A crop growth model optimization method and system based on phenotypic curve
By integrating land surface mode and single point scale model, combining spatiotemporal clustering and phenotypic curve model, the problem that existing crop models cannot achieve continuous regional simulation is solved, and the multi-environmental adaptability and accuracy of crop growth models are improved, reducing the number of crop growth models. System errors were processed and spatially.
Patent Information
- Application Number
- CN202410658550.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-05-27
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2044-05-27
AI Technical Summary
The existing crop models are based on point scales and cannot achieve continuous simulation of the regional range, resulting in large system errors and direct use of spatial data to transmit errors, which is separated from the crop model's own characteristics.
By integrating the land surface model and single-point scale crop model, the environmental parameters are initialized using the spatiotemporal clustering algorithm, the phenotypic curve model corrects the genetic parameters, avoiding the direct use of spatial data, and interpolation processing is performed through the Kriging algorithm to achieve regional scope continuous simulation and spatialization.
It improves the multi-environmental adaptability and simulation accuracy of the crop growth model, reduces system errors, and realizes continuous simulation and spatial processing in the region.
Smart Images

Figure CN118313160B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of data processing, and in particular relates to a crop growth model optimization method and system based on a phenotypic curve. Background Art
[0002] Crop models can help understand many complex crop growth processes through simulation. At the beginning of development, crop models were mainly single-point models designed according to specific spatial scales or farmland areas. However, as new spatial data are provided, they are increasingly used in different regions and spatial scales. There are two main aspects to the need for crop spatialization. On the one hand, there is the cross-domain migration of models. For example, in crop management applications at specific locations, different plots need to be managed differently and in a variable manner due to the needs of precision agriculture. On the other hand, there is the cross-scale application of models. For example, when revealing and understanding spatial heterogeneity, crop models are limited to specific spatial scales, so upsampling and downsampling are used to understand the simulation process at different spatial scales. Therefore, the spatialization of crop models directly affects the migration application and scalability of crop models. In order to spatialize crop models, spatialization can be performed at the input and output stages of the model; or the process framework of the crop model can be modified to complete the spatialization.
[0003] However, existing crop models are point-scale models that cannot achieve continuous simulation at a regional scale. Most existing crop models use spatialized input and output methods, which are essentially predictions of geographic data and are divorced from the characteristics of the crop models themselves, so they have large systematic errors. The main reason is that existing crop models directly use spatialized data, and the errors of spatialized data will be directly transmitted to the model. Summary of the invention
[0004] In view of the above deficiencies in the prior art, the purpose of the invention is to provide a crop growth model optimization method and system based on phenotypic curves, which can integrate the land surface pattern model with the crop model at a single point scale to construct a crop growth model, and can realize continuous simulation in a regional range. The characteristics of the crop model itself are taken into account, and the environmental parameters of the crop growth model are initialized through a spatiotemporal clustering algorithm, the model error is corrected through a phenotypic curve model, and the crop genetic parameters of the crop growth model are initialized. After that, the crop growth simulation results are output through the initialized crop growth model, thereby avoiding the direct use of spatialized data and reducing system errors.
[0005] In a first aspect of the present invention, a crop growth model optimization method based on phenotypic curves is proposed, comprising:
[0006] S1, integrate the land surface model with the crop model at a single point scale to construct a crop growth model;
[0007] S2, initializing the environmental parameters of the crop growth model by using a spatiotemporal clustering algorithm;
[0008] S3, correcting the model error through the phenotypic curve model, and initializing the crop genetic parameters of the crop growth model;
[0009] S4, outputting crop growth simulation results through the initialized crop growth model;
[0010] S5, performing interpolation processing on the crop growth simulation result by using a Kriging algorithm.
[0011] A second aspect of the present invention provides a crop growth model optimization system based on phenotypic curves, comprising: a memory and one or more processors.
[0012] One or more application programs are stored in the memory, and the one or more application programs are suitable for being executed by one or more processors to implement the crop growth model optimization method based on phenotype curves of the first aspect.
[0013] According to a third aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. The computer program can be loaded and executed by a processor to implement the crop growth model optimization method based on phenotypic curves according to the first aspect.
[0014] The beneficial effects of the present invention are as follows:
[0015] The method and system of the present invention fuse the land surface pattern model with the crop model at a single point scale to construct a crop growth model, which can realize continuous simulation in a regional scope, take into account the characteristics of the crop model itself, initialize the environmental parameters of the crop growth model through a spatiotemporal clustering algorithm, realize the spatialization of parameters at a regional scale, and improve the multi-environment adaptability of the model, correct the model error through a phenotypic curve model, initialize the crop genetic parameters of the crop growth model, and then output the crop growth simulation results through the initialized crop growth model, avoid directly using spatialized data, reduce system errors, and perform spatial continuous interpolation processing on discrete crop growth simulation results through a Kriging algorithm, thereby realizing the spatialization of simulation results. BRIEF DESCRIPTION OF THE DRAWINGS
[0016] The accompanying drawings are only used to illustrate specific embodiments and are not considered to limit the present invention. In the entire drawings, the same reference symbols represent the same components. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.
[0017] Figure 1A schematic diagram of a process flow of a crop growth model optimization method based on phenotypic curves provided in an embodiment of the present invention;
[0018] Figure 2 A schematic diagram of the structure of a crop growth model optimization system based on phenotypic curves provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0019] In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. It should be understood that these descriptions are only exemplary and are not intended to limit the scope of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without making creative work should fall within the scope of protection of the present invention.
[0020] Furthermore, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessarily obscuring the concepts disclosed in the present invention.
[0021] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the orientation or position relationship indicated by the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inside", "outside", etc. is based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation of the present invention. In addition, the terms "first", "second", and "third" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. The terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be a connection between the two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0022] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Instead, they are merely examples of methods and systems consistent with some aspects of the present invention as detailed in the appended claims.
[0023] The present invention proposes a crop growth model optimization method and system based on phenotypic curves, which solves the problem that the existing crop model is a point-scale model and cannot achieve continuous simulation in a regional range. Most of the existing crop models use a spatial input and output method, which is essentially a prediction of geographic data and is separated from the characteristics of the crop model itself, so it has problems such as large system errors.
