Crop photosynthetic parameter inversion method and device, medium and product

Abstract

The invention discloses a crop photosynthetic parameter inversion method and device, a medium and a product, and relates to the technical field of phenotype omics, and the method comprises the steps: obtaining images of all characteristic wave bands of a to-be-predicted crop; determining the reflectivity of the corresponding characteristic wave band based on the image of each characteristic wave band; preprocessing the reflectivity of the corresponding characteristic wave band by using the optimal preprocessing strategy corresponding to each characteristic wave band to obtain the optimally preprocessed reflectivity of the corresponding characteristic wave band; sequentially inputting the predicted values of the prediction parameters corresponding to the photosynthetic parameters into the corresponding photosynthetic parameter inversion models according to a preset sequence to obtain the predicted values of the corresponding photosynthetic parameters; the photosynthetic parameters comprise A1200, SPAD, chlorophyll a, chlorophyll b and leaf nitrogen content. According to the method, multiple photosynthetic parameters of crops can be determined quickly, losslessly and accurately.

Description

Technical Field

[0001] The present application relates to the technical field of phenomics, and in particular to a method, device, medium and product for inverting crop photosynthetic parameters. Background Art

[0002] The yield of crops is determined by the light energy capture efficiency, economic coefficient and photosynthetic efficiency. At present, both the light energy capture efficiency and economic coefficient have reached the theoretical maximum value, and cultivating high-light-efficiency crops has become the fundamental way to further significantly increase crop yields. Photosynthetic parameters are the specific manifestations of crop photosynthetic efficiency, among which A1200, SPAD, chlorophyll a, chlorophyll b and leaf nitrogen content are important photosynthetic parameters. In traditional photosynthetic parameter measurement methods, different photosynthetic parameters often require different instruments to measure, and the measurement process is time-consuming and complicated to operate. In addition, photosynthetic physiological parameters such as chlorophyll content and nitrogen content are usually measured in a way that is destructive to crops. At present, phenotypic studies use different sensors to measure the geometric and physiological shapes of crops, including the use of radar, high / multispectral cameras, infrared cameras, and optical imaging sensors to observe crops. Among them, hyperspectral imaging equipment can obtain many continuous and narrow spectral bands from visible light to infrared spectra. These band images contain rich information about the spectral and spatial distribution of different surface materials, which can realize non-destructive detection of plant physiological processes that are invisible to the naked eye. This characteristic makes non-destructive detection of crop photosynthetic parameters possible.

[0003] Traditional spectral analysis work usually focuses on a single target and performs independent analysis on each, but multiple photosynthetic parameters of crops are usually interrelated and work synergistically. Single-target analysis may affect the understanding of complex systems due to the lack of a global perspective on the overall system. Summary of the invention

[0004] The purpose of this application is to provide a crop photosynthetic parameter inversion method, device, medium and product, which can quickly, non-destructively and accurately determine multiple photosynthetic parameters of crops.

[0005] To achieve the above objectives, this application provides the following solutions:

[0006] In a first aspect, the present application provides a method for inverting crop photosynthetic parameters, comprising:

[0007] Acquire images of all characteristic bands of the crops to be predicted;

[0008] Determine the reflectivity of the corresponding characteristic band based on the image of each characteristic band;

[0009] The reflectance of the corresponding characteristic band is preprocessed by using the optimal preprocessing strategy corresponding to each characteristic band to obtain the reflectance after the optimal preprocessing of the corresponding characteristic band;

[0010] The predicted values ​​of the prediction parameters corresponding to each photosynthetic parameter are input into the corresponding photosynthetic parameter inversion model in sequence according to a preset order to obtain the predicted values ​​of the corresponding photosynthetic parameters; the predicted values ​​of the prediction parameters are the first parameter or the second parameter; the first parameter corresponding to the current photosynthetic parameter includes: the optimal pre-processed reflectance of all characteristic bands corresponding to the current photosynthetic parameter; the second parameter corresponding to the current photosynthetic parameter includes: the optimal pre-processed reflectance of all characteristic bands corresponding to the current photosynthetic parameter and the predicted value of the inversion parameter; the inversion parameter includes: one or more photosynthetic parameters sequentially located before the current photosynthetic parameter, and the predicted value of the inversion parameter includes: the predicted value of the photosynthetic parameter output by the photosynthetic parameter inversion model corresponding to the inversion parameter; the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter is determined by using the single parameter prediction model corresponding to the current photosynthetic parameter, the optimal pre-processed reflectance of all characteristic bands and the true value of the inversion parameter; the photosynthetic parameters include: A1200, SPAD, chlorophyll a, chlorophyll b and leaf nitrogen content.

[0011] Optionally, the process of determining the best preprocessing strategy corresponding to any current photosynthetic parameter includes:

[0012] Obtaining a full-band image of the plant used for training and the corresponding true value of the current photosynthetic parameter; the full-band includes characteristic bands corresponding to all photosynthetic parameters;

[0013] Determine the reflectance of the corresponding band based on the image of each band in the full band;

[0014] Based on multiple preprocessing methods, multiple initial preprocessing strategies are constructed; the preprocessing methods include: baseline correction, scatter correction and scale scaling, the baseline correction includes: Savitzky-Golay first-order differential, Savitzky-Golay second-order differential and continuous wavelet transform; the scatter correction includes: multivariate scatter correction and standard normal transformation; the scale scaling includes: mean centering and standardization;

[0015] The reflectance of each band is processed respectively using each initial preprocessing strategy to obtain the preprocessed reflectance of each band under the corresponding initial preprocessing strategy;

[0016] Using the partial least squares cross-validation method, based on the preprocessed reflectance of all bands under each initial preprocessing strategy, the predicted value of the current photosynthetic parameter of the training plant under the corresponding initial preprocessing strategy is determined;

[0017] Based on the true value of the current photosynthetic parameter of the training plant and the predicted value of the current photosynthetic parameter of the training plant under each initial pretreatment strategy, the regression coefficient under the corresponding initial pretreatment strategy is calculated;

[0018] The initial preprocessing strategy corresponding to the maximum regression coefficient is determined as the optimal preprocessing strategy corresponding to the current photosynthetic parameters.

