances data processing efficiency and information volume. Thirdly, through a simulation algorithm combined with LiDAR data and water, pest and disease regulatory factors, the crop plant height growth fitting function under different growth states is obtained, which realizes the accurate modeling of crop morphological changes. These regulatory factors are multiplied with the vegetation index of the corresponding growth stage to obtain the real growth vegetation index space, and combined with the standard growth status data to generate the real growth status space of the standard staged growth period, ensuring that the basic data of the prediction model is closer to the actual situation. Finally, a reinforced prediction adjustment model was constructed through reinforcement learning algorithm for training, and the vegetation index and plant height growth fitting function used in the prediction were dynamically adjusted using real-time crop growth data to predict the results. This closed-loop feedback mechanism not only takes into account the differences between different growth stages, but also quantifies the specific impact of water and pests and diseases on crops at different growth stages, greatly improving the prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 This is a flow chart of a crop growth prediction method based on multispectral UAV monitoring according to Example 1 of the present invention;

[0061] Figure 2 This is a module diagram of a crop growth prediction system based on multispectral UAV monitoring according to Example 2 of the present invention. DETAILED DESCRIPTION

[0062] Example 1

[0063] Growth refers to the condition and trend of crop growth. Crop growth can be described by individual and group characteristics. A reasonable group composed of healthy and robust individuals is a crop area with good growth. Individual characteristics can be described by the characteristics of stems, leaves, roots and ears. Such as plant height, number of stalks, number, shape, color of leaves, root development, etc. Group characteristics can be described by group density, leaf area index, layout and dynamics. Existing algorithms only use single NDVI and image data to predict the growth status of crops, but do not take into account that the NDVI corresponding to crops at different growth stages is different, and do not take into account the specific effects of water, diseases and pests on growth status, resulting in low prediction accuracy. For this reason, please refer to Figure 1 The present invention provides an embodiment of a crop growth prediction method based on multispectral UAV monitoring, comprising the following steps:

[0064] S1. Configure crop growth status data for each growth stage under a standard growth environment, and obtain multispectral data, crop LiDAR data, and hyperspectral data of real crops at different growth stages, different soil moisture content, and pest and disease conditions;

[0065] Furthermore, in this embodiment, the growth status data of the crops in the standard growth environment in different growth stages are obtained in the laboratory using the growth environment required by the corresponding crop standards, and can also be obtained from corresponding research papers.

[0066] Furthermore, in this embodiment, data acquisition includes:

[0067] Standard growth environment data: By consulting professional research papers, we can obtain wheat growth status data at different stages under a standard growth environment, such as plant height, leaf number, biomass, and other indicators at each stage, including the tillering stage, jointing stage, and heading stage.

[0068] Real-world crop data: Wheat was grown in experimental fields under varying soil moisture levels (e.g., drought, normal, and moist) and pest and disease conditions (no pests and diseases, mild pests and diseases, and severe pests and diseases). Multispectral cameras, LiDAR equipment, and hyperspectral instruments were used to obtain data at different growth stages.

[0069] S2. Obtain a multi-vegetation index space, and calculate the contribution of each vegetation index to the growth simulation of the corresponding type of crop at different growth stages based on a factor analysis algorithm, and obtain the matching vegetation index for each growth stage of the corresponding type of crop;

[0070] Existing data show that the leaves of green plants are the basic organs of photosynthesis. Under the same conditions, the larger the leaf area, the stronger the photosynthesis. The stronger the photosynthesis, the more dry matter accumulation, and the higher the biomass. This physiological mechanism can be indirectly confirmed from the analysis of remote sensing data through the relationship between different bands in the plant reflectance spectrum, such as the algebraic combination of red light and near-infrared band reflectance information to form various vegetation indices. Numerous studies have shown that the Normalized Difference Vegetation Index (NDVI) has a good correlation with the leaf area index, biomass, and chlorophyll content of plants, and is an important indicator of plant growth status and vegetation spatial distribution density. Therefore, in this embodiment, the contribution of different vegetation index spaces to different crops at different biological stages is analyzed to obtain a vegetation index that matches each type of crop at each growth stage, thereby improving the accuracy of the prediction.