[0024] Method Embodiment
[0025] Reference Manual Attached Figure 1 , showing a schematic flow chart of a crop growth model optimization method based on phenotypic curves provided in an embodiment of the present invention.
[0026] An embodiment of the present invention provides a crop growth model optimization method based on a phenotypic curve, comprising:
[0027] Specifically, the method includes steps S1 to S5.
[0028] S1, the land surface model is integrated with the crop model at a single point scale to construct a crop growth model.
[0029] Among them, the Land Surface Model (LSM) is a numerical model used to simulate surface processes. It is widely used in climate, hydrology and ecosystem research. The LSM model aims to describe the energy, moisture and momentum exchange processes on the surface, as well as the interaction between the surface and the atmosphere.
[0030] Among them, the crop model CGM (CERES crop growth model) is a classic crop growth model. The CGM model is designed to simulate the growth, development and yield of major food crops (such as wheat, corn, rice, etc.).
[0031] It should be noted that the traditional crop model does not consider the impact of crops on land-atmosphere interaction, which will have a greater impact on crop simulation in cross-regional expansion, while the land surface model realizes the key land-atmosphere interaction process, and the land surface model is a continuous simulation of the surface space. Therefore, in order to meet the spatial needs of the crop model, the crop model is combined with the land surface model.
[0032] In a possible implementation, S1 specifically includes: based on the Noah-PM model, taking the vegetation area index as the simulation process quantity, introducing the dynamic growth process of the CERES crop model, and constructing a crop growth model.
[0033] Among them, the Noah-PM model is a land surface process model that aims to simulate and predict surface energy and water exchange processes, including soil water movement, vegetation growth, evaporation and transpiration, etc. The model combines the Noah land surface model of the National Center for Climate Change (NCEP) and the Penman-Monteith formula of the United States Department of Agriculture, and can accurately simulate surface energy and water processes. It is a commonly used component in the Earth System Model (ESM).
[0034] Among them, the Leaf Area Index (LAI) is an important indicator used to describe the density of vegetation. It refers to the sum of the leaf area of vegetation per unit area, usually expressed as the ratio of leaf area to ground area. The value range of LAI is usually between 0 and infinity, indicating the surface area of vegetation leaves per unit area of ground.
[0035] Among them, the CERES crop model is a classic crop growth model. CERES stands for "Crop Estimation through Resource and Environment Synthesis". The model is designed to simulate the growth, development and yield of major food crops (such as wheat, corn, rice, etc.).
[0036] Furthermore, the crop growth model calculates the vegetation area index, which in turn directly affects surface radiation, photosynthesis process, soil moisture and surface energy flux. This interaction between crops and land-climate reflects the characteristics of farmland climate in different regions. The Noah-PM model is used to further adjust the land surface pattern of the farmland environment, and finally the spatialization of the crop model is realized.
[0037] In the present invention, traditional crop models usually only consider the internal growth process of crops, while ignoring the interaction between crops and the surface environment and atmosphere. After integrating the LSM of the land surface model, the crop model can more accurately consider the impact of crops on surface energy, moisture and momentum, thereby more realistically reflecting the growth of crops. At the same time, the LSM model has the ability to continuously simulate the surface space and can simulate the energy and moisture distribution of the surface in a regional range. By integrating the crop model with the LSM, the crop simulation can be extended to the spatial scale, and continuous simulation of crop growth in different regions can be achieved, so as to better understand and predict the growth characteristics of crops in different geographical environments.
[0038] In summary, the integration of crop models with land surface model can comprehensively consider the interaction between crops and the surface environment and atmosphere, realize the spatial simulation of crop growth, improve the simulation accuracy, better respond to challenges such as climate change, and provide more scientific decision-making support for agricultural production and management.
[0039] S2, initialize the environmental parameters of the crop growth model through the spatiotemporal clustering algorithm.
[0040] Among them, the spatiotemporal clustering algorithm is a clustering method for processing spatiotemporal data. It takes into account the temporal and spatial information of data points in order to divide them into groups with similar characteristics. This algorithm has a wide range of applications in many fields, including geographic information systems (GIS), meteorology, environmental science, etc.
[0041] It should be noted that the model needs to be regionalized according to the spatial variation characteristics of different regions. Therefore, zoning is required before calibrating the model parameters. In addition, since the coupled model contains complex parameters of crop physiology, physical models and surface processes, the parameterization of the model is extremely challenging. According to the first law of geography, the geographical environment within a region has similar geographical characteristics. Therefore, meteorological and remote sensing data are first used for spatial clustering, and the computational cost of adjusting the model parameters for the entire space is reduced by dividing homogeneous planting areas as representative samples.
[0042] In a possible implementation, S2 specifically includes sub-steps S201 to S207:
[0043] S201, performing region division by using a spatiotemporal clustering algorithm.
[0044] In a possible implementation, S201 specifically includes:
[0045] S2011, obtain spatiotemporal data.
[0046] Among them, spatiotemporal data include: meteorological data and remote sensing data.
[0047] S2012: Determine the central object and the neighborhood objects according to the spatiotemporal data.
[0048] S2013, calculate the characteristic distance between the current neighborhood object and the central object:
[0049]
[0050] Among them, Dist represents the characteristic distance between the central object and the neighboring objects. represents the eigenvalue of the central object, Represents the eigenvalue of the neighborhood object, i represents the number of neighbors in the x-axis direction, j represents the number of neighbors in the y-axis direction, and k represents the number of neighbors in the z-axis direction.
[0051] S2014: Determine whether the characteristic distance between the current neighborhood object and the central object is less than a preset distance. If so, classify the current neighborhood object into the region category of the central object.
[0052] Optionally, the shapes of the final clusters are roughly divided into three types: cylindrical clusters, pie-shaped clusters, and spherical clusters.
[0053] Among them, the cylindrical clusters have a smaller range in the spatial direction but a larger range in the temporal direction, which represents a situation that is relatively stable in space but lasts for a long time; the pie-shaped clusters have a larger range in the spatial direction but a smaller range in the temporal direction, which represents a situation that is widely distributed in space but lasts for a short time; the spherical clusters have a relatively uniform distribution in both spatial and temporal directions.
