CN117744898A - Construction method of annual prediction model of yield of field grain crops - Google Patents
Construction method of annual prediction model of yield of field grain crops Download PDFInfo
- Publication number
- CN117744898A CN117744898A CN202410191321.8A CN202410191321A CN117744898A CN 117744898 A CN117744898 A CN 117744898A CN 202410191321 A CN202410191321 A CN 202410191321A CN 117744898 A CN117744898 A CN 117744898A
- Authority
- CN
- China
- Prior art keywords
- yield
- month
- index
- average
- growth stage
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000010276 construction Methods 0.000 title claims description 9
- 230000012010 growth Effects 0.000 claims abstract description 90
- 238000000034 method Methods 0.000 claims abstract description 49
- 235000016709 nutrition Nutrition 0.000 claims abstract description 22
- 230000035764 nutrition Effects 0.000 claims abstract description 22
- 230000001850 reproductive effect Effects 0.000 claims abstract description 19
- 230000000875 corresponding effect Effects 0.000 claims description 27
- 238000004364 calculation method Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 8
- 230000015572 biosynthetic process Effects 0.000 claims description 7
- 238000004458 analytical method Methods 0.000 claims description 6
- 230000008859 change Effects 0.000 claims description 6
- 230000001419 dependent effect Effects 0.000 claims description 6
- 238000002310 reflectometry Methods 0.000 claims description 5
- 230000003698 anagen phase Effects 0.000 claims description 4
- 230000002596 correlated effect Effects 0.000 claims description 4
- 238000005259 measurement Methods 0.000 claims description 4
- 238000012545 processing Methods 0.000 claims description 4
- 230000009105 vegetative growth Effects 0.000 claims description 4
- 238000007781 pre-processing Methods 0.000 claims description 3
- 230000001105 regulatory effect Effects 0.000 claims description 3
- 238000003306 harvesting Methods 0.000 abstract description 5
- 238000012271 agricultural production Methods 0.000 abstract description 3
- 230000005855 radiation Effects 0.000 description 13
- 235000013339 cereals Nutrition 0.000 description 12
- 238000001556 precipitation Methods 0.000 description 10
- 101150105133 RRAD gene Proteins 0.000 description 6
- 241000209140 Triticum Species 0.000 description 6
- 235000021307 Triticum Nutrition 0.000 description 6
- 230000000694 effects Effects 0.000 description 6
- 238000011161 development Methods 0.000 description 5
- 230000018109 developmental process Effects 0.000 description 5
- 101001057166 Homo sapiens Protein EVI2A Proteins 0.000 description 4
- 102100027246 Protein EVI2A Human genes 0.000 description 4
- 238000012795 verification Methods 0.000 description 4
- 238000011835 investigation Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000004069 differentiation Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000013179 statistical model Methods 0.000 description 2
- 230000017260 vegetative to reproductive phase transition of meristem Effects 0.000 description 2
- 240000007594 Oryza sativa Species 0.000 description 1
- 235000007164 Oryza sativa Nutrition 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002301 combined effect Effects 0.000 description 1
- 230000001276 controlling effect Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013524 data verification Methods 0.000 description 1
- 230000008641 drought stress Effects 0.000 description 1
- 238000009472 formulation Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 235000015097 nutrients Nutrition 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000029553 photosynthesis Effects 0.000 description 1
- 238000010672 photosynthesis Methods 0.000 description 1
- 235000009566 rice Nutrition 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the field of crop yield prediction, and discloses a method for constructing an annual prediction model of yield of field grain crops, which comprises the following steps: s1: acquiring remote sensing data corresponding to a vegetation index optimal point and a key nutrition growth stage P1 and a reproductive growth stage P2 in a crop growing period; s2: acquiring month-average relevant meteorological indexes of each growth stage of crops; s3: layering and nesting meteorological data and satellite remote sensing data to construct a yield prediction model of multi-layer regression; s4: and obtaining global optimal parameters of the yield prediction model according to the actually measured yield data set to obtain a final yield prediction model. The method comprehensively considers the influence of crop growth rules and climate factors, combines satellite observation data and regional meteorological information, has strong universality, good cross-regional property and simple input, is suitable for large-space-scale yield prediction, can realize accurate prediction of yield three to four weeks before harvesting, provides decision basis and scientific guidance for agricultural production, and promotes grain income.
Description
Technical Field
The invention belongs to the field of crop yield prediction, and particularly relates to a method for constructing an annual prediction model of the yield of field grain crops.