[0071] Furthermore, the multi-vegetation index space in this embodiment includes ratio vegetation index, normalized vegetation index, difference vegetation index, enhanced vegetation index, soil-adjusted vegetation index, greenness normalized vegetation index, triangular vegetation index, vertical vegetation index, red edge vegetation index, etc.;

[0072] Furthermore, in this embodiment, a vegetation index matrix is constructed: band reflectance data is extracted from the multispectral data, and multiple vegetation indices such as the ratio vegetation index (RVI), normalized difference vegetation index (NDVI), and difference vegetation index (DVI) are calculated to construct a standardized vegetation index matrix containing N = 10 vegetation indices. For example, when calculating NDVI, assuming the reflectance of the near-infrared band is 0.6 and the reflectance of the red band is 0.2, the NDVI is 0.5;

[0073] Furthermore, in this embodiment, the steps of obtaining the vegetation index matching each growth stage include:

[0074] S201. Extracting the band reflectance data required for calculating each vegetation index by acquiring multispectral data of real crops at different growth stages, different soil moisture contents, and pest and disease conditions, calculating different vegetation indices, and constructing a standardized vegetation index matrix containing N vegetation indices;

[0075] S202: Set each category of crops to include m growth stages, and select q crop samples from each stage. The sample space of

[0076] S203, according to The sample space and standardized vegetation index matrix are input into the factor analysis algorithm to obtain N vegetation indices. The covariance matrix C of the sample points;

[0077] S204, calculating the eigenvector and eigenvalue of the corresponding vegetation index using the characteristic equation according to the obtained covariance matrix C;

[0078] S205. Calculate the contribution rate and factor loading matrix of the nth vegetation index to the kth type of crop at the mth growth stage based on the eigenvector and eigenvalue of the corresponding vegetation index;

[0079] S206: Setting a contribution rate threshold. Based on the contribution rates of the N vegetation indices obtained for the crops in the corresponding category at the mth growth stage and the contribution rate threshold, the vegetation index corresponding to the crop with a cumulative contribution rate greater than the contribution rate threshold is retained as the vegetation index corresponding to the current growth stage of the crop.

[0080] S207, according to the factor loading matrix corresponding to the retained vegetation index, calculating the contribution of the corresponding vegetation index by summing the squares of the corresponding elements in the factor loading matrix;

[0081] S208. Calculate the corresponding vegetation index weighting coefficient using the contribution degree and contribution rate of the corresponding vegetation index, and use the calculated vegetation index weighting coefficient and the vegetation index to obtain a comprehensive matching vegetation index for the current growth stage of the corresponding crop.

[0082] This process uses a sophisticated factor analysis algorithm to achieve accurate acquisition of vegetation indices matching different growth stages of crops, significantly improving the accuracy and reliability of crop growth status prediction. Specifically, first, the band reflectance is extracted from real crop multispectral data and multiple vegetation indices are calculated to construct a standardized vegetation index matrix to ensure data consistency and comparability; secondly, by setting the sample space and inputting the factor analysis algorithm, the covariance matrix, eigenvectors and eigenvalues are obtained, which not only considers the diversity of different growth stages, but also quantifies the importance of each vegetation index, solving the problem in existing technologies that a single NDVI cannot adapt to different growth stages; thirdly, by calculating the contribution rate and factor loading matrix, and according to the contribution The contribution threshold screens out key vegetation indices, ensuring that the selected vegetation index has the highest explanatory power. In particular, the contribution is calculated by summing the squares of the elements in the factor loading matrix, which further enhances the scientificity and rationality of the selection. Finally, the weighted coefficient and contribution of the vegetation index are used to obtain a comprehensive matching vegetation index, achieving the best vegetation index matching for each growth stage. This refined matching method not only improves the prediction accuracy, but also dynamically reflects the specific impact of water and pests and diseases, greatly improving the adaptability and practicality of the model. In summary, this process solves the problem in the existing technology that a single NDVI cannot adapt to different growth stages through the systematic application of the factor analysis algorithm, and achieves high-precision vegetation index matching.