[0054] In the present invention, different types of clusters represent different spatiotemporal patterns. By dividing the region into these different types of clusters, the patterns and regularities of spatiotemporal data can be better understood. At the same time, dividing the region into several different types of clusters can simplify the interpretation of spatiotemporal data. By identifying these basic patterns, it is easier to interpret the complex spatiotemporal relationships in the data, providing more intuitive and clear information for decision making.
[0055] Furthermore, by clustering, regions with the same geographical environment characteristics can be subjected to unified model simulation, which can reduce the cost of parameter adjustment and calculation, and accurately obtain local model parameters from the clustering data.
[0056] S202, for each divided area, estimate the posterior probability distribution of the environmental parameters by using the Bayesian rule:
[0057]
[0058] Among them, θ represents the environmental parameter, represents the posterior probability distribution of the observed data yo from time 0 to T, represents the likelihood of model simulation under the condition of random distribution of environmental parameters, Represents the initialization prior probability distribution of environmental parameters.
[0059] Among them, Bayesian rule is a probabilistic inference method based on Bayesian theorem, which is used to update prior probability to posterior probability and is an important tool in probability statistics.
[0060] In the present invention, the Bayesian rule is used to combine the observed data with the prior knowledge, which can fully consider the uncertainty and obtain the posterior probability distribution of the environmental parameters. This helps to better understand the uncertainty of parameter estimation and improve the reliability of the model.
[0061] S203, creating different combinations of environmental parameters through Latin square supersampling according to the posterior probability distribution of the environmental parameters.
[0062] Among them, Latin Hypercube Sampling is a method for designing experimental samples, which is used to uniformly select sample points in a multidimensional parameter space to represent different combinations in the parameter space. This sampling method has been widely used in engineering, statistics, computer simulation, and optimization.
[0063] In the present invention, the Latin square supersampling method can uniformly select sample points in the multi-dimensional parameter space, thereby efficiently exploring the parameter space, ensuring that a variety of parameter combinations are covered, and helping to find the global optimal solution.
[0064] S204, using the environmental parameter combination as an input parameter of a crop growth model, and generating crop growth simulation data through the crop growth model.
[0065] S205, constructing a cost function based on the error between the crop growth simulation data and the actual crop growth observation data:
[0066]
[0067] Where C represents the cost function, exp( ) represents the exponential function with the natural constant e as the base, RMSE represents the root mean square error between the simulated value and the true value, σ 0 Represents the observation error when collecting crop growth observation data.
[0068] In the present invention, by constructing a cost function using the difference between the simulated data generated by the crop growth model and the actual observed data, the degree of fit of the model can be effectively evaluated, and the parameters can be optimized through the Metropolis-Hastings algorithm to better fit the actual situation.
[0069] S206, determining optimal environmental parameters according to the convergence rate through the Metropolis-Hastings algorithm.
[0070] The specific convergence rate is:
[0071]
[0072] Among them, α represents the convergence rate, θ crepresents the candidate environment parameters created when iterating the model, θ i Represents the environmental parameters at the i-th iteration of the model.
[0073] Among them, the Metropolis-Hastings algorithm is a Markov chain Monte Carlo (MCMC) method for extracting samples from complex probability distributions.
[0074] It should be noted that the difference between the actual crop growth observation data and the crop growth simulation data is calculated by the cost function C. Then, the convergence rate calculated by the cost function is used to determine whether to accept the reshuffled environmental parameters. Based on the data generated by different combination parameters, the Metropolis-Hastings algorithm is used to iteratively calculate the optimal environmental parameters.
[0075] In the present invention, the Metropolis-Hastings algorithm determines the optimal environmental parameters according to the convergence rate through iterative calculation, and can search quickly and effectively in the parameter space and gradually approach the optimal solution.
[0076] S207, initializing the environmental parameters of the crop growth model according to the optimal environmental parameters.
[0077] In the present invention, the environmental parameters of the crop growth model are initialized according to the optimal environmental parameters, which helps to improve the accuracy and stability of the model.
[0078] S3, correcting the model error through the phenotypic curve model and initializing the crop genetic parameters of the crop growth model.
[0079] In a possible implementation, S3 specifically includes sub-steps S301 to S308:
[0080] S301, constructing a remote sensing curve.
[0081] In a possible implementation, S301 specifically includes:
[0082] S3011, obtain multi-source remote sensing data.
[0083] Among them, remote sensing data can come from Sentinel-2A / B—MSI sensors and Landsat-8 / 9—OLI sensors.
[0084] Among them, Sentinel-2A / B—MSI (Multi-Spectral Instrument) is a pair of satellite sensors of the European Space Agency (ESA). They are equipped with MSI sensors and can provide high spatial resolution, multi-spectral, and global coverage of earth surface image data.
[0085] Among them, Landsat-8 and Landsat-9 are equipped with OLI (Operational Land Imager) sensors, a pair of satellite missions jointly carried out by the United States Geological Survey (USGS) and NASA, designed to provide global land observation data to support applications in earth science research, resource management, environmental monitoring, and response to natural disasters.
[0086] In the present invention, by combining the data of Sentinel-2A / B-MSI and Landsat-8 / 9-OLI sensors, more comprehensive and high-quality remote sensing information can be obtained, thereby improving the accuracy and reliability of monitoring.
[0087] S3012, collaboratively integrate multi-source remote sensing data to obtain remote sensing time series data.
[0088] It should be noted that the L2-level surface reflectance data of Sentinel-2A / B is fused with the L2-level surface reflectance data of Landsat-8 / 9, and the reflectance data is corrected for BRDF according to the sensor's own bidirectional reflectance distribution function (BRDF). Then, a regression relationship between the spectral response functions (SRFs) of the two sensors is established, and the reflectance data of the OLI sensor is coordinated to the band data of the MSI sensor. Since the smallest unit of the present invention is a field, the reflectance data is aggregated according to the field range. Ultimately, each field will generate two time series data, and the two reflectances from the MSI sensor and the OLI sensor will be fused into one time series data according to the time dimension.
[0089] In the present invention, a time series data is obtained by collaboratively fusing multi-source data, which helps to maintain continuity and consistency and facilitates subsequent processing and analysis.
[0090] S3013, filtering and processing the remote sensing time series data.
[0091] It should be noted that due to the influence of sensor instability, atmospheric environment, surface environment and other factors, remote sensing will produce abnormal data, and correspondingly produce unreasonable fluctuations in time series data. In order to eliminate the interference of abnormal values and make the changes in time series data conform to the characteristic laws of crop growth, the present invention constructs a remote sensing time series data reconstruction method for the remote sensing NDVI index.
[0092] Among them, NDVI (Normalized Difference Vegetation Index) is a common indicator used to evaluate the coverage and growth status of surface vegetation. It is based on remote sensing data, especially the reflectivity of infrared and near-infrared bands, and can effectively reflect the photosynthesis and growth of vegetation.
[0093] Optionally, the filtering process includes performing envelope filtering on the time series and performing harmonic filtering on the time series.
[0094] Among them, envelope filtering of the time series is to filter out abnormal data caused by cloud cover by creating a curve envelope, retaining high-quality data reflecting crop growth information, and avoiding the sudden drop of the normalized vegetation index.
[0095] Among them, the time series is harmonically filtered, and after the fusion of multi-source data, the length of the time series in the 3-4 month growth cycle of the crop is greater than 15. The sixth harmonic is used to build a harmonic fitting model to ensure that the fitting results better take into account the overall characteristics and detailed characteristics of the original data, and the threshold method is used in the fitting process to filter out the subtle oscillations caused by surface environmental factors and sensor photosensitive elements.
[0096] In the present invention, a time series data reconstruction method based on the NDVI index is adopted, combined with envelope filtering and harmonic filtering techniques, which can effectively eliminate the interference of outliers and improve the reliability and stability of the data.
[0097] S3014, performing double logic curve fitting on the filtered remote sensing time series data:
[0098]
[0099] Among them, s represents the remote sensing curve, t represents time, v min Indicates the low NDVI value before crop planting, v 1 It is the difference between the highest NDVI value during the crop growth period and the lowest value before planting. 2 It is the difference between the highest NDVI value during the crop growth period and the lowest value after harvest, m 1 and n 1 are the peak position and slope of the time series change curve of the crop rapid growth stage, m 2 and n 2 are the peak position and slope of the crop senescence stage, respectively.
[0100] It should be noted that the full growth of crops includes vegetative growth, reproductive growth and senescence stages. In the vegetative growth stage, crops will grow rapidly from the second to the fourth phenological stage, and the NDVI curve will respond with a rapid rise; in the reproductive growth stage, the biomass and leaf area index of crops will gradually stabilize at the highest level and will be maintained for 2-3 phenological periods. The response of the NDVI curve is that the highest value will last for a long time; in the senescence stage, the crop biomass will gradually decrease, and the vegetation area index will accelerate the decline with the aging process. The response of the NDVI curve is to decrease from the highest value and the rate of decrease will increase. In order to characterize these characteristics, this paper uses a double logistic curve to fit the filtered data.
[0101] In the present invention, by performing dual logic curve fitting on the filtered remote sensing time series data, the characteristics and change patterns of different stages in the crop growth process can be more accurately described, providing a more reliable basis for monitoring and analyzing the crop growth status.
[0102] Furthermore, the integration of multiple remote sensing data and analysis methods can achieve comprehensive monitoring and analysis of the crop growth process, providing important support and reference for agricultural production, resource management, environmental monitoring and other fields.
[0103] S302, calibrating the remote sensing curve.
[0104] It should be noted that the remote sensing curve constructed by the present invention has basically reflected the variation characteristics of LAI throughout the growth period, but in order to better reflect the growth characteristics of the crop itself and the phenotypic characteristics (G×M) under the influence of the environment, the phenotypic curve was further constructed and corrected.
[0105] In a possible implementation, S302 specifically includes:
[0106] S3021, fit the growth rate model according to the remote sensing curve:
[0107]
[0108] Among them, LAI represents the vegetation area index, LAI max Indicates the maximum value of vegetation area index, LAI 0 represents the starting value of the vegetation area index, and r represents the growth rate.
[0109] S3022, Correction of growth rate model for radiation effects:
[0110]
[0111] Where A represents the amplitude, Indicates phase.
[0112] It should be noted that light is one of the limiting factors for crop growth. The daily LAI change growth rate is represented by a variable with amplitude A and phase The radiation effect correction is performed using a sinusoidal function.
[0113] S3023, Correct the growth rate model for temperature effects:
[0114]
[0115]
[0116] Among them, f T ( ) represents the temperature correction function, T ( ) represents the temperature function, T L Indicates the minimum temperature for crop growth, T H Indicates the highest temperature for crop growth, T AL Indicates T L The corresponding Arrhenius temperature, T AH Indicates T H The corresponding Arrhenius temperature.
[0117] S3024, correct the growth rate model for the influence of soil moisture:
[0118]
[0119] Among them, c, n, and v represent parameters related to crop absorption of soil moisture, c represents the maximum soil moisture absorption rate of crops, n represents the activity parameter of crop water channel protease, v represents the ratio of the surface of crop water absorption structure to its volume, and W () represents the soil moisture function.
[0120] It should be noted that soil moisture is one of the limiting factors for crop growth. Crops absorb water from the soil, which is limited by the activity and diffusion rate of aquaporins, so the influence of soil moisture is described by the Michaelis-Menten function.
[0121] Among them, the Michaelis-Menten function is used to describe the effect of soil moisture on crop growth rate. By substituting the soil moisture function into the Michaelis-Menten function, the limiting effect of soil moisture on crop growth can be described.
[0122] In the present invention, by fitting the growth rate model, correcting the radiation effect, temperature effect and soil moisture effect, the obtained growth model is closer to the actual situation. Such a model can more accurately predict the growth of crops and provide a more accurate decision-making basis for agricultural production. At the same time, the correction of the growth rate model takes into account the effects of various environmental factors such as light, temperature and soil moisture on crop growth. In this way, various external conditions in the crop growth process can be fully considered, making the model more reliable and applicable.
[0123] S303, constructing a phenotypic curve model based on crop genetic parameters according to the corrected remote sensing curve:
[0124]
[0125] Among them, g represents the phenotypic curve model based on crop genetic parameters, k represents spatial location, t represents time, and x represents the t represents the time series covariate, A vector representing the crop genetic parameters required by the crop model, s k (t) represents the corrected remote sensing curve at spatial position k, Indicates an approximate symbol.
[0126] Crop genetic parameters include: growth period trait parameters and phenotypic quantitative characteristic parameters.
[0127] Specifically, growth period trait parameters may include phenological period, fastest growth time point, flowering time, etc. Phenotypic quantitative characteristic parameters may include the number of tillers at jointing, the average canopy coverage between tillers and jointing, etc.
[0128] Optionally, the phenotypic quantitative characteristic parameter is specifically a range of variation in different growth stages:
[0129]
[0130] in, represents the parameter set of the phenotypic quantitative characteristic k, T(a) and T(b) represent the ath and bth time points, , represents the corresponding key growth period node, f k (t) represents the time function of the quantitative characteristic k of the phenotype.
[0131] Among them, the growth period trait parameters are based on the change rate characteristics of the crop phenotypic curve. The present invention constructs an empirical comparison relationship between the crop phenological period nodes and the phenotypic curve characteristics. Taking corn as an example, 4 minimum curvature points and 3 maximum curvature points are calculated for the constructed phenotypic curve. The 7 characteristic points correspond to the three-leaf stage, jointing stage, 15-leaf stage, flowering stage, milk stage, wax stage, and harvest stage of corn in sequence. The remaining key growth period nodes are determined according to the phenological period duration database.
[0132] The phenotypic quantitative characteristic parameters of crop growth describe the variation range of the phenotypic parameters in each growth stage, and are also the cumulative values of this growth stage. The present invention targets the phenotypic parameter of leaf area index, which represents the variation range of the phenotypic quantitative characteristic parameters of crop growth in different growth stages.
[0133] It should be noted that the crop model has spatial heterogeneity in different regions. In order to accurately obtain this feature, it is necessary to establish a connection between the observed characteristics and the mechanistic characteristics of the crop model through the phenotypic curve. In terms of the observed characteristics, the observed phenotypic curve constructed in the previous step is used; in terms of the mechanistic characteristics of the crop model, the leaf area index module of the CERES crop growth model is used to simulate the crop growth process.
[0134] In the present invention, the corrected remote sensing curve is combined with the mechanism characteristics of the crop model to achieve effective fusion of the observed data and the model. In this way, the measured data and the physical mechanism of the model can be comprehensively considered, and the credibility and applicability of the simulation results can be improved. At the same time, the crop model may have spatial heterogeneity characteristics in different regions, that is, the same crop may show different growth patterns in different regions. Through the phenotypic curve model based on crop genetic parameters, the spatial heterogeneity characteristics of different regions can be accurately obtained, making the crop growth simulation more accurate and reliable.
[0135] Furthermore, by introducing crop genetic parameters, the growth characteristics of crops are incorporated into the model for description. This parametric modeling method makes the model more flexible and adjustable, and can adapt to the needs of different crop varieties and different growth environments. At the same time, by accurately defining and describing the key growth period nodes in the crop growth process, the characteristics and change patterns of crops at different growth stages can be more accurately captured, providing a more reliable reference and decision-making basis for agricultural production.
[0136] S304, improve the phenotypic curve model based on crop genetic parameters through the covariate response characteristic model of crop phenotype:
[0137]
[0138] in, The covariate response characteristic model represents the crop phenotype, and the covariate response characteristic model is specifically the CERES crop growth model. represents the covariate parameter.
[0139] It should be noted that the crop growth process will be affected by various environmental covariate factors, and the response characteristics of crops to environmental factors are the key to the mechanism of crop growth models.
[0140] S305, using crop genetic parameters as input parameters of the phenotypic curve model, and generating a crop growth curve by constructing the phenotypic curve model.
[0141] S306, determining the optimal crop genetic parameters at each spatial position with the goal of minimizing the error between the generated crop growth curve and the improved remote sensing curve.
[0142] S307, introducing observation errors to correct the optimal crop genetic parameters.
[0143] It should be noted that after solving the optimal crop genetic parameters, it is necessary to analyze the numerical uncertainty of the remote sensing curve and phenotypic curve model. The observation error of remote sensing data and the temporal and spatial variation of crop model parameters and the influence of environmental covariates will lead to differences between the observed values and the simulated values. Therefore, it is necessary to introduce observation errors to correct the optimal crop genetic parameters.
[0144] In a possible implementation manner, S307 specifically includes:
[0145] S3071, determine the response of the remote sensing curve to the kth spatial position in the jth year:
[0146]
[0147]
[0148] in, represents the response of the remote sensing curve to the kth spatial position in the jth year, μ represents the intercept, g represents the result of the phenotypic curve model, and v j represents the main effect of year j, represents the interaction between the jth year and the crop model, represents the observation error, represents the spatial error, represents the time series observation error, Represents the inter-annual error.
[0149] Among them, time series observation error refers to the random error introduced by measurement error or uncertainty in the observation process. Even under the same conditions, repeated observations of the same object may produce different observations. Time series observation error is usually used to describe the random deviation between the observed value and its true value.
[0150] Among them, interannual error refers to the random component in the model residuals caused by interannual variability that is not considered or modeled. In time series data analysis, there may be interannual correlation or trend between adjacent time points, but the model fails to capture this interannual variability, resulting in an interannual structure in the model residuals. Interannual error is usually used to describe the interannual correlation or trend in such residuals.
[0151] Among them, time series observation error refers to the random error introduced by measurement error or uncertainty in the observation process. Even under the same conditions, repeated observations of the same object may produce different observations. Time series observation error is usually used to describe the random deviation between the observed value and its true value.
[0152] S3072, for the observed data at time t in year j, a smooth bivariate surface model is constructed using the SpATS algorithm to determine the main effects of the crop model and the random effects in the row and column directions, and the spatial effects of the phenotypic data are applied to the crop parameters:
[0153]
[0154]
[0155] in, Respectively represent the independent effects in the spatial row and column directions, represents the row-column interaction effect. They represent the influence of the time at moment t on the row and column directions, represents the row-column interaction response at the tth moment, represents the estimated value of the nth crop parameter at the kth spatial position in the jth year, represents the initial value of the nth crop parameter in the jth year, Respectively represent the effect of the nth crop parameter in the spatial row and column directions, represents the interaction effect of the nth crop parameter, represents the spatial error of the phenotype, Represents the spatial error of the nth parameter. S3073, quantifies the instability caused by inter-annual differences, and aggregates the crop parameters of the observed fields into one value , and then use the ASReml-R algorithm to correct the inter-annual differences:
[0156]
[0157] in, represents the nth crop parameter in the jth year, represents the intercept of the fitting equation for the nth crop parameter, represents the main effect of the nth crop parameter in the jth year, represents the nth crop parameter, represents the main effect of the nth crop parameter, represents the interaction effect between the nth crop parameter and the jth year, Represents the statistical residual of the nth crop parameter in the jth year.