Background
In recent years, crop yield prediction has become a research hotspot in the field of agricultural science, and has a key effect on solving the problem of grain production. Therefore, the accurate prediction of the yield and the promotion of the yield development of the field grain crops are new requirements for implementing accurate agriculture in China, and the accurate and timely prediction of the yield of the crops has great significance for the national formulation of related grain policies.
Conventional yield estimation methods include statistical methods based on historical yield, sample investigation methods based on sample data analysis and expert experience methods based on field investigation, so that a yield prediction mode is realized. However, the conventional production estimation method is large in workload and low in efficiency. The sample method is 'point-substituted surface', and the expert experience method is high in subjectivity, practical significance and the achievable estimated yield target can not meet corresponding requirements, so that the large-scale development of crop estimated yield is hindered.
The existing stage has feasibility of predicting crop yield in seasons by using remote sensing data. The remote sensing estimation means which are realized at present can be divided into crop models and experience models. The crop model based on the physiological and ecological mechanism can simulate and generate the yield by simulating and predicting the growth and development of crops or combining a great amount of remote sensing data obtained efficiently by the crop model and an assimilation algorithm. Although the crop model takes mechanization as an advantage, parameters in prediction are complex, a large amount of data from field investigation is needed, the localization of the parameters is complex, the acquisition difficulty is high, time and labor are consumed, and the model operation efficiency is low. Statistical models are therefore more widely used in large-scale yield predictions than process-based crop models. The regression model based on the remote sensing information and the machine learning algorithm combining the remote sensing information and the environmental factors have simple structures, and can achieve a good prediction effect. But is limited to a specific research area and growing season, the former is easily affected by vegetation index saturation, applicability is poor, and the latter is computationally complex and labor-consuming because of the input of a large amount of data.
Therefore, although the mode of predicting the yield based on the remote sensing technology is mature gradually, the model is limited by the data requirement of the model itself or by the space-time expansion of the model which cannot be met by the model architecture, and generalization on the annual and regional scale is difficult on the premise of ensuring the estimated yield accuracy. In order to meet the application requirements of large areas, how to rely on satellite remote sensing data to construct an efficient yield model crossing the annual and area becomes one of the focus of crop yield prediction research.
Disclosure of Invention
In order to solve the technical problems, the invention comprehensively considers the influence of crop growth rules and climate factors, combines satellite observation data and regional weather information, provides a method for constructing an annual prediction model of the yield of the field grain crops, has strong universality, good cross-territory property and simple input, is suitable for large-space-scale yield prediction, can realize accurate prediction of the yield three to four weeks before harvesting, provides decision basis and scientific guidance for agricultural production, and promotes grain income increase.
The technical scheme adopted in the invention is as follows:
a construction method of annual prediction model of yield of field grain crops comprises the following specific steps:
s1: acquiring satellite remote sensing data, and preprocessing to obtain remote sensing data corresponding to a vegetation index optimal point and a key nutrition growth stage P1 and a reproductive growth stage P2 in a crop growing period;
s2: acquiring historical meteorological data and forecast data, and processing to obtain month-average relevant meteorological indexes of each growth stage of crops;
s3: layering and nesting meteorological data and satellite remote sensing data to construct a yield prediction model of multi-layer regression;
s4: and obtaining global optimal parameters of the yield prediction model according to the actually measured yield data set to obtain a final yield prediction model.
Preferably, the specific method of step S1 is as follows:
s1-1: according to the annual growth characteristics of the crops, determining the dates corresponding to a key nutrition growth stage P1 and a reproduction growth stage P2 in the key growth stages of the crops;
s1-2: acquiring a surface reflectivity image data set of crops, calculating a vegetation index VI based on the surface reflectivity image data set, and acquiring the vegetation index VI in a selected time range;
s1-3: fitting the obtained vegetation index VI into a time-series index curve, wherein the starting time of the index curve is the starting date of the key nutrition growth phase P1 of the crop, and the ending time is the estimated ending date of the reproduction growth phase P2 of the crop;
s1-4: selecting the maximum value of vegetation index on the index curve, and recording as VI max Representing the maximum growth of crops;
s1-5: calculating the average value of the vegetation indexes of the key nutrition growth stage P1 and the reproductive growth stage P2 respectively, wherein the average value of the vegetation indexes of the key nutrition growth stage P1 is recorded as VI mean1 Representing the growth vigor degree of the early-stage crops; the average value of vegetation index in the reproductive growth stage P2 is recorded as VI mean2 Indicating the extent of post crop yield formation.