[0083] S3. Input the hyperspectral data of different soil moisture contents and pest and disease status into the principal component analysis algorithm for dimensionality reduction, and fuse the hyperspectral data after dimensionality reduction with the corresponding multispectral data to obtain the multispectral resolution feature space of different growth stages;

[0084] In this embodiment, the rich spectral information of hyperspectral data is integrated into the multispectral image to enhance the spectral resolution and feature expression ability of the image; for example, the hyperspectral data is fused with the multispectral data after dimensionality reduction through methods such as principal component analysis, or the mapping relationship between hyperspectral and multispectral data is automatically learned using a deep learning algorithm to generate a fused high-resolution and high-spectral resolution image.

[0085] S4. Utilize the multispectral resolution feature space of different growth stages to construct vegetation index water and pest and disease regulation factors corresponding to different growth stages. At the same time, through crop LiDAR data and water and pest and disease regulation factors, a simulation algorithm is used to obtain the actual plant height growth fitting function of crops under different growth states.

[0086] Furthermore, the steps for constructing the water and pest and disease regulating factors include:

[0087] S401, select from the multi-spectral resolution feature space containing m growth stages according to the S3 process The corresponding samples under different moisture conditions are selected, and the growth evaluation scores under the corresponding moisture conditions are obtained through the evaluation algorithm; p represents p soil moisture levels; the selection process is selected by those skilled in the art based on expert experience, and will not be further described here;

[0088] S402. Calculate vegetation indices corresponding to different moisture conditions based on the selected samples. In this embodiment, the vegetation indices of the selected samples are calculated using the standard vegetation index calculation formula corresponding to the corresponding growth stage obtained in S201-S208.

[0089] S403, obtaining a regression function of the effect of different water conditions on crop growth using a multi-source linear regression algorithm based on vegetation indices corresponding to different water conditions and growth assessment scores under the corresponding water conditions;

[0090] S404, using the coefficient of the regression function for crop growth as a regulating factor of water on crop growth to obtain a water regulating factor sequence;

[0091] S405, inputting the water regulation factor sequence into a genetic algorithm for optimization search to obtain the optimal water regulation factor sequence corresponding to different growth stages;

[0092] Furthermore, let's take an example of the above process:

[0093] Constructing moisture regulating factors:

[0094] Sample selection and evaluation: From the multispectral resolution feature space containing m = 5 growth stages, 30 samples are selected for each of the three water states: drought, normal, and wet. Through an evaluation algorithm (such as a comprehensive score based on crop growth indicators), the growth evaluation score under the corresponding water state is obtained. Assume that the drought state scores 60, the normal state scores 80, and the wet state scores 70.

[0095] Calculate vegetation index: Calculate the vegetation index corresponding to different moisture conditions based on the selected samples.

[0096] Regression calculation and sequence generation: Through the multi-source linear regression algorithm, with vegetation index as the independent variable and growth assessment score as the dependent variable, the regression function of different water conditions on crop growth is obtained. Assuming the regression function , then the coefficient of the regression function, 0.5, is used as the regulating factor of water on crop growth, and the water regulating factor sequence is obtained. x is the independent variable of the regression function.

[0097] S406, select from the multi-spectral resolution feature space containing m growth stages according to the S3 process Different disease states and The corresponding samples under different pest conditions are obtained through the evaluation algorithm, and the growth evaluation scores under the corresponding disease conditions and pest conditions are obtained; where u represents u types of disease state samples, and v represents v types of pest state samples;

[0098] S407, repeating the processes of S402-S405 to obtain the optimal disease regulating factor sequence and the optimal pest regulating factor sequence under different disease states and pest states corresponding to different growth stages;

[0099] S408. Construct a water and pest regulation factor space based on the optimal water regulation factor sequence, the optimal disease regulation factor sequence, and the optimal pest regulation factor sequence.