[0019] Optionally, the process of determining a single parameter prediction model corresponding to any current photosynthetic parameter includes:

[0020] The reflectance of each band in the full band of the training plant is preprocessed by using the optimal preprocessing strategy corresponding to the current photosynthetic parameters to obtain the reflectance of each band in the full band after optimal preprocessing;

[0021] The optimal pre-processed reflectances of all bands in the full band are respectively input into three first pre-training models to obtain the predicted values ​​of the current photosynthetic parameters output by the corresponding first pre-training models; the three first pre-training models are respectively trained by the same rounds of three initial models using the initial model training set, wherein the initial model training set includes: the optimal pre-processed reflectances of all bands in the full band corresponding to the current photosynthetic parameters of the plant used for training and the corresponding true values ​​of the current photosynthetic parameters, and the three initial models are: least squares regression, support vector regression and random forest;

[0022] Based on the true value of the current photosynthetic parameter of the training plant and the predicted value of the current photosynthetic parameter output by each first pre-training model, calculating the regression coefficient corresponding to the first pre-training model;

[0023] The first pre-trained model corresponding to the maximum regression coefficient is determined as the single-parameter prediction model corresponding to the current photosynthetic parameter.

[0024] Optionally, the process of determining a characteristic band corresponding to any current photosynthetic parameter includes:

[0025] Four band selection methods are used to screen each band in the full band, and the initial screening band corresponding to the current photosynthetic parameters under the corresponding band selection method is obtained;

[0026] The reflectance of each initial screening band under each band selection method is preprocessed respectively using the optimal preprocessing strategy corresponding to the current photosynthetic parameters to obtain the reflectance after the optimal preprocessing of the corresponding initial screening band;

[0027] The optimal pre-processed reflectances of all the initial screening bands under each band selection method are respectively input into the single parameter prediction model corresponding to the current photosynthetic parameter to obtain the predicted value of the current photosynthetic parameter under the corresponding band selection method;

[0028] Based on the true value of the current photosynthetic parameter and the predicted value of the current photosynthetic parameter under each band selection method, the regression coefficient under the corresponding band selection method is calculated;

[0029] The initial screening band of the current photosynthetic parameter under the band selection method corresponding to the maximum regression coefficient is determined as the characteristic band corresponding to the current photosynthetic parameter.

[0030] Optionally, the process of determining the inversion parameter and the photosynthetic parameter inversion model corresponding to any current photosynthetic parameter includes:

[0031] Numbering each photosynthetic parameter according to the preset sequence to obtain a sequence number of each photosynthetic parameter; the sequence numbers of the photosynthetic parameters are 1, 2, 3, 4 and 5 respectively;

[0032] When the sequence number of the current photosynthetic parameter is 1, the single parameter prediction model of the current photosynthetic parameter is determined as the photosynthetic parameter inversion model of the current photosynthetic parameter;

[0033] When the sequence number of the current photosynthetic parameter is not 1, the inversion parameter and the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter are determined based on the single parameter prediction model of the current photosynthetic parameter and the predicted values ​​of each photosynthetic parameter whose sequence number is smaller than the current photosynthetic parameter.

[0034] Optionally, based on the single parameter prediction model of the current photosynthetic parameter and the predicted values ​​of each photosynthetic parameter whose sequence number is less than the current photosynthetic parameter, determining the inversion parameter and the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter includes:

[0035] Inputting the optimal preprocessed reflectance of all characteristic bands corresponding to the current photosynthetic parameter into the single parameter prediction model corresponding to the current photosynthetic parameter to obtain a separate prediction value of the current photosynthetic parameter;

[0036] Calculate the individual regression coefficient of the current photosynthetic parameter based on the true value and the individual predicted value of the current photosynthetic parameter;

[0037] Arrange and combine the predicted values ​​of the photosynthetic parameters whose serial numbers are smaller than the current photosynthetic parameters to obtain a plurality of groups of predicted values ​​of inversion parameters corresponding to the current photosynthetic parameters; a group of inversion parameters includes one or more photosynthetic parameters;

[0038] Determine any set of inversion parameters corresponding to the current photosynthetic parameter as the current inversion parameter, and determine the predicted value of the current inversion parameter and the optimal preprocessed reflectance of all characteristic bands corresponding to the current photosynthetic parameter as the predicted value of the current combination parameter;

[0039] Inputting the predicted value of the current combination parameter into the second pre-trained model corresponding to the current combination parameter to obtain the joint predicted value of the current photosynthetic parameter under the current combination parameter; the second pre-trained model corresponding to the current combination parameter is obtained by training the single parameter prediction model corresponding to the current photosynthetic parameter using the true value of the current combination parameter of a plurality of training plants, and the training rounds of each second pre-trained model are the same;

[0040] Based on the joint predicted value of the current photosynthetic parameter under the current combination parameter and the true value of the current photosynthetic parameter, calculating the joint regression coefficient of the current photosynthetic parameter under the current combination parameter;

[0041] The combined parameter corresponding to the maximum value of the joint regression coefficient of the current photosynthetic parameter under all combined parameters and the individual regression coefficient of the current photosynthetic parameter or the optimal preprocessed reflectivity of all characteristic bands corresponding to the current photosynthetic parameter is determined as the prediction parameter, and the second pre-trained model corresponding to the prediction parameter is determined as the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter.

