Field intelligent prediction method for water content of corn kernels

Abstract

The invention relates to the field of corn water content prediction, in particular to a corn kernel water content field intelligent prediction method, which comprises the following steps of: acquiring a corn plant spectral image and corn kernel water content data, and performing stability screening on spectral pixel reflectivity data to obtain a preliminary spectral band screening result; performing band weight analysis on the preliminary spectral band screening result to obtain the importance degree of each spectral band in the preliminary spectral band screening result; performing weighted optimization on the spectral pixel reflectivity data through the importance degree of each spectral band, performing PLSR model establishment through the optimized spectral pixel reflectivity data and the corn kernel water content data, and performing corn kernel water content prediction on the corn field spectral data through a corn kernel water content prediction model. Therefore, the prediction precision of the water content of corn kernels is improved.

Description

Technical Field

[0001] The invention relates to the technical field, and in particular to a field intelligent prediction method for corn grain moisture content. Background Art

[0002] Corn is an important food crop in my country, and its grain moisture content is a key factor affecting yield, quality, storage and processing. Accurate and rapid monitoring of corn grain moisture content is of great significance for guiding field management (such as irrigation and harvesting), optimizing storage conditions, and improving processing efficiency. Traditional methods for measuring corn grain moisture content (such as drying) are time-consuming and lossy, and cannot achieve large-scale, rapid, and non-destructive monitoring. In recent years, with the development of remote sensing technology, drones equipped with multispectral sensors have acquired spectral information of corn canopies, and then inverted crop physiological parameters, which has become an important precision agriculture technology.

[0003] Multispectral remote sensing technology reflects the growth status and physiological parameters of crops by obtaining the reflectance information of crops in several specific bands (usually visible light, near infrared and short-wave infrared bands). Partial least squares regression (PLSR) is a commonly used multivariate statistical analysis method that can effectively deal with multicollinearity problems and has been widely used in spectral data analysis and modeling. Some studies have attempted to combine the PLSR method with multispectral remote sensing data to predict the moisture content or other physiological parameters of crops.

[0004] Although partial least squares regression ( ) method has been widely used in spectral data analysis, and multispectral remote sensing technology has shown great potential in monitoring crop physiological parameters. However, in the specific application scenario of using drone multispectral images to predict corn kernel moisture content, the existing The algorithm and its commonly used variable selection methods still have limitations. This results in limited model prediction accuracy, generalization ability, and interpretability. Specifically, although the number of bands of drone multispectral images is relatively small, they still contain multiple bands, some of which are closely related to the moisture content of corn kernels, while others contain noise, redundant information, or have a weak relationship with moisture content. Traditional The algorithm treats all bands equally and does not fully consider the contribution differences between bands, which makes the model vulnerable to noise and redundant information. More importantly, existing variable selection methods (such as uninformative variable elimination) , Competitive Adaptive Reweighted Sampling , variable projection importance When applied to multispectral data, it is often difficult to take into account the linear relationship between variables and water content, the nonlinear relationship, and the correlation between variables. For example, The method mainly selects variables based on the stability of regression coefficients, which tends to ignore variables that are nonlinearly related to the target variable; Although this method simulates the principle of "survival of the fittest", the selection process has a certain degree of randomness, which can easily lead to unstable results; The method can measure the contribution of variables to the overall model, but does not consider the correlation between variables, resulting in redundancy in variable selection. The corn kernel moisture content prediction model based on multispectral data has problems such as overfitting, insufficient generalization ability, and poor model interpretability. It is difficult to meet the needs of precision agriculture for high-precision and high-reliability monitoring. Therefore, how to improve The variable selection method in the algorithm enables it to more effectively extract information related to corn kernel moisture content from multispectral data. Reducing the impact of noise and redundant information and taking into account the linear and nonlinear relationship between variables and moisture content are the keys to improving the accuracy and interpretability of the prediction model. Summary of the invention

[0005] In view of this, the present invention aims to propose a field intelligent prediction method for corn kernel moisture content to optimize the variable screening of corn spectral data and thus improve the accuracy of the prediction model.

[0006] To achieve the above object, the technical solution of the present invention is achieved as follows:

[0007] A method for intelligent field prediction of corn kernel moisture content comprises the following steps: collecting corn plant spectral images and corn kernel moisture content data to obtain spectral pixel reflectance data and corn kernel moisture content data corresponding to the pixels; performing stability screening on the spectral pixel reflectance data to obtain preliminary spectral band screening results; performing band weight analysis on the preliminary spectral band screening results to obtain the importance of each spectral band in the preliminary spectral band screening results; performing weighted optimization on the spectral pixel reflectance data according to the importance of each spectral band to obtain optimized spectral pixel reflectance data; and establishing a PLSR model using the optimized spectral pixel reflectance data and corn kernel moisture content data to obtain a corn kernel moisture content prediction model; and performing corn kernel moisture content prediction on corn field spectral data using the corn kernel moisture content prediction model.

[0008] Further, according to the stability screening of the spectral pixel reflectance data, a preliminary spectral band screening result is obtained, and the specific steps include:

[0009] Acquire spectral pixel reflectance data of corn plants and corn kernel moisture content data corresponding to each pixel, and perform nonlinear transformation on the spectral pixel reflectance data of corn plants to obtain transformed spectral pixel reflectance data of corn plants;

[0010] The regression coefficients of each band in the spectral pixel reflectance data of corn plants and the spectral pixel reflectance data of corn plants after transformation are evaluated respectively, and the regression coefficients of each band in the spectral pixel reflectance data of corn plants and the spectral pixel reflectance data of corn plants after transformation are obtained respectively;

[0011] By evaluating the stability of the regression coefficient of each band in the spectral pixel reflectance data of corn plants and the spectral pixel reflectance data of corn plants after transformation, the stability of each band in the spectral pixel reflectance data of corn plants and the spectral pixel reflectance data of corn plants after transformation is obtained respectively;

[0012] By comparing and analyzing the stability of each band in the spectral pixel reflectance data of the corn plant and the transformed spectral pixel reflectance data of the corn plant, a stability difference evaluation of each band in the spectral pixel reflectance data of the corn plant is obtained;

[0013] By mapping and analyzing the stability difference evaluation results, a band screening threshold is obtained, and the stability of each band in the corn plant spectral pixel reflectance data is compared with the band screening threshold to obtain a preliminary spectral band screening result.

