Method and device for determining crop variety parameters
By acquiring data from multiple observation areas over a large region, a leaf area index estimation model and a crop growth model were established. Combining global optimization and spatial clustering algorithms, the spatial heterogeneity problem of crop variety parameters over a large region was solved, and high-precision determination of crop variety parameters was achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-05
- Publication Date
- 2026-03-10
AI Technical Summary
Existing technologies struggle to accurately account for the spatial heterogeneity of crop growth when acquiring crop variety parameters over large regions, and farmers' planting preferences are not fully reflected, resulting in poor parameter representativeness.
By obtaining data from multiple observation areas, including environmental data, crop leaf area index, and remote sensing data, a leaf area index estimation model was established. Combined with a crop growth model, a global optimization algorithm and a spatial clustering algorithm were used to determine the parameters of the target variety, taking into account spatial heterogeneity and farmer preferences.
It enables the determination of spatially representative crop variety parameters over a large area, accurately characterizing crop growth and improving the representativeness and accuracy of the parameters.
Smart Images
Figure CN121638672A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of biotechnology, and more specifically, to a method and apparatus for determining crop variety parameters. Background Technology
[0002] Process-based crop models, which comprehensively consider the interactions between environmental factors, crop variety information, and field management practices, are often used to develop crop irrigation and fertilization strategies to improve crop yield and water and fertilizer use efficiency. However, crop models are typically developed at a single-point scale. To meet the operational need for developing irrigation and fertilization strategies over large regions using crop models, the spatial heterogeneity of crop variety parameters must be considered. However, accurately obtaining information on the crop varieties grown by farmers within a large region is quite challenging.
[0003] Within this field, there have been explorations into obtaining crop variety parameters for large regions. One approach is to use a set of variety parameters to represent the varieties within a specific region. Another approach is to divide a large region into several sub-regions when the region is too large, and to determine a set of variety parameters for each sub-region. However, the above methods, which use a single set of variety parameters to characterize regional crop variety information, have poor representativeness and do not consider the tendency of farmers in the same area to choose similar crop varieties. This problem arises because the determination of a set of variety parameters is based on limited single-point experiments and does not take into account the growth conditions of crops throughout the region. Summary of the Invention
[0004] In view of the above, this application provides the following technical solution:
[0005] The first aspect of this application provides a method for determining crop variety parameters, including:
[0006] Data from observation sample areas are obtained. There are multiple observation sample areas. The data from the observation sample areas include environmental data, crop leaf area index measurements, phenological information, and yield. The crop leaf area index measurements and the yield are the average values of the corresponding values of several sampling points in the observation sample areas.
[0007] Obtain remote sensing data of the observed sample area, the remote sensing data including backscattering intensity and incident angle under polarization conditions;
[0008] The first model is trained based on the crop leaf area index measurement, the backscattering intensity, and the incident angle to obtain a leaf area index estimation model that meets the convergence condition.
[0009] The environmental data, the leaf area index estimated by the leaf area index estimation model, the phenological information, and the yield are input into the crop growth model to obtain the target variety parameters.
[0010] In a possible implementation, the obtaining the remote sensing data of the observation sample area comprises:
[0011] obtaining a dual-polarization C-band synthetic aperture radar image of the observation sample area;
[0012] performing filtering processing on the dual-polarization C-band synthetic aperture radar image to obtain a processed image;
[0013] extracting, from the processed image, backscattering intensity and incident angle under vertical transmission-vertical reception and vertical transmission-horizontal reception polarization conditions.
[0014] In a possible implementation, the training of the first model based on the crop leaf area index measurement value, the backscattering intensity and the incident angle to obtain a leaf area index estimation model meeting a convergence condition comprises:
[0015] establishing a correlation between the backscattering intensity, the incident angle, the crop leaf area index measurement value and the leaf area index measurement time by using a support vector regression method;
[0016] based on the correlation, taking the crop leaf area index measurement value and corresponding backscattering intensity and incident angle as a training set and a test set;
[0017] based on the training set and the test set, obtaining a leaf area index estimation model meeting a convergence condition.
[0018] In a possible implementation, the environmental data comprises geographical position, meteorological data, soil data and field management data.
