A method and system for unmanned aerial vehicle multispectral prediction of maize canopy nitrogen nutrition parameters

Through the multi-spectral remote sensing technology and machine learning method of drone, a regression prediction model of corn nitrogen nutrition parameters was constructed, which solved the problem that traditional methods were difficult to monitor the nitrogen nutrition parameters of spring corn, and achieved the accuracy of nitrogen fertilizer management and environmental protection.

CN118887549BActive Publication Date: 2024-11-29SHENYANG INST OF APPL ECOLOGY CHINESE ACAD OF SCI
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411363641.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-28
Publication Date
2024-11-29
Estimated Expiration
2044-09-28

AI Technical Summary

Technical Problem

Traditional methods are difficult to efficiently and accurately monitor and manage the nitrogen nutrition parameters of spring corn under no-till conditions, resulting in unscientific nitrogen fertilizer usage and risk of environmental pollution.

Method used

Using drone multispectral remote sensing technology and machine learning methods, a regression prediction model of the nitrogen concentration and chlorophyll relative content of corn canopy is constructed through multispectral image processing and correlation analysis, and a rapid and lossless monitoring of nitrogen nutrition parameters is achieved.

Benefits of technology

Accurate estimation and dynamic monitoring of nitrogen nutrition parameters of no-till spring corn has been achieved, which improves the scientificity and efficiency of nitrogen fertilizer management and reduces the risk of environmental pollution.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118887549B_ABST
    Figure CN118887549B_ABST
Patent Text Reader

Abstract

The present invention belongs to the technical field of combining unmanned aerial vehicle remote sensing with agriculture, and in particular provides a method and system for unmanned aerial vehicle multispectral prediction of nitrogen nutrition parameters of corn canopy, wherein the range of corn field to be measured is set as field plots with different nitrogen application amounts, and the nitrogen nutrition parameters of corn canopy leaves in each plot are measured; multispectral images of corn field plots are collected to obtain vegetation index and texture index; vegetation index and texture index with high correlation are screened; the above steps are repeated to form a data set with the vegetation index and texture index screened each time; a regression prediction model of corn nitrogen nutrition parameters is constructed using linear functions and machine learning algorithms, and the best regression prediction model of each nitrogen nutrition parameter is screened out; an inversion remote sensing monitoring image of corn nitrogen nutrition parameters is constructed to realize information visualization of corn nitrogen nutrition parameters. The present invention can realize high-throughput rapid non-destructive detection of nitrogen nutrition parameters for field planting, thereby providing a basis for precise nitrogen management.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of combining unmanned aerial vehicle remote sensing with agriculture, and in particular provides an unmanned aerial vehicle multi-spectral prediction method and system for nitrogen nutrition parameters of corn canopy. Background Art

[0002] At present, no-tillage and full-scale straw mulching management has become an important technical measure for the protective utilization of black soil in the Northeast Corn Belt. Scientific application of nitrogen fertilizer can ensure stable and high yield of corn. On the contrary, excessive nitrogen fertilizer application will lead to nitrogen loss and environmental pollution. The parameters related to spring corn nitrogen nutrition mainly include canopy nitrogen concentration, plant nitrogen accumulation, chlorophyll relative content SPAD and nitrogen nutrition index. Traditional nitrogen concentration determination is mainly through destructive sampling-chemical analysis. The inherent low-throughput key parameter acquisition of this method, as well as destructive sampling and destructive testing limit its application in actual production. In addition, in actual production, production personnel rely on experience to diagnose whether plants lack nutrients through appearance, which lacks strict data support and has great uncertainty. Therefore, high-throughput and accurate monitoring of spring corn nitrogen nutrition parameters under no-tillage management conditions is of great significance for scientific management of spring corn growth and regulation of nitrogen fertilizer application.

[0003] The quantitative remote sensing method of UAV has the advantages of high flexibility, high resolution, high throughput and rapid non-destructive monitoring, and is the preferred method for future smart agriculture and field precision planting. At present, the quantitative remote sensing technology of UAV has begun to be used in the monitoring methods and technology research and development of nitrogen and chlorophyll and other related physical and chemical parameters of some field crops (winter wheat, summer corn and rice, etc.). For example, based on the high-definition image data of winter wheat and summer corn canopies collected by UAVs, the nitrogen nutrition orthophoto diagnosis method of the two crops was established. The nitrogen content, aboveground biomass and nitrogen nutrition index of rice leaves were inverted by machine learning methods. By establishing a hyperspectral inversion model of winter wheat canopy nitrogen concentration, the nitrogen concentration of winter wheat leaves at different growth stages was successfully estimated. The above results provide a scientific basis for the efficient use of UAV remote sensing data to quantitatively monitor the growth of field crops and accurately diagnose crop nutrition. However, there is a lack of research on the estimation of spring corn nitrogen nutrition parameters (canopy nitrogen concentration and chlorophyll relative content SPAD) under no-tillage conditions in Northeast China using UAV multispectral remote sensing technology.