[0014] Further, according to the above, by evaluating the regression coefficients of each band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation, the regression coefficients of each band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation are obtained respectively, and the specific steps include:

[0015] Acquire the spectral pixel reflectance data of the corn plant, the spectral pixel reflectance data of the corn plant after transformation, and the corn kernel moisture content data corresponding to each pixel, establish a first PLSR model through the spectral pixel reflectance data of the corn plant and the corn kernel moisture content data corresponding to each pixel, and obtain the regression coefficient vector of the first PLSR model; establish a second PLSR model through the spectral pixel reflectance data of the corn plant after transformation and the corn kernel moisture content data corresponding to each pixel, and obtain the regression coefficient vector of the second PLSR model; by repeatedly establishing the first PLSR model and establishing the second PLSR model, obtain multiple groups of regression coefficient vectors of the first PLSR model and multiple groups of regression coefficient vectors of the second PLSR model, calculate the mean of the multiple groups of regression coefficient vectors of the first PLSR model and the multiple groups of regression coefficient vectors of the second PLSR model, obtain the first regression coefficient vector of the first PLSR model and the first regression coefficient vector of the second PLSR model, thereby obtaining the regression coefficient of each band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation.

[0016] Further, according to the stability evaluation of the regression coefficient of each band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation, the stability corresponding to each band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation is obtained, and the specific steps include:

[0017] The absolute value calculation result of the mean of the regression coefficient of any band in the spectral pixel reflectance data of the corn plant is used as the first mean of the band, and the standard deviation of the regression coefficient of any band in the spectral pixel reflectance data of the corn plant is divided by the first mean of the band as the stability of the band in the spectral pixel reflectance data of the corn plant; the absolute value calculation result of the mean of the regression coefficient of any band in the spectral pixel reflectance data of the corn plant after transformation is used as the first mean of the band, and the standard deviation of the regression coefficient of any band in the spectral pixel reflectance data of the corn plant after transformation is divided by the first mean of the band as the stability of the band in the spectral pixel reflectance data of the corn plant after transformation.

[0018] Further, according to the comparative analysis of the stability corresponding to each band in the spectral pixel reflectance data of the corn plant and the transformed spectral pixel reflectance data of the corn plant, the stability difference evaluation of each band in the spectral pixel reflectance data of the corn plant is obtained, and the specific steps include:

[0019] The stability of any band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation is obtained, and the stability of any band in the spectral pixel reflectance data of the corn plant after transformation is divided by the stability of the band in the spectral pixel reflectance data of the corn plant as a calculation result of the stability difference evaluation of the band in the spectral pixel reflectance data of the corn plant.

[0020] Further, according to the mapping analysis of the stability difference evaluation result, a band screening threshold is obtained, and the stability of each band in the corn plant spectral pixel reflectance data is compared with the band screening threshold to obtain a preliminary spectral band screening result, the specific steps of which include:

[0021] Obtaining a stability difference assessment of each band in the spectral pixel reflectance data of the corn plant, setting a first mapping parameter and a set adjustment coefficient, taking the mean of the stability difference assessments of all bands in the spectral pixel reflectance data of the corn plant as a second mean, taking a calculation result of subtracting the stability difference assessment of each band in the spectral pixel reflectance data of the corn plant from the second mean as a first mean difference, performing an exponential mapping with a natural constant as the base on the inverse of a calculation result of multiplying the first mapping parameter by the first mean difference to obtain a first exponential mapping; taking a calculation result of adding a constant 1 to the first exponential mapping as a threshold optimization factor;

[0022] Obtaining the stability of each band in the spectral pixel reflectance data of the corn plant, taking the mean of the stability of all bands in the spectral pixel reflectance data of the corn plant as the third mean, taking the standard deviation of the stability of all bands in the spectral pixel reflectance data of the corn plant as the first standard deviation, and taking the third mean plus the adjustment coefficient multiplied by the first standard deviation as the basic screening threshold;

[0023] The calculation result of multiplying the threshold optimization factor by the basic screening threshold is used as the band screening threshold; the stability of each band in the spectral pixel reflectance data of the corn plant is compared with the band screening threshold; if the stability of the band in the spectral pixel reflectance data of the corn plant is less than or equal to the band screening threshold, the band data is retained to obtain a preliminary spectral band screening result.

[0024] Further, according to the band weight analysis of the preliminary spectral band screening result, the importance of each spectral band in the preliminary spectral band screening result is obtained, and the specific steps include:

[0025] The spectral band screening results of the spectral pixel reflectance data of the corn plant are obtained, and the spectral band screening results are integrated into a spectral matrix. By performing correlation analysis on the spectral matrix, a correlation coefficient matrix of the spectral matrix is ​​obtained; by performing eigenvalue analysis on the correlation coefficient matrix of the spectral matrix, the eigenvalue of each spectral band in the spectral matrix is ​​obtained; by performing feature difference evaluation on the characteristics of each spectral band in the spectral matrix, the importance of each spectral band in the preliminary spectral band screening results is obtained.

