A method, equipment, and medium for inverting the physicochemical parameters of rice based on satellite imagery

Abstract

This application discloses a method, device, and medium for inverting rice physicochemical parameters based on satellite imagery, relating to the field of agricultural remote sensing. The method includes: performing SRF simulation on Sentinel-2A satellite multispectral imagery based on UAV hyperspectral imagery of rice in the target phenological stage within the target area to obtain simulated broadband reflectance and construct band indices; obtaining band index combinations through correlation analysis; training a preset machine learning combination model using the band index combinations as input and measured rice physicochemical parameters as labels to obtain a rice physicochemical parameter inversion model; acquiring the Sentinel-2A satellite multispectral imagery to be processed, and inputting the data corresponding to the band index combinations into the rice physicochemical parameter inversion model to obtain the spatial distribution of rice physicochemical parameters. This application can improve the accuracy of inverting rice physicochemical parameters from satellite imagery and achieve high-precision monitoring of rice growth over large areas.

Description

Technical Field

[0001] This application relates to the field of agricultural remote sensing, and in particular to a method, equipment, and medium for inverting the physicochemical parameters of rice based on satellite imagery. Background Technology

[0002] During the critical growth stages of rice, chlorophyll content and leaf nitrogen content (LNC) are core phenotypic parameters characterizing the physiological state of rice, directly affecting photosynthetic efficiency, nitrogen allocation, and canopy structure, thus influencing rice yield. The most common traditional method for detecting leaf chlorophyll is spectrophotometry, which not only damages leaf structure but is also extremely time-consuming and labor-intensive. Furthermore, chlorophyll is structurally unstable and easily decomposes in light, leading to errors between measured values ​​and actual content. Studies have shown that the relative chlorophyll content (SPAD value) measured by a portable chlorophyll meter is positively correlated with chlorophyll content; therefore, the SPAD value can replace chlorophyll content detected by chemical analysis methods, achieving efficient and non-destructive detection of leaf chlorophyll. LNC typically relies on laboratory Kjeldahl nitrogen determination. While these agronomic parameter measurement methods are reliable, they require significant time and manpower and are sometimes affected by weather or external factors.

[0003] Satellite remote sensing is widely used for quantitative remote sensing monitoring of crop growth indicators over large areas due to its advantages of wide coverage, speed, and low cost. However, satellite remote sensing still has significant limitations in extracting and estimating crop information at the field scale. This is due to its relatively low spatial and spectral resolution, long revisit periods, and susceptibility to interference from weather conditions such as clouds and atmospheric conditions. Satellite imagery, due to resolution limitations, struggles to capture subtle differences in crop growth, thus limiting its sensitivity to changes in crop (such as rice) physiological parameters, and the accuracy of inversion needs further improvement. Summary of the Invention

[0004] The purpose of this application is to provide a method, equipment, and medium for inverting the physicochemical parameters of rice based on satellite imagery, which can improve the accuracy of inverting the physicochemical parameters of rice from satellite imagery and realize high-precision monitoring of rice growth over a large area.

[0005] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a method for inverting the physicochemical parameters of rice based on satellite imagery, including: Acquire UAV hyperspectral images of rice in the target area during the target phenological stage, measured physicochemical parameters of rice, and corresponding Sentinel-2A satellite multispectral images; Based on the UAV hyperspectral imagery, SRF simulation was performed on the Sentinel-2A satellite multispectral imagery to obtain the simulated broadband reflectance. Based on the simulated wideband reflectivity, multiple band indices are constructed; Correlation analysis was performed on multiple band indices and preset vegetation indices, with the measured physicochemical parameters of rice as the target variable, to obtain a combination of band indices. Using the band index combination as input and the measured rice physicochemical parameters as labels, a preset machine learning combination model is trained to obtain a rice physicochemical parameter inversion model; Acquire multispectral images of the Sentinel-2A satellite to be processed, and input the data corresponding to the band index combination into the rice physicochemical parameter inversion model to obtain the spatial distribution of rice physicochemical parameters.

[0006] Secondly, this application provides a computer device, 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 method for inverting the physicochemical parameters of rice based on satellite imagery.

[0007] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for inverting the physicochemical parameters of rice based on satellite imagery.