[0042] Optionally, determining the reflectivity of the corresponding characteristic band based on the image of each characteristic band includes:

[0043] The reflectivity calculation formula is used to determine the reflectivity of the corresponding characteristic band according to the image of each characteristic band; the reflectivity calculation formula is:

[0044]

[0045] Among them, F α is the reflectivity of band α; r α is the original average reflectivity of the image in band α; a is the reflectivity of the dark current; b is the average reflectivity of the white reference plate; c is the true reflectivity of the white reference plate.

[0046] In a second aspect, the present application provides a computer device, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the crop photosynthetic parameter inversion methods described above.

[0047] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned crop photosynthetic parameter inversion methods.

[0048] In a fourth aspect, the present application provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned crop photosynthetic parameter inversion methods.

[0049] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0050] The present application discloses a crop photosynthetic parameter inversion method, device, medium and product. First, images of all characteristic bands of the crop to be predicted are obtained; then, the reflectance of the corresponding characteristic band is determined based on the image of each characteristic band; the reflectance of the corresponding characteristic band is preprocessed using the optimal preprocessing strategy corresponding to each characteristic band to obtain the reflectance of the corresponding characteristic band after optimal preprocessing; finally, the predicted values ​​of the prediction parameters corresponding to each photosynthetic parameter are input into the corresponding photosynthetic parameter inversion model in sequence according to a preset order to obtain the predicted values ​​of the corresponding photosynthetic parameters; the predicted values ​​of the prediction parameters are the first parameters or the second parameters; the first parameter corresponding to the current photosynthetic parameter includes: all characteristic bands corresponding to the current photosynthetic parameter; The optimal pre-processed reflectance of the characteristic band; the second parameter corresponding to the current photosynthetic parameter includes: the optimal pre-processed reflectance of all characteristic bands corresponding to the current photosynthetic parameter and the predicted value of the inversion parameter; the inversion parameter includes: one or more photosynthetic parameters sequentially located before the current photosynthetic parameter, and the predicted value of the inversion parameter includes: the predicted value of the photosynthetic parameter output by the photosynthetic parameter inversion model corresponding to the inversion parameter; the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter is determined by using the single parameter prediction model corresponding to the current photosynthetic parameter, the optimal pre-processed reflectance of all characteristic bands and the true value of the inversion parameter; the photosynthetic parameters include: A1200, SPAD, chlorophyll a, chlorophyll b and leaf nitrogen content. In the present application, the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter is determined by using the single parameter prediction model corresponding to the current photosynthetic parameter, the optimal pre-processed reflectance of all characteristic bands and the true value of the inversion parameter, and the current photosynthetic parameter is predicted in combination with other photosynthetic parameters, so that multiple photosynthetic parameters of the crop can be determined quickly, non-destructively and accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0052] Figure 1 A schematic diagram of a process flow of a crop photosynthetic parameter inversion method provided in an embodiment of the present application;

[0053] Figure 2 Flowchart for selecting the best pretreatment strategy based on a multi-stage experimental design;

[0054] Figure 3 The basic flow chart of the stepwise regression chain model for correlation;

[0055] Figure 4 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0057] The purpose of this application is to provide a crop photosynthetic parameter inversion method, device, medium and product, aiming to quickly, non-destructively and accurately determine multiple photosynthetic parameters of crops.

[0058] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0059] In an exemplary embodiment, Figure 1 As shown, the crop photosynthetic parameter inversion method in this embodiment includes:

[0060] Step 1: Obtain images of all characteristic bands of the crop to be predicted.

[0061] Step 2: Determine the reflectivity of the corresponding characteristic band based on the image of each characteristic band.

[0062] As an optional implementation, step 2 includes:

[0063] The reflectivity calculation formula is used to determine the reflectivity of the corresponding characteristic band according to the image of each characteristic band; the reflectivity calculation formula is:

[0064]

[0065] Among them, F α is the reflectivity of band α; r α is the original average reflectivity of the image in band α, that is, the average reflectivity of all pixels in the image in band α; a is the reflectivity of dark current, that is, the reflectivity generated by the hyperspectral camera when there is no light; b is the average reflectivity of the white reference plate, that is, the average value of the reflectivity of all pixels in the image obtained by hyperspectral imaging of the white reference plate in the current imaging environment; c is the true reflectivity of the white reference plate.

[0066] Step 3: Preprocess the reflectance of the corresponding characteristic band using the optimal preprocessing strategy corresponding to each characteristic band to obtain the reflectance after the optimal preprocessing of the corresponding characteristic band.

[0067] Step 4: Input the predicted values ​​of the prediction parameters corresponding to each photosynthetic parameter into the corresponding photosynthetic parameter inversion model in sequence according to a preset order to obtain the predicted values ​​of the corresponding photosynthetic parameters.