[0026] Furthermore, according to the feature difference evaluation of the features of each spectral band in the spectral matrix, the importance of each spectral band in the preliminary spectral band screening result is obtained, and the specific steps include:

[0027] Acquire the correlation coefficient matrix and the importance adjustment parameter; perform matrix eigenvector extraction on the correlation coefficient matrix to obtain the eigenvalue vector of the correlation coefficient matrix; use the difference between the eigenvalue of any spectral band in the eigenvalue vector and the mean of all eigenvalues ​​in the eigenvalue vector of the correlation coefficient matrix as the second mean difference; perform exponential mapping with a natural constant as the base on the result of multiplying the importance adjustment parameter by the second mean difference to obtain a second exponential mapping; use the inverse of the result of adding the constant 1 to the second exponential mapping as the importance of the spectral band.

[0028] Compared with the prior art, the present invention has the following advantages:

[0029] The present invention discloses a method for intelligent field prediction of corn kernel moisture content. The method comprises the following steps: collecting corn plant spectral images and corn kernel moisture content data to obtain spectral pixel reflectance data and corn kernel moisture content data corresponding to the pixels, and performing stability screening on the spectral pixel reflectance data to obtain preliminary spectral band screening results; performing band weight analysis on the preliminary spectral band screening results to obtain the importance of each spectral band in the preliminary spectral band screening results; performing weighted optimization on the spectral pixel reflectance data according to the importance of each spectral band to obtain optimized spectral pixel reflectance data, and establishing a PLSR model with the optimized spectral pixel reflectance data and corn kernel moisture content data to obtain a corn kernel moisture content prediction model; and performing corn kernel moisture content prediction on corn field spectral data using the corn kernel moisture content prediction model. The spectral pixel reflectance data are screened for stability, and the bands that are greatly affected by environmental interference are eliminated to improve the reliability of the input data. Then, the contribution of each spectral band to the prediction of grain moisture content is quantified through weight analysis of each band, and weighted optimization is performed accordingly to enhance the influence of key spectral features on the prediction model, thereby improving the prediction accuracy of the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the accompanying drawings:

[0031] Figure 1 A method flow chart of a method for intelligent field prediction of corn grain moisture content according to an embodiment of the present invention; DETAILED DESCRIPTION

[0032] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0033] In the description of the present invention, it should be noted that the directions or positional relationships indicated by the terms "upper", "lower", "inner", "back", etc. are based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0034] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0035] See also Figure 1 , is a method flow chart of a method for intelligent field prediction of corn grain moisture content provided by the first embodiment of the present invention, such as Figure 1 The method for intelligent field prediction of corn kernel moisture content may include:

[0036] Step S101, collecting spectral images of corn plants and corn kernel moisture content data to obtain spectral pixel reflectance data and corn kernel moisture content data corresponding to the pixels.

[0037] Using drones equipped with multispectral sensors, aerial photography of the experimental fields was carried out during the critical period of corn growth. After preprocessing such as radiation calibration, atmospheric correction, and geometric correction, the reflectance value of each sampling point in each band was obtained to form a multispectral data matrix.

[0038] Ground sampling was carried out synchronously during the flight of the UAV. Corn kernel samples were collected and their moisture content was measured by drying method to obtain moisture content data, and the corresponding relationship between the moisture content data and the multispectral data matrix was determined.

[0039] Specifically, for the multispectral data matrix, each data in the matrix corresponds to the spectral data of a pixel point (pixel) in the multispectral image. For the moisture content data of the corn kernel samples, the average moisture content data of the corn kernel samples at the corresponding position of each pixel point in each spectral image is obtained as the actual measured value of the moisture content of the corn kernel corresponding to the pixel point.

[0040] Step S102, obtaining a preliminary spectral band screening result by performing stability screening on the spectral pixel reflectance data.

[0041] When using drone multispectral images to predict corn kernel moisture content, the existing The algorithm faces the challenge of variable selection. The traditional PLSR algorithm treats all spectral bands (variables) equally, without considering the difference in the contribution of different bands to water content prediction, which may cause the model to contain a lot of noise and redundant information, affecting the prediction accuracy, generalization ability and interpretability. To solve this problem, an effective variable selection method is needed. Uninformative variable elimination method ( ) is a commonly used The variable selection method is a method for model regression coefficient stability, but it mainly focuses on the linear relationship between the variable and the target variable. However, there is often a complex nonlinear relationship between spectral reflectance and crop physiological parameters, and the standard UVE method may mistakenly exclude variables that are nonlinearly related to water content.

[0042] To overcome the standard The limitation of the method is that it is necessary to identify variables that are nonlinearly related to water content by comparing the stability of the variables in the original spectral data and the spectral data after nonlinear transformation. In specific implementation, the original spectral data is first subjected to appropriate nonlinear transformation (such as logarithmic transformation, square root transformation, inverse transformation, etc., the specific choice depends on the characteristics of the data and prior knowledge) to transform the potential nonlinear relationship into a linear relationship, or enhance the existing nonlinear relationship. Then, the original spectral data and the transformed spectral data are respectively established. The model is constructed and the stability of each variable is calculated in both cases (by multiple modeling, the ratio of the standard deviation of the regression coefficient to the mean is calculated). If a variable shows higher stability on the transformed data than on the original data, it indicates that there is a nonlinear relationship between the variable and the water content and it should be retained. This method can evaluate the importance of variables more comprehensively and accurately, laying the foundation for the subsequent construction of a better PLSR prediction model.