[0004] In summary, by establishing a multispectral prediction method for nitrogen nutrition parameters of spring corn in Northeast China based on the multispectral data of UAV from V11 in the tasseling stage to R1 in the silking stage, a model for accurately estimating the canopy nitrogen concentration and relative chlorophyll content SPAD of no-till spring corn was constructed, which will be of great significance to improving the nitrogen nutrition diagnosis ability of no-till spring corn and precise nitrogen fertilizer management. Summary of the invention

[0005] In order to solve the above problems, the present invention provides a method and system for unmanned aerial vehicle multispectral prediction of maize canopy nitrogen nutrition parameters. The multispectral remote sensing data is obtained by unmanned aerial vehicle, and the rapid, non-destructive and accurate estimation of nitrogen nutrition parameters of no-till spring maize is achieved through quantitative remote sensing monitoring and machine learning technology of unmanned aerial vehicle.

[0006] The present invention is implemented by providing a method for unmanned aerial vehicle multi-spectral prediction of nitrogen nutrition parameters of corn canopy, comprising the following steps:

[0007] Step 1: Set the range of the corn field to be tested as field plots with different nitrogen application rates, and measure the nitrogen nutrition parameters of the corn canopy leaves in each plot. The nitrogen nutrition parameters are the nitrogen concentration and chlorophyll relative content SPAD of the canopy leaves;

[0008] Step 2: Plan the flight path of the drone according to the range of the corn field to be tested, collect multispectral images of the corn field through the multispectral camera carried by the drone, perform multispectral image standardization processing, and obtain vegetation index and texture index;

[0009] Step 3: Conduct correlation analysis between vegetation index and texture index and nitrogen nutrition parameters of corn, and select vegetation index and texture index with high correlation from each nitrogen nutrition parameter;

[0010] The specific steps of step 3 are:

[0011] Step 301: using an environmental visualization program to determine plots with different nitrogen application rates, and extracting multispectral images of the plots;

[0012] Step 302: Determine the vegetation index and texture index with high correlation with nitrogen concentration and chlorophyll relative content SPAD through correlation analysis, and extract sensitive vegetation index and texture index from all spectral parameters using traditional feature selection;

[0013] Step 303: using the Boruta algorithm to filter the sensitive vegetation index and texture index extracted in step 302, and further obtain the vegetation index and texture index with higher correlation with nitrogen concentration and chlorophyll relative content SPAD;

[0014] Step 4: Repeat steps 1 to 3 to form a data set with the vegetation index and texture index screened each time;

[0015] Step 5: Based on the data set in step 4, linear functions and machine learning algorithms are used to construct regression prediction models for corn nitrogen nutrition parameters, and the best regression prediction model for each nitrogen nutrition parameter is selected;

[0016] Step 6: Based on the optimal regression prediction model, construct the inversion remote sensing monitoring image of corn nitrogen nutrition parameters to realize the information visualization of corn nitrogen nutrition parameters.

[0017] Preferably, in step 5, the specific steps of screening the best regression prediction model for each nitrogen nutrition parameter are:

[0018] Step 501: Determination coefficient R 2 , root mean square error RMSE and normalized root mean square error nRMSE to evaluate the performance of the regression prediction model of maize nitrogen nutrition parameters constructed by linear function and machine learning algorithm, and select the best regression prediction model;

[0019] Step 502: Fitting the measured values ​​of each corn nitrogen concentration and chlorophyll relative content SPAD with the vegetation index and texture index with high correlation, and constructing a regression prediction model of the nitrogen concentration and chlorophyll relative content SPAD of corn canopy leaves;

[0020] Step 503: Compare the measured values ​​of nitrogen concentration and chlorophyll relative content SPAD in the verification set with the estimated values ​​of the regression prediction model of nitrogen concentration and chlorophyll relative content SPAD in corn canopy leaves to verify the stability of the regression prediction model.

[0021] Further preferably, the machine learning method includes random forest, support vector machine, K nearest neighbor regression, and extreme learning machine.

[0022] Further preferably, in the step 3, in the V11~R1 stage of corn, the vegetation indices with the highest correlation with the nitrogen concentration of corn canopy leaves are NREI (Normalized Red edge Index), NDRE (Normalized Difference RedEdge), and DATT, and the texture indexes are R1m (texture index of red edge 1 band: mean), Gm (texture index of green band: mean), and Rm (texture index of red band: mean); the vegetation indices with the highest correlation with the relative chlorophyll content SPAD are GLI (GreenLeaf Index), NEXG (Normalized Excess green Index), and NGBDI (Normalized Green-BlueDifference Index), and the texture indices are R2m (texture index of red edge 2 band: mean), Gm (texture index of green band: mean), and Nm (texture index of near infrared band: mean).

[0023] Further preferably, in step 1, the corn field to be tested is divided into rectangular areas with a side length of 10 m*13 m.

[0024] Further preferably, in step 2, the flight altitude of the UAV is 90-150 m; the field of view angle of the multispectral camera is 49.6°, the focal length is 5.2 mm, and the aperture is F / 22.

[0025] Further preferably, the corn is Northeast spring corn grown under no-tillage management conditions.

[0026] The present invention also provides a maize canopy nitrogen nutrition parameter UAV multi-spectral prediction system, based on the above-mentioned maize canopy nitrogen nutrition parameter UAV multi-spectral prediction method, the system comprises:

[0027] The nitrogen nutrition parameter measurement module is used to set the range of the corn field to be tested as field plots with different nitrogen application rates, and measure the nitrogen nutrition parameters of the corn canopy leaves in each plot;

[0028] The vegetation index and texture index acquisition module is used to plan the flight path of the UAV according to the range of the corn field to be tested, collect multispectral images of the corn field through the multispectral camera carried by the UAV, perform multispectral image standardization processing, and obtain the vegetation index and texture index;

[0029] A highly correlated vegetation index and texture index screening module is used to perform correlation analysis on the vegetation index and texture index with the nitrogen nutrition parameters of corn, and screen the vegetation index and texture index with high correlation from each nitrogen nutrition parameter;

[0030] A data set forming module, used for forming a data set from the vegetation index and texture index screened each time;

[0031] A regression prediction model construction and screening module is used to construct a regression prediction model of corn nitrogen nutrition parameters using a linear function and a machine learning algorithm according to the data set formed by the data set forming module, and screen out the best regression prediction model for each nitrogen nutrition parameter;

[0032] The information visualization module of corn nitrogen nutrition parameters is used to construct the inversion remote sensing monitoring image of corn nitrogen nutrition parameters based on the optimal regression prediction model, and realize the information visualization of corn nitrogen nutrition parameters.

[0033] Compared with the prior art, the present invention has the following advantages:

[0034] 1. The present invention quantitatively estimates the dynamic change characteristics of corn nitrogen nutrition parameter indicators in different growth stages through the study of the dynamic laws of corn nitrogen nutrition parameters, and applies the inversion model to the field to the plot scale. By adjusting the model optimization, the quantitative monitoring of nitrogen nutrition parameters and growth conditions of spring corn in Northeast China at different growth stages under no-till conditions is achieved.

[0035] 2. The present invention obtains the spectral information of the corn canopy by using the multispectral camera carried by the drone, combined with the simultaneous determination of leaf nitrogen nutrient parameters. A regression prediction model is established to study the correlation between the vegetation index and texture index of the Northeast spring corn canopy under no-tillage conditions and the nitrogen nutrition parameter indicators, and an inversion model of the spring corn canopy nitrogen concentration and chlorophyll relative content SPAD based on drone multispectral data is constructed. It can achieve non-destructive and effective detection of the nitrogen nutrition parameter indicators of Northeast spring corn under no-tillage management conditions, thereby providing an important basis for improving the nitrogen fertilizer utilization efficiency and precise nitrogen fertilizer management of Northeast spring corn. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A flow chart of a method for unmanned aerial vehicle multi-spectral prediction of maize canopy nitrogen nutrition parameters provided by the present invention;

[0037] Figure 2 Correlation coefficients between spectral parameters and nitrogen concentration (a) and chlorophyll relative content SPAD (b);

[0038] Figure 3 The stability results of SPAD linear and machine learning estimation models for nitrogen concentration and relative chlorophyll content in maize canopy leaves, where (a) is the fitting of canopy nitrogen concentration and NREI, (b) is the fitting of canopy nitrogen concentration and R1m, (c) is the fitting of chlorophyll relative content SPAD and GLI, (d) is the fitting of chlorophyll relative content SPAD and R2m; (e) is the random forest estimation of canopy nitrogen concentration, and (f) is the random forest estimation of chlorophyll relative content SPAD;

[0039] Figure 4 Multispectral remote sensing inversion images of nitrogen concentration (a) and chlorophyll relative content SPAD (b) of no-till spring corn at the silking stage. DETAILED DESCRIPTION

[0040] The technical solution of the present invention is further described in detail with reference to the following specific examples.

[0041] Example

[0042] A method for unmanned aerial vehicle multispectral prediction of maize canopy nitrogen nutrition parameters, the process is shown in Figure 1 , including the following steps:

[0043] Step 1: Set the maize field to be tested as field plots with different nitrogen application rates, and measure the nitrogen nutrition parameters of the maize canopy leaves in each plot, that is, measure the nitrogen concentration and relative chlorophyll content SPAD of the canopy leaves in the plots with different nitrogen application rates. The statistical description of the measured indicators is shown in Table 1.

[0044] Table 1: Statistical description of measured indicators

[0045]

[0046] Step 2: Plan the flight path of the drone according to the range of the corn field to be tested, collect multispectral images of the corn field through the multispectral camera carried by the drone, perform multispectral image standardization, and obtain vegetation index and texture index:

[0047] (1) Determine the spatial scope of the test area and plan the flight path of the drone. A quadrotor drone (D2000, Inc., China) equipped with a multispectral camera (D-MSPC2000, Yusense, Inc., Qingdao, China) was used to collect multispectral images of no-till spring corn during its growth period. The camera had a field of view of 49.6° and a flight altitude of 90-150 m. The multispectral camera used was equipped with six bands: 450 nm, 555 nm, 660 nm, 720 nm, 750 nm, and 840 nm. Each channel used a 1.2 mp high dynamic range global shutter CMOS detector. The ground spatial resolution was 8.65 cm@h=120 m. The sensor had 1.2 million pixels and a resolution of 1280*960.

[0048] (2) Yusense map software (V1.0, Yusense, Inc., Qingdao, China) was used to preprocess the collected UAV remote sensing images to obtain vegetation index and texture index.

[0049] Step 3: Conduct correlation analysis between vegetation index and texture index and nitrogen nutrition parameters of corn, and select vegetation index and texture index with high correlation from each nitrogen nutrition parameter;

[0050] Step 301: using an environmental visualization program to determine plots with different nitrogen application rates, and extracting multispectral images of the plots;

[0051] ArcGIS 10.8, ENVI 5.2 and Python 3.6.1 were used to batch extract the average reflectance of 6 bands, 82 vegetation indices and 48 texture indices.

[0052] Step 302: Determine the vegetation index and texture index with high correlation with nitrogen concentration and chlorophyll relative content SPAD through correlation analysis, and extract sensitive vegetation index and texture index from all spectral parameters using traditional feature selection;

[0053] According to the existing spectral parameters, three vegetation indices and three texture indices that are highly correlated with canopy nitrogen concentration and chlorophyll relative content SPAD were screened out by correlation analysis. The vegetation indices with the highest correlation with canopy nitrogen concentration are NREI, NDRE, and DATT, and the texture indices are R1m, Gm, and Rm; the vegetation indices with the highest correlation with chlorophyll relative content SPAD are GLI, NEXG, and NGBDI, and the texture indices are R2m, Gm, and Nm ( Figure 2 ).

[0054] Step 303: In order to prevent the use of too many spectral parameters during modeling, which may lead to overfitting of the model, the sensitive vegetation index and texture index extracted in step 302 are screened using the Boruta algorithm to reduce the characteristic parameters with low importance, and further obtain the vegetation index and texture index with higher SPAD correlation with nitrogen concentration and chlorophyll relative content, so as to improve the computational efficiency of subsequent modeling using machine learning (SVM, RF, KNN, ELM).

[0055] Step 4: To ensure the validity and reliability of the data, repeat steps 1 to 3 to obtain the multispectral and corn nitrogen nutrition parameter index data of the UAV during the V11-R1 period, and form a data set with the vegetation index and texture index selected each time;

[0056] Step 5: Based on the data set in step 4, linear functions and machine learning algorithms are used to construct regression prediction models for corn nitrogen nutrition parameters, and the best regression prediction model for each nitrogen nutrition parameter is selected;

[0057] The implementation of traditional regression is relatively simple and has high computational efficiency. Based on the three highly correlated vegetation indices and texture indexes selected, the regression model was established using linear functions with canopy nitrogen concentration and chlorophyll relative content SPAD. At the same time, machine learning (SVM, KNN, ELM and RF) methods were also used to establish regression prediction models and select the best model. When establishing the model, all data sets were divided into 67% training sets and 33% test sets.

[0058] The specific steps for screening the best regression prediction model for each nitrogen nutrition parameter are:

[0059] Step 501: Determination coefficient R 2 , root mean square error RMSE and normalized root mean square error nRMSE evaluate the performance of the regression prediction model of maize nitrogen nutrition parameters constructed using linear functions and machine learning algorithms, and select the best regression prediction model; the larger the R2, the smaller the RMSE and nRMSE, the better the performance of the model. R2, RMSE and nRMSE are calculated by the following formulas:

[0060]

[0061] Where n is the number of samples; y i is the measured value of the sample, y q is the simulated value of the sample; y m is the mean of the measured values ​​of the sample.

[0062] Step 502: Fitting the measured values ​​of nitrogen concentration and chlorophyll relative content SPAD of each corn canopy with the vegetation index and texture index with high correlation, and constructing a regression prediction model of nitrogen concentration and chlorophyll relative content SPAD of the corn canopy;

[0063] Tables 2 and 3 show the established linear and machine learning (SVM, RF, KNN, and ELM) models, respectively.

[0064] Table 2: Linear regression model

[0065]

[0066] Table 3: SVM, RF, KNN and ELM models

[0067]

[0068] Step 503: In order to evaluate the inversion accuracy of each model, the measured values ​​of nitrogen concentration and chlorophyll relative content SPAD in the validation set are compared with the estimated values ​​of the regression prediction model of nitrogen concentration and chlorophyll relative content SPAD in the corn canopy to verify the stability of the regression prediction model. 2 , RMSE and nRMSE are close to the training set, indicating that the model has good stability ( Figure 3 ).

[0069] Step 6: Based on the optimal regression prediction model, construct the inversion remote sensing monitoring image of corn nitrogen nutrition parameters ( Figure 4 ), to achieve information visualization of corn nitrogen nutrition parameters.

[0070] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for a person skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to replace some of the technical features therein by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions claimed to be protected by the present invention.

Claims

1. A method for unmanned aerial vehicle multispectral prediction of corn canopy nitrogen nutrition parameters, characterized in that: The steps include: Step 1: Set the range of the corn field to be tested as field plots with different nitrogen application rates, and measure the nitrogen nutrition parameters of the corn canopy leaves in each plot. The nitrogen nutrition parameters are the nitrogen concentration and chlorophyll relative content SPAD of the canopy leaves; Step 2: Plan the flight path of the drone according to the range of the corn field to be tested, collect multispectral images of the corn field through the multispectral camera carried by the drone, perform multispectral image standardization processing, and obtain vegetation index and texture index; Step 3: Conduct correlation analysis between vegetation index and texture index and nitrogen nutrition parameters of corn, and select vegetation index and texture index with high correlation from each nitrogen nutrition parameter; The specific steps of step 3 are: Step 301: using an environmental visualization program to determine plots with different nitrogen application rates, and extracting multispectral images of the plots; Step 302: Determine the vegetation index and texture index with high correlation with nitrogen concentration and chlorophyll relative content SPAD through correlation analysis, and extract sensitive vegetation index and texture index from all spectral parameters using traditional feature selection; Step 303: using the Boruta algorithm to filter the sensitive vegetation index and texture index extracted in step 302, and further obtain the vegetation index and texture index with higher correlation with nitrogen concentration and chlorophyll relative content SPAD; In the V11-R1 period of maize, the vegetation indices with the highest correlation with the nitrogen concentration of maize canopy leaves were NREI, NDRE, and DATT, and the texture indices were R1m, Gm, and Rm; the vegetation indices with the highest correlation with the relative chlorophyll content SPAD were GLI, NEXG, and NGBDI, and the texture indices were R2m, Gm, and Nm; Step 4: Repeat steps 1 to 3 to form a data set with the vegetation index and texture index screened each time; Step 5: Based on the data set in step 4, linear functions and machine learning algorithms are used to construct regression prediction models for corn nitrogen nutrition parameters, and the best regression prediction model for each nitrogen nutrition parameter is selected; Step 6: Based on the optimal regression prediction model, construct the inversion remote sensing monitoring image of corn nitrogen nutrition parameters to realize the information visualization of corn nitrogen nutrition parameters.

2. The method for unmanned aerial vehicle multispectral prediction of corn canopy nitrogen nutrition parameters according to claim 1, characterized in that: In step 5, the specific steps of screening the best regression prediction model for each nitrogen nutrition parameter are: Step 501: Determination coefficient R 2 , root mean square error RMSE and normalized root mean square error nRMSE to evaluate the performance of the regression prediction model of maize nitrogen nutrition parameters constructed by linear function and machine learning algorithm, and select the best regression prediction model; Step 502: Fitting the measured values ​​of each corn nitrogen concentration and chlorophyll relative content SPAD with the vegetation index and texture index with high correlation, and constructing a regression prediction model of the nitrogen concentration and chlorophyll relative content SPAD of corn canopy leaves; Step 503: Compare the measured values ​​of nitrogen concentration and chlorophyll relative content SPAD in the verification set with the estimated values ​​of the regression prediction model of nitrogen concentration and chlorophyll relative content SPAD in corn canopy leaves to verify the stability of the regression prediction model.

3. The method for unmanned aerial vehicle multispectral prediction of corn canopy nitrogen nutrition parameters according to claim 2, characterized in that: The machine learning methods include random forest, support vector machine, K nearest neighbor regression, and extreme learning machine.

4. The method for unmanned aerial vehicle multispectral prediction of corn canopy nitrogen nutrition parameters according to claim 1, characterized in that: In the step 1, the corn field to be tested is divided into rectangular areas with a side length of 10 m*13 m.

5. The method for unmanned aerial vehicle multi-spectral prediction of corn canopy nitrogen nutrition parameters according to claim 1, characterized in that: In step 2, the flight altitude of the UAV is 90-150 m; the field of view angle of the multispectral camera is 49.6°, the focal length is 5.2 mm, and the aperture is F / 22.

6. The method for unmanned aerial vehicle multi-spectral prediction of corn canopy nitrogen nutrition parameters according to claim 1, characterized in that: The corn is Northeast spring corn grown under no-till management conditions.

7. A UAV multi-spectral prediction system for corn canopy nitrogen nutrition parameters, characterized in that: Based on the method for unmanned aerial vehicle multi-spectral prediction of corn canopy nitrogen nutrition parameters according to claim 1, the system comprises: The nitrogen nutrition parameter measurement module is used to set the range of the corn field to be tested as field plots with different nitrogen application rates, and measure the nitrogen nutrition parameters of the corn canopy leaves in each plot; The vegetation index and texture index acquisition module is used to plan the flight path of the UAV according to the range of the corn field to be tested, collect multispectral images of the corn field through the multispectral camera carried by the UAV, perform multispectral image standardization processing, and obtain the vegetation index and texture index; A highly correlated vegetation index and texture index screening module is used to perform correlation analysis on the vegetation index and texture index with the nitrogen nutrition parameters of corn, and screen the vegetation index and texture index with high correlation from each nitrogen nutrition parameter; A data set forming module, used for forming a data set from the vegetation index and texture index screened each time; A regression prediction model construction and screening module is used to construct a regression prediction model of corn nitrogen nutrition parameters using a linear function and a machine learning algorithm according to the data set formed by the data set forming module, and screen out the best regression prediction model for each nitrogen nutrition parameter; The information visualization module of corn nitrogen nutrition parameters is used to construct the inversion remote sensing monitoring image of corn nitrogen nutrition parameters based on the optimal regression prediction model, and realize the information visualization of corn nitrogen nutrition parameters.

Citation Information

Patent Citations

  • Rice overground portion biomass estimation method based on multispectral images of unmanned aerial vehicle

    CN109459392A

  • Beet nitrogen nutrition detection method and system based on unmanned aerial vehicle multispectral data

    CN112986158A

  • Winter wheat SPAD spatio-temporal change monitoring method

    CN117760984A

  • Estimating nitrogen content using hyperspectral and multispectral images

    US20180035605A1

  • Method for estimating aboveground biomass of rice based on multi-spectral images of unmanned aerial vehicle

    US20200141877A1