Preferably, the specific method of step S2 is as follows:
s2-1: acquiring an analysis data set and weather forecast data which comprise meteorological types influencing crop yield, and acquiring current month-average related meteorological data;
s2-2: calculating the average value of the month-to-month relevant meteorological data of the month corresponding to the last ten years according to the months of the critical growth stage of the crops;
s2-3: dividing the current month-to-month relevant weather data quantity by the ten-year month-to-month relevant weather data quantity average value to obtain a final month-to-month relevant weather index;
s2-4: and acquiring month-average related meteorological indexes corresponding to the key nutrition growth stage P1 and the reproduction growth stage P2 respectively.
Preferably, in step S3, the vegetation index and the weather index related to the lunar average are combined, layered nesting is performed, the weather data is used as an influencing factor, the relation between the vegetation index and the yield of crops in different places is regulated, a layered linear model HLM is introduced to analyze the influence of different layered prediction variables on the prediction value, and a yield prediction model of a multilayer structure is constructed.
Preferably, the specific structure of the yield prediction model constructed in step S3 is as follows: model Level-1 layer including vegetation index VI max 、VI mean1 And VI mean2 The method is specifically expressed as follows:(1) The method comprises the steps of carrying out a first treatment on the surface of the In the formula, YIeld represents crop Yield, VI mean1 Representing the growth condition of the early-stage crops in the key nutrition growth stage P1; VI (VI) max Indicating the maximum degree of crop growth; VI (VI) mean2 Represents the degree of post crop yield formation in the reproductive growth stage P2,the intercept is indicated as the intercept and,、、the regression coefficient of the model is represented, and r represents the random error of the model Level-1 layer; model Level-2 layer, level-1 layer、、、The parameters are dependent variables, and the dependent variables are adjusted by month-average relevant meteorological indexes:(2) The method comprises the steps of carrying out a first treatment on the surface of the In the method, in the process of the invention,respectively in Level-1、、、；Representing the intercept;representing a random error, the random error is represented,represent the firstiMonth-to-month correlated meteorological indexN represents the number of month-average related meteorological indexes MI; for the followingThe change of j causes the weather index MI related to both n and month to change correspondingly, forAnd VI max Corresponding coefficientsThe month-average related meteorological index MI is a month-to-month average related meteorological index of a key growth stage; for VI mean1 Corresponding coefficientsThe month-average relevant weather index MI refers to the month-average relevant weather index of the critical vegetative growth phase P1; for VI mean2 Corresponding coefficientsThe month-average related meteorological index MI refers to the reproductive growth stageMonth-average relevant meteorological index of section P2. Preferably, the parameter solving method of the yield prediction model in step S4 is as follows: the indicators in each piece of data in the measured yield dataset include: yield, VI max， VI mean1， VI mean2 And month-average relevant meteorological indexes in the key growth stage, based on the data set, adopting extremum solution of nonlinear programming, setting target errors and iterative calculation times, and realizing the construction of a yield prediction model by inputting an actual measurement yield data set and determining globally optimal parameters through iteration.
The beneficial effects are that: the invention provides a method for constructing an annual prediction model of yield of field grain crops, which has the following advantages compared with the prior art: (1) The method reflects the deviation degree of the weather conditions from the normal conditions by comparing the weather data with the annual average values based on satellite remote sensing and regional weather data, fully considers the influence effect of the weather conditions on the crop growth process, and realizes large-scale high-precision prediction of crop yield by coupling the remote sensing and the weather data in a layering manner.
(2) The model constructed by the method is easy to obtain the required parameters, is simple and convenient to calculate, is particularly rich in available remote sensing data sources, and has good application prospect and huge market value along with the continuous shortening of satellite remote sensing revisit period and the continuous updating of sensing technology, the yield prediction model constructed by the method is high in universality and simple in structure, a better solution is provided for timely and efficient yield prediction, decision basis and scientific guidance can be provided for agricultural production by predicted yield information, and the increase of grain in the field is promoted.
Drawings
FIG. 1 is a timing curve fitting chart of the remote sensing vegetation index of example 1;
FIG. 2 is a statistical graph of the prediction accuracy of the yield model of example 1;
FIG. 3 is a statistical graph of model accuracy under different yield conditions for example 1;
fig. 4 is a statistical graph of model accuracy under the influence of drought conditions in example 1.
Detailed Description
In order to better understand the technical solutions in the present application, the following description will clearly and completely describe the technical solutions in the embodiments of the present application, and it is obvious that the described embodiments are only some embodiments of the present application, not all embodiments. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments herein without making any inventive effort, shall fall within the scope of the present application.