[0008] According to the specific embodiments provided in this application, the following technical effects are disclosed: Based on UAV hyperspectral images of rice in the target phenological stage within the target area, this application performs SRF simulation on Sentinel-2A satellite multispectral images to obtain simulated broadband reflectance; based on this, multiple band indices are constructed; then, a combination of band indices is obtained through correlation analysis; using the combination of band indices as input and measured rice physicochemical parameters as labels, a preset machine learning combination model is trained to obtain a rice physicochemical parameter inversion model. Finally, the rice physicochemical parameter inversion model is used for Sentinel-2A satellite multispectral images to invert the spatial distribution of rice physicochemical parameters. Through the above processing, the spatial and spectral detail advantages of UAVs are combined with the wide coverage characteristics of satellites, which can effectively improve the monitoring accuracy of rice physicochemical parameters in small areas and achieve high-precision monitoring of rice growth over large areas. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is an application environment diagram of the rice physicochemical parameter inversion method based on satellite imagery in one embodiment of this application.

[0011] Figure 2 This is a flowchart illustrating a method for inverting the physicochemical parameters of rice based on satellite imagery in one embodiment of this application.

[0012] Figure 3(a) is a schematic diagram showing the consistency of UAV hyperspectral, S2 original spectrum and SRF simulated spectrum during the tillering stage of rice.

[0013] Figure 3(b) shows a schematic diagram comparing the consistency of UAV hyperspectral, S2 original spectrum and SRF simulated spectrum during the rice booting stage.

[0014] Figure 4(a) shows the scatter plot of measured and predicted SPAD values ​​for the optimal SRF and S2 models during the tillering stage.

[0015] Figure 4(b) shows the scatter plot of measured and predicted SPAD values ​​during the heading stage for the optimal SRF and S2 models.

[0016] Figure 5 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0018] To compensate for the low spatial resolution of satellite imagery, unmanned aerial vehicle (UAV) near-Earth remote sensing technology has become a supplementary method. As an important form of low-altitude remote sensing, UAVs offer significant advantages such as ease of operation, flexible takeoff conditions, high maneuverability, strong timeliness, and continuous aerial photography. They can acquire images with extremely high spatial and spectral resolution, making them more suitable for high-precision monitoring of small areas and field plots. Combining the spatial and spectral detail advantages of UAVs with the wide coverage of satellites can effectively improve the monitoring accuracy of chlorophyll in the canopy of small areas of farmland, enabling large-scale, high-precision monitoring of crop growth.

[0019] This application combines UAV hyperspectral imagery with Sentinel-2A satellite multispectral imagery, along with synchronous field surveys and sampling. It comprehensively utilizes methods such as spectral data simulation and fusion, feature analysis, and machine learning regression modeling to invert SPAD and nitrogen content during key rice growth stages. This enables large-scale upscaling and rapid monitoring, providing a theoretical basis and regional reference for rice growth monitoring and refined management.

[0020] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0021] The method for inverting the physicochemical parameters of rice based on satellite imagery provided in this application can be applied to, for example... Figure 1 In the application environment shown, terminal 101 communicates with server 102 via a network. A data storage system can store the data that server 102 needs to process. The data storage system can be set up independently, integrated into server 102, or placed in the cloud or on another server. Terminal 101 can send UAV hyperspectral images of rice in the target phenological stage, measured rice physicochemical parameters, and corresponding Sentinel-2A satellite multispectral images to server 102. After receiving the images, server 102 performs SRF simulation on the Sentinel-2A satellite multispectral images to obtain simulated broadband reflectance; based on the simulated broadband reflectance, multiple band indices are constructed; through correlation analysis, a combination of band indices is obtained; using the band indices combination as input and measured rice physicochemical parameters as labels, a preset machine learning combination model is trained to obtain a rice physicochemical parameter inversion model. Server 102 can feed back the obtained rice physicochemical parameter inversion model to terminal 101, where further applications are executed. In addition, in some embodiments, the method for inverting the physicochemical parameters of rice based on satellite imagery can also be implemented separately by the server 102 or the terminal 101.

[0022] The terminal 101 can be, but is not limited to, various desktop computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, and smart in-vehicle devices. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted devices. The server 102 can be implemented using a standalone server or a server cluster composed of multiple servers, or it can be a cloud server.

[0023] In one exemplary embodiment, such as Figure 2As shown, a method for inverting the physicochemical parameters of rice based on satellite imagery is provided. This method is executed by computer equipment, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is applied to... Figure 1 Taking server 102 as an example, the explanation includes the following steps 201 to 206.