[0019] In a possible implementation, the inputting, into a crop growth model, of the environmental data, the leaf area index estimated by the leaf area index estimation model, the phenology information and the yield to obtain target variety parameters comprises:
[0020] inputting, into a crop growth model, meteorological data, soil data, original target variety parameters and field management data;
[0021] assimilating, into the crop growth model, the leaf area index estimated by the leaf area index estimation model, the phenology information and the yield;
[0022] based on the crop growth model, iteratively adjusting the original target variety parameters by using a global optimization algorithm until the iteratively adjusted target variety parameters are in a preset normalized range and a cost function of the crop growth model decreases by less than a set value in continuous N times of iteration.
[0023] In a possible implementation, the method further includes: dividing the observation plots into simulation grid cells, each simulation grid cell obtaining a set of target variety parameters based on the crop growth model;
[0024] The multiple sets of target variety parameters are clustered using a spatial clustering algorithm, and the clustering features include the target variety parameters and geographical positions;
[0025] For each category obtained through the clustering, the mean value of all parameter values in the category is taken as the variety parameter of the current category.
[0026] In a possible implementation, the meteorological data of each simulation grid cell is determined based on known site meteorological data in combination with an interpolation method.
[0027] In a possible implementation, the soil data of each simulation grid cell is determined based on known soil grid data in combination with an inverse distance weighted difference method.
[0028] The second aspect of the application provides a device for determining crop variety parameters, including:
[0029] A first data obtaining module is configured to obtain observation plot data, the observation plot data including environmental data, crop leaf area index measurement values, phenology information, and yield, wherein the crop leaf area index measurement values and the yield are the mean values of corresponding values of a plurality of sampling points in the observation plot;
[0030] A second data obtaining module is configured to obtain remote sensing data of the observation plot, the remote sensing data including backscattering intensity and incident angle under polarization conditions.
[0031] A model training module is configured to train a first model based on the crop leaf area index measurement values, the backscattering intensity, and the incident angle, to obtain a leaf area index estimation model that meets a convergence condition.
[0032] A parameter obtaining module is configured to input the environmental data, the leaf area index estimated by the leaf area index estimation model, the phenology information, and the yield into a crop growth model, to obtain target variety parameters.
[0033] In a possible implementation, the device further includes:
[0034] A grid processing unit is configured to divide the observation plot into simulation grid cells, each simulation grid cell obtaining a set of target variety parameters based on the crop growth model.
[0035] The clustering processing module is configured to perform clustering processing on the multiple groups of target variety parameters by using a spatial clustering algorithm, and the clustering features include the target variety parameters and geographical positions; for each category obtained by the clustering processing, the mean value of all parameter values in the category is taken as the variety parameter of the current category.
[0036] According to the technical solution, the embodiment of the application discloses a method and device for determining crop variety parameters, the method comprising: obtaining observation plot data, the observation plot data including environmental data, crop leaf area index measurement values, phenology information, and yield; obtaining remote sensing data of the observation plot, the remote sensing data including backscattering intensity and incident angle under polarization conditions; training a first model based on the crop leaf area index measurement values, the backscattering intensity, and the incident angle to obtain a leaf area index estimation model that meets a convergence condition; and inputting the environmental data, the leaf area index estimated by the leaf area index estimation model, the phenology information, and the yield into a crop growth model to obtain target variety parameters. The method does not rely on limited single-point experiments to determine a group of variety parameters, but rather assimilates remote sensing information and regional investigation information (observation plot data) that can represent crop growth in a large area into a crop model, and the crop variety parameters determined by the method are spatially representative and take spatial heterogeneity into account, and can well represent crop variety parameters. BRIEF DESCRIPTION OF DRAWINGS
[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments or the prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only embodiments of the present application, and for those skilled in the art, other drawings can be obtained without creative labor on the basis of the provided drawings.
[0038] Figure 1 A flowchart of a method for determining crop variety parameters disclosed in the embodiments of the present application;
[0039] Figure 2 A flowchart of obtaining target variety parameters disclosed in the embodiments of the present application;
[0040] Figure 3 A flowchart of crop variety parameter clustering processing disclosed in the embodiments of the present application;
[0041] Figure 4 A structural schematic diagram of a device for determining crop variety parameters disclosed in the embodiments of the present application. DETAILED DESCRIPTION
[0042] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.
[0043] Figure 1 A flow chart of a method for determining crop variety parameters is disclosed in the embodiments of the present application. Referring to FIG. 1, Figure 1 The method for determining crop variety parameters can include the following steps.