Examples
A construction method of annual prediction model of yield of field grain crops comprises the following specific steps:
s1: acquiring satellite remote sensing data, and preprocessing to obtain remote sensing data corresponding to a vegetation index optimal point and a key nutrition growth stage P1 and a reproductive growth stage P2 in a crop growing period; in this embodiment 1, the acquisition and processing of satellite remote sensing data are all completed by the GEE (Google Earth Engine) platform. The specific method comprises the following steps:
s1-1: in the embodiment 1, the satellite image data selects a Sentinel-2 data set, and the corresponding month of the key growth stage of the crop is determined according to the annual growth characteristics of the crop (for example, the period from the node pulling period to the flowering period of winter wheat is concentrated for 3-4 months and is marked as a key nutrition growth stage P1, the period from the flowering period to the harvesting period is concentrated for 5 months and is marked as a reproductive growth stage P2);
s1-2: selecting and storing a surface reflectivity image data set 'COPERNICUS/S2_SR_HARMONIZED' of a Sentinel-2Level-2A Level in a GEE platform, calculating a vegetation index VI (Vegetation index) after the data are subjected to terrain and atmosphere correction, and obtaining the vegetation index of a selected time range (namely, the month corresponding to the critical growth stage of the crop determined in the step S1-1);
s1-3: fitting the obtained vegetation index VI in the selected time range into a time-series index curve, as shown in figure 1, wherein the starting time of the index curve is the starting date of the key nutrient growth stage P1 of the crop, and the ending time is the estimated ending date of the reproductive growth stage P2 of the crop; s1-4: selecting the maximum value of vegetation index on the index curve, and recording as VI max Representing the maximum growth of crops;
s1-5: calculating the average value of the vegetation indexes of the key nutrition growth stage P1 and the reproductive growth stage P2 respectively, wherein the average value of the vegetation indexes of the key nutrition growth stage P1 is recorded as VI mean1 Representing the growth vigor degree of the early-stage crops; the average value of vegetation index in the reproductive growth stage P2 is recorded as VI mean2 Indicating the extent of post crop yield formation.
S2: and acquiring historical meteorological data and forecast data, and processing to obtain meteorological indexes of each growth stage of the crops. The specific method comprises the following steps:
s2-1: acquiring an analysis data set and weather forecast data which comprise meteorological types influencing crop yield, and acquiring current month-average related meteorological data; in this example 1, the meteorological data selects the analysis data set ERA5 of ECMWF, and the era5_land data set "ECMWF/era5_land/horly" with a time-division rate per hour is used in the GEE platform. The analysis data set ERA5 contains three meteorological types of rainfall, solar radiation and temperature, and the month average rainfall Pre, month average solar radiation Rad and month average temperature Tem are calculated based on the analysis data set ERA5, and the resolution is 0.1 degree x 0.1 degree. In addition, for the weather conditions of the unknown key growth stage, future weather data and current known data are obtained through weather forecast data of a weather bureau to participate in calculation, and month average weather data of the last key month is obtained.
S2-2: according to months of the critical growth stage of crops, calculating month-by-month average precipitation average value, month average solar radiation average value and month average temperature average value of months corresponding to the last ten years;
s2-3: dividing the current corresponding month average precipitation, month average solar radiation and month average temperature with a month average precipitation average, month average solar radiation average and month average temperature average of ten years respectively to obtain a final month average precipitation index rPre, month average solar radiation index rRAd and month average temperature index rTem which are used for reflecting the degree of deviation of meteorological data from normal values;
s2-4: and acquiring month-average related meteorological indexes corresponding to the key nutrition growth stage P1 and the reproduction growth stage P2 respectively.