[0043] The specific preliminary spectral band screening process includes:

[0044] The spectral pixel reflectance data of corn plants and the corn kernel moisture content data corresponding to each pixel are obtained, and the spectral pixel reflectance data of corn plants are nonlinearly transformed to obtain the spectral pixel reflectance data of corn plants after transformation; the regression coefficients of the spectral pixel reflectance data of corn plants and each band in the spectral pixel reflectance data of corn plants after transformation are evaluated respectively, and the regression coefficients of the spectral pixel reflectance data of corn plants and each band in the spectral pixel reflectance data of corn plants after transformation are obtained respectively; the regression coefficients of the spectral pixel reflectance data of corn plants and each band in the spectral pixel reflectance data of corn plants after transformation are evaluated respectively. A stability assessment was performed to obtain the corresponding stability of each band in the spectral pixel reflectance data of corn plants and the transformed spectral pixel reflectance data of corn plants; a stability difference assessment of each band in the spectral pixel reflectance data of corn plants was obtained by comparing and analyzing the corresponding stability of each band in the spectral pixel reflectance data of corn plants and the transformed spectral pixel reflectance data of corn plants; a band screening threshold was obtained by mapping and analyzing the stability difference assessment results, and the stability of each band in the spectral pixel reflectance data of corn plants was compared with the band screening threshold to obtain a preliminary spectral band screening result.

[0045] Among them, by evaluating the regression coefficient of each band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation, the regression coefficient of each band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation is obtained, including:

[0046] The spectral pixel reflectance data of corn plants are obtained, and the spectral pixel reflectance data of corn plants after transformation and the corn kernel moisture content data corresponding to each pixel are used to establish a first PLSR model through the spectral pixel reflectance data of corn plants and the corn kernel moisture content data corresponding to each pixel, and the regression coefficient vector of the first PLSR model is obtained; the second PLSR model is established through the spectral pixel reflectance data of corn plants after transformation and the corn kernel moisture content data corresponding to each pixel, and the regression coefficient vector of the second PLSR model is obtained; by repeatedly establishing the first PLSR model and establishing the second PLSR model, multiple groups of regression coefficient vectors of the first PLSR model and multiple groups of regression coefficient vectors of the second PLSR model are obtained, and the regression coefficient vector of the first PLSR model and the first regression coefficient vector of the second PLSR model are obtained by performing mean calculation on the regression coefficient vectors of the multiple groups of first PLSR models and the regression coefficient vectors of the multiple groups of second PLSR models, thereby obtaining the regression coefficient of each band in the spectral pixel reflectance data of corn plants and the spectral pixel reflectance data of corn plants after transformation.

[0047] It should be noted that these two The model is not used for the final water content prediction, but serves as an intermediate step for variable selection. Its role is to provide a preliminary assessment to help determine which variables have a strong linear or nonlinear relationship with water content. These two models provide relatively reliable information about the stability of the variables and support preliminary variable screening. and the corresponding measured water content, by calling the existing The algorithm library establishes the first Then, the original spectral data A nonlinear transformation (such as logarithmic transformation, square root transformation, etc.) is performed. In this embodiment, a nonlinear transformation process is performed by logarithmic transformation to obtain a transformed spectral data matrix; the transformed spectral data matrix and the measured water content are used to establish a second Model, and extract regression coefficient vectors from the two models respectively. In order to ensure the stability of the regression coefficient, the method of multiple modeling and averaging is adopted, that is, the above modeling process is repeated multiple times and the regression coefficient vector obtained each time is averaged, and finally the regression coefficient of each band in the spectral pixel reflectance data of corn plants and the transformed spectral pixel reflectance data of corn plants is obtained.

[0048] After obtaining the regression coefficient of each band in the spectral pixel reflectance data of the corn plant and the transformed spectral pixel reflectance data of the corn plant, it is necessary to evaluate the stability of the regression coefficient of each band in the spectral pixel reflectance data of the corn plant and the transformed spectral pixel reflectance data of the corn plant, and obtain the stability corresponding to each band in the spectral pixel reflectance data of the corn plant and the transformed spectral pixel reflectance data of the corn plant, respectively. The specific steps include:

[0049] The absolute value calculation result of the mean of the regression coefficient of any band in the spectral pixel reflectance data of the corn plant is used as the first mean of the band, and the standard deviation of the regression coefficient of any band in the spectral pixel reflectance data of the corn plant is divided by the first mean of the band as the stability of the band in the spectral pixel reflectance data of the corn plant; the absolute value calculation result of the mean of the regression coefficient of any band in the spectral pixel reflectance data of the corn plant after transformation is used as the first mean of the band, and the standard deviation of the regression coefficient of any band in the spectral pixel reflectance data of the corn plant after transformation is divided by the first mean of the band as the stability of the band in the spectral pixel reflectance data of the corn plant after transformation.

[0050] In one embodiment, it is assumed that the first The regression coefficient of each band is , the spectral pixel reflectance data of corn plants after transformation The regression coefficient of each band is , then the spectral pixel reflectance data of corn plants is The stability of each band is related to the transformed spectral pixel reflectance data of corn plants. The calculation expression of the stability of each band is:

[0051]

[0052]

[0053] in, Represents the spectral pixel reflectance data of corn plants The stability of each band; Represents the first pixel in the spectral pixel reflectance data of the corn plant after transformation The stability of each band; Represents the spectral pixel reflectance data of corn plants The standard deviation of the regression coefficients for each band; Represents the first pixel in the spectral pixel reflectance data of the corn plant after transformation The standard deviation of the regression coefficients for each band; Represents the spectral pixel reflectance data of corn plants The mean of the regression coefficients of the bands; Represents the first pixel in the spectral pixel reflectance data of the corn plant after transformation The mean of the regression coefficients for the bands.

[0054] It should be noted that and The stability of each variable on the original spectral data and the spectral data after nonlinear transformation was calculated. The idea of ​​the method is to use the ratio of the standard deviation of the regression coefficient to the mean (i.e., the coefficient of variation) to measure the stability of the variable. The coefficient of variation can eliminate the influence of the dimensions between different variables, making the stability of different variables comparable. The reason why the stability of the variables in the original data and the transformed data is calculated is that there is often a nonlinear relationship between spectral reflectance and water content. By introducing nonlinear transformation, the potential nonlinear relationship can be transformed into a linear relationship, or the strength of the nonlinear relationship can be enhanced to make it The model is better able to capture these relationships.

[0055] After obtaining the stability of each band in the spectral pixel reflectance data of the corn plant before and after the transformation, the stability of each band in the spectral pixel reflectance data of the corn plant can be compared and analyzed with that in the spectral pixel reflectance data of the corn plant after the transformation, and the stability difference evaluation of each band in the spectral pixel reflectance data of the corn plant can be obtained, including:

[0056] The stability of any band in the spectral pixel reflectance data of corn plants and the transformed spectral pixel reflectance data of corn plants is obtained, and the stability of any band in the spectral pixel reflectance data of corn plants after transformation is divided by the stability of the band in the spectral pixel reflectance data of corn plants, and the result is used as the stability difference evaluation of the band in the spectral pixel reflectance data of corn plants.

[0057] Finally, the band screening threshold was obtained by mapping and analyzing the stability difference evaluation results. The stability of each band in the spectral pixel reflectance data of corn plants was compared with the band screening threshold to obtain preliminary spectral band screening results, including:

[0058] Obtaining the stability difference evaluation of each band in the spectral pixel reflectance data of the corn plant, setting the first mapping parameter and the set adjustment coefficient, taking the mean of the stability difference evaluation of all bands in the spectral pixel reflectance data of the corn plant as the second mean, taking the calculation result of subtracting the stability difference evaluation of each band in the spectral pixel reflectance data of the corn plant from the second mean as the first mean difference, performing an exponential mapping with a natural constant as the base on the inverse of the calculation result of multiplying the first mapping parameter by the first mean difference to obtain a first exponential mapping; taking the calculation result of adding the constant 1 to the first exponential mapping as the threshold optimization factor;

[0059] Obtain the stability of each band in the spectral pixel reflectance data of the corn plant, take the mean of the stability of all bands in the spectral pixel reflectance data of the corn plant as the third mean, take the standard deviation of the stability of all bands in the spectral pixel reflectance data of the corn plant as the first standard deviation, and take the third mean plus the adjustment coefficient multiplied by the first standard deviation as the basic screening threshold;

[0060] The calculation result of multiplying the threshold optimization factor by the basic screening threshold is used as the band screening threshold; the stability of each band in the spectral pixel reflectance data of the corn plant is compared with the band screening threshold. If the stability of the band in the spectral pixel reflectance data of the corn plant is less than or equal to the band screening threshold, the band data is retained to obtain the preliminary spectral band screening results.

[0061] In one embodiment, it is assumed that the first The stability difference of the bands is evaluated as , then the spectral pixel reflectance data of corn plants is The calculation expression of the band screening threshold of a band is:

[0062]

[0063] in, Represents the spectral pixel reflectance data of corn plants The band filtering threshold for each band; represents the first mapping parameter; Represents the spectral pixel reflectance data of corn plants Stability difference assessment of each band; It represents the mean value of the stability difference assessment of all bands in the spectral pixel reflectance data of corn plants; It indicates the mean value of the stability of all variables on the original data; represents the adjustment factor; It indicates the standard deviation of all variables on the original data.

[0064] It should be noted that and The stability ratio of the variables is further calculated based on . It indicates the degree of change in the stability of the variable on the transformed data relative to the stability on the original data. A larger value indicates that the stability of the variable on the transformed data is significantly improved, which means that there is a high probability of a nonlinear relationship between the variable and the water content. This method only judges the importance of variables based on their stability on the original data and ignores nonlinear relationships. , taking nonlinear relationships into account in variable selection. Transforming into a variable selection threshold requires the introduction of an exponential mapping function ( ). It can The value is mapped to a When Much larger than the offset parameter When , the output of the exponential mapping function is close to ;when near When ;when Much smaller than When By introducing the exponential mapping function, The nonlinear transformation of the threshold The change of is smoother. In addition, the steepness parameter of the exponential mapping function Controlled threshold The sensitivity of The larger the value, the higher the threshold The more sensitive the offset parameter is. It can be fine-tuned according to the actual situation. , , parameter settings can be adjusted according to actual scenarios and are not required.

[0065] Finally, The output value of the function is multiplied by a baseline threshold to obtain the final variable selection threshold The benchmark threshold is the average value of the stability of all variables on the original data. and standard deviation Joint decision, i.e. The average value reflects the overall level and the standard deviation reflects the degree of dispersion. Adjustment coefficient Used to control the overall level of the threshold, The larger the value, the higher the threshold and the fewer variables are selected. For variables that are linearly related to water content, near , close to the benchmark threshold; for variables that are nonlinearly related to water content, Greater than , will be above the baseline threshold, making these variables more likely to be selected.

[0066] In this step, the nonlinear characteristics in the spectral data are enhanced through nonlinear transformation, and the responsiveness of the spectral information to the change of water content is improved; then, the contribution of different bands to the prediction of grain water content is identified by evaluating the regression coefficients of the original spectral data and the transformed spectral data, and the spectral bands with robust contribution to the prediction are further screened out through stability analysis to reduce noise interference. In addition, through stability difference evaluation and mapping analysis, the band screening threshold is accurately set to ensure that only spectral bands with stable and important contributions to water content prediction are retained, avoiding errors introduced by redundant bands or low-contribution bands, thereby improving the generalization ability of the model. Finally, a high-precision PLSR water content prediction model is established based on the optimized and screened spectral data, which can effectively improve the prediction accuracy of the spatial distribution of corn grain water content in the field, provide a scientific basis for precise harvesting and water management, reduce food losses caused by water differences, and improve the level of intelligence in agricultural production.

[0067] Step S103 , performing band weight analysis on the preliminary spectral band screening result to obtain the importance of each spectral band in the preliminary spectral band screening result.

[0068] After preliminary spectral band screening, some noise and variables (spectral bands) with weak correlation with water content have been eliminated. However, there may still be high correlation and redundancy between the remaining variables, which is mainly reflected in the following two aspects:

[0069] 1. Information redundancy caused by spectral continuity: Although the number of spectral bands of drone multispectral images is not as many as that of hyperspectral images, they also have a certain degree of continuity. The physical properties of adjacent bands (such as absorption and reflection of light) are often very similar, resulting in their spectral values ​​being highly correlated. This correlation means that there is a lot of redundancy in the information carried by adjacent bands. If these highly correlated bands are included The model will become too complex and prone to overfitting.

[0070] 2. Information overlap caused by latent factors: Even non-adjacent bands will show a certain correlation because they are all affected by some common factors (such as light, atmosphere, etc.) or are related to some latent variables (such as chlorophyll content, cell structure, etc.). This means that there is a problem of spectral information overlap in different bands. That is, multiple bands reflect the changes in the same latent factor. This information overlap will also lead to overfitting of the model.

[0071] The above two problems (information redundancy caused by spectral continuity and information overlap caused by latent factors) will lead to overfitting of the PLSR model. Therefore, in order to solve the above problems, it is necessary to further refine the variables after preliminary screening to reduce the correlation and redundancy between variables.

[0072] The method used in the present invention is to analyze the correlation coefficient matrix between variables. The correlation coefficient matrix describes the linear correlation between variables. If the correlation coefficient of two variables is close to or , indicating that there is a strong positive or negative correlation between them. If a variable has a strong correlation with multiple other variables, then this variable is redundant. However, variables cannot be eliminated simply based on the size of the correlation coefficient. Because the correlation coefficient can only reflect the relationship between two variables, but cannot reflect the complex relationship between multiple variables. Therefore, in order to more accurately evaluate the correlation and redundancy between variables, it is necessary to analyze the overall structure of the correlation coefficient matrix. Eigenvalue decomposition can decompose the correlation coefficient matrix into a set of eigenvectors and corresponding eigenvalues. Each eigenvector represents a variable combination, and each eigenvalue represents the variance of the variable combination. The larger the eigenvalue, the larger the variance of the variable combination, indicating that it contains more information; the smaller the eigenvalue, the smaller the variance of the variable combination, indicating that it contains less information, pointing to noise or redundant information.

[0073] Finally, the independent variable combinations are identified by analyzing the eigenvalues ​​and eigenvectors of the correlation coefficient matrix. For those variable combinations with small eigenvalues, the original variables corresponding to them are likely to be highly correlated with other variables, indicating that they are redundant. Based on this criterion, the weights of these variables are reduced.

[0074] Specifically, a band weight analysis is performed on the preliminary spectral band screening results to obtain the importance of each spectral band in the preliminary spectral band screening results, including:

[0075] The spectral band screening results of the spectral pixel reflectance data of corn plants are obtained, and the spectral band screening results are integrated into a spectral matrix. The correlation coefficient matrix of the spectral matrix is ​​obtained by performing correlation analysis on the spectral matrix. The eigenvalue analysis is performed on the correlation coefficient matrix of the spectral matrix to obtain the eigenvalue of each spectral band in the spectral matrix. The importance of each spectral band in the preliminary spectral band screening results is obtained by performing feature difference evaluation on the characteristics of each spectral band in the spectral matrix.

[0076] Among them, by evaluating the characteristic differences of each spectral band in the spectral matrix, the importance of each spectral band in the preliminary spectral band screening results is obtained, including:

[0077] Obtain a correlation coefficient matrix and an importance adjustment parameter; extract matrix eigenvectors from the correlation coefficient matrix to obtain an eigenvalue vector of the correlation coefficient matrix; use the difference between the eigenvalue of any spectral band in the eigenvalue vector and the mean of all eigenvalues ​​in the eigenvalue vector of the correlation coefficient matrix as a second mean difference; perform an exponential mapping with a natural constant as the base on the result of multiplying the importance adjustment parameter by the second mean difference to obtain a second exponential mapping; use the inverse of the result of adding the constant 1 to the second exponential mapping as the importance of the spectral band.

[0078] In one embodiment, assuming that The characteristic value of the band is , then the first The calculation expression of the importance of each band is:

[0079]

[0080] in, Indicates the spectral matrix The importance of each band; represents the importance adjustment parameter; Indicates the spectral matrix The characteristic value of each band; Represents the mean of all eigenvalues ​​in the eigenvalue vector of the correlation coefficient matrix; Represents a constant ; Represents a natural constant.

[0081] It should be noted that the eigenvector represents the linear combination of variables, and the eigenvalue represents the variance of these linear combinations. The larger the eigenvalue, the larger the variance of the corresponding variable combination, and the more information it contains; the smaller the eigenvalue, the smaller the variance of the corresponding variable combination, and the less information it contains. At the same time, the weight is not directly determined based on the size of the eigenvalue, but the average value of the eigenvalue is calculated. , and then calculate each eigenvalue and The difference reflects the degree of deviation of the eigenvalue from the average level. Less than , indicating that the corresponding variable combination contains less information, and the weights of these variables need to be reduced. Finally, the weight of each variable By formula Calculate so that sensitive to the difference between the eigenvalue and the mean, The smaller, The closer ; The bigger, The closer In this way, variables with smaller eigenvalues ​​(usually corresponding to redundant information) are given lower weights, while variables with larger eigenvalues ​​(usually containing more valid information) are given higher weights.

[0082] Step S104, weighted optimization is performed on the spectral pixel reflectance data according to the importance of each spectral band to obtain optimized spectral pixel reflectance data, and a PLSR model is established using the optimized spectral pixel reflectance data and corn kernel moisture content data.

[0083] First, the obtained band importance is used to weight the spectral reflectance corresponding to each band in the spectral matrix to obtain a weighted spectral data matrix. Then, the weighted spectral data matrix and the corresponding measured water content vector are used , build Model. The model construction process belongs to the existing technology and will not be described here. The spectral data matrix and the water content vector are decomposed into a linear combination of several latent variables. These latent variables can both explain the variation of the spectral data to the greatest extent and are most related to the water content. The model determines the number of latent variables and the load vector corresponding to each latent variable through iterative calculation to establish the regression relationship between spectral data and water content. In order to avoid overfitting of the model, cross validation is required to determine The optimal number of latent variables in the model.

[0084] This step evaluates the importance of each band in the spectral matrix, and then weights the reflectance data of each band in the spectral matrix according to the importance in the subsequent steps, so as to eliminate the information redundancy caused by spectral continuity in the spectral matrix and the information overlap caused by potential factors, improve the prediction accuracy of the PLSR model, and predict the moisture content of corn kernels more accurately.

[0085] Step S105, predicting the moisture content of corn kernels based on the corn field spectral data using a corn kernel moisture content prediction model.

[0086] For a new sample, first obtain its multispectral data and perform the same preliminary spectral band screening and band importance evaluation and weighting as the above steps to obtain weighted spectral data. Then input it into the constructed Specifically, for each pixel in the image, its multispectral reflectance value is extracted and variable selection and weighting are performed according to the method of the present invention to obtain weighted spectral data. The weighted spectral data of each pixel is input into The corresponding predicted value of grain moisture content is obtained from the model. Then, the predicted value of moisture content of each pixel is arranged according to its spatial position in the image to generate a spatial distribution map of corn grain moisture content, and different colors or gray levels are used to represent different moisture content ranges. This moisture content distribution map can provide an important reference for precision agricultural production. According to the moisture content distribution map, the difference in moisture content in different areas of the field can be identified to guide precision irrigation. The difference in maturity of corn grains in different areas of the field can also be determined based on the moisture content distribution to guide timely harvesting. Areas where the moisture content has reached the harvest standard should be harvested first, and the harvest time should be appropriately postponed for areas where the moisture content is still high, thereby improving harvesting efficiency, reducing grain loss, and ensuring corn quality.

[0087] In summary, the field intelligent prediction method for corn kernel moisture content described in the present invention collects corn plant spectral images and corn kernel moisture content data to obtain spectral pixel reflectance data and corn kernel moisture content data corresponding to the pixel, and obtains preliminary spectral band screening results by stability screening of the spectral pixel reflectance data; performs band weight analysis on the preliminary spectral band screening results to obtain the importance of each spectral band in the preliminary spectral band screening results; performs weighted optimization on the spectral pixel reflectance data according to the importance of each spectral band to obtain optimized spectral pixel reflectance data, and establishes a PLSR model through the optimized spectral pixel reflectance data and corn kernel moisture content data to obtain a corn kernel moisture content prediction model; and predicts corn kernel moisture content on corn field spectral data through the corn kernel moisture content prediction model. The spectral pixel reflectance data are screened for stability, and the bands that are greatly affected by environmental interference are eliminated to improve the reliability of the input data. Then, the contribution of each spectral band to the prediction of grain moisture content is quantified through weight analysis of each band, and weighted optimization is performed accordingly to enhance the influence of key spectral features on the prediction model, thereby improving the prediction accuracy of the model.

[0088] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A field intelligent prediction method for corn kernel moisture content, comprising: The spectral image of corn plants and the moisture content data of corn kernels are collected to obtain spectral pixel reflectance data and the moisture content data of corn kernels corresponding to the pixels, wherein the method is characterized in that the spectral pixel reflectance data is subjected to stability screening to obtain a preliminary spectral band screening result; the preliminary spectral band screening result is subjected to band weight analysis to obtain the importance of each spectral band in the preliminary spectral band screening result; the spectral pixel reflectance data is weightedly optimized according to the importance of each spectral band to obtain optimized spectral pixel reflectance data; a PLSR model is established by using the optimized spectral pixel reflectance data and the corn kernel moisture content data to obtain a corn kernel moisture content prediction model; and the corn kernel moisture content is predicted for corn field spectral data using the corn kernel moisture content prediction model.

2. The method for intelligent field prediction of corn grain moisture content according to claim 1, characterized in that: According to the above, by performing stability screening on the spectral pixel reflectance data, a preliminary spectral band screening result is obtained, and the specific steps include: Acquire spectral pixel reflectance data of corn plants and corn kernel moisture content data corresponding to each pixel, and perform nonlinear transformation on the spectral pixel reflectance data of corn plants to obtain transformed spectral pixel reflectance data of corn plants; The regression coefficients of each band in the spectral pixel reflectance data of corn plants and the spectral pixel reflectance data of corn plants after transformation are evaluated respectively, and the regression coefficients of each band in the spectral pixel reflectance data of corn plants and the spectral pixel reflectance data of corn plants after transformation are obtained respectively; By evaluating the stability of the regression coefficient of each band in the spectral pixel reflectance data of corn plants and the spectral pixel reflectance data of corn plants after transformation, the stability of each band in the spectral pixel reflectance data of corn plants and the spectral pixel reflectance data of corn plants after transformation is obtained respectively; By comparing and analyzing the stability of each band in the spectral pixel reflectance data of the corn plant and the transformed spectral pixel reflectance data of the corn plant, a stability difference evaluation of each band in the spectral pixel reflectance data of the corn plant is obtained; By mapping and analyzing the stability difference evaluation results, a band screening threshold is obtained, and the stability of each band in the corn plant spectral pixel reflectance data is compared with the band screening threshold to obtain a preliminary spectral band screening result.

3. The method for intelligent field prediction of corn grain moisture content according to claim 2, characterized in that: According to the method, by evaluating the regression coefficients of each band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation, the regression coefficients of each band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation are obtained respectively, and the specific steps include: Acquire the spectral pixel reflectance data of the corn plant, the spectral pixel reflectance data of the corn plant after transformation, and the corn kernel moisture content data corresponding to each pixel, establish a first PLSR model through the spectral pixel reflectance data of the corn plant and the corn kernel moisture content data corresponding to each pixel, and obtain the regression coefficient vector of the first PLSR model; establish a second PLSR model through the spectral pixel reflectance data of the corn plant after transformation and the corn kernel moisture content data corresponding to each pixel, and obtain the regression coefficient vector of the second PLSR model; by repeatedly establishing the first PLSR model and establishing the second PLSR model, obtain multiple groups of regression coefficient vectors of the first PLSR model and multiple groups of regression coefficient vectors of the second PLSR model, calculate the mean of the multiple groups of regression coefficient vectors of the first PLSR model and the multiple groups of regression coefficient vectors of the second PLSR model, obtain the first regression coefficient vector of the first PLSR model and the first regression coefficient vector of the second PLSR model, thereby obtaining the regression coefficient of each band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation.

4. The method for intelligent field prediction of corn grain moisture content according to claim 2, characterized in that: According to the method of evaluating the stability of the regression coefficient of each band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation, the stability corresponding to each band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation is obtained, and the specific steps include: The absolute value calculation result of the mean of the regression coefficient of any band in the spectral pixel reflectance data of the corn plant is used as the first mean of the band, and the standard deviation of the regression coefficient of any band in the spectral pixel reflectance data of the corn plant is divided by the first mean of the band as the stability of the band in the spectral pixel reflectance data of the corn plant; the absolute value calculation result of the mean of the regression coefficient of any band in the spectral pixel reflectance data of the corn plant after transformation is used as the first mean of the band, and the standard deviation of the regression coefficient of any band in the spectral pixel reflectance data of the corn plant after transformation is divided by the first mean of the band as the stability of the band in the spectral pixel reflectance data of the corn plant after transformation.

5. The method for intelligent field prediction of corn grain moisture content according to claim 2, characterized in that: According to the comparative analysis of the stability corresponding to each band in the spectral pixel reflectance data of the corn plant and the transformed spectral pixel reflectance data of the corn plant, the stability difference evaluation of each band in the spectral pixel reflectance data of the corn plant is obtained, and the specific steps include: The stability of any band in the spectral pixel reflectance data of the corn plant and the spectral pixel reflectance data of the corn plant after transformation is obtained, and the stability of any band in the spectral pixel reflectance data of the corn plant after transformation is divided by the stability of the band in the spectral pixel reflectance data of the corn plant as a calculation result of the stability difference evaluation of the band in the spectral pixel reflectance data of the corn plant.

6. The method for intelligent field prediction of corn grain moisture content according to claim 2, characterized in that: According to the mapping analysis of the stability difference evaluation results, a band screening threshold is obtained, and the stability of each band in the corn plant spectral pixel reflectance data is compared with the band screening threshold to obtain a preliminary spectral band screening result, the specific steps of which include: Obtaining a stability difference assessment of each band in the spectral pixel reflectance data of the corn plant, setting a first mapping parameter and a set adjustment coefficient, taking the mean of the stability difference assessments of all bands in the spectral pixel reflectance data of the corn plant as a second mean, taking a calculation result of subtracting the stability difference assessment of each band in the spectral pixel reflectance data of the corn plant from the second mean as a first mean difference, performing an exponential mapping with a natural constant as the base on the inverse of a calculation result of multiplying the first mapping parameter by the first mean difference to obtain a first exponential mapping; taking a calculation result of adding a constant 1 to the first exponential mapping as a threshold optimization factor; Obtaining the stability of each band in the spectral pixel reflectance data of the corn plant, taking the mean of the stability of all bands in the spectral pixel reflectance data of the corn plant as the third mean, taking the standard deviation of the stability of all bands in the spectral pixel reflectance data of the corn plant as the first standard deviation, and taking the third mean plus the adjustment coefficient multiplied by the first standard deviation as the basic screening threshold; The calculation result of multiplying the threshold optimization factor by the basic screening threshold is used as the band screening threshold; the stability of each band in the spectral pixel reflectance data of the corn plant is compared with the band screening threshold; if the stability of the band in the spectral pixel reflectance data of the corn plant is less than or equal to the band screening threshold, the band data is retained to obtain a preliminary spectral band screening result.

7. The method for intelligent field prediction of corn grain moisture content according to claim 1, characterized in that: According to the band weight analysis of the preliminary spectral band screening result, the importance of each spectral band in the preliminary spectral band screening result is obtained, and the specific steps include: The spectral band screening results of the spectral pixel reflectance data of the corn plant are obtained, and the spectral band screening results are integrated into a spectral matrix. By performing correlation analysis on the spectral matrix, a correlation coefficient matrix of the spectral matrix is ​​obtained; by performing eigenvalue analysis on the correlation coefficient matrix of the spectral matrix, the eigenvalue of each spectral band in the spectral matrix is ​​obtained; by performing feature difference evaluation on the characteristics of each spectral band in the spectral matrix, the importance of each spectral band in the preliminary spectral band screening results is obtained.

8. The method for intelligent field prediction of corn grain moisture content according to claim 7, characterized in that: According to the feature difference evaluation of the features of each spectral band in the spectral matrix, the importance of each spectral band in the preliminary spectral band screening result is obtained, and the specific steps include: Acquire the correlation coefficient matrix and the importance adjustment parameter; perform matrix eigenvector extraction on the correlation coefficient matrix to obtain the eigenvalue vector of the correlation coefficient matrix; use the difference between the eigenvalue of any spectral band in the eigenvalue vector and the mean of all eigenvalues ​​in the eigenvalue vector of the correlation coefficient matrix as the second mean difference; perform exponential mapping with a natural constant as the base on the result of multiplying the importance adjustment parameter by the second mean difference to obtain a second exponential mapping; use the inverse of the result of adding the constant 1 to the second exponential mapping as the importance of the spectral band.

Images