[0044] Step 101: Obtain observation plot data, the observation plot has multiple observation plots, and the observation plot data includes environmental data, crop leaf area index measurement values, phenological information, and yield, wherein the crop leaf area index measurement values and the yield are the average values of corresponding values of a plurality of sampling points in the observation plot.
[0045] In the embodiments of the present application, the crop in the observation plot is corn, but this does not affect the implementation logic of the scheme for determining crop variety parameters disclosed in the present application. In actual application, the crop variety for which the variety parameters need to be determined is not fixedly limited. In order to ensure the objectivity and comprehensiveness of the data, multiple observation plots can be selected. The environmental data of the observation plot can include but is not limited to meteorological data, soil data, field management data, etc.
[0046] In the actual data collection work of the observation plot, that is, during the development of the scheme of the present application, 23 plots with a size of 10 × 10 m² were selected in a certain county area to monitor the growth of corn in 2020. These plots are far away from roads and trees, and the center positions are determined by a global positioning system device. In each plot, three 1 × 1 m² sampling points are selected, which are uniformly distributed in the plot. The LAI (leaf area index) determination time is from July 16-18, 2020, July 30-August 1, 2020, August 20-22, 2020, September 6-8, 2020, and September 14-16, 2020. The yield is determined when the corn matures, and the male flowering stage and the mature stage are recorded. The LAI and yield values of each plot are the average values of the three sampling points. In addition, the sowing date, sowing depth, plant spacing, row spacing, irrigation method, irrigation date, irrigation water volume, fertilization method, fertilization time, and fertilization amount of each plot are also collected.
[0047] In addition, 56 sample plots of 10 x 10 m2were selected in a certain plain area to monitor the growth of summer maize in 2021. In these sample plots, LAI measurements were taken on July 15-17, July 26-28, August 5-7, and August 23-25, 2021. Yield was measured after the maize was harvested, and the male stage and maturity stage were recorded. In addition, various field management parameters were collected, such as irrigation date, water quantity, fertilization time, and fertilization quantity.
[0048] In each growth season, LAI was measured using a measuring instrument with a precision of 1 mm. The specific method is to measure the length (from the leaf sheath to the leaf tip) and width (the widest part of the leaf) of all the leaves of the selected plant. Then, the single leaf area and LAI are calculated according to the following formula, respectively:
[0049]
[0050]
[0051] In the formula, the single leaf area (ASA) is expressed in square meters (m2), L and W are the length and width of the leaf (meters), respectively, and the coefficient 0.75 is a correction coefficient; the total area of all leaves of a single corn plant (TA) is expressed in square meters, and the planting density (PD) is expressed in the number of plants per hectare.
[0052] Step 102: Obtain remote sensing data of the observation sample plot, the remote sensing data including backscattering intensity and incident angle under polarization conditions.
[0053] In implementation, in order to determine the planting area of corn, a random forest classifier can be used on satellite data to identify and determine. The ground data collected by the satellite has usually been radiometrically and geometrically corrected, and the classification features (including normalized vegetation index NDVI, normalized water index NDWI, and normalized building index NDBI) extracted based on the ground data collected by the satellite, as well as the category information (corn and non-corn categories) obtained through field investigation, use the random forest classifier to extract the corn planting area. The random forest classifier shows good accuracy in the identification of corn and non-corn samples, with a Kappa coefficient of 0.91 and an overall accuracy of 95.52%.
[0054] On the basis of the extracted corn planting area, the leaf area index (LAI) of corn was estimated using the backscatter information of satellites. The satellite data collected included dual-polarized C-band synthetic aperture radar (SAR) images (frequency of 5.4 GHz) with a revisit period of 6 days. The microwave remote sensing system is independent of solar radiation, and therefore, the SAR images are particularly suitable for crop growth monitoring in cloudy and rainy areas. The stored relevant SAR images have been subjected to various pre-processing, including thermal noise removal, radiometric correction, and terrain correction. To reduce the coherent speckle noise, the collected SAR images can be further subjected to an improved Lee filter. In addition, from the stored SAR images, the backscatter intensity and incidence angle information under the conditions of vertical transmission-vertical reception polarization and vertical transmission-horizontal reception polarization are extracted.