S3: and combining the vegetation indexes with the weather indexes related to the lunar average, performing multi-level nesting, taking weather data as an influence factor, regulating and controlling the relation between the vegetation indexes and the yield of crops at different places, introducing a hierarchical linear model HLM to analyze the influence of different hierarchical prediction variables on the prediction value, and constructing a yield prediction model of a multi-layer structure. In this embodiment 1, a hierarchical linear model HLM is introduced to analyze the influence of two-layer prediction variables on the predicted values, and a yield prediction model of a two-layer structure is constructed, and the specific model structure is as follows:
introducing a hierarchical linear model HLM to analyze the influence of different hierarchy prediction variables on a prediction value, and constructing a yield prediction model of the multilayer structure; in the model Level-2 layer, the Level-1 layer、、、The parameters are dependent variables, and the dependent variables are adjusted by annual and regional difference factors, namely, month-average relevant meteorological indexes:(2) The method comprises the steps of carrying out a first treatment on the surface of the In the method, in the process of the invention,respectively in Level-1、、、；Representing the intercept;representing a random error, the random error is represented,represent the firstiMonth-to-month correlated meteorological indexN represents the number of month-average related meteorological indexes MI; in this example 1, there are three month-average related meteorological indexes MI, specifically including a month-average precipitation index rpe, a month-average solar radiation index rRad, and a month-average temperature index rTem. Wherein,the change of j causes the weather index MI related to both n and month to change correspondingly, forAnd VI max Corresponding coefficientsThe month-average related meteorological index MI is a month-by-month average precipitation index rPre, a month-average solar radiation index rRAd and a month-average temperature index rTem of a key growth stage; for VI mean1 Corresponding coefficientsThe month-average related meteorological index MI refers to a month-average precipitation index rPre, a month-average solar radiation index rAdd and a month-average temperature index rTem of the key nutrition growth stage P1; for VI mean2 Corresponding coefficientsThe month-average related meteorological index MI refers to a month-average precipitation index rpe, a month-average solar radiation index rRad, and a month-average temperature index rTem of the reproductive growth stage P2. S4: and obtaining the global optimal parameters of the yield prediction model to obtain the final yield prediction model. In this example 1, a hierarchical linear model was basedAll unknown parameters in the yield prediction model of (c) can be calculated from the existing measured yield dataset. Each piece of data in the measured yield dataset includes the following: yield, VI max， VI mean1， VI mean2， The month average precipitation index rpe, month average solar radiation index rRad and month average temperature index rTem of the key growth stage. The process of solving the unknown parameters can also be called as extremum solving of nonlinear programming, setting target errors (i.e. acceptable minimum errors) and iterative calculation times, and determining globally optimal parameters in multiple iterations by inputting the actually measured yield data set, so that the construction of a yield prediction model is realized.
In example 1, the yield was measured in the field, and during the harvest period of the crop, a five-point sampling method was used to select 5 evenly grown sample points (1 m 2 Spot) to perform real harvest measurement, requiring the sampling point to be more than 10 meters from the peripheral boundary of the field to reduce the influence caused by edge effect. And (5) carrying out standard measurement and calculation on dry weight, water content and the like on the collected seeds to obtain the yield of the field. Meanwhile, the VI corresponding to the point is calculated according to the steps S1 and S2 max， VI mean1， VI mean2， The month average precipitation index rpe, month average solar radiation index rRad and month average temperature index rTem of the key growth stage. In the present invention, the acquisition and selection of the actual yield dataset can be determined by those skilled in the art according to the actual requirements, and are conventional technical means, and therefore will not be described in detail.
In the invention, the selection of the type and the number of the relevant meteorological data can be selectively designed by a person skilled in the art according to the actual requirements of the crop types and the like predicted by actual requirements.
Yield prediction model performance verification in example 1:
the results of model verification on two types of field grain crops of wheat and rice show that the yield prediction model constructed by using the HLM method to couple the remote sensing data and the meteorological data in the crop growth process has errors lower than 15% in the yield results of 4 years continuously in a provincial range, has better precision and stability between the years and the areas, and shows that the use of the HLM for predicting the yield in the season has great potential and can realize expansion in time and area scale.
Specific application of example 1:
(a) And (3) data acquisition:
taking a Shandong province wheat yield model as an example, according to actual measured wheat yield data acquired by a plurality of fields in Shandong province for 4 years continuously, generating a time sequence curve of EVI2 vegetation indexes of 3 months, 1 day and 5 months and 31 days of each field corresponding to the year in a GEE platform by recording the longitude and latitude of the fields, and acquiring EVI2 from the curve max Mainly distributed in the last ten days of 4 months and 5 months, and EVI2 is obtained by calculation according to P1 and P2 mean1 And EVI2 mean2 . In the GEE platform, pre03, pre04, pre05, rad03, rad04, rad05, tem03, tem04 and Tem05 in ERA5 are calculated, and nine meteorological parameters are divided by meteorological average values of the last ten years of the corresponding month respectively to obtain rPre03, rPre04, rPre05, rRAD03, rRAD04, rRAD05, rTem03, rTem04 and rTem05.
(b) Model construction:
the specific form of the constructed model is as follows:(3) The method comprises the steps of carrying out a first treatment on the surface of the As in the formula (1),、、andthe intercept and regression coefficients of the first layer regression equation are respectively calculated by the meteorological data of the model Level-2 layer:the method comprises the steps of carrying out a first treatment on the surface of the As in the formula (2),representation ofFirst, theiMonth-to-month correlated meteorological indexN represents the number of month-average related meteorological indexes MI; for the followingAndmonth-average related meteorological index participating in calculationComprises 9 parameters of rPre03, rPre04, rPre05, rRAd03, rRAd04, rRAd05, rTem03, rTem04 and rTem 05; for the followingMonth-average related meteorological index participating in calculationComprises 6 parameters of rPre03, rPre04, rRAd03, rRAd04, rTem03 and rTem 04; for the followingMonth-average related meteorological index participating in calculationIncludes rPre05, rRAd05, rTem05 total 3 parameters.