[0024] Step 201: Acquire UAV hyperspectral images of rice in the target area during the target phenological stage, measured physicochemical parameters of rice, and corresponding Sentinel-2A satellite multispectral images; wherein the target phenological stage is the tillering stage or the booting stage, and the measured physicochemical parameters of rice include the measured SPAD value and the measured LNC value of rice.

[0025] The acquisition of UAV hyperspectral images of rice in the target phenological period within the target area includes: (1) Using a UAV equipped with a hyperspectral sensor, raw hyperspectral images of rice in the target area are collected during the target phenological period; the hyperspectral sensor can acquire 64 band images in the wavelength range of 500nm-900nm, with a spectral resolution of 6nm and a single image size of 1010×1010 pixels. The flight altitude is 50m, and raw hyperspectral images of rice during the tillering and booting stages are collected. (2) For any of the raw hyperspectral images, dark current correction, radiometric calibration, orthorectification, and SG (Savitzky-Golay smoothing filter) convolution smoothing are performed to obtain the corresponding UAV hyperspectral image. The UAV hyperspectral image preprocessing is completed through dark current correction, radiometric calibration, and orthorectification. In order to reduce or eliminate the influence of noise on the spectral characteristics of ground objects, the SG convolution smoothing method is used to process the raw spectral reflectance.

[0026] The process of obtaining the measured physicochemical parameters of rice included: the main vegetation parameters collected on the ground were the SPAD / LNC values ​​during the tillering and booting stages. Each small experimental plot was set to 1.5m × 1.5m, with a 15m interval between plots, ensuring uniform distribution. The rice in the experimental plots showed uniform overall growth, representing the average physiological state of homogeneous fields at the Sentinel-2A satellite scale.

[0027] To ensure consistency with UAV data, this application selects Sentinel-2A satellite multispectral imagery from the same period as the field sampling and performs band fusion, choosing the B2-B8A bands that match the UAV's bands.

[0028] Step 202: Based on the UAV hyperspectral image, perform SRF simulation on the Sentinel-2A satellite multispectral image to obtain the simulated broadband reflectance.

[0029] Differences in band spacing, band range, and resolution among different sensors lead to inconsistent spectral dimensions, affecting data fusion and analysis. SRF is defined as the ratio of the received radiance at each wavelength to the incident radiance. This method avoids the problem of negative or excessively small coefficients that may occur with the least squares method. The determination of spatial detail modulation parameters comprehensively considers the global variance of the image, reducing spectral distortion.

[0030] In one application, based on the UAV hyperspectral imagery, SRF simulation is performed on the Sentinel-2A satellite multispectral imagery to obtain simulated broadband reflectance, including: calculating the simulated broadband reflectance using the following formula: .

[0031] Where R is the simulated broadband reflectivity, and The starting and ending wavelengths (nm) of the Sentinel-2A satellite multispectral image. For Sentinel-2A satellite multispectral imagery in The spectral response function value of wavelength, SRF is defined as follows: The ratio of the received radiance at a given wavelength to the incident radiance; For UAV hyperspectral imagery in Reflectivity of wavelength.

[0032] Figure 3(a) shows a comparison of the consistency between the UAV hyperspectral, the original S2 spectrum, and the SRF simulated spectrum during the rice tillering stage; Figure 3(b) shows a comparison of the consistency between the UAV hyperspectral, the original S2 spectrum, and the SRF simulated spectrum during the rice booting stage. Analysis shows a significant difference between the original Sentinel-2A spectrum and the UAV hyperspectral spectrum, while the SRF simulated Sentinel-2A spectrum is closer to the UAV hyperspectral. This is because SRF simulation can significantly improve the quality of the original Sentinel-2A spectral data by considering ground reflectance characteristics and more accurately capture vegetation-related spectral features. This improvement not only effectively enhances the spectral information quality of satellite images but also strengthens their feasibility in inversion.

[0033] Step 203: Based on the simulated wideband reflectivity, construct multiple band indices.

[0034] Step 204: For multiple band indices and preset vegetation indices, a correlation analysis is performed with the measured physicochemical parameters of rice as the target variable to obtain a combination of band indices.

[0035] By traversing all bands, we thoroughly searched for band combinations with high correlation, which were then used as a supplement to 15 traditional vegetation indices to analyze the correlation between UAV hyperspectral reflectance, Sentinel-2A satellite reflectance, and simulated broadband reflectance based on Sentinel-2A and SPAD values. In this way, parameters with high correlation to the above agronomic variables were selected as the final model input variables.