[0055] Based on the above, obtaining the remote sensing data of the observation sample area can include: obtaining dual-polarized C-band synthetic aperture radar images of the observation sample area. The dual-polarized C-band synthetic aperture radar images are subjected to filtering processing to obtain processed images. From the processed images, the backscatter intensity and incidence angle under the conditions of vertical transmission-vertical reception and vertical transmission-horizontal reception polarization are extracted.
[0056] Step 103: training a first model based on the crop leaf area index measurement value, the backscatter intensity, and the incidence angle to obtain a leaf area index estimation model that meets the convergence condition.
[0057] To estimate the regional LAI of summer corn, the support vector regression (SVR) method can be used to establish the relationship between the backscatter intensity, incidence angle, LAI measurement date, and the 2021 measured LAI, and the LAI observation date is converted into the day sequence in a year. The SVR model has been widely used in crop LAI estimation, and even with a small number of training samples, it can also show good performance. The observed LAI and the corresponding microwave remote sensing data are organized into a training set (n = 130) and a test set (n = 94) for training and verification of the LAI estimation model. In the training set, the LAI predicted by the established LAI estimation model is highly consistent with the measured value; the corresponding root mean square error (RMSE), average relative error (ARE), and bias are 0.49 m² / m², 36.03%, and 0.05 m² / m², respectively. In the test set, the RMSE, ARE, and bias are 0.59 m² / m², 41.89%, and 0.02 m² / m², respectively. In actual experiments, the LAI measured value and the predicted value are highly consistent, indicating that the established LAI estimation model can predict the LAI with high accuracy according to the microwave remote sensing data.
[0058] In combination with the foregoing, training the first model based on the crop leaf area index measurement value, the backscattering intensity and the incident angle to obtain a leaf area index estimation model meeting a convergence condition can include: using a support vector regression method to establish a correlation between the backscattering intensity, the incident angle, the crop leaf area index measurement value and the leaf area index measurement time; based on the correlation, taking the crop leaf area index measurement value and the corresponding backscattering intensity and incident angle resistance value as a training set and a test set; based on the training set and the test set, obtaining a leaf area index estimation model meeting a convergence condition.
[0059] Step 104: inputting the environment data, the leaf area index estimated by the leaf area index estimation model, the phenology information and the yield into a crop growth model to obtain target variety parameters.
[0060] The crop growth model can be a CERES-Maize model. The CERES-Maize model is a process-based model that can be used to simulate the growth and development of corn under different climatic conditions, soil conditions and field management measures. The model uses photosynthetically active radiation intercepted by the crop canopy to calculate the potential crop growth; but the actual growth of the crop on a certain day will be limited by soil water stress, nutrient stress and temperature stress. In order to determine the daily soil water balance, the CERES-Maize model uses a one-dimensional soil water balance method called "tilting bucket". This method is used to predict the migration process of soil water and the water absorption of roots in up to 10 soil layers.
[0061] The present application uses the CERES-Maize model to simulate the spatial growth of corn. In addition, field management information is obtained through field measurement and interviews with local farmers. Existing research through parameter sensitivity analysis shows that parameters P2 (photosynthetic sensitivity parameter) and P5 (accumulated temperature required from silking to maturity) and G3 (potential grain filling rate during linear grain filling) have a significant impact on leaf area index LAI and yield, respectively. Therefore, only P2, P5 and G3 are re-estimated, and the data used include remote sensing estimated LAI, regional recorded phenology information and measured yield.
[0062] The method for determining crop variety parameters described in the present embodiment no longer relies on limited single-point tests to determine a set of variety parameters, but rather assimilates remote sensing information and regional survey information (observation sample data) that can characterize the growth of crops in a large area into a crop model. The crop variety parameters determined by this method are spatially representative and take into account spatial heterogeneity, and can well characterize crop variety parameters.
[0063] In the above embodiments, the environment data can include geographic location, meteorological data, soil data, field management data, etc.Figure 2 A flowchart for obtaining target variety parameters is disclosed in embodiments of the present application. Referring to FIG. 1, the flowchart for obtaining target variety parameters comprises the following steps: Figure 2 As shown, the inputting the environmental data, the leaf area index estimated by the leaf area index estimation model, the phenology information and the yield into the crop growth model to obtain the target variety parameters can comprise:
[0064] Step 201: inputting meteorological data, soil data, original target variety parameters and field management data into the crop growth model.
[0065] The meteorological data of each simulation grid cell is determined based on known site meteorological data combined with an interpolation method. The soil data of each simulation grid cell is determined based on known soil grid data combined with an inverse distance weighted difference method. Details will be described in the embodiments below and will not be described here.