[0158] In the present invention, the introduction of observation error, spatial error and inter-annual error can more accurately reflect the characteristics of actual observation data and random variability not considered in the model. By correcting these errors, the fitting accuracy of the crop growth model to the actual observation data can be improved, thereby enhancing the accuracy and robustness of the model. At the same time, considering the observation error, spatial error and inter-annual error, remote sensing data and other environmental data can be more comprehensively utilized, the information in the data can be fully mined, and the utilization efficiency and integrity of the data can be improved.
[0159] S308, initializing the crop genetic parameters of the crop growth model using the corrected optimal crop genetic parameters.
[0160] In the present invention, by correcting and optimizing the optimal crop genetic parameters, the role and influence of different parameters in the crop growth model can be better understood, providing more operational guidance and decision support for agricultural production. At the same time, it can be flexibly adjusted and adapted according to different crop types, growth environments and observation conditions, has strong applicability and versatility, and can be applied and promoted in different agricultural production practices.
[0161] In summary, by introducing observation errors, spatial errors and inter-annual errors, and correcting and optimizing the optimal crop genetic parameters, the reliability, accuracy and applicability of crop growth models can be improved, providing more effective support and guidance for agricultural production.
[0162] S4, outputting crop growth simulation results through the initialized crop growth model.
[0163] S5, interpolating the crop growth simulation results through the Kriging algorithm.
[0164] Among them, the Kriging algorithm is a geostatistical method used for spatial interpolation and geological modeling. It is based on point data samples in geographic space and predicts the values of unknown locations by modeling the spatial correlation between these sample data. The main goal of this algorithm is to estimate the values of unknown locations through the data values of known points, thereby achieving spatial interpolation.
[0165] In a possible implementation, S5 specifically includes sub-steps S501 to S503:
[0166] S501, using the vegetation area index in the crop growth simulation results as the main variable, and the inverted vegetation area index as the covariate, as the reference data for the Kriging algorithm to fit the semivariogram, the spatial and temporal semivariograms are constructed:
[0167]
[0168]
[0169] in, represents the spatial semivariogram, represents the temporal semivariogram, Represents the number of space-time matrix changes in the spatial direction, Represents the number of time matrices in the time-varying direction, represents the value of the space-time matrix at position s at time t, represents the spatial variation, represents the time variation, Represents the matrix at time t The value at the position, Represents the matrix in The value at the position at time s. S502, according to the spatial and temporal semivariogram, a weight matrix is obtained by fitting and calculating through the Kriging algorithm.
[0170] It should be noted that the Kriging algorithm fits the semivariogram inversion covariance matrix in the spatial and temporal dimensions, and the LAI data inverted by remote sensing is used to fit the semivariogram in space and time. At the same time, the size of the space-time window is determined according to the critical upper limit of the fitting function.
[0171] Optionally, the space-time window used is a 5-day sliding window with a range of 850 m, and each window contains an average of 200 Kriging main variables.
[0172] S503, interpolating the crop growth simulation results according to the weight matrix:
[0173]
[0174] in, Indicates t 0 At position x 0 The predicted value at, T represents the number of image phases of the main variable, N i represents the number of effective pixels of the main variable, V represents the number of image phases of the covariate, and M k represents the number of effective pixels of the covariate, represents the weight of the main variable on the jth valid pixel in the i-th phase, represents the weight of the covariate on the lth effective pixel in the kth phase, Z 1 represents the main variable, Z 2 represents the covariate, Indicates the main variable at t i At position S ij The value at Represents the covariate in v k At position u at all times kl The value at .
[0175] It should be noted that the present invention uses daily 10m resolution vegetation area index phenotype data as the covariate of the Kriging algorithm, and obtains a weight matrix through semivariogram fitting calculation to perform spatiotemporal interpolation on the simulation results of the crop growth model.
[0176] In the present invention, the Kriging algorithm can use the data values of known points in geographic space to estimate the values of unknown locations, thereby realizing the spatial continuity interpolation of crop growth simulation results. This can fill in blank areas, generate a continuous spatial surface, and better show the spatial distribution law of crop growth. At the same time, by performing spatiotemporal interpolation of crop growth simulation results through the Kriging algorithm, the temporal consistency of the interpolation results can be ensured. This helps to generate continuous time series data, better reflect the temporal change trend in the crop growth process, and provide more accurate time information for decision-making.
[0177] Furthermore, by interpolating crop growth simulation results from different sources, at different times, and with different spatial resolutions through the Kriging algorithm, data integration and unification can be achieved, which helps to integrate information from different data sources into the same spatial and temporal framework, thereby improving the comparability and availability of data.
[0178] The beneficial effects of the present invention are as follows:
[0179] The method and system of the present invention fuse the land surface pattern model with the crop model at a single point scale to construct a crop growth model, which can realize continuous simulation in a regional scope, take into account the characteristics of the crop model itself, initialize the environmental parameters of the crop growth model through a spatiotemporal clustering algorithm, realize the spatialization of parameters at a regional scale, and improve the multi-environment adaptability of the model, correct the model error through a phenotypic curve model, initialize the crop genetic parameters of the crop growth model, and then output the crop growth simulation results through the initialized crop growth model, avoid directly using spatialized data, reduce system errors, and perform spatial continuous interpolation processing on discrete crop growth simulation results through a Kriging algorithm, thereby realizing the spatialization of simulation results.
[0180] System Example
[0181] Reference Manual Attached Figure 2 , showing a schematic structural diagram of a crop growth model optimization system based on phenotypic curves provided in an embodiment of the present invention.
[0182] The present invention proposes a crop growth model optimization system 30 based on phenotypic curves, comprising: a memory 303 and one or more processors 301 .
[0183] One or more application programs are stored in the memory 303 , and the one or more application programs are suitable for being executed by the one or more processors 301 to implement the crop growth model optimization method based on phenotypic curves described in the method embodiment.
[0184] The crop growth model optimization system 30 based on phenotype curves includes: a processor 301 and a memory 303. The processor 301 and the memory 303 are connected, for example, via a bus 302.
[0185] The structure of the crop growth model optimization system 30 based on phenotype curves does not constitute a limitation on the embodiment of the present invention.
[0186] The beneficial effects of the present invention are as follows:
[0187] The method and system of the present invention fuse the land surface pattern model with the crop model at a single point scale to construct a crop growth model, which can realize continuous simulation in a regional scope, take into account the characteristics of the crop model itself, initialize the environmental parameters of the crop growth model through a spatiotemporal clustering algorithm, realize the spatialization of parameters at a regional scale, and improve the multi-environment adaptability of the model, correct the model error through a phenotypic curve model, initialize the crop genetic parameters of the crop growth model, and then output the crop growth simulation results through the initialized crop growth model, avoid directly using spatialized data, reduce system errors, and perform spatial continuous interpolation processing on discrete crop growth simulation results through a Kriging algorithm, thereby realizing the spatialization of simulation results.
[0188] Processor 301 may be a CPU, a general purpose processor, a DSP, an ASIC, an FPGA or other programmable logic device, a transistor logic device, a hardware component or any combination thereof. It may implement or execute various exemplary logic blocks, modules and circuits described in conjunction with the disclosure of the present invention. Processor 301 may also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc.
[0189] The bus 302 may include a path to transmit information between the above components. The bus 302 may be a PCI bus or an EISA bus, etc. The bus 302 may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.
[0190] The memory 303 can be a ROM or other type of static storage device that can store static information and instructions, a RAM or other type of dynamic storage device that can store information and instructions, or an EEPROM, a CD-ROM or other optical disk storage, an optical disk storage (including a compressed optical disk, a laser disk, an optical disk, a digital versatile disk, a Blu-ray disk, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited to these.
[0191] Computer Readable Storage Medium Embodiments
[0192] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program can be loaded and executed by a processor to implement the crop growth model optimization method based on phenotypic curves as described in the first aspect.
[0193] The applicant of the present invention has made a detailed explanation and description of the implementation examples of the present invention in conjunction with the drawings in the specification. However, those skilled in the art should understand that the above implementation examples are only preferred implementation schemes of the present invention, and the detailed description is only to help readers better understand the spirit of the present invention, but not to limit the scope of protection of the present invention. On the contrary, any improvements or modifications based on the inventive spirit of the present invention should fall within the scope of protection of the present invention.
[0194] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of the present invention, rather than to limit them. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the above embodiments, or replace some of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. Any changes or replacements that can be easily thought of by those skilled in the art within the technical scope disclosed in the present invention should be covered within the protection scope of the present invention.
Claims
1. A crop growth model optimization method based on phenotypic curves, characterized in that: include: S1, integrate the land surface model with the crop model at a single point scale to construct a crop growth model; S2, initializing the environmental parameters of the crop growth model by using a spatiotemporal clustering algorithm; S3, correcting the model error through the phenotypic curve model, and initializing the crop genetic parameters of the crop growth model; S4, outputting crop growth simulation results through the initialized crop growth model; S5, performing interpolation processing on the crop growth simulation result by using a Kriging algorithm; Wherein, the S2 specifically includes: S201, performing regional division by using a spatiotemporal clustering algorithm; S202, for each divided area, estimate the posterior probability distribution of the environmental parameters by using the Bayesian rule: ; Among them, θ represents the environmental parameter, represents the posterior probability distribution of the observed data y from time 0 to T, represents the likelihood of model simulation under the condition of random distribution of environmental parameters, Represents the initialization prior probability distribution of environmental parameters; S203, creating different combinations of environmental parameters through Latin hypercube sampling according to the posterior probability distribution of the environmental parameters; S204, using the environmental parameter combination as an input parameter of the crop growth model, and generating crop growth simulation data through the crop growth model; S205, constructing a cost function based on the error between the crop growth simulation data and the actual crop growth observation data: ; Where C represents the cost function, exp( ) represents the exponential function with the natural constant e as the base, RMSE represents the root mean square error between the simulated value and the true value, and σ0 represents the observation error when collecting crop growth observation data; S206, determining optimal environmental parameters according to the convergence rate through the Metropolis-Hastings algorithm; The convergence rate is specifically: ; Among them, α represents the convergence rate, θ c represents the candidate environment parameters created when iterating the model, θ i represents the environmental parameters at the i-th iteration of the model; S207, initializing the environmental parameters of the crop growth model according to the optimal environmental parameters; Wherein, the S3 specifically includes: S301, construct remote sensing curve; S302, calibrating the remote sensing curve; S303, constructing a phenotypic curve model based on crop genetic parameters according to the corrected remote sensing curve: ; Among them, g represents the phenotypic curve model based on crop genetic parameters, k represents spatial location, t represents time, and x represents the t represents the time series covariate, A vector representing the crop genetic parameters required by the crop model, s k (t) represents the corrected remote sensing curve at spatial position k, Indicates approximate symbols; The crop genetic parameters include: growth period trait parameters and phenotypic quantitative characteristic parameters; The phenotypic quantitative characteristic parameters specifically refer to the range of variation in different growth stages: ; in, represents the parameter set of the phenotypic quantitative characteristic e, T(a) and T(b) represent the ath and bth time points, , represents the corresponding key growth period node, f e (t) represents the time function of the quantitative characteristic e of the phenotype; S304, improve the phenotypic curve model based on crop genetic parameters through the covariate response characteristic model of crop phenotype: ; in, The covariate response characteristic model represents the crop phenotype, and the covariate response characteristic model is specifically the CERES crop growth model. represents the covariate parameter; S305, using crop genetic parameters as input parameters of the phenotypic curve model, and generating a crop growth curve through the phenotypic curve model; S306, determining the optimal crop genetic parameters at each spatial position with the goal of minimizing the error between the generated crop growth curve and the improved remote sensing curve; S307, introducing observation errors to correct the optimal crop genetic parameters; S308, initializing the crop genetic parameters of the crop growth model using the corrected optimal crop genetic parameters.
2. The crop growth model optimization method based on phenotypic curve according to claim 1, characterized in that: S1 specifically includes: based on the Noah-PM model, taking the vegetation area index as the simulation process quantity, introducing the dynamic growth process of the CERES crop model, and constructing a crop growth model; The Noah-PM model is a land surface process model that combines the Noah land surface model and the Penman-Monteith formula.
3. The crop growth model optimization method based on phenotypic curve according to claim 1, characterized in that: The S201 specifically includes: S2011, acquiring spatiotemporal data, the spatiotemporal data including: meteorological data and remote sensing data; S2012, determining a central object and neighboring objects according to the spatiotemporal data; S2013, calculating the characteristic distance between the current neighborhood object and the central object; S2014, determining whether the characteristic distance between the current neighborhood object and the central object is less than a preset distance; if so, classifying the current neighborhood object into the region category of the central object.
4. The crop growth model optimization method based on phenotypic curve according to claim 1, characterized in that: The S301 specifically includes: S3011, Acquisition of multi-source remote sensing data; S3012, collaboratively integrate multi-source remote sensing data to obtain remote sensing time series data; S3013, filtering the remote sensing time series data; S3014, performing double logic curve fitting on the filtered remote sensing time series data: ; Among them, s represents the remote sensing curve, t represents time, v min It represents the low NDVI value before crop planting, v1 is the difference between the highest NDVI value during the crop growth period and the low value before planting, v2 is the difference between the highest NDVI value during the crop growth period and the value after harvest, m1 and n1 are the peak position and slope of the time series change curve in the rapid growth stage of the crop, respectively, and m2 and n2 are the peak position and slope in the aging stage of the crop, respectively.
5. The crop growth model optimization method based on phenotypic curve according to claim 1, characterized in that: The S302 specifically includes: S3021, fitting a growth rate model according to the remote sensing curve: ; Among them, LAI represents the vegetation area index, LAI max represents the maximum value of vegetation area index, LAI0 represents the starting value of vegetation area index, and r represents the growth rate; S3022, Correction of growth rate model for radiation effects: ; Where A represents the amplitude, Indicates phase; S3023, Correct the growth rate model for temperature effects: ; ; Among them, f T ( ) represents the temperature correction function, T ( ) represents the temperature function, T L Indicates the minimum temperature for crop growth, T H Indicates the highest temperature for crop growth, T AL Indicates T L The corresponding Arrhenius temperature, T AH Indicates T H The corresponding Arrhenius temperature; S3024, correct the growth rate model for the influence of soil moisture: ; Among them, c, d, and v represent parameters related to crop absorption of soil moisture, c represents the maximum soil moisture absorption rate of crops, d represents the activity parameter of crop water channel protease, v represents the ratio of the surface of the crop water absorption structure to its volume, and W () represents the soil moisture function.
6. The crop growth model optimization method based on phenotypic curve according to claim 1, characterized in that: The S307 specifically includes: S3071, determine the response of the remote sensing curve to the kth spatial position in the jth year: ; ; in, represents the response of the remote sensing curve to the kth spatial position in the jth year, μ represents the intercept, g represents the result of the phenotypic curve model, and v j represents the main effect of year j, represents the observation error, represents the spatial error, represents the time series observation error, represents the inter-annual error; S3072, for the observed data at time t in year j, a smooth bivariate surface model is constructed using the SpATS algorithm to determine the main effects of the crop model and the random effects in the row and column directions, and the spatial effects of the phenotypic data are applied to the crop parameters: ; ; in, Respectively represent the independent effects in the spatial row and column directions, represents the row-column interaction effect, They represent the influence of the time at moment t on the row and column directions, represents the row-column interaction effect at time t, represents the estimated value of the nth crop parameter at the kth spatial position in the jth year, represents the initial value of the nth crop parameter in the jth year, Respectively represent the effect of the nth crop parameter in the spatial row and column directions, represents the row-column interaction effect of the nth crop parameter, represents the spatial error, represents the spatial error of the nth crop parameter; S3073, quantify the instability caused by inter-annual differences, and aggregate the crop parameters of the observed fields into one value , and then use the ASReml-R algorithm to correct the inter-annual differences: ; in, represents the nth crop parameter in the jth year, represents the intercept of the fitting equation for the nth crop parameter, represents the main effect of the nth crop parameter in the jth year, represents the nth crop parameter, represents the main effect of the nth crop parameter, represents the interaction effect between the nth crop parameter and the jth year, Represents the statistical residual of the nth crop parameter in the jth year.
7. The crop growth model optimization method based on phenotypic curve according to claim 1, characterized in that: The S5 specifically includes: S501, using the vegetation area index in the crop growth simulation results as the main variable, and the inverted vegetation area index as the covariate, as the reference data for fitting the semivariogram using the Kriging algorithm, the spatial and temporal semivariograms are constructed: ; ; in, represents the spatial semivariogram, represents the temporal semivariogram, Represents the number of space-time matrix changes in the spatial direction, Represents the number of time matrices in the time-varying direction, represents the value of the space-time matrix at position p at time t, represents the spatial variation, represents the time variation, Represents the matrix at time t The value at the position, Represents the matrix in The value at position p at time; S502, according to the spatial and temporal semivariograms, a weight matrix is obtained by fitting and calculating through a Kriging algorithm; S503, performing interpolation processing on the crop growth simulation result according to the weight matrix: ; in, represents the predicted value at position x0 at time t0, T represents the number of image phases of the main variable, N i represents the number of effective pixels of the main variable, V represents the number of image phases of the covariate, and M k represents the number of effective pixels of the covariate, represents the weight of the main variable on the jth valid pixel in the i-th phase, represents the weight of the covariate on the lth effective pixel in the kth phase, Z1 represents the main variable, Z2 represents the covariate, Indicates the main variable at t i At the time position p ij The value at Represents the covariate in v k At position u at all times kl The value at .
8. A crop growth model optimization system based on phenotypic curves, characterized in that: include: memory and one or more processors; One or more application programs are stored in the memory, and the one or more application programs are suitable for being executed by the one or more processors to implement the crop growth model optimization method based on phenotypic curves according to any one of claims 1 to 7.