[0036] Correlation analysis was performed using Pearson correlation analysis. For multiple band indices and preset vegetation indices, with the measured rice physicochemical parameters as the target variables, correlation analysis was conducted to obtain band index combinations. This included: constructing a random forest regression model; inputting multiple band indices and preset vegetation indices into the random forest regression model to obtain predicted rice physicochemical parameters; calculating the Pearson correlation coefficient between each input and the predicted rice physicochemical parameters; and selecting and determining band index combinations based on the Pearson correlation coefficients. p represents the significance value; the Pearson correlation coefficient ranges from -1 to 1. The closer the value is to -1 or 1, the higher the correlation between the two variables, and vice versa.

[0037] Three optimized spectral indices—Ratio Spectral Index (RSI), Normalized Difference Spectral Index (NDSI), and Difference Spectral Index (DSI)—were selected as the sensitivity band index (BI). The band index combination includes RSI, NDSI, and DSI, and the calculation formula is as follows: .

[0038] .

[0039] .

[0040] in, , Let be the simulated broadband reflectivity at any band i and j within the wavelength range.

[0041] Step 205: Using the band index combination as input and the measured rice physicochemical parameters as labels, train the preset machine learning combination model to obtain the rice physicochemical parameter inversion model.

[0042] The preset machine learning combination model is constructed based on a combination of partial least squares regression (PLSR), support vector regression (SVR), and CatBoost regression model.

[0043] All models were implemented in Python 3.10. Training and testing samples were divided in a 7:3 ratio, and 5-fold cross-validation and grid search were used for parameter tuning. The number of latent variables in the PLSR was set to 1, and the optimal number of principal components was determined by cross-validation. The SVR used an rbf kernel with C ranging from 0.1 to 100, g from 0.01 to 1, and epsilon from 0.01 to 0.5. CatBoost used 100 to 500 iterations, a maximum depth of 4 to 10, and a learning rate of 0.01 to 0.2. Subsample and colsample_bylevel were both set to 0.8 to 1.0, and the L2 regularization parameter was 1 to 10. The optimal configuration was finally determined by grid search.

[0044] Figures 4(a) and 4(b) show the predicted and measured values ​​of 25 validation samples. Analysis of the relationship between the predicted SPAD and the measured values ​​of the 25 validation samples shows that the SRF optimal inversion model has higher prediction accuracy for the two phenotypic parameters than the satellite optimal inversion model, and can fit the rice phenotypic agronomic parameters of the study area very well.

[0045] As shown in Figure 4(a) of the SPAD validation results during the tillering stage, the scatter plots (red) predicted by the SRF model are more closely clustered around the 1:1 standard line compared to the scatter plots (blue) directly modeled by the satellite. This demonstrates that SRF simulation data, as the upper limit for simulating satellite inversion, exhibits stronger physical consistency than directly using satellite data for modeling when dealing with data such as tillering stage data, which has strong surface background noise.

[0046] During the heading stage, as plant biomass reaches its peak and the canopy closes, the prediction accuracy of both SRF-based models improves. In the SPAD validation shown in Figure 4(b), the predictions from the Sentinel-2A direct model and the SRF transfer model maintain good linear fit. Due to its underlying spectral mechanism, the SRF-trained model exhibits higher stability and prediction accuracy in cross-scale transfer applications.

[0047] Step 206: Obtain the Sentinel-2A satellite multispectral image to be processed, and input the data corresponding to the band index combination into the rice physicochemical parameter inversion model to obtain the spatial distribution of rice physicochemical parameters.

[0048] In summary, this application collected measured SPAD data from ground sampling points during the rice tillering and booting stages. Subsequently, based on contemporaneous Sentinel-2A multispectral imagery, Random Forest Regression (RFR) was used to screen sensitive single bands and 15 commonly used vegetation indices, and Pearson correlation analysis was used to optimize three sensitive band indices (NDSI, DSI, and RSI). Regarding models, three inversion models—Partial Least Squares Regression (PLSR), Support Vector Regression (SVR), and CatBoost Regression—were constructed, and the prediction accuracy of different models and spectral parameters was compared. The optimal model was applied to Sentinel-2A imagery to retrieve the spatial distribution of SPAD.

[0049] This application improves the accuracy of Sentinel-2A rice SPAD retrieval through UAV hyperspectral image acquisition and preprocessing, field-measured rice SPAD values, Sentinel-2A satellite image processing, Sentinel-2A data simulation based on SRF method and UAV data, parameter optimization of sensitive band index, model construction and model application, thereby achieving high-precision monitoring of rice growth over large areas.

[0050] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 5 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and databases. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When the computer program is executed by the processor, it implements a method for inverting the physicochemical parameters of rice based on satellite imagery.

[0051] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0052] In one exemplary embodiment, a computer device is also 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 the steps in the above-described method embodiments.

[0053] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0054] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0055] 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, data stored, data displayed, 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 the relevant data must comply with relevant regulations.

[0056] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this 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 can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0057] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0058] The technical features of the above embodiments can be combined in any way. For the sake of brevity, 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.

[0059] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for inverting the physicochemical parameters of rice based on satellite imagery, characterized in that, The method includes: Acquire UAV hyperspectral images of rice in the target area during the target phenological stage, measured physicochemical parameters of rice, and corresponding Sentinel-2A satellite multispectral images; Based on the UAV hyperspectral imagery, SRF simulation was performed on the Sentinel-2A satellite multispectral imagery to obtain the simulated broadband reflectance. Based on the simulated wideband reflectivity, multiple band indices are constructed; Correlation analysis was performed on multiple band indices and preset vegetation indices, with the measured physicochemical parameters of rice as the target variable, to obtain a combination of band indices. Using the band index combination as input and the measured rice physicochemical parameters as labels, a preset machine learning combination model is trained to obtain a rice physicochemical parameter inversion model; Acquire multispectral images of the Sentinel-2A satellite to be processed, and input the data corresponding to the band index combination into the rice physicochemical parameter inversion model to obtain the spatial distribution of rice physicochemical parameters.

2. The method for inverting the physicochemical parameters of rice based on satellite imagery according to claim 1, characterized in that, Acquire UAV hyperspectral images of rice within the target area during the target phenological stage, including: Using drones equipped with hyperspectral sensors, raw hyperspectral images of rice in the target area are collected during the target phenological period; For any of the original hyperspectral images, perform dark current correction, radiometric calibration, orthorectification, and SG convolution smoothing to obtain the corresponding UAV hyperspectral image.

3. The method for inverting the physicochemical parameters of rice based on satellite imagery according to claim 1, characterized in that, The target phenological stage is the tillering stage or the heading stage; The measured physicochemical parameters of rice include the measured SPAD value and the measured LNC value.

4. The method for inverting the physicochemical parameters of rice based on satellite imagery according to claim 1, characterized in that, Based on the UAV hyperspectral imagery, SRF simulation was performed on the Sentinel-2A satellite multispectral imagery to obtain simulated broadband reflectivity, including: The simulated broadband reflectivity is calculated using the following formula: ； Where R is the simulated broadband reflectivity, and The starting and ending wavelengths of the Sentinel-2A satellite multispectral image. For Sentinel-2A satellite multispectral imagery in The spectral response function value of wavelength, SRF is defined as follows: The ratio of the received radiance at a given wavelength to the incident radiance; For UAV hyperspectral imagery in Reflectivity of wavelength.

5. The method for inverting the physicochemical parameters of rice based on satellite imagery according to claim 1, characterized in that, The band index combination includes the ratio spectral index RSI, the normalized spectral index NDSI, and the difference index DSI, and the calculation formula is as follows: ； ； ； in, , Let be the simulated broadband reflectivity at any band i and j within the wavelength range.

6. The method for inverting the physicochemical parameters of rice based on satellite imagery according to claim 1, characterized in that, The preset machine learning combination model is constructed based on partial least squares regression, support vector regression, and CatBoost regression.

7. The method for inverting the physicochemical parameters of rice based on satellite imagery according to claim 1, characterized in that, The correlation analysis used Pearson correlation analysis.

8. The method for inverting the physicochemical parameters of rice based on satellite imagery according to claim 7, characterized in that, Correlation analysis was performed on multiple band indices and preset vegetation indices, using the measured physicochemical parameters of rice as the target variable, to obtain band index combinations, including: Construct a random forest regression model; Multiple band indices and preset vegetation indices are used as inputs and sent to the random forest regression model to obtain predicted physicochemical parameters of rice. Calculate the Pearson correlation coefficient between each input and the predicted physicochemical parameters of rice; Based on the Pearson correlation coefficient, the band index combination was selected and determined.

9. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the method for inverting the physicochemical parameters of rice based on satellite imagery as described in any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the method for inverting the physicochemical parameters of rice based on satellite imagery as described in any one of claims 1-8.

Images