[0066] Step 202: assimilating the leaf area index estimated by the leaf area index estimation model, the phenology information and the yield into the crop growth model.
[0067] Step 203: based on the crop growth model, using a global optimization algorithm to iteratively adjust the original target variety parameters until the iteratively adjusted target variety parameters are within a preset normalized range and the cost function of the crop growth model decreases by less than a set value in consecutive N iterations.
[0068] Specifically, the present application uses the CERES-Maize model to simulate the spatial growth of corn. Therefore, the input data required by the crop model (including meteorological data, soil data, variety parameters and field management data) set needs to be regionalized. In addition, field management information is also obtained through field measurement and interviews with local farmers.
[0069] In the implementation, the SCE-UA optimization algorithm can be used to iteratively adjust the variety parameters P2, P5 and G3, assimilate the remotely sensed LAI, the recorded phenology information (tasseling stage, mature stage) and the measured yield data into the CERES-Maize model, and minimize the following cost function:
[0070]
[0071] where J represents the cost function, X is a vector consisting of P2, P5, G3, LAIsim is the simulated LAI using CERES-Maize model, LAIrs is the estimated LAI from remote sensing data, Q is the number of remote sensing data points, K is the number of simulation units. Antsim and Matsim represent the simulated tasseling and maturing dates, Antmea and Matmea represent the observed tasseling and maturing dates, Yieldsim and Yieldmea represent the simulated and measured yields, respectively. When a simulation unit lacks the measured data of phenological dates and yield, the data of its nearest simulation unit is used as a substitute. The SCE-UA (Soil and Water Assessment Tool-Calibration using Universal Approximation) optimization algorithm of the University of Arizona is used to minimize the cost function in the above formula. This algorithm is a global optimization algorithm that combines the advantages of deterministic search, random search and competitive evolution, and is not sensitive to the initial parameter value. When the optimized parameter value is in the preset normalized range and the decline of the cost function in the last 10 iterations is not more than 0.01%, it is considered that the SCE-UA algorithm has converged.
[0072] In other implementations, considering the cost of calculation, the observed sample area can be divided into simulation units in a 3 x 3 km² grid, and the total area of the simulation units containing corn pixels is 5724 km². In order to convert the site-scale meteorological data (daily precipitation, maximum temperature, minimum temperature and solar radiation) into grid point data, the ordinary Kriging interpolation method is used. These site meteorological data come from the China Meteorological Data Service Center. Soil grid data with a spatial resolution of 5' has been provided in the field. In order to make the spatial resolution of the soil data consistent with the meteorological data, it is converted into 3 x 3 km² resolution soil data by inverse distance weighted interpolation.
[0073] In this method, 636 (5724 ÷ 9) sets of variety parameters (P2, P5, G3) data were obtained. In regional simulation, if each set of data is input into the crop model separately, the calculation cost of the model will be significantly increased. To solve this problem, efficient data mining techniques are needed to extract important features from massive data. Clustering techniques can divide a large data set into several subsets with different levels of similarity. Researchers have developed a variety of clustering algorithms, among which the density-based spatial clustering algorithm is a widely used algorithm. The advantage of this algorithm is that it can identify clusters of arbitrary shape and noise points (outliers) in the data, which traditional clustering algorithms such as k-means cannot do. Spatial clustering algorithms do not need to specify the number of clusters in advance, and the cluster shape does not have to be spherical, so they are particularly suitable for this study. Therefore, in this study, spatial clustering algorithms were used to analyze the clustering of 636 summer corn simulation units, and the clustering characteristics included P2, P5, G3, and geographic location (longitude and latitude). The average value of each cluster was used as the variety parameter of the class.
[0074] Based on the above, the observation plot is divided into simulation grid cells, and each simulation grid based on the crop growth model obtains a set of target variety parameters. The method for determining crop variety parameters also includes a clustering process. Figure 3 The flowchart of the crop variety parameter clustering process disclosed in the embodiments of the present application is shown in Figure 3 As shown, the clustering process can include:
[0075] Step 301: Use a spatial clustering algorithm to cluster multiple sets of target variety parameters, and the clustering characteristics include target variety parameters and geographic location.
[0076] Step 302: For each class obtained by the clustering process, take the average value of all parameter values in the class as the variety parameter of the current class.