[0100] This process constructs accurate water and pest regulation factors through principal component analysis (PCA) dimensionality reduction and multi-source linear regression algorithm, combined with genetic algorithm optimization, significantly improving the accuracy and adaptability of crop growth status prediction; specifically, first, the hyperspectral data of different soil moisture content and pest and disease status are PCA-reduced and fused with multispectral data to generate a multispectral resolution feature space, ensuring efficient data processing and information retention; secondly, for samples under different water conditions, the growth assessment score is obtained through the evaluation algorithm, and the regression function of water on crop growth is calculated using the multi-source linear regression algorithm, and the regression coefficient is finally used as the water regulation factor. The introduction of genetic algorithms further optimized the sequence of water regulating factors, ensuring their optimal performance in each growth period. This process not only quantified the specific impact of water on crop growth, but also improved the robustness and accuracy of the model; thirdly, the optimal sequence of disease regulating factors and the optimal sequence of pest regulating factors were obtained respectively, comprehensively considering the impact of diseases and pests on crop growth; finally, the optimal water, disease and pest regulating factors were integrated to construct a complete space of water and disease and pest regulating factors, realizing all-round regulation of the crop growth environment; through systematic data analysis and optimization algorithms, the problem that a single factor in existing technologies cannot fully reflect the crop growth status is solved, and more accurate and dynamic regulating factors are provided, which greatly improves the accuracy of crop growth status prediction.

[0101] Furthermore, in this embodiment, the steps of constructing the real plant height growth fitting function include:

[0102] S409, constructing a three-dimensional model of standard heights corresponding to different growth stages of the crop using a three-dimensional algorithm based on LiDAR data of different growth stages of the crop and growth status data of different growth stages of the crop under a standard growth environment;

[0103] S410, based on the three-dimensional model of standard heights corresponding to different growth stages of crops and the space of water and pest and disease regulation factors corresponding to different growth stages, obtain a fitting function of the actual plant height growth corresponding to different growth stages of the corresponding type of crops through a simulation algorithm and a nonlinear fitting function;

[0104] S411. Obtain a prediction error based on the actual plant height data and the predicted data of the actual plant height growth fitting function, and feed the prediction error back to the simulation algorithm for adjustment to obtain a actual plant height growth fitting function with a prediction error less than a configured prediction error threshold, as the actual plant height growth fitting function for the corresponding growth stage of the current type of crop.

[0105] This process constructs an accurate fitting function for real plant height growth through three-dimensional modeling, simulation algorithms and error feedback mechanisms, significantly improving the accuracy and reliability of crop growth status predictions. First, using LiDAR data of crops at different growth stages and staged growth status data under a standard growth environment, a three-dimensional model of standard height for crops at different growth stages is constructed through a three-dimensional algorithm. This model provides an accurate structural reference for subsequent simulations, ensuring the accuracy of the simulation results.

[0106] Secondly, combining the standard height three-dimensional model and the space of water and pest and disease regulation factors, the real plant height growth fitting function of the corresponding type of crops in different growth stages was obtained through simulation algorithms and nonlinear fitting functions. This method of comprehensively considering the influence of multiple factors not only improves the explanatory power of the model, but also can dynamically reflect changes in the actual growth environment.

[0107] S5. Multiply the vegetation index moisture and pest regulation factors corresponding to different growth stages with the vegetation index matching each growth stage to obtain a real growth vegetation index space, and multiply the real growth vegetation index space with the growth state data of the growth period in the standard growth environment to obtain the real growth state space of the standard growth period in each stage;

[0108] The real growth vegetation index space is obtained by multiplying the vegetation index corresponding to different growth stages, including moisture and pest regulation factors, with the matching vegetation index for each growth stage. For example, at the jointing stage, assuming the comprehensive matching vegetation index is 0.6 and the moisture regulation factor is 0.8, the multiplication results in a value of 0.48 in the real growth vegetation index space for this stage.