[0068] Among them, the predicted value of the prediction parameter is the first parameter or the second parameter; the first parameter corresponding to the current photosynthetic parameter includes: the reflectance of all characteristic bands corresponding to the current photosynthetic parameter after the best preprocessing; the second parameter corresponding to the current photosynthetic parameter includes: the reflectance of all characteristic bands corresponding to the current photosynthetic parameter after the best preprocessing and the predicted value of the inversion parameter; the inversion parameter includes: one or more photosynthetic parameters sequentially located before the current photosynthetic parameter, and the predicted value of the inversion parameter includes: the predicted value of the photosynthetic parameter output by the photosynthetic parameter inversion model corresponding to the inversion parameter; the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter is determined by using the single parameter prediction model corresponding to the current photosynthetic parameter, the reflectance of all characteristic bands after the best preprocessing and the true value of the inversion parameter; the photosynthetic parameters include: A1200, SPAD, chlorophyll a, chlorophyll b and leaf nitrogen content.

[0069] As an optional implementation, the process of determining the best preprocessing strategy corresponding to any current photosynthetic parameter includes:

[0070] Step 11: Obtain the full-band image of the plant used for training and the corresponding true value of the current photosynthetic parameter; the full-band includes the characteristic bands corresponding to all photosynthetic parameters.

[0071] Specifically, the training crops are placed on the hyperspectral data acquisition system for overhead push-scanning photography. The hyperspectral data acquisition system is controlled by an electronic system, with multiple halogen lamps installed on the top to provide supplementary lighting for the crops. The hyperspectral camera is installed on the top horizontal slide rail to perform horizontal push-scanning photography of the leaves of the training crops, and real-time hyperspectral images of the crops are collected. The hyperspectral images include full-band images.

[0072] The ENVI software was used to manually identify the location of the Regions of Interest (ROI) in the hyperspectral image, and the hyperspectral image in the ROI area was used as the full-band image of the object for training.

[0073] Step 12: Determine the reflectivity of the corresponding band based on the image of each band in the full band.

[0074] Specifically, the reflectivity calculation formula is used to determine the reflectivity of the corresponding band according to the image of each band in the full band.

[0075] Step 13: Based on multiple preprocessing methods, construct multiple initial preprocessing strategies; the preprocessing methods include: baseline correction, scatter correction and scale scaling. Baseline correction includes: Savitzky-Golay first-order differential, Savitzky-Golay second-order differential and continuous wavelet transform; scatter correction includes: multivariate scatter correction and standard normal transformation; scale scaling includes: mean centering and standardization.

[0076] Step 14: Use each initial preprocessing strategy to process the reflectivity of each band respectively, and obtain the preprocessed reflectivity of each band under the corresponding initial preprocessing strategy.

[0077] Step 15: Using the partial least squares cross-validation method, based on the preprocessed reflectance of all bands under each initial preprocessing strategy, determine the predicted value of the current photosynthetic parameter of the training plant under the corresponding initial preprocessing strategy.

[0078] Step 16: Based on the true value of the current photosynthetic parameter of the training plant and the predicted value of the current photosynthetic parameter of the training plant under each initial pretreatment strategy, the regression coefficient under the corresponding initial pretreatment strategy is calculated.

[0079] Step 17: Determine the initial preprocessing strategy corresponding to the maximum regression coefficient as the optimal preprocessing strategy corresponding to the current photosynthetic parameters.

[0080] Specifically, Figure 2 As shown in Figure 2, the optimal pretreatment strategy is determined based on the principle of multi-stage experimental design. The treatment process includes:

[0081] S171: The specific process of multi-stage design includes three stages, including execution order, inter-class interaction and specific method strategies. The preprocessing methods used include baseline correction, scatter correction and scale scaling. Baseline correction includes Savitzky-Golay first-order differential, Savitzky-Golay second-order differential and continuous wavelet transform; scatter correction includes multivariate scatter correction and standard normal transformation; scale scaling includes mean centering and standardization. The three types of preprocessing are applied in the order of baseline correction, scatter correction and scale scaling.

[0082] S172: For various preprocessing methods in S171, 36 initial preprocessing strategies are designed by permutation and combination, and the best preprocessing strategy is selected by partial least squares cross validation.

[0083] S173: Determine whether each initial preprocessing strategy has the best performance (ie, the regression coefficient is the largest). If so, select the initial preprocessing strategy as the best preprocessing strategy; otherwise, remove the initial preprocessing strategy.

[0084] As an optional implementation, the process of determining a single parameter prediction model corresponding to any current photosynthetic parameter includes:

[0085] Step 21: Preprocess the reflectance of each band in the full band of the training plant using the optimal preprocessing strategy corresponding to the current photosynthetic parameters to obtain the optimal preprocessed reflectance of each band in the full band.

[0086] Step 22: input the optimal pre-processed reflectance of all bands in the full band into three first pre-trained models respectively to obtain the predicted value of the current photosynthetic parameter output by the corresponding first pre-trained model; the three first pre-trained models are obtained by training the three initial models in the same round using the initial model training set, and the initial model training set includes: the optimal pre-processed reflectance of all bands in the full band corresponding to the current photosynthetic parameters of the plant used for training and the corresponding true value of the current photosynthetic parameter, and the three initial models are: least squares regression, support vector regression and random forest.

[0087] Step 23: Based on the true value of the current photosynthetic parameter of the training plant and the predicted value of the current photosynthetic parameter output by each first pre-training model, the regression coefficient of the corresponding first pre-training model is calculated.