(c) Model accuracy verification
The model was verified using 163 total yield data in total in 4 years, as shown in fig. 2, the verification accuracy of the model was 89.71%, and the estimated yield error was 39.78 kg/mu. For wheat with a yield below 467 kg/mu (7 t/ha), the estimated yield error was 52.23 kg/mu as shown in FIG. 2. For wheat subjected to drought stress in the early stage, as shown in FIG. 3, the estimated yield error is 35.33 kg/mu. In the growth and yield formation process of crops, meteorological factors such as rainfall, solar radiation, temperature and the like can influence the growth of roots and leaves and the differentiation of stem nodes in the vegetative growth stage of crops, can influence the differentiation of spikes in the combined stage of vegetative growth and reproductive growth, and can influence the crop development by influencing photosynthesis. Meanwhile, the weather environment has different effects on yield in different growth stages, for example, in the early growth stage (key nutrition growth stage P1), the rainfall excessively limits the crop growth, and in the later growth stage (reproduction growth stage P2), the crop development is promoted by proper temperature and rainfall. Thus, crop growth is independent of the combined effects of the various meteorological environmental conditions during the reproductive cycle. Therefore, the invention aims at the limitation of different area environments in the current crop yield prediction, combines the vegetation index and the meteorological data according to the influence effect of the meteorological conditions on the crop growth and yield formation, uses the meteorological data to adjust the relation between the remote sensing vegetation index and the yield, and establishes a yield prediction model based on a layered linear model. The actual data verification can well realize the yield prediction of farm crops in a provincial range for years, overcomes the limitation of inconsistency of the relation between the remote sensing vegetation index and the yield in cross-region and cross-annual dimensions, has better interpretation than a black box model of machine learning, and can realize time and space expansion than a traditional statistical model.
The foregoing is merely a preferred embodiment of the present invention and it should be noted that modifications and adaptations to those skilled in the art may be made without departing from the principles of the present invention, which are intended to be comprehended within the scope of the present invention.
Claims (6)
1. The annual prediction model construction method for the yield of the field grain crops is characterized by comprising the following specific steps:
s1: acquiring satellite remote sensing data, and preprocessing to obtain remote sensing data corresponding to a vegetation index optimal point and a key nutrition growth stage P1 and a reproductive growth stage P2 in a crop growing period;
s2: acquiring historical meteorological data and forecast data, and processing to obtain month-average relevant meteorological indexes of each growth stage of crops;
s3: layering and nesting meteorological data and satellite remote sensing data to construct a yield prediction model of multi-layer regression;
s4: and obtaining global optimal parameters of the yield prediction model according to the actually measured yield data set to obtain a final yield prediction model.
2. The method for constructing an annual prediction model of yield of field food crops according to claim 1, wherein the specific method of step S1 is as follows:
s1-1: according to the annual growth characteristics of the crops, determining the dates corresponding to a key nutrition growth stage P1 and a reproduction growth stage P2 in the key growth stages of the crops;
s1-2: acquiring a surface reflectivity image data set of crops, calculating a vegetation index VI based on the surface reflectivity image data set, and acquiring the vegetation index VI in a selected time range;
s1-3: fitting the obtained vegetation index VI into a time-series index curve, wherein the starting time of the index curve is the starting date of the key nutrition growth phase P1 of the crop, and the ending time is the estimated ending date of the reproduction growth phase P2 of the crop;
s1-4: selecting the maximum value of vegetation index on the index curve, and recording as VI max Representing the maximum growth of crops;
s1-5: calculating the average value of the vegetation indexes of the key nutrition growth stage P1 and the reproductive growth stage P2 respectively, wherein the average value of the vegetation indexes of the key nutrition growth stage P1 is recorded as VI mean1 Representing the growth vigor degree of the early-stage crops; the average value of vegetation index in the reproductive growth stage P2 is recorded as VI mean2 Indicating the extent of post crop yield formation.