[0109] S6. Input the actual growth state space of the standard staged growth period and the actual plant height growth fitting function of the crop under different growth states into the reinforcement prediction and adjustment model constructed by the reinforcement learning algorithm for training. A trained reinforcement prediction and adjustment model is obtained, and the crop growth state is predicted using the real-time crop growth data. Furthermore, in this embodiment, the prediction results include the predicted plant height at the heading stage, the number of leaves, and the final number of ears of crop.

[0110] Furthermore, in this embodiment, the steps of building and training the enhanced prediction adjustment model include:

[0111] S601, constructing the current input state based on the acquired water and pest and disease regulation factor space, the growth state of the staged growth period under the standard growth environment, and the actual plant height growth fitting function;

[0112] S602: Set the current moment trigger information according to the current moment input state, specifically:

[0113] If the current crop is the kth type of crop in the mth growth stage, the action execution information is triggered. That is, the water and pest regulation factors of the kth type of crop in the mth growth stage are called from the water and pest regulation factor space, and the plant height growth fitting function and the corresponding comprehensive matching vegetation index are adjusted and corrected. The corrected plant height growth fitting function and the corresponding comprehensive matching vegetation index are built into the corresponding time series prediction sub-model to predict the current crop growth status;

[0114] S603: Constructing an action to be executed at the current moment based on the trigger information at the current moment, and constructing a reward function based on the error between the predicted result and the actual growth state, the plant height growth fitting function, and the correction loss of the corresponding comprehensive matching vegetation index;

[0115] S604: Input the current input state, current trigger information, current execution action, and reward function into the reinforcement prediction and adjustment model for training, obtain a trained reinforcement prediction and adjustment model, and output the crop growth state corresponding to the current moment.

[0116] This process achieves high-precision dynamic prediction of crop growth status through the reinforced prediction and adjustment model constructed by the reinforcement learning algorithm, significantly improving the scientificity and efficiency of agricultural management. Specifically, first, the vegetation index moisture and pest and disease regulation factors of different growth stages are multiplied with the matching vegetation index to obtain the real growth vegetation index space, and further multiplied with the growth status data of the staged growth period under the standard growth environment to generate the standard staged growth period real growth status space; this process ensures that the vegetation index and growth status of each growth stage are accurately adjusted to reflect the actual growth conditions. Secondly, the reinforced prediction and adjustment model constructed by the reinforcement learning algorithm is trained in combination with the real growth state space of the standard staged growth period and the real plant height growth fitting function. It not only takes into account the influence of multiple factors, but also makes dynamic adjustments through real-time crop growth data, thereby improving the accuracy and adaptability of the prediction. In particular, when the crop is in a specific growth stage, the model will call the corresponding water and pest and disease regulation factors to correct the plant height growth fitting function and the comprehensive matching vegetation index, and build them into the time series prediction sub-model for prediction. This closed-loop feedback mechanism continuously adjusts the model parameters by comparing the error between the prediction results and the actual growth state to ensure that the prediction accuracy is always better than the set threshold.

[0117] Example 2

[0118] See also Figure 2 , another embodiment provided by the present invention: a crop growth prediction system based on multispectral UAV monitoring, comprising: a data acquisition module, an index matching module, a simulation module, a growth state space module and a prediction module;

[0119] The data acquisition module is used to configure the growth status data of crops in different growth stages under a standard growth environment, and to obtain multispectral data, crop LiDAR data and hyperspectral data of real crops under different growth stages, different soil moisture content and pest and disease conditions;

[0120] The index matching module is used to obtain the matching vegetation index of each growth stage of the corresponding type of crop; the index matching module includes an index space construction unit and a matching analysis unit;

[0121] The index space construction unit is used to obtain different types of vegetation indices and construct a multi-vegetation index space;

[0122] The matching analysis unit calculates the contribution of each vegetation index to the growth simulation of different growth stages of the corresponding type of crops through a factor analysis algorithm, and obtains the matching vegetation index of each growth stage of the corresponding type of crops;

[0123] The simulation module obtains the actual plant height growth fitting function of crops under different growth conditions through simulation algorithm;

[0124] The simulation module includes a feature fusion unit, a factor construction unit and a function simulation unit;

[0125] The feature fusion unit is used to input the hyperspectral data of different soil moisture contents and pest and disease status into the principal component analysis algorithm for dimensionality reduction, and fuse the hyperspectral data after dimensionality reduction with the corresponding multispectral data to obtain the multispectral resolution feature space of different growth stages;

[0126] The factor construction unit uses the multi-spectral resolution feature space of different growth stages to construct vegetation index moisture and pest and disease regulation factors corresponding to different growth stages;

[0127] The function simulation unit obtains the actual plant height growth fitting function of crops under different growth conditions through a simulation algorithm based on crop LiDAR data and water and pest and disease regulation factors;

[0128] The growth state space module is used to multiply the vegetation index moisture and pest and disease regulation factors corresponding to the same growth stage with the vegetation index matching each growth stage to obtain a real growth vegetation index space, and multiply the real growth vegetation index space with the growth state data of the staged growth period under the standard growth environment to obtain the real growth state space of the standard staged growth period;

[0129] The prediction module is used to input the real growth state space of the standard staged growth period and the real plant height growth fitting function of crops under different growth states into the reinforcement prediction adjustment model constructed by the reinforcement learning algorithm for training, obtain the trained reinforcement prediction adjustment model, and predict the crop growth state through the crop growth data obtained in real time.

[0130] Example 3

[0131] An electronic device includes a memory and a processor, wherein the memory stores a computer program, and the processor implements a crop growth prediction method based on multispectral unmanned aerial vehicle monitoring when executing the computer program.

[0132] A computer-readable storage medium stores computer instructions, which, when executed, execute a crop growth prediction method based on multispectral unmanned aerial vehicle monitoring.

[0133] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also change, modify, replace and modify the above-mentioned embodiments without departing from the scope of protection of the purpose of the present invention and the claims, and all of these are protected by the present invention.

Claims

1. A crop growth prediction method based on multispectral UAV monitoring is characterized by: include: S1. Configure crop growth status data for each growth stage under a standard growth environment, and obtain multispectral data, crop LiDAR data, and hyperspectral data of real crops at different growth stages, different soil moisture content, and pest and disease conditions; S2. Obtain a multi-vegetation index space, and calculate the contribution of each vegetation index to the growth simulation of the corresponding type of crop at different growth stages based on a factor analysis algorithm, and obtain the matching vegetation index for each growth stage of the corresponding type of crop; S3. Input the hyperspectral data of different soil moisture contents and pest and disease status into the principal component analysis algorithm for dimensionality reduction, and fuse the hyperspectral data after dimensionality reduction with the corresponding multispectral data to obtain the multispectral resolution feature space of different growth stages; S4. Utilize the multispectral resolution feature space of different growth stages to construct vegetation index water and pest and disease regulation factors corresponding to different growth stages. At the same time, through crop LiDAR data and water and pest and disease regulation factors, a simulation algorithm is used to obtain the actual plant height growth fitting function of crops under different growth states. S5. Multiply the vegetation index moisture and pest regulation factors corresponding to different growth stages with the vegetation index matching each growth stage to obtain a real growth vegetation index space, and multiply the real growth vegetation index space with the growth state data of the growth period in the standard growth environment to obtain the real growth state space of the standard growth period in each stage; S6. Input the real growth state space of the standard staged growth period and the real plant height growth fitting function of crops under different growth states into the reinforcement prediction and adjustment model constructed by the reinforcement learning algorithm for training, obtain the trained reinforcement prediction and adjustment model, and predict the crop growth state through the crop growth data obtained in real time.

2. The crop growth prediction method based on multispectral UAV monitoring according to claim 1, characterized in that: The steps of obtaining the vegetation index matching each growth stage include: S201. Extracting the band reflectance data required for calculating each vegetation index by acquiring multispectral data of real crops at different growth stages, different soil moisture contents, and pest and disease conditions, calculating different vegetation indices, and constructing a standardized vegetation index matrix containing N vegetation indices; S202: Set each category of crops to include m growth stages, and select q crop samples from each stage. The sample space of S203, according to The sample space and standardized vegetation index matrix are input into the factor analysis algorithm to obtain N vegetation indices. The covariance matrix C of the sample points; S204 , calculating the eigenvector and eigenvalue of the corresponding vegetation index using the characteristic equation according to the obtained covariance matrix C.

3. The crop growth prediction method based on multispectral UAV monitoring according to claim 2, characterized in that: The step of obtaining the vegetation index matching each growth stage also includes: S205. Calculate the contribution rate and factor loading matrix of the nth vegetation index to the kth type of crop at the mth growth stage based on the eigenvector and eigenvalue of the corresponding vegetation index; S206: Setting a contribution rate threshold. Based on the contribution rates of the N vegetation indices obtained for the crops in the corresponding category at the mth growth stage and the contribution rate threshold, the vegetation index corresponding to the crop with a cumulative contribution rate greater than the contribution rate threshold is retained as the vegetation index corresponding to the current growth stage of the crop. S207, according to the factor loading matrix corresponding to the retained vegetation index, calculating the contribution of the corresponding vegetation index by summing the squares of the corresponding elements in the factor loading matrix; S208. Calculate the corresponding vegetation index weighting coefficient using the contribution degree and contribution rate of the corresponding vegetation index, and use the calculated vegetation index weighting coefficient and the vegetation index to obtain a comprehensive matching vegetation index for the current growth stage of the corresponding crop.

4. The crop growth prediction method based on multispectral UAV monitoring according to claim 3, characterized in that: The steps of constructing the moisture and pest and disease regulating factors include: S401, select from the multi-spectral resolution feature space containing m growth stages according to the S3 process The corresponding samples under different moisture conditions are obtained through the evaluation algorithm, and the growth evaluation scores under the corresponding moisture conditions are obtained; p represents p soil moisture levels; S402. Calculate vegetation indices corresponding to different moisture conditions based on the selected samples; S403, obtaining a regression function of the effect of different water conditions on crop growth using a multi-source linear regression algorithm based on vegetation indices corresponding to different water conditions and growth assessment scores under the corresponding water conditions; S404, using the coefficient of the regression function for crop growth as a regulating factor of water on crop growth to obtain a water regulating factor sequence; S405: Input the water regulation factor sequence into a genetic algorithm for optimization search to obtain the optimal water regulation factor sequence corresponding to different growth stages.

5. The crop growth prediction method based on multispectral UAV monitoring according to claim 4, characterized in that: The step of constructing the moisture and pest and disease regulating factor also includes: S406, select from the multi-spectral resolution feature space containing m growth stages according to the S3 process Different disease states and The corresponding samples under different pest conditions are obtained through the evaluation algorithm, and the growth evaluation scores under the corresponding disease conditions and pest conditions are obtained; where u represents u types of disease state samples, and v represents v types of pest state samples; S407, repeating the processes of S402-S405 to obtain the optimal disease regulating factor sequence and the optimal pest regulating factor sequence under different disease states and pest states corresponding to different growth stages; S408. Construct a water and pest regulation factor space based on the optimal water regulation factor sequence, the optimal disease regulation factor sequence, and the optimal pest regulation factor sequence.

6. The crop growth prediction method based on multispectral UAV monitoring according to claim 5, characterized in that: The steps of constructing the true plant height growth fitting function include: S409, constructing a three-dimensional model of standard heights corresponding to different growth stages of the crop using a three-dimensional algorithm based on LiDAR data of different growth stages of the crop and growth status data of different growth stages of the crop under a standard growth environment; S410, based on the three-dimensional model of standard heights corresponding to different growth stages of crops and the space of water and pest and disease regulation factors corresponding to different growth stages, obtain a fitting function of the actual plant height growth corresponding to different growth stages of the corresponding type of crops through a simulation algorithm and a nonlinear fitting function; S411. Obtain a prediction error based on the actual plant height data and the predicted data of the actual plant height growth fitting function, and feed the prediction error back to the simulation algorithm for adjustment to obtain a actual plant height growth fitting function with a prediction error less than a configured prediction error threshold, as the actual plant height growth fitting function for the corresponding growth stage of the current type of crop.

7. The crop growth prediction method based on multispectral UAV monitoring according to claim 6, characterized in that: The steps of building and training the enhanced prediction adjustment model include: S601, constructing the current input state based on the acquired water and pest and disease regulation factor space, the growth state of the staged growth period under the standard growth environment, and the actual plant height growth fitting function; S602: Set the current moment trigger information according to the current moment input state, specifically: If the current crop is the kth type of crop in the mth growth stage, the action execution information is triggered. That is, the water and pest regulation factors of the kth type of crop in the mth growth stage are called from the water and pest regulation factor space, and the plant height growth fitting function and the corresponding comprehensive matching vegetation index are adjusted and corrected. The corrected plant height growth fitting function and the corresponding comprehensive matching vegetation index are built into the corresponding time series prediction sub-model to predict the current crop growth status; S603: Constructing an action to be executed at the current moment based on the trigger information at the current moment, and constructing a reward function based on the error between the predicted result and the actual growth state, the plant height growth fitting function, and the correction loss of the corresponding comprehensive matching vegetation index; S604: Input the current input state, current trigger information, current execution action, and reward function into the reinforcement prediction and adjustment model for training, obtain a trained reinforcement prediction and adjustment model, and output the crop growth state corresponding to the current moment.

8. A crop growth prediction system based on multispectral UAV monitoring, which is used to implement the crop growth prediction method based on multispectral UAV monitoring according to any one of claims 1 to 7, characterized in that: include: Data acquisition module, index matching module; The data acquisition module is used to configure the growth status data of crops in different growth stages under a standard growth environment, and to obtain multispectral data, crop LiDAR data and hyperspectral data of real crops under different growth stages, different soil moisture content and pest and disease conditions; The index matching module includes an index space construction unit and a matching analysis unit; The index space construction unit is used to obtain different types of vegetation indices and construct a multi-vegetation index space; The matching analysis unit calculates the contribution of each vegetation index to the growth simulation of different growth stages of the corresponding type of crops through a factor analysis algorithm, and obtains the matching vegetation index of each growth stage of the corresponding type of crops.

9. The crop growth prediction system based on multispectral UAV monitoring according to claim 8, characterized in that: The system also includes a simulation module; The simulation module includes a feature fusion unit, a factor construction unit and a function simulation unit; The feature fusion unit is used to input the hyperspectral data of different soil moisture contents and pest and disease status into the principal component analysis algorithm for dimensionality reduction, and fuse the hyperspectral data after dimensionality reduction with the corresponding multispectral data to obtain the multispectral resolution feature space of different growth stages; The factor construction unit utilizes the multi-spectral resolution feature space of different growth stages to construct vegetation index moisture and pest and disease regulation factors corresponding to different growth stages.

10. The crop growth prediction system based on multispectral UAV monitoring according to claim 9, characterized in that: The system also includes a growth state space module and a prediction module; The growth state space module is used to multiply the vegetation index moisture and pest and disease regulation factors corresponding to the same growth stage with the vegetation index matching each growth stage to obtain a real growth vegetation index space, and multiply the real growth vegetation index space with the growth state data of the staged growth period under the standard growth environment to obtain the real growth state space of the standard staged growth period; The prediction module is used to input the real growth state space of the standard staged growth period and the real plant height growth fitting function of crops under different growth states into the reinforcement prediction adjustment model constructed by the reinforcement learning algorithm for training, obtain the trained reinforcement prediction adjustment model, and predict the crop growth state through the crop growth data obtained in real time.

Images