[0088] Step 24: Determine the first pre-trained model corresponding to the maximum regression coefficient as the single parameter prediction model corresponding to the current photosynthetic parameter.

[0089] As an optional implementation manner, the process of determining a characteristic band corresponding to any current photosynthetic parameter includes:

[0090] Step 31: using the four band selection methods to screen each band in the full band, respectively, to obtain the initial screening band corresponding to the current photosynthetic parameter under the corresponding band selection method.

[0091] Step 32: Preprocess the reflectance of each initial screening band under each band selection method using the optimal preprocessing strategy corresponding to the current photosynthetic parameters to obtain the optimal preprocessed reflectance of the corresponding initial screening band.

[0092] Step 33: inputting the optimal pre-processed reflectance of all the initial screening bands under each band selection method into the single parameter prediction model corresponding to the current photosynthetic parameter to obtain the predicted value of the current photosynthetic parameter under the corresponding band selection method.

[0093] Step 34: Based on the true value of the current photosynthetic parameter and the predicted value of the current photosynthetic parameter under each band selection method, the regression coefficient under the corresponding band selection method is calculated.

[0094] Step 35: Determine the initial screening band of the current photosynthetic parameter under the band selection method corresponding to the maximum regression coefficient as the characteristic band corresponding to the current photosynthetic parameter.

[0095] As an optional implementation, Figure 3 As shown, the process of determining the inversion parameters and photosynthetic parameter inversion model corresponding to any current photosynthetic parameter includes:

[0096] Step 41: numbering each photosynthetic parameter in a preset order to obtain a sequence number of each photosynthetic parameter; the sequence numbers of the photosynthetic parameters are 1, 2, 3, 4 and 5 respectively.

[0097] Specifically, the process of determining the preset order includes:

[0098] (1) Calculate the correlation coefficients between the two photosynthetic parameters respectively to obtain a covariance matrix with a dimension of 5×5. The calculation formula of the correlation coefficient is:

[0099]

[0100] Among them, C XY is the correlation coefficient between photosynthetic parameters X and Y; I is the number of crops when determining the correlation coefficient; X i is the true value of photosynthetic parameter X of crop i; is the average value of photosynthetic parameter X; Y i is the true value of photosynthetic parameter Y of crop i; Y is the average value of photosynthetic parameter Y.

[0101] (2) Sum all the correlation coefficients corresponding to each photosynthetic parameter to obtain the total correlation coefficient of the corresponding photosynthetic parameter.

[0102] (3) The order of the total correlation coefficients from large to small is determined as a preset order.

[0103] Step 42: When the sequence number of the current photosynthetic parameter is 1, the single parameter prediction model of the current photosynthetic parameter is determined as the photosynthetic parameter inversion model of the current photosynthetic parameter.

[0104] Step 43: When the serial number of the current photosynthetic parameter is not 1, based on the single parameter prediction model of the current photosynthetic parameter and the predicted values ​​of each photosynthetic parameter whose serial number is smaller than the current photosynthetic parameter, determine the inversion parameter and the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter.

[0105] As an optional implementation, in step 43, based on the single parameter prediction model of the current photosynthetic parameter and the predicted values ​​of the photosynthetic parameters whose serial numbers are smaller than the current photosynthetic parameter, the inversion parameter and the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter are determined, including:

[0106] Step 431: inputting the optimal preprocessed reflectance of all characteristic bands corresponding to the current photosynthetic parameter into the single parameter prediction model corresponding to the current photosynthetic parameter to obtain a single predicted value of the current photosynthetic parameter.

[0107] Step 432: Calculate the individual regression coefficient of the current photosynthetic parameter based on the true value and the individual predicted value of the current photosynthetic parameter.

[0108] Step 433: Arrange and combine the predicted values ​​of the photosynthetic parameters whose serial numbers are smaller than the current photosynthetic parameters to obtain a plurality of groups of predicted values ​​of inversion parameters corresponding to the current photosynthetic parameters; a group of inversion parameters includes one or more photosynthetic parameters.

[0109] Step 434: determine any set of inversion parameters corresponding to the current photosynthetic parameters as the current inversion parameters, and determine the predicted value of the current inversion parameters and the optimal preprocessed reflectivity of all characteristic bands corresponding to the current photosynthetic parameters as the predicted value of the current combination parameter.

[0110] Step 435: input the predicted value of the current combination parameter into the second pre-trained model corresponding to the current combination parameter to obtain the joint predicted value of the current photosynthetic parameter under the current combination parameter; the second pre-trained model corresponding to the current combination parameter is obtained by training the single parameter prediction model corresponding to the current photosynthetic parameter using the true values ​​of the current combination parameters of multiple training plants, and the training rounds of each second pre-trained model are the same.

[0111] Step 436: Calculate the joint regression coefficient of the current photosynthetic parameter under the current combination parameter based on the joint predicted value of the current photosynthetic parameter under the current combination parameter and the true value of the current photosynthetic parameter.

[0112] Step 437: Determine the combined parameter corresponding to the maximum value of the joint regression coefficient of the current photosynthetic parameter under all combined parameters and the individual regression coefficient of the current photosynthetic parameter or the optimal preprocessed reflectivity of all characteristic bands corresponding to the current photosynthetic parameter as the prediction parameter, and determine the second pre-trained model corresponding to the prediction parameter as the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter.