3. The method for constructing an annual prediction model of yield of field food crops according to claim 2, wherein the specific method of step S2 is as follows:
s2-1: acquiring an analysis data set and weather forecast data which comprise meteorological types influencing crop yield, and acquiring current month-average related meteorological data;
s2-2: calculating the average value of the month-to-month relevant meteorological data of the month corresponding to the last ten years according to the months of the critical growth stage of the crops;
s2-3: dividing the current month-to-month relevant weather data quantity by the ten-year month-to-month relevant weather data quantity average value to obtain a final month-to-month relevant weather index;
s2-4: and acquiring month-average related meteorological indexes corresponding to the key nutrition growth stage P1 and the reproduction growth stage P2 respectively.
4. The method for constructing the annual prediction model of the yield of the field grain crops according to claim 3, wherein in the step S3, the vegetation indexes and the weather indexes related to the month average are combined, layered nesting is carried out, the weather data are used as influencing factors, the relation between the vegetation indexes and the yield of the crops in different places is regulated and controlled, and a layered linear model HLM is introduced to analyze the influence of different layered prediction variables on the prediction value, so that the yield prediction model of the multilayer structure is constructed.
5. The method for constructing an annual prediction model of yield of field food crops according to claim 4, wherein the specific structure of the yield prediction model constructed in step S3 is as follows: model Level-1 layer including vegetation index VI max 、VI mean1 And VI mean2 The method is specifically expressed as follows:(1) The method comprises the steps of carrying out a first treatment on the surface of the In the formula, YIeld represents crop Yield, VI mean1 Representing the growth condition of the early-stage crops in the key nutrition growth stage P1; VI (VI) max Indicating the maximum degree of crop growth; VI (VI) mean2 Represents the degree of post crop yield formation in the reproductive growth stage P2,the intercept is indicated as the intercept and,、、the regression coefficient of the model is represented, and r represents the random error of the model Level-1 layer; model Level-2 layer, level-1 layer、、、The parameters are dependent variables, and the dependent variables are adjusted by month-average relevant meteorological indexes:(2) The method comprises the steps of carrying out a first treatment on the surface of the In the method, in the process of the invention,respectively in Level-1、、、；Representing the intercept;representing a random error, the random error is represented,represent the firstiMonth-to-month correlated meteorological indexN represents the number of month-average related meteorological indexes MI; for the followingThe change of j causes the weather index MI related to both n and month to change correspondingly, forAnd VI max Corresponding coefficientsThe month-average related meteorological index MI is a month-to-month average related meteorological index of a key growth stage; for VI mean1 Corresponding coefficientsThe month-average relevant weather index MI refers to the month-average relevant weather index of the critical vegetative growth phase P1; for VI mean2 Corresponding coefficientsThe month-average relevant weather index MI refers to the month-average relevant weather index of the reproductive growth stage P2.
6. The method for constructing an annual prediction model of yield of field food crops according to claim 5, wherein the parameter solving method of the yield prediction model in step S4 is as follows: the indicators in each piece of data in the measured yield dataset include: yield, VI max ，VI mean1 ，VI mean2 And month-average relevant meteorological indexes in the key growth stage, based on the data set, adopting extremum solution of nonlinear programming, setting target errors and iterative calculation times, and realizing the construction of a yield prediction model by inputting an actual measurement yield data set and determining globally optimal parameters through iteration.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410191321.8A CN117744898B (en) | 2024-02-21 | 2024-02-21 | Construction method of annual prediction model of yield of field grain crops |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410191321.8A CN117744898B (en) | 2024-02-21 | 2024-02-21 | Construction method of annual prediction model of yield of field grain crops |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117744898A true CN117744898A (en) | 2024-03-22 |
| CN117744898B CN117744898B (en) | 2024-05-28 |
Family
ID=90259569
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410191321.8A Active CN117744898B (en) | 2024-02-21 | 2024-02-21 | Construction method of annual prediction model of yield of field grain crops |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117744898B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109829234A (en) * | 2019-01-30 | 2019-05-31 | 北京师范大学 | A kind of across scale Dynamic High-accuracy crop condition monitoring and yield estimation method based on high-definition remote sensing data and crop modeling |