[0077] The following table lists the variety parameters obtained by this method. A total of six clusters were obtained, and most of the parameters were distributed in the first, second, fifth, and sixth classes.
[0078] Regional parameter estimation results
[0079] Cluster P2(-) P5 (°C d) G3 (mg / d) P1 (°C d) G2 (number) PHINT (°C d) CAGCs-1 1.054 614.7 11.02 CAGCs -2 0.509 636.2 9.93 CAGCs -3 1.002 735 8.92 CAGCs -4 1.145 715.2 9 CAGCs -5 1.197 717.1 11.12 CAGCs -6 1.083 769.5 11.2
[0080] In the above table, P2: photosynthetic sensitivity parameter; P5: accumulated temperature required from silking to maturity; G3: potential grain filling rate during linear grain filling; P1: accumulated temperature from emergence to end of seedling; G2: potential grain number per plant; PHINT: accumulated temperature required for continuous leaf tip appearance.
[0081] Based on the foregoing scheme, the research team simulated the LAI values, yields, male plucking periods and maturation periods in 2020 and 2021 respectively to verify the accuracy thereof, and the accuracy was high, the RMSE (root mean square error) of the simulated LAI in 2020 and 2021 was 0.42 and 0.53 m respectively 2 / m 2 , and the RMSE of the simulated yield was 1060.28 kg / ha and 945.77 kg / ha respectively.
[0082] The method for determining crop variety parameters is disclosed in the embodiments of the application, which innovatively integrates a spatial clustering algorithm into the assimilation of the leaf area index obtained by remote sensing and the crop model. In the implementation, a set of variety parameters is determined for each pixel. Considering that determining a set of variety parameters for each pixel is not conducive to subsequent field water and fertilizer management and does not take into account the reality that farmers in the same region often choose similar crop varieties, the variety parameters and geographic location information are further used as clustering features, and finally a set of variety parameters is assigned to each class. The crop variety parameters determined by the method have spatial representativeness, consider spatial heterogeneity, and consider farmer preferences.
[0083] For the method embodiments described above, in order to simply describe, they are all expressed as a series of action combinations, but those skilled in the art should know that the application is not limited by the order of the described actions, because according to the application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions and modules involved are not necessarily required by the application.
[0084] The method is described in detail in the embodiments disclosed in the application, and the method of the application can be implemented in various forms of devices, so the application also discloses a device, and specific embodiments are given below to be described in detail.
[0085] Figure 4 A structural schematic diagram of a device for determining crop variety parameters disclosed in the embodiments of the application is shown in FIG. 4. As shown in FIG. 4, the device 40 for determining crop variety parameters can include: Figure 4
[0086] A first data obtaining module 401 is configured to obtain observation plot data, the observation plot has a plurality of observation plots, and the observation plot data includes environmental data, crop leaf area index measurement values, phenological information and yield, wherein the crop leaf area index measurement values and the yield are the average values of the corresponding values of a plurality of sampling points in the observation plot.
[0087] A second data obtaining module 402 is configured to obtain remote sensing data of the observation plot, and the remote sensing data includes backscattering intensity and incident angle under polarization conditions.
[0088] The model training module 403 is configured to train a first model based on the crop leaf area index measurement value, the backscattering intensity, and the incident angle, to obtain a leaf area index estimation model meeting a convergence condition.
[0089] The parameter obtaining module 404 is configured to input the environmental data, the leaf area index estimated by the leaf area index estimation model, the phenology information, and the yield into a crop growth model, to obtain target variety parameters.
[0090] The device for determining crop variety parameters in the embodiment does not rely on limited single-point tests to determine a set of variety parameters, but rather assimilates remote sensing information and regional investigation information capable of representing crop growth conditions in a large area into a crop model. The crop variety parameters determined by the method have spatial representativeness and consider spatial heterogeneity, and can well represent crop variety parameters.
[0091] In one implementation, the second data obtaining module can be specifically configured to: obtain a dual-polarized C-band synthetic aperture radar image of the observation sample area; perform filtering processing on the dual-polarized C-band synthetic aperture radar image to obtain a processed image; and extract, from the processed image, backscattering intensities and incident angles under vertical transmission-vertical reception and vertical transmission-horizontal reception polarization conditions.
[0092] In one implementation, the model training module can be specifically configured to: establish a correlation between backscattering intensity, incident angle, crop leaf area index measurement value, and leaf area index measurement time by using a support vector regression method; based on the correlation, take the crop leaf area index measurement value and corresponding backscattering intensity and incident angle as a training set and a test set; and based on the training set and the test set, obtain a leaf area index estimation model meeting a convergence condition.
[0093] In one implementation, the environmental data includes geographic location, meteorological data, soil data, and field management data.
[0094] In one implementation, the parameter obtaining module can be specifically configured to: input meteorological data, soil data, original target variety parameters, and field management data into a crop growth model; assimilate the leaf area index estimated by the leaf area index estimation model, the phenology information, and the yield into the crop growth model; based on the crop growth model, iteratively adjust the original target variety parameters by using a global optimization algorithm until the iteratively adjusted target variety parameters are in a preset normalized range, and a cost function of the crop growth model decreases by less than a set value in consecutive N iterations.
[0095] In one implementation, the apparatus can further include a grid processing unit configured to divide the observation plot into simulation grid cells, each simulation grid cell being associated with a set of target variety parameters based on the crop growth model; and a clustering processing module configured to perform clustering processing on the sets of target variety parameters using a spatial clustering algorithm, the clustering features including the target variety parameters and geographical locations; and for each cluster obtained by the clustering processing, taking the mean of all parameter values in the cluster as the variety parameter of the current cluster.
[0096] In one implementation, the meteorological data of each simulation grid cell is determined based on known site meteorological data in combination with an interpolation method.
[0097] In one implementation, the soil data of each simulation grid cell is determined based on known soil grid data in combination with an inverse distance weighted difference method.
[0098] The specific implementation of the above-described apparatus for determining crop variety parameters and each module included therein, as well as other possible implementations, can be found in the content of the corresponding part of the method embodiments, which will not be repeated here.
[0099] The apparatus for determining crop variety parameters in any of the above-described embodiments includes a processor and a memory, and the first data obtaining module, the second data obtaining module, the model training module, the parameter obtaining module, the grid processing unit, the clustering processing module, etc. in the above-described embodiments are stored in the memory as program modules, and the corresponding functions are implemented by the processor executing the above-described program modules stored in the memory.
[0100] The processor includes a core, and the core retrieves the corresponding program modules from the memory. The core can be set to one or more, and the processing of the revisit data is implemented by adjusting the core parameters.
[0101] The memory can include a non-permanent memory in a computer readable medium, a random access memory (RAM) and / or a non-volatile memory such as a read-only memory (ROM) or a flash memory (flash RAM), and the memory includes at least one memory chip.
[0102] In an exemplary embodiment, a computer readable storage medium directly loadable into the internal memory of a computer is also provided, which contains software codes. The computer program, when loaded and executed by a computer, can implement the steps shown in any of the above-described embodiments of the method for determining crop variety parameters.
[0103] In an exemplary embodiment, a computer program product directly loadable into the internal memory of a computer is also provided, which contains software codes. The computer program, when loaded and executed by a computer, can implement the steps shown in any of the above-described embodiments of the method for determining crop variety parameters.
[0104] The various embodiments described in this specification are presented by way of example, and are not intended to limit the scope of the application. The various embodiments described in this specification can be combined in any combination. Various embodiments of the application are described in detail below.
[0105] It is also important to note that the use of relative terms, such as first and second, in describing the various embodiments is not an indication of a relative importance or a relative size. Rather, such terms are used only to distinguish one embodiment from another. Furthermore, the use of the term "including" as well as other forms such as "include", "includes", and "included" is intended to cover a non-exclusive inclusion, such that items not explicitly listed are also to be considered as included. The use of the term "comprising" as well as other forms such as "comprise", "comprises", and "comprised of" is intended to cover a non-exclusive inclusion, such that items not explicitly listed are also to be considered as included. The use of the term "consisting of" as well as other forms such as "consist of", "consists of", and "consisting of is intended to cover an exclusive inclusion, such that items not explicitly listed are not to be considered as included.
[0106] The steps of a method or algorithm described in connection with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module can reside in RAM, flash memory, ROM, electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such that the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium can be integral to the processor. The processor and the storage medium can reside in an ASIC.
[0107] The above description of disclosed embodiments is intended to be illustrative and not restrictive. Many embodiments of the application will be apparent to those of skill in the art upon reviewing the above description. The scope of the application should, therefore, be determined not with reference to the above description, but instead with reference to the appended claims, along with their full scope of equivalents.
Claims
1. A method of determining crop variety parameters, characterized by, The application relates to a method for obtaining target variety parameters, comprising the following steps: obtaining observation sample area data, the observation sample area has a plurality of observation sample areas, the observation sample area data comprises environmental data, crop leaf area index measurement values, phenology information and yield, wherein the crop leaf area index measurement values and the yield are the average values of corresponding values of a plurality of sampling points in the observation sample area; obtaining remote sensing data of the observation sample area, the remote sensing data comprises backscattering intensity and incident angle under polarization conditions; training a first model based on the crop leaf area index measurement values, the backscattering intensity and the incident angle, so as to obtain a leaf area index estimation model meeting a convergence condition; inputting the environmental data, the leaf area index estimated by the leaf area index estimation model, the phenology information and the yield into a crop growth model, so as to obtain target variety parameters. The application further relates to a method for obtaining target variety parameters, comprising the following steps: obtaining dual-polarization C-band synthetic aperture radar images of the observation sample area; performing filtering processing on the dual-polarization C-band synthetic aperture radar images, so as to obtain processed images; and obtaining backscattering intensity and incident angle under vertical emission-vertical reception and vertical emission-horizontal reception polarization conditions from the processed images. The application further relates to a method for obtaining target variety parameters, comprising the following steps: establishing a correlation between backscattering intensity, incident angle, crop leaf area index measurement values and leaf area index measurement time by using a support vector regression method; taking the crop leaf area index measurement values and corresponding backscattering intensity and incident angle values as a training set and a test set based on the correlation; and obtaining a leaf area index estimation model meeting a convergence condition based on the training set and the test set. The application further relates to a method for obtaining target variety parameters, comprising the following steps: inputting meteorological data, soil data, original target variety parameters and field management data into a crop growth model; and assimilating leaf area index values estimated by the leaf area index estimation model, the phenology information and the yield into the crop growth model. The application further relates to a method for obtaining target variety parameters, comprising the following steps: dividing the observation sample area into simulation grid units, each simulation grid unit obtains a group of target variety parameters based on the crop growth model; adopting a spatial clustering algorithm to perform clustering processing on the groups of target variety parameters, and the clustering features comprise target variety parameters and geographical positions; and taking the average values of all parameter values in each category as the variety parameters of the current category.
2. The method of determining crop variety parameters according to claim 1, wherein, The meteorological data of each simulation grid unit is determined based on known station meteorological data and an interpolation method. 3. The method of determining crop variety parameters according to claim 1, wherein, 4. The method of determining crop variety parameters according to claim 1, characterized in that, 5. The method of determining crop variety parameters according to claim 4, wherein, 6. The method of determining crop variety parameters according to claim 1, wherein, 7. The method of determining crop variety parameters according to claim 6, wherein, 8. The method of determining crop variety parameters according to claim 6, wherein, Soil data for each simulation grid cell is determined based on known soil grid data in combination with inverse distance weighted interpolation.
9. An apparatus for determining crop variety parameters, characterized by Comprise: A first data obtaining module is configured to obtain observation plot data, the observation plot has a plurality of observation plots, and the observation plot data comprises environmental data, crop leaf area index measurement values, phenology information, and yield, wherein the crop leaf area index measurement values and the yield are mean values of corresponding values of a plurality of sampling points in the observation plots; A second data obtaining module is configured to obtain remote sensing data of the observation plots, the remote sensing data comprising backscattering intensity and incident angle under polarization conditions; A model training module is configured to train a first model based on the crop leaf area index measurement values, the backscattering intensity, and the incident angle, to obtain a leaf area index estimation model satisfying a convergence condition; A parameter obtaining module is configured to input the environmental data, leaf area index estimated by the leaf area index estimation model, the phenology information, and the yield into a crop growth model, to obtain target variety parameters.
10. The apparatus for determining crop variety parameters according to claim 9, wherein, Further comprise: A grid processing unit is configured to divide the observation plots into simulation grid cells, and each simulation grid cell obtains a set of target variety parameters based on the crop growth model; A clustering processing module is configured to perform clustering processing on a plurality of sets of target variety parameters by using a spatial clustering algorithm, and the clustering characteristics comprise target variety parameters and geographical positions; For each category obtained by the clustering processing, the mean value of all parameter values in the category is taken as the variety parameter of the current category.