[0113] Specifically, when the current photosynthetic parameter is the photosynthetic parameter with sequence number 3, the process of determining the inversion parameter corresponding to the current photosynthetic parameter and the photosynthetic parameter inversion model includes:

[0114] 1) The optimal preprocessed reflectance of all characteristic bands corresponding to the photosynthetic parameter numbered 3 is input into the single parameter prediction model corresponding to the photosynthetic parameter numbered 3 to obtain a single predicted value of the photosynthetic parameter numbered 3.

[0115] 2) Based on the true value and individual predicted value of the photosynthetic parameter No. 3, calculate the individual regression coefficient R3 of the photosynthetic parameter No. 32 .

[0116] 3) The predicted values ​​of the photosynthetic parameters with serial numbers 1 and 2 are arranged and combined to obtain the predicted values ​​of three groups of inversion parameters corresponding to the current photosynthetic parameters. The three groups of inversion parameters are respectively recorded as A1, A2 and A3; the predicted value of inversion parameter A1 includes: the predicted value of the photosynthetic parameter with serial number 1, the predicted value of inversion parameter A2 includes: the predicted value of the photosynthetic parameter with serial number 2, and the predicted value of inversion parameter A3 includes: the predicted values ​​of the photosynthetic parameters with serial numbers 1 and 2.

[0117] 4) The predicted value of the inversion parameter A1 and the optimal pre-processed reflectance of all characteristic bands corresponding to the photosynthetic parameter with serial number 3 are determined as the predicted value of the combination parameter B1, the predicted value of the inversion parameter A2 and the optimal pre-processed reflectance of all characteristic bands corresponding to the photosynthetic parameter with serial number 3 are determined as the predicted value of the combination parameter B2, and the predicted value of the inversion parameter A3 and the optimal pre-processed reflectance of all characteristic bands corresponding to the photosynthetic parameter with serial number 3 are determined as the predicted value of the combination parameter B3.

[0118] 5) Inputting the predicted value of the combination parameter B1 into the second pre-trained model corresponding to the combination parameter B1 to obtain the joint predicted value of the current photosynthetic parameters under the combination parameter B1; inputting the predicted value of the combination parameter B2 into the second pre-trained model corresponding to the combination parameter B2 to obtain the joint predicted value of the current photosynthetic parameters under the combination parameter B2; inputting the predicted value of the combination parameter B3 into the second pre-trained model corresponding to the combination parameter B3 to obtain the joint predicted value of the current photosynthetic parameters under the combination parameter B3.

[0119] 6) Based on the predicted value of the photosynthetic parameter numbered 3 and the joint predicted value of the current photosynthetic parameter under the combination parameter B1, the joint regression coefficient of the photosynthetic parameter numbered 3 under the joint predicted value of the current photosynthetic parameter under the combination parameter B1 is calculated Based on the predicted value of the photosynthetic parameter with sequence number 3 and the joint predicted value of the current photosynthetic parameter under the combination parameter B2, the joint predicted value of the current photosynthetic parameter under the combination parameter B2 is calculated. The joint regression coefficient of the photosynthetic parameter with sequence number 3 Based on the predicted value of the photosynthetic parameter with sequence number 3 and the joint predicted value of the current photosynthetic parameter under the combination parameter B3, the joint predicted value of the current photosynthetic parameter under the combination parameter B3 is calculated. The joint regression coefficient of the photosynthetic parameter with sequence number 3

[0120] 7) and The combined parameter with the largest median value or the optimal preprocessed reflectivity of all characteristic bands corresponding to the photosynthetic parameter with serial number 3 is determined as the prediction parameter corresponding to the photosynthetic parameter with serial number 3, and the second pre-trained model corresponding to the prediction parameter corresponding to the photosynthetic parameter with serial number 3 is determined as the photosynthetic parameter inversion model corresponding to the photosynthetic parameter with serial number 3.

[0121] Further, in order to reduce the amount of calculation, on the basis of keeping steps 1) to 4) unchanged, steps 5) to 7) can be replaced by steps S1 to S3:

[0122] Step S1: Calculate the true value of the combination parameter B1 and input it into the second pre-trained model corresponding to the combination parameter B1 to obtain the joint prediction value of the current photosynthetic parameter under the combination parameter B1. Based on the true value of the photosynthetic parameter No. 3 and the joint prediction value of the current photosynthetic parameter under the combination parameter B1, calculate the joint prediction value of the current photosynthetic parameter under the combination parameter B1 and the joint regression coefficient of the photosynthetic parameter No. 3. Compare and When When is larger, execute step S2. If it is larger, execute step S3.

[0123] Step S2: input the true value of the combination parameter B3 into the second pre-trained model corresponding to the combination parameter B3 to obtain the joint prediction value of the current photosynthetic parameter under the combination parameter B3; based on the true value of the photosynthetic parameter numbered 3 and the joint prediction value of the current photosynthetic parameter under the combination parameter B3, calculate the joint prediction value of the current photosynthetic parameter under the combination parameter B3 and the joint regression coefficient of the photosynthetic parameter numbered 3 Compare and When When the value is larger, the combined parameter B1 is determined as the prediction parameter corresponding to the photosynthetic parameter with the serial number 3, and the second pre-trained model corresponding to the combined parameter B1 is determined as the photosynthetic parameter inversion model corresponding to the photosynthetic parameter with the serial number 3. When it is larger, the combination parameter B3 is determined as the prediction parameter corresponding to the photosynthetic parameter with serial number 3, and the second pre-training model corresponding to the combination parameter B3 is determined as the photosynthetic parameter inversion model corresponding to the photosynthetic parameter with serial number 3.

[0124] Step S3: input the true value of the combination parameter B2 into the second pre-trained model corresponding to the combination parameter B2 to obtain the joint prediction value of the current photosynthetic parameter under the combination parameter B2; based on the true value of the photosynthetic parameter numbered 3 and the joint prediction value of the current photosynthetic parameter under the combination parameter B2, calculate the joint prediction value of the current photosynthetic parameter under the combination parameter B2 and the joint regression coefficient of the photosynthetic parameter numbered 3 Compare and When When the value is larger, the optimal pre-processed reflectance of all characteristic bands corresponding to the photosynthetic parameter numbered 3 is determined as the prediction parameter corresponding to the photosynthetic parameter numbered 3. The corresponding second pre-trained model is determined to be the photosynthetic parameter inversion model corresponding to the photosynthetic parameter numbered 3. When it is larger, the combination parameter B2 is determined as the prediction parameter corresponding to the photosynthetic parameter with serial number 3, and the second pre-training model corresponding to the combination parameter B2 is determined as the photosynthetic parameter inversion model corresponding to the photosynthetic parameter with serial number 3.

[0125] In an exemplary embodiment, a computer device is provided, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a crop photosynthetic parameter inversion method.

[0126] In an exemplary embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, a method for inverting crop photosynthetic parameters is implemented.

[0127] In an exemplary embodiment, a computer program product is provided, including a computer program, which implements a crop photosynthetic parameter inversion method when executed by a processor.

[0128] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 4 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for inverting crop photosynthetic parameters is implemented.

[0129] Those skilled in the art will understand that Figure 4The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0130] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0131] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0132] The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0133] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0134] This article uses specific examples to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand the method and core ideas of this application. At the same time, for those skilled in the art, according to the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A crop photosynthetic parameter inversion method, characterized in that: The crop photosynthetic parameter inversion method comprises: Acquire images of all characteristic bands of the crops to be predicted; Determine the reflectivity of the corresponding characteristic band based on the image of each characteristic band; The reflectance of the corresponding characteristic band is preprocessed by using the optimal preprocessing strategy corresponding to each characteristic band to obtain the reflectance after the optimal preprocessing of the corresponding characteristic band; The predicted values ​​of the prediction parameters corresponding to each photosynthetic parameter are input into the corresponding photosynthetic parameter inversion model in sequence according to a preset order to obtain the predicted values ​​of the corresponding photosynthetic parameters; the predicted values ​​of the prediction parameters are the first parameter or the second parameter; the first parameter corresponding to the current photosynthetic parameter includes: the optimal pre-processed reflectance of all characteristic bands corresponding to the current photosynthetic parameter; the second parameter corresponding to the current photosynthetic parameter includes: the optimal pre-processed reflectance of all characteristic bands corresponding to the current photosynthetic parameter and the predicted value of the inversion parameter; the inversion parameter includes: one or more photosynthetic parameters sequentially located before the current photosynthetic parameter, and the predicted value of the inversion parameter includes: the predicted value of the photosynthetic parameter output by the photosynthetic parameter inversion model corresponding to the inversion parameter; the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter is determined by using the single parameter prediction model corresponding to the current photosynthetic parameter, the optimal pre-processed reflectance of all characteristic bands and the true value of the inversion parameter; the photosynthetic parameters include: A1200, SPAD, chlorophyll a, chlorophyll b and leaf nitrogen content.

2. The crop photosynthetic parameter inversion method according to claim 1, characterized in that: The process of determining the best pretreatment strategy corresponding to any current photosynthetic parameter includes: Obtaining a full-band image of the plant used for training and the corresponding true value of the current photosynthetic parameter; the full-band includes characteristic bands corresponding to all photosynthetic parameters; Determine the reflectance of the corresponding band based on the image of each band in the full band; Based on multiple preprocessing methods, multiple initial preprocessing strategies are constructed; the preprocessing methods include: baseline correction, scatter correction and scale scaling, the baseline correction includes: Savitzky-Golay first-order differential, Savitzky-Golay second-order differential and continuous wavelet transform; the scatter correction includes: multivariate scatter correction and standard normal transformation; the scale scaling includes: mean centering and standardization; The reflectance of each band is processed respectively using each initial preprocessing strategy to obtain the preprocessed reflectance of each band under the corresponding initial preprocessing strategy; Using the partial least squares cross-validation method, based on the preprocessed reflectance of all bands under each initial preprocessing strategy, the predicted value of the current photosynthetic parameter of the training plant under the corresponding initial preprocessing strategy is determined; Based on the true value of the current photosynthetic parameter of the training plant and the predicted value of the current photosynthetic parameter of the training plant under each initial pretreatment strategy, the regression coefficient under the corresponding initial pretreatment strategy is calculated; The initial preprocessing strategy corresponding to the maximum regression coefficient is determined as the optimal preprocessing strategy corresponding to the current photosynthetic parameters.

3. The crop photosynthetic parameter inversion method according to claim 2, characterized in that: The process of determining a single parameter prediction model corresponding to any current photosynthetic parameter includes: The reflectance of each band in the full band of the training plant is preprocessed by using the optimal preprocessing strategy corresponding to the current photosynthetic parameters to obtain the reflectance of each band in the full band after optimal preprocessing; The optimal pre-processed reflectances of all bands in the full band are respectively input into three first pre-training models to obtain the predicted values ​​of the current photosynthetic parameters output by the corresponding first pre-training models; the three first pre-training models are respectively trained by the same rounds of three initial models using the initial model training set, wherein the initial model training set includes: the optimal pre-processed reflectances of all bands in the full band corresponding to the current photosynthetic parameters of the plant used for training and the corresponding true values ​​of the current photosynthetic parameters, and the three initial models are: least squares regression, support vector regression and random forest; Based on the true value of the current photosynthetic parameter of the training plant and the predicted value of the current photosynthetic parameter output by each first pre-training model, calculating the regression coefficient corresponding to the first pre-training model; The first pre-trained model corresponding to the maximum regression coefficient is determined as the single-parameter prediction model corresponding to the current photosynthetic parameter.

4. The crop photosynthetic parameter inversion method according to claim 3, characterized in that: The process of determining the characteristic band corresponding to any current photosynthetic parameter includes: Four band selection methods are used to screen each band in the full band, and the initial screening band corresponding to the current photosynthetic parameters under the corresponding band selection method is obtained; The reflectance of each initial screening band under each band selection method is preprocessed respectively using the optimal preprocessing strategy corresponding to the current photosynthetic parameters to obtain the reflectance after the optimal preprocessing of the corresponding initial screening band; The optimal pre-processed reflectances of all the initial screening bands under each band selection method are respectively input into the single parameter prediction model corresponding to the current photosynthetic parameter to obtain the predicted value of the current photosynthetic parameter under the corresponding band selection method; Based on the true value of the current photosynthetic parameter and the predicted value of the current photosynthetic parameter under each band selection method, the regression coefficient under the corresponding band selection method is calculated; The initial screening band of the current photosynthetic parameter under the band selection method corresponding to the maximum regression coefficient is determined as the characteristic band corresponding to the current photosynthetic parameter.

5. The crop photosynthetic parameter inversion method according to claim 4, characterized in that: The process of determining the inversion parameters and photosynthetic parameter inversion model corresponding to any current photosynthetic parameter includes: Numbering each photosynthetic parameter according to the preset sequence to obtain a sequence number of each photosynthetic parameter; the sequence numbers of the photosynthetic parameters are 1, 2, 3, 4 and 5 respectively; When the sequence number of the current photosynthetic parameter is 1, the single parameter prediction model of the current photosynthetic parameter is determined as the photosynthetic parameter inversion model of the current photosynthetic parameter; When the sequence number of the current photosynthetic parameter is not 1, the inversion parameter and the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter are determined based on the single parameter prediction model of the current photosynthetic parameter and the predicted values ​​of each photosynthetic parameter whose sequence number is smaller than the current photosynthetic parameter.

6. The crop photosynthetic parameter inversion method according to claim 5, characterized in that: Based on the single parameter prediction model of the current photosynthetic parameter and the predicted values ​​of each photosynthetic parameter whose serial number is less than the current photosynthetic parameter, the inversion parameter and the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter are determined, including: Inputting the optimal preprocessed reflectance of all characteristic bands corresponding to the current photosynthetic parameter into the single parameter prediction model corresponding to the current photosynthetic parameter to obtain a separate prediction value of the current photosynthetic parameter; Calculate the individual regression coefficient of the current photosynthetic parameter based on the true value and the individual predicted value of the current photosynthetic parameter; Arrange and combine the predicted values ​​of the photosynthetic parameters whose serial numbers are smaller than the current photosynthetic parameters to obtain a plurality of groups of predicted values ​​of inversion parameters corresponding to the current photosynthetic parameters; a group of inversion parameters includes one or more photosynthetic parameters; Determine any set of inversion parameters corresponding to the current photosynthetic parameter as the current inversion parameter, and determine the predicted value of the current inversion parameter and the optimal preprocessed reflectance of all characteristic bands corresponding to the current photosynthetic parameter as the predicted value of the current combination parameter; Inputting the predicted value of the current combination parameter into the second pre-trained model corresponding to the current combination parameter to obtain the joint predicted value of the current photosynthetic parameter under the current combination parameter; the second pre-trained model corresponding to the current combination parameter is obtained by training the single parameter prediction model corresponding to the current photosynthetic parameter using the true value of the current combination parameter of a plurality of training plants, and the training rounds of each second pre-trained model are the same; Based on the joint predicted value of the current photosynthetic parameter under the current combination parameter and the true value of the current photosynthetic parameter, calculating the joint regression coefficient of the current photosynthetic parameter under the current combination parameter; The combined parameter corresponding to the maximum value of the joint regression coefficient of the current photosynthetic parameter under all combined parameters and the individual regression coefficient of the current photosynthetic parameter or the optimal preprocessed reflectivity of all characteristic bands corresponding to the current photosynthetic parameter is determined as the prediction parameter, and the second pre-trained model corresponding to the prediction parameter is determined as the photosynthetic parameter inversion model corresponding to the current photosynthetic parameter.

7. The crop photosynthetic parameter inversion method according to claim 1, characterized in that: Determine the reflectivity of the corresponding characteristic band based on the image of each characteristic band, including: The reflectivity calculation formula is used to determine the reflectivity of the corresponding characteristic band according to the image of each characteristic band; the reflectivity calculation formula is: Among them, F α is the reflectivity of band α; r α is the original average reflectivity of the image in band α; a is the reflectivity of the dark current; b is the average reflectivity of the white reference plate; c is the true reflectivity of the white reference plate.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the crop photosynthetic parameter inversion method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the crop photosynthetic parameter inversion method described in any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the crop photosynthetic parameter inversion method described in any one of claims 1 to 7 is implemented.

Images