| CN113554232A (en) * | 2021-07-26 | 2021-10-26 | 吉林大学 | Crop yield prediction method and system |
| CN115496999A (en) * | 2022-09-05 | 2022-12-20 | 北京市农林科学院信息技术研究中心 | Method and device for estimating field straw yield |
| CN116956036A (en) * | 2023-07-26 | 2023-10-27 | 阳光农业相互保险公司 | Crop yield estimation method based on multi-source data |
-
2024
- 2024-02-21 CN CN202410191321.8A patent/CN117744898B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109829234A (en) * | 2019-01-30 | 2019-05-31 | 北京师范大学 | A kind of across scale Dynamic High-accuracy crop condition monitoring and yield estimation method based on high-definition remote sensing data and crop modeling |
| CN113554232A (en) * | 2021-07-26 | 2021-10-26 | 吉林大学 | Crop yield prediction method and system |
| US20230024846A1 (en) * | 2021-07-26 | 2023-01-26 | Jilin University | Crop yield prediction method and system |
| CN115496999A (en) * | 2022-09-05 | 2022-12-20 | 北京市农林科学院信息技术研究中心 | Method and device for estimating field straw yield |
| CN116956036A (en) * | 2023-07-26 | 2023-10-27 | 阳光农业相互保险公司 | Crop yield estimation method based on multi-source data |
Non-Patent Citations (4)
| Title |
|---|
| BINGXUE ZHU等: "A Regional Maize Yield Hierarchical Linear Combining Landsat 8 Vegetative Indices and Meteorological Data:Case Study in Jilin Provinces", 《REMOTE SENSING》, 21 January 2021 (2021-01-21) * |
| YU ZHAO等: "ChinaWheatYield30m: a 30 m annual winter wheat yield dataset from 2016 to 2021 in China", 《EARTH SYSTEM SCIENCE DATA》, 13 September 2023 (2023-09-13) * |
| 尹朝静;李谷成;高雪;: "气候因素对水稻单产影响的实证分析――基于湖北农户层面的分层模型", 自然资源学报, no. 08, 15 August 2017 (2017-08-15) * |
| 韩少宇: "基于多平台遥感数据的冬小麦长势监测和产量预测", 《中国优秀博士论文 农业科技》, 15 February 2024 (2024-02-15), pages 59 - 74 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117744898B (en) | 2024-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110751094B (en) | Crop yield estimation method based on GEE comprehensive remote sensing image and deep learning method | |
| Nilsson et al. | Full carbon account for Russia | |
| Chen et al. | A neural network integrated approach for rice crop monitoring | |
| Kucharik | Evaluation of a process-based agro-ecosystem model (Agro-IBIS) across the US Corn Belt: Simulations of the interannual variability in maize yield | |
| Wang et al. | Estimating near future regional corn yields by integrating multi-source observations into a crop growth model | |
| CN110909933B (en) | Agricultural drought rapid diagnosis and evaluation method coupling crop model and machine learning language | |
| CN109800921B (en) | Regional winter wheat yield estimation method based on remote sensing phenological assimilation and particle swarm optimization | |
| Srivastava et al. | Effect of sowing date distributions on simulation of maize yields at regional scale–A case study in Central Ghana, West Africa | |
| Raes et al. | AquaCrop-The FAO crop model to simulate yield response to water | |
| CN110826797B (en) | Method for determining optimal agricultural planting system based on multi-target comprehensive evaluation system | |
| CN111898922A (en) | Multi-scale crop yield assessment method and system | |
| Soundharajan et al. | Sensitivity analysis and auto-calibration of ORYZA2000 using simulation-optimization framework | |
| Tang et al. | Prediction of field winter wheat yield using fewer parameters at middle growth stage by linear regression and the BP neural network method | |
| CN116341724A (en) | Carbon absorption pre-estimating method based on global climate mode driven carbon circulation mechanism model | |
| CN117688404B (en) | Intelligent fertilization formula decision method, intelligent fertilization formula decision device, intelligent fertilization formula decision equipment, intelligent fertilization formula decision medium and program product | |
| Chaudhari et al. | Spatial wheat yield prediction using crop simulation model, GIS, remote sensing and ground observed data | |
| CN116595333B (en) | Soil-climate intelligent rice target yield and nitrogen fertilizer consumption determination method | |
| Cui et al. | Estimating maize yield in the black soil region of Northeast China using land surface data assimilation: integrating a crop model and remote sensing | |
| CN117744898B (en) | Construction method of annual prediction model of yield of field grain crops | |
| Wu et al. | Crop yield estimation and irrigation scheduling optimization using a root-weighted soil water availability based water production function | |
| CN116579872A (en) | Accurate irrigation decision-making method based on crop growth model and weather forecast | |
| Li et al. | Crop model data assimilation with particle filter for yield prediction using leaf area index of different temporal scales | |
| Howell et al. | Integrating multiple irrigation technologies for overall improvement in irrigation management | |
| Clouse | Spatial application of a cotton growth model for analysis of site-specific irrigation in the Texas High Plains | |
| Yuqi et al. | Temporal Variation Analysis of Rice Yield in the Jiangsu Province, China: Application of Decision Support System for Agrotechnology Transfer Model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |