Crop growth period prediction method based on ground temperature and related equipment
By constructing a machine learning model of annual average ground temperature sequence and crop growth period nodes, using temperature and ground temperature data to predict the growth period of crops, the problem of ground temperature data in the existing technology failing to effectively predict the growth period of crops, and achieving crop growth Accurate control.
Patent Information
- Application Number
- CN202010854948.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-08-24
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2040-08-24
AI Technical Summary
现有技术未能系统地利用地温数据预测农作物的生长期,导致对农作物生长的影响不够直观和准确。
By obtaining continuous temperature data and discontinuous ground temperature data from multiple adjacent areas, an annual average ground temperature sequence for the whole region is constructed, and a machine learning algorithm is used to generate a crop growth prediction model, and training is carried out to predict crop growth. period node.
It realizes the prediction of the growth period nodes of crops through ground temperature data, effectively controls the growth of crops, and improves the accuracy and controllability of predictions.
Smart Images

Figure CN112001543B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of intelligent agricultural technology, and in particular to a method for predicting the growth period of crops based on ground temperature and related equipment. Background Art
[0002] With the development of agricultural science and technology, more and more studies have been conducted on the impact of natural weather on agricultural production, specifically on the atmosphere, temperature, moisture, wind and disastrous weather. In order to utilize favorable climate resources, overcome the impact of unfavorable climate factors, provide agricultural climate resources for agricultural production development and agricultural economic adjustment, and at the same time have an important impact on other agricultural environmental factors and agricultural natural resources.
[0003] There are currently various natural climate change related data and corresponding crop growth data records to predict the growth period of crops, the most important of which include air temperature and accumulated temperature. For example, when the accumulated temperature of rapeseed is 111.7℃, it enters the germination and emergence stage, when the accumulated temperature is 505.9℃, it enters the seedling stage, when the accumulated temperature is 810.1℃, it enters the budding and bolting stage, when the accumulated temperature is 956.1℃, it enters the flowering stage, and when the accumulated temperature is 1355.6℃, it enters the silique development and maturity stage. However, the root system of crops grows in the soil, and the influence of soil temperature (ground temperature) on the growth of crops is more intuitive than air temperature, accumulated temperature, etc., and may even reverse the influence of air temperature, accumulated temperature, etc. on the growth of crops. However, in the prior art, there is still no systematic analysis of ground temperature data to predict the growth period of crops. Summary of the invention
[0004] The main purpose of the present invention is to solve the technical problem that there is no way to predict the growth period of crops through ground temperature data in the prior art.
[0005] A first aspect of the present invention provides a method for predicting the growth period of crops based on ground temperature, comprising:
[0006] Obtain N consecutive years of temperature data and M non-continuous years of ground temperature data from multiple adjacent observation stations in different regions, where N is greater than M and the same base year is used;
[0007] Based on the air temperature data and the ground temperature data, construct an annual average ground temperature sequence for the entire region for N consecutive years;
[0008] The annual average ground temperature sequence and the growing period nodes of crops for N consecutive years are used as training samples, and a preset machine learning algorithm is used for training to generate a crop growing period prediction model;
[0009] According to the annual average ground temperature sequence, the annual average ground temperature corresponding to the year to be predicted is estimated;
[0010] The annual average ground temperature is input into the crop growth period prediction model for processing, and the growth period nodes of the crops in the year to be predicted are output.
[0011] Optionally, in a first implementation of the first aspect of the present invention, after the step of obtaining N consecutive years of air temperature data and M non-continuous years of ground temperature data from multiple adjacent and different regional observation stations, the step further includes:
[0012] Conduct homogeneity tests on the temperature data of each observation station for M consecutive years, and retain the temperature data that pass the homogeneity test;
[0013] Using the temperature data of multiple adjacent observation stations that have passed the homogeneity check for M consecutive years, a stepwise multivariate linear regression analysis is performed on the non-continuous ground temperature data of a single observation station for M consecutive years, and a regression equation is established between the ground temperature data of a single observation station and the temperature data of multiple adjacent observation stations;
[0014] Taking the M-year continuous air temperature data of the plurality of adjacent observation stations that have passed the homogeneity check as reference values, the M-year discontinuous ground temperature data of the single observation station are interpolated based on the regression equation to obtain the M-year continuous ground temperature data of each observation station.
[0015] Optionally, in a second implementation of the first aspect of the present invention, constructing an annual average ground temperature sequence for N consecutive years in the entire region based on the air temperature data and the ground temperature data includes:
[0016] Select one or more observation stations as local stations, obtain monthly temperature data of the local stations for M consecutive years and calculate the average value year by year to obtain the first local annual average temperature data for M consecutive years;
[0017] Obtain the monthly ground temperature data of each observation station for M consecutive years and calculate the average value year by year to obtain the first annual average ground temperature data of the whole area for M consecutive years;
[0018] Conduct linear regression analysis on the first local annual average temperature data and the first regional annual average ground temperature data for M consecutive years, and establish a linear regression equation between the local annual average temperature data and the regional annual average ground temperature data;
[0019] Obtain the monthly temperature data of the local station for consecutive (NM) years and calculate the average year by year to obtain the second local annual average temperature data for consecutive (NM) years, wherein the consecutive (NM) years correspond to the years with temperature data but without ground temperature data;
[0020] Calculate the second regional average annual ground temperature data for consecutive (NM) years according to the linear regression equation and the second local average annual temperature data;
[0021] The first average annual geothermal data of the whole region for M consecutive years is spliced with the second average annual geothermal data of the whole region for (NM) consecutive years to obtain the average annual geothermal sequence of the whole region for N consecutive years.
[0022] Optionally, in a third implementation of the first aspect of the present invention, performing a homogeneity test on the temperature data of each observation station for M consecutive years and retaining the temperature data that passes the homogeneity test comprises:
[0023] The preset first mutation year verification algorithm and the second mutation year verification algorithm are respectively used to verify the temperature data of each observation station for M consecutive years, and the first predicted mutation year and the second predicted mutation year are obtained accordingly;
[0024] Determining whether the first predicted mutation year and the second predicted mutation year are the same;
[0025] If they are the same, the temperature data of the corresponding year will be eliminated;
[0026] If different, determining whether the first predicted mutation year or the second predicted mutation year is marked with a preset mutation label;
[0027] If it is marked with a preset mutation label, the temperature data of the corresponding year will be eliminated, and the temperature data that have not been eliminated will be determined to pass the homogeneity test and be retained.
[0028] A second aspect of the present invention provides a device for predicting the growth period of crops based on ground temperature, comprising:
[0029] An acquisition module is used to obtain N consecutive years of temperature data and M non-continuous years of ground temperature data from multiple adjacent observation stations in different regions, where N is greater than M and the same base year is used;
[0030] A construction module is used to construct an annual average ground temperature sequence for the entire region for N consecutive years based on the air temperature data and the ground temperature data;
[0031] A training module, for using the annual average ground temperature sequence of N consecutive years and the growing period nodes of crops as training samples, using a preset machine learning algorithm for training, and generating a crop growing period prediction model;
[0032] An estimation module, used for estimating the annual average ground temperature corresponding to the year to be predicted based on the annual average ground temperature sequence;
[0033] The prediction module is used to input the annual average ground temperature into the crop growth period prediction model for processing, and output the growth period node of the crop in the year to be predicted.
[0034] Optionally, in a first implementation of the second aspect of the present invention, the method and device for predicting crop growth period based on ground temperature further includes:
[0035] The test module is used to perform a homogeneity test on the temperature data of each observation station for M consecutive years, and retain the temperature data that pass the homogeneity test;
[0036] An analysis module is used to use the temperature data of multiple adjacent observation stations that have passed the homogeneity check for M consecutive years to perform stepwise multivariate linear regression analysis on the non-continuous ground temperature data of a single observation station for M consecutive years, and to establish a regression equation between the ground temperature data of a single observation station and the temperature data of multiple adjacent observation stations;
[0037] The interpolation module is used to use the temperature data of multiple adjacent observation stations that have passed the uniformity check for M consecutive years as reference values, and based on the regression equation, interpolate the discontinuous ground temperature data of the single observation station for M consecutive years to obtain the ground temperature data of each observation station for M consecutive years.
[0038] Optionally, in a second implementation of the second aspect of the present invention, the building module includes:
[0039] The first mean value calculation unit is used to select a single or multiple observation stations as local stations, obtain the monthly temperature data of the local station for M consecutive years and perform mean calculation year by year to obtain the first local annual average temperature data for M consecutive years; and to obtain the monthly ground temperature data of each observation station for M consecutive years and perform mean calculation year by year to obtain the first annual average ground temperature data of the whole area for M consecutive years;
[0040] An analysis unit is used to perform a linear regression analysis on the first local annual average temperature data and the first regional annual average ground temperature data for M consecutive years, and establish a linear regression equation between the local annual average temperature data and the regional annual average ground temperature data;
[0041] A second mean value calculation unit is used to obtain the monthly temperature data of the local station for consecutive (NM) years and perform mean calculation year by year to obtain the second local annual mean temperature data for consecutive (NM) years, wherein the consecutive (NM) years correspond to years with temperature data but without ground temperature data;
[0042] A ground temperature data calculation unit, used for calculating the second regional average annual ground temperature data for consecutive (NM) years according to the linear regression equation and the second local average annual air temperature data;
[0043] The splicing unit is used to splice the first average annual geothermal data of the whole area for M consecutive years with the second average annual geothermal data of the whole area for (NM) consecutive years to obtain the average annual geothermal sequence of the whole area for N consecutive years.
[0044] Optionally, in a third implementation of the second aspect of the present invention, the inspection module includes:
[0045] A first judging unit, used for judging whether the first predicted mutation year and the second predicted mutation year are the same;
[0046] A first screening unit, configured to remove temperature data of a corresponding year if the first predicted mutation year and the second predicted mutation year are the same;
[0047] a second determination unit, configured to determine whether the first predicted mutation year or the second predicted mutation year is marked with a preset mutation label if the first predicted mutation year and the second predicted mutation year are different;
[0048] The second screening unit is used to eliminate the temperature data of the corresponding year if it is marked with a preset mutation label, and to ensure that the temperature data that has not been eliminated passes the homogeneity test and is retained.
[0049] The third aspect of the present invention provides a ground temperature-based crop growth period prediction device, comprising: a memory and at least one processor, wherein the memory stores instructions; the at least one processor calls the instructions in the memory so that the ground temperature-based crop growth period prediction device executes the above-mentioned ground temperature-based crop growth period prediction method.
[0050] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the above-mentioned method for predicting the growth period of crops based on ground temperature.
[0051] In the technical solution provided by the present invention, firstly, multiple linear regression is performed through non-continuous M-year ground temperature data, missing ground temperature data is interpolated, and ground temperature data of continuous M years is obtained; then, ground temperature data of NM years is predicted through continuous N-year air temperature data and continuous M-year ground temperature data, and ground temperature data of continuous N years is obtained; then, ground temperature data of continuous N years and related crop growth period nodes are trained to generate a crop growth period prediction model; finally, the average annual ground temperature is predicted, and then the crop growth period prediction model is input to predict the growth period of crops. The present invention realizes the prediction of crop generation period nodes through ground temperature, which can effectively control the growth of crops. BRIEF DESCRIPTION OF THE DRAWINGS
[0052] Figure 1 It is a schematic diagram of a first embodiment of a method for predicting the growth period of crops based on ground temperature according to the present invention;
[0053] Figure 2 It is a schematic diagram of a second embodiment of the crop growth period prediction method based on ground temperature of the present invention;
[0054] Figure 3It is a schematic diagram of a third embodiment of the crop growth period prediction method based on ground temperature of the present invention;
[0055] Figure 4 It is a schematic diagram of a fourth embodiment of the method for predicting the growing period of crops based on ground temperature of the present invention;
[0056] Figure 5 It is a schematic diagram of an embodiment of a device for predicting the growth period of crops based on ground temperature of the present invention;
[0057] Figure 6 It is a schematic diagram of another embodiment of the crop growth period prediction device based on ground temperature of the present invention;
[0058] Figure 7 It is a schematic diagram of an embodiment of the crop growth period prediction device based on ground temperature of the present invention. DETAILED DESCRIPTION
[0059] The embodiment of the present invention provides a method and related equipment for predicting the growing period of crops based on ground temperature, obtaining N consecutive years of air temperature data and M discontinuous years of ground temperature data from multiple adjacent and different regional observation stations, wherein N is greater than M and the same base year is used; based on the air temperature data and ground temperature data, constructing the annual average ground temperature sequence for N consecutive years in the whole region; using the annual average ground temperature sequence for N consecutive years and the growing period nodes of crops as training samples, using a preset machine learning algorithm for training, generating a crop growing period prediction model; estimating the annual average ground temperature corresponding to the year to be predicted according to the annual average ground temperature sequence; inputting the annual average ground temperature into the crop growing period prediction model for processing, and outputting the growing period nodes of the crops in the year to be predicted. The present invention realizes the prediction of the crop generation period nodes by ground temperature, and can effectively control the growth of crops.
[0060] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0061] Before explaining the embodiments of the present invention, the main idea of the present invention is first introduced as follows:
[0062] Possibly due to technological development or historical reasons, there are two problems in the monitoring and recording of historical geothermal data in various regions: first, the historical recording time of geothermal data is short, resulting in few available data samples; second, the geothermal data records are incomplete, resulting in discontinuous available data samples.
[0063] Therefore, the crop growth period prediction method based on ground temperature of the present invention is divided into two main parts: ground temperature data completion and crop growth period prediction. Ground temperature data completion includes local ground temperature data interpolation and global ground temperature data correction; crop growth period prediction includes crop growth period prediction model training and crop growth period node prediction. First, the two main parts and four branch parts of the present invention are briefly introduced as follows:
[0064] (I) Local geothermal data interpolation
[0065] If we need N consecutive years of geothermal data, but there are only N consecutive years of air temperature data with discrete observation sequences and good continuity, and M discontinuous years (M<N) of geothermal data, and the geothermal data of years a, b, c, d, and e are missing, then we can perform a multivariate linear regression analysis on the discontinuous M years of geothermal data using the continuous M years of air temperature data to obtain a regression equation related to geothermal data and air temperature data. Substituting the air temperature data of years a, b, c, d, and e into the regression equation can obtain the missing geothermal data and interpolate the discontinuous M years of geothermal data.
[0066] (II) Global correction of ground temperature data
[0067] After obtaining the geothermal data for M consecutive years, calculate the annual average air temperature data and annual average geothermal data for M consecutive years respectively, and calculate the linear regression equation between the two; then substitute the annual average air temperature data for NM consecutive years into the linear regression equation, and calculate the annual average geothermal data for NM consecutive years; splice the annual average geothermal data for M consecutive years and NM consecutive years to obtain the annual average geothermal sequence for N consecutive years.
[0068] (III) Crop growth period prediction model training
[0069] The crop growth period prediction model is trained by comparing the annual average ground temperature sequence for N consecutive years with the various growth period nodes of the crops in the corresponding years, where the input is the annual average ground temperature in the prediction year and the output is the various growth period nodes of the crops.
[0070] (IV) Forecasting of crop growth period nodes
[0071] The annual average ground temperature sequence can be used to predict the annual average ground temperature that was not recorded in the past or in a certain year in the future; then the predicted annual average ground temperature is input into the crop growth prediction model to obtain the various growth period nodes of the crop in the predicted year.
[0072] Next, the specific process of the embodiment of the present invention is described. Figure 1 The first embodiment of the crop growth period prediction method based on ground temperature in the embodiment of the present invention includes:
[0073] 101. Obtain N consecutive years of temperature data and M non-continuous years of ground temperature data from multiple adjacent observation stations in different regions, where N is greater than M and the same base year is used;
[0074] It is understandable that the execution subject of the present invention may be a device for predicting crop growth period based on ground temperature, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking a server as the execution subject as an example.
[0075] In this embodiment, multiple adjacent and different regional observation stations refer to multiple observation stations in multiple different regions within the entire area; among them, the multiple observation stations can be divided into multiple groups of observation stations according to different regions, and each group of observation stations takes one of the observation stations as a reference, and the altitude difference and straight-line distance between them and other observation stations do not exceed the preset values A and B.
[0076] In this embodiment, N consecutive years of temperature data means that temperature data is recorded in each of the specified N years; non-continuous M years of geothermal data means that among the specified M years, geothermal data is recorded for many years, but geothermal data for one or more years is missing.
[0077] Among them, geothermal data and air temperature data refer to the geothermal temperature and air temperature recorded in each preset period in the corresponding year. The preset period can be half a year, a quarter, a month, a week, a day, etc.; geothermal data can include temperatures at different depths such as 0 cm on the surface and 5 cm, 10 cm, 15 cm, 20 cm underground.
[0078] In addition, M is the maximum number of years with recorded ground temperature data in the observation station, so N is greater than M in order to ensure that the subsequent temperature data of NM years can be used to calculate and predict the ground temperature data of NM years that are not recorded. The same base year is used for N and M to ensure comparability between the two.
[0079] 102. Based on the air temperature data and the ground temperature data, construct an annual average ground temperature sequence for the entire region for N consecutive years;
[0080] In this embodiment, the annual average geothermal sequence refers to: the average periodic geothermal temperature recorded in the geothermal data of a specified year. For example, if the geothermal data records the geothermal temperature of each month in a specified year, the geothermal temperatures of each month are added together and then divided by 12 months to obtain the annual average geothermal sequence. Therefore, it is necessary to first obtain the geothermal data of the entire area for N consecutive years.
[0081] In this embodiment, the acquired ground temperature data contains two problems:
[0082] 1. Lack of ground temperature data for one or more years within M years;
[0083] 2. Lack of ground temperature data for NM years;
[0084] Therefore, it is necessary to first fill in the missing geothermal data of one or more years within M years, as well as the geothermal data of NM years. The geothermal data can be filled in by the following methods:
[0085] 1. Complete the missing geothermal data of one or more years within M years: Use the continuous M years of air temperature data to conduct stepwise multivariate linear regression analysis on the non-continuous M years of geothermal data, and establish a regression equation between the two; then substitute the years to be completed into the regression equation to obtain the missing geothermal data of one or more years.
[0086] 2. Complete the ground temperature data for NM years: Perform linear regression analysis on the annual average air temperature data and annual average ground temperature data for M consecutive years, establish a linear regression equation between the two, and then substitute NM years into the linear regression equation to obtain the annual average ground temperature data for NM years.
[0087] 103. Using the annual average ground temperature sequence of N consecutive years and the growing period nodes of crops as training samples, a preset machine learning algorithm is used for training to generate a crop growing period prediction model;
[0088] In this embodiment, the growth period nodes of crops, such as the growth period nodes of rapeseed, include germination and emergence, seedling stage, bud and bolting stage, flowering stage, and silique development and maturity stage. The growth period nodes of crops here refer to the time of each growth period node of crops in the whole area (i.e., the ground temperature monitoring area) for N consecutive years, in days, that is, recorded in the form of year / month / day. Among them, the crops for which the ground temperature data is used to predict the growth period nodes are preferably dryland crops.
[0089] In this embodiment, the annual average ground temperature sequence and the growing period nodes of crops are labeled, and then the relationship between the two is trained through a preset machine learning algorithm, wherein the machine learning algorithm can be: Spark machine learning, XGBoost (eXtreme Gradient Boosting, extreme gradient boosting) machine learning, to predict the growing period nodes of crops corresponding to different annual average ground temperatures.
[0090] Among them, the annual average ground temperature series can be preprocessed by low-pass filtering to obtain the interannual variation characteristics of the annual average low temperature series.
[0091] 104. Estimate the annual average ground temperature corresponding to the year to be predicted based on the annual average ground temperature sequence;
[0092] In this embodiment, the annual average ground temperature sequence can be represented by a linear regression equation, and the year to be predicted can be input into the linear regression equation to obtain the annual average ground temperature corresponding to the year to be predicted. The year to be predicted can include the annual average ground temperature that has not been recorded N years ago and the annual average ground temperature of future years.
[0093] 105. Input the annual average ground temperature into the crop growth period prediction model for processing, and output the growth period nodes of the crops in the year to be predicted.
[0094] In this embodiment, the trained crop growth period prediction model is directly used, and the annual average ground temperature of the year to be predicted is input to predict the growth period nodes of the crops in the year to be predicted.
[0095] In the embodiment of the present invention, firstly, multiple linear regression is performed through non-continuous M-year ground temperature data, missing ground temperature data is interpolated, and ground temperature data of continuous M years is obtained; then, ground temperature data of NM years is predicted through continuous N-year air temperature data and continuous M-year ground temperature data, and ground temperature data of continuous N years is obtained; then, ground temperature data of continuous N years and related crop growth period nodes are trained to generate a crop growth period prediction model; finally, the average annual ground temperature is predicted, and then the crop growth period prediction model is input to predict the growth period of crops. The present invention realizes the prediction of crop generation period nodes through ground temperature, which can effectively control the growth of crops.
[0096] See also Figure 2 The second embodiment of the crop growth period prediction method based on ground temperature in the embodiment of the present invention includes:
[0097] 201. Obtain N consecutive years of temperature data and M non-continuous years of ground temperature data from multiple adjacent observation stations in different regions, where N is greater than M and the same base year is used;
[0098] In this embodiment, the air temperature data and ground temperature data of multiple adjacent and different regional observation stations are used to ensure that the annual average air temperature data and annual average ground temperature data obtained by subsequent calculation are representative and comprehensive in the entire region.
[0099] 202. Conduct a homogeneity test on the temperature data of each observation station for M consecutive years, and retain the temperature data that pass the homogeneity test;
[0100] In this embodiment, the uniformity test refers to checking the temperature data with abnormal difference from the average temperature data level in M consecutive years; the temperature data with abnormal difference from the average temperature data level includes: abnormal monitoring temperature data and sudden change temperature data caused by the migration of the detection station. Other temperature data are at the average temperature data level and are considered to have passed the uniformity test. For example, if an observation station is moved from a city to a mountainous area, the temperature data before and after the migration year will cause a sudden change in temperature data due to environmental changes, so the temperature data monitored in that year cannot pass the uniformity test and will not be retained.
[0101] In this embodiment, the temperature data is subjected to a uniformity test, and only the temperature data that passes the uniformity test is retained. The purpose of this test is to increase the fit between the subsequently established regression equation and the changing trend of the actual temperature data, and to avoid the influence of abnormal temperature data on the regression equation.
[0102] Specifically, the Penalized Maximal F Test (PMFT) and the Penalized Maximal T Test (PMT) can be used to test the homogeneity of the temperature data of each observation station for M consecutive years.
[0103] 203. Using the temperature data of multiple adjacent observation stations that have passed the homogeneity check for M consecutive years, a stepwise multivariate linear regression analysis is performed on the non-continuous ground temperature data of a single observation station for M consecutive years, and a regression equation is established between the ground temperature data of a single observation station and the temperature data of multiple adjacent observation stations;
[0104] In this embodiment, the stepwise multiple linear regression analysis of the non-continuous M-year ground temperature data of a single observation station is specifically performed as follows:
[0105] (1) Construct a model containing only constant terms;
[0106] (2) Calculate the absolute value of the correlation coefficient between each air temperature data and ground temperature data, and select the two air temperature data with the largest absolute value of the correlation coefficient to input into the model;
[0107] (3) Conduct a significance test on the temperature data and eliminate the temperature data that does not meet the significance test;
[0108] (4) According to the order of the absolute value of the correlation coefficient, select another temperature data input model and execute step (3);
[0109] (5) Steps (4) and (3) are executed in a loop until all the temperature data in the model meet the significance test, and every temperature data in the model does not meet the significance test, and the regression equation between the temperature data and the ground temperature data is obtained.
[0110] 204. Using the temperature data of the plurality of adjacent observation stations that have passed the homogeneity check for M consecutive years as reference values, interpolating the non-continuous ground temperature data of the single observation station for M consecutive years based on the regression equation to obtain the ground temperature data of each observation station for M consecutive years;
[0111] In this embodiment, the missing geothermal data year in the non-continuous M-year geothermal data is determined; the air temperature data corresponding to the missing geothermal data year is determined; the air temperature data is substituted into the regression equation to calculate the geothermal data of the corresponding year; the calculated geothermal data is spliced into the corresponding year to obtain the geothermal data of each observation station for M consecutive years. The regression equation is as follows:
[0112]
[0113] Among them, T (j) is the missing ground temperature in the jth year, T ij is the temperature of the i-th observation station in the j-th year, R i is the relationship coefficient between the ith observation station and the detection station that monitors ground temperature data, and k is the number of adjacent observation stations.
[0114] 205. Based on the air temperature data and the ground temperature data, construct an annual average ground temperature sequence for the entire region for N consecutive years;
[0115] 206. Using the annual average ground temperature sequence of N consecutive years and the growing period nodes of crops as training samples, a preset machine learning algorithm is used for training to generate a crop growing period prediction model;
[0116] 207. Estimate the annual average ground temperature corresponding to the year to be predicted based on the annual average ground temperature sequence;
[0117] 208. Input the annual average ground temperature into the crop growth period prediction model for processing, and output the growth period nodes of the crops in the year to be predicted.
[0118] The embodiment of the present invention introduces in detail how to interpolate the missing parts of non-continuous geothermal data. Specifically, a regression equation related to the existing continuous air temperature data and non-continuous geothermal data is fitted, and the missing geothermal data is interpolated by substituting the air temperature data of the missing years into the regression equation to obtain continuous geothermal data.
[0119] See also Figure 3 The third embodiment of the crop growth period prediction method based on ground temperature in the embodiment of the present invention includes:
[0120] 301. Obtain N consecutive years of temperature data and M non-continuous years of ground temperature data from multiple adjacent observation stations in different regions, where N is greater than M and the same base year is used;
[0121] 302. Select one or more observation stations as local stations, obtain monthly temperature data of the local stations for M consecutive years, and calculate the average value year by year to obtain the first local annual average temperature data for M consecutive years;
[0122] In this embodiment, the global adjustment of station value (GAoSV) can be used to construct the annual average geothermal sequence of the whole area for N consecutive years, which means that the local (M years) geothermal data are used to complete the global (NM years) geothermal data.
[0123] In this embodiment, the local station refers to one or more observation stations with the highest correlation with the temperature data (and ground temperature data) collected by all observation stations in the whole region. Its function is to represent the temperature data of all observation stations in the whole region through the temperature data of the local station. The specific selection process is as follows:
[0124] (1) Calculate S for each observation station i Correlation coefficient R relative to all observation stations i , and for R i Sort from largest to smallest;
[0125] (2) According to R i The sorting order is from S i Select one or more observation stations as local stations;
[0126] For a single local station, it is only necessary to add up the monthly temperature data of the local station in each of the M consecutive years and divide by 12 to obtain the first local annual average temperature data for the M consecutive years; for multiple local stations, it is necessary to calculate the mean of the annual average temperature data of each local station to obtain the first local annual average temperature data for the M consecutive years.
[0127] 303. Obtain the monthly ground temperature data of each observation station for M consecutive years and calculate the average value year by year to obtain the first annual average ground temperature data of the whole area for M consecutive years;
[0128] In this embodiment, since the geothermal data collected by each observation station are relatively small, and the geothermal data for M consecutive years are obtained through interpolation to predict the missing geothermal data, in order to ensure the comprehensiveness and representativeness of the annual average geothermal data in the entire area, the geothermal data of the entire area is used here to calculate the first annual average geothermal data of the entire area for M consecutive years.
[0129] For each of the M consecutive years, the monthly temperature data of each monitoring station is added up and divided by 12 to obtain the first annual average ground temperature data of the entire region for M consecutive years.
[0130] 304. Perform linear regression analysis on the first local annual average temperature data and the first regional annual average ground temperature data for M consecutive years, and establish a linear regression equation between the local annual average temperature data and the regional annual average ground temperature data;
[0131] In this embodiment, a regression model is established for the first local annual average temperature data and the first regional annual average ground temperature data for M consecutive years to perform linear regression analysis on them, and a linear regression equation between the two can be output. Specifically, linear fitting can be performed through Excel or Matlab.
[0132] The linear regression equation is T s =aT c +b, where T s The first area's annual average ground temperature data T c It is the first local average temperature data.
[0133] 305. Obtain the monthly temperature data of the local station for consecutive (NM) years and perform average calculation year by year to obtain the second local annual average temperature data for consecutive (NM) years, wherein the consecutive (NM) years correspond to years with temperature data but without ground temperature data;
[0134] 306. Calculate the second regional average annual ground temperature data for consecutive (NM) years according to the linear regression equation and the second local average annual temperature data;
[0135] In this embodiment, there is a lack of geothermal data in the NM year, and the linear regression equation is used to express the relationship between the local annual average air temperature data and the annual average geothermal data of the entire region; therefore, here the second local annual average air temperature data of each NM year can be calculated and substituted into the linear regression equation respectively, to obtain the corresponding second annual average geothermal data of the entire region for each NM year.
[0136] 307. Connect the first average annual ground temperature data of the whole region for M consecutive years with the second average annual ground temperature data of the whole region for (NM) consecutive years to obtain an average annual ground temperature sequence of the whole region for N consecutive years;
[0137] 308. Using the annual average ground temperature sequence of N consecutive years and the growing period nodes of crops as training samples, a preset machine learning algorithm is used for training to generate a crop growing period prediction model;
[0138] 309. Estimate the annual average ground temperature corresponding to the year to be predicted based on the annual average ground temperature sequence;
[0139] 310. Input the annual average ground temperature into the crop growth period prediction model for processing, and output the growth period nodes of the crops in the year to be predicted.
[0140] In an embodiment of the present invention, the prediction and completion of geothermal data for years that have not been collected are introduced in detail. Specifically, a linear regression equation related to continuous local annual average air temperature data and continuous regional annual average geothermal data is fitted. The local annual average air temperature data for years that have not been collected is substituted into the linear regression equation to obtain the corresponding regional annual average geothermal data, thereby realizing the prediction and completion of geothermal data for years that have not been collected.
[0141] See also Figure 4 The fourth embodiment of the crop growth period prediction method based on ground temperature in the embodiment of the present invention includes:
[0142] 401. Obtain N consecutive years of temperature data and M non-continuous years of ground temperature data from multiple adjacent observation stations in different regions, where N is greater than M and the same base year is used;
[0143] 402. Use the preset first mutation year verification algorithm and the preset second mutation year verification algorithm to verify the temperature data of each observation station for M consecutive years, and obtain the first predicted mutation year and the second predicted mutation year accordingly;
[0144] In this embodiment, the first mutation year verification algorithm and the second mutation year verification algorithm may be: PMFT and PMT.
[0145] Among them, PMFT does not need to consider the order of temperature data, and can use the temperature data of the detection station to be tested or multiple nearby detection stations for calculation; and PMT uses multiple nearby detection stations at the same time to ensure that the normal distribution sequence of temperature data is uniform and representative of the region, and calculates through the temperature data of PMFT. The selection criteria for multiple nearby detection stations are: taking the detection station to be tested as a reference, the correlation coefficient, altitude difference, and distance between the detection station to be tested and other detection stations must meet the preset conditions; for example, the correlation coefficient between the detection station to be tested and other detection stations must be greater than r; the altitude difference does not exceed h, and the distance does not exceed d.
[0146] In addition, PMFT focuses on the average level of temperature data to test the mutation years; PMT focuses on the trend characteristics of temperature data to detect the mutation years.
[0147] 403. Determine whether the first predicted mutation year and the second predicted mutation year are the same;
[0148] 404. If they are the same, the temperature data of the corresponding year will be eliminated;
[0149] In this embodiment, if the first predicted mutation year and the second predicted mutation year are the same, it means that the average level and trend characteristics of the temperature data in that year are consistent with the characteristics of the mutation temperature data, and the homogeneity test fails.
[0150] 405. If different, determining whether the first predicted mutation year or the second predicted mutation year is marked with a preset mutation label;
[0151] 406. If the preset mutation label is marked, the temperature data of the corresponding year is removed, and the temperature data that is not removed is determined to pass the homogeneity test and is retained;
[0152] In this embodiment, when the detection station is migrated and the geographical environment changes significantly, such as migrating from a forest to a sandy land, the collected ground temperature data will obviously increase, and the year in which the detection station is migrated is marked with a preset mutation label; for the temperature data that only passes one uniformity test, it is determined whether the detection station is migrated in that year; and for the abnormal temperature data caused by the migration of the detection station, it is also determined that it does not pass the uniformity test and is therefore eliminated.
[0153] 407. Using the temperature data of multiple adjacent observation stations that have passed the homogeneity check for M consecutive years, a stepwise multivariate linear regression analysis is performed on the ground temperature data of a single observation station for M non-continuous years, and a regression equation between the ground temperature data of a single observation station and the temperature data of multiple adjacent observation stations is established;
[0154] 408. Using the temperature data of the plurality of adjacent observation stations that have passed the homogeneity check for M consecutive years as reference values, interpolating the non-continuous ground temperature data of the single observation station for M consecutive years based on the regression equation to obtain the ground temperature data of each observation station for M consecutive years;
[0155] 409. Based on the air temperature data and the ground temperature data, construct an annual average ground temperature sequence for the entire region for N consecutive years;
[0156] 410. Using the annual average ground temperature sequence of N consecutive years and the growing period nodes of crops as training samples, a preset machine learning algorithm is used for training to generate a crop growing period prediction model;
[0157] 411. Estimate the annual average ground temperature corresponding to the year to be predicted based on the annual average ground temperature sequence;
[0158] 412. Input the annual average ground temperature into the crop growth period prediction model for processing, and output the growth period nodes of the crops in the year to be predicted.
[0159] In an embodiment of the present invention, the abnormally collected temperature data is eliminated through uniformity judgment, and the temperature data with higher uniformity is used for fitting the subsequent regression equation and the linear regression equation, with higher goodness of fit and more accurate prediction of the ground temperature data.
[0160] The above is a preliminary description of the crop growth period prediction method based on ground temperature of the present invention. In order to make the present invention easier to understand, an optimal embodiment is selected below for further explanation, wherein N=37 (1981-2018), M=27 (1991-2018), and the air temperature data and ground temperature data are derived from 68 detection stations in area A.
[0161] (I) Local geothermal data interpolation
[0162] In this embodiment, the air temperature data of 68 detection stations in area A from 1981 to 2018 are obtained, and the ground temperature data of 68 detection stations in area A from 1991 to 2018 are obtained (the ground temperature data for 4 years from 1991 to 2018 are missing).
[0163] In this embodiment, the air temperature data and ground temperature data collected by each detection station from 1991 to 2018 are tested for homogeneity, and the confidence level is set to 0.95.
[0164] Preferably, PMFT and PMT are used to calculate the mutation year, and the air temperature data and ground temperature data collected in the mutation year are eliminated, and the remaining air temperature data and ground temperature data are retained to pass the uniformity test.
[0165] The neighboring temperature data and ground temperature data that passed the homogeneity test were subjected to stepwise multivariate linear regression analysis to obtain the corresponding regression equation. Then, the temperature data for the four years from 1991 to 2018 when ground temperature data was missing were found and substituted into the regression equation to obtain the ground temperature data for these four years. The data were then interpolated into the missing ground temperature data from 1991 to 2018 to obtain continuous ground temperature data from 1991 to 2018.
[0166] (II) Global correction of ground temperature data
[0167] In this embodiment, the correlation coefficient between each detection station and all 68 detection stations is calculated through the temperature data from 1991 to 2018; assuming that the correlation coefficients between detection stations No. 2 and No. 3 and all 68 detection stations are the highest, which are 0.98 and 0.97 respectively, the first local average annual temperature data of detection stations No. 2 and No. 3 are calculated; then the first full-area average annual ground temperature data from 1991 to 2018 is calculated.
[0168] In this embodiment, the first local annual average temperature data of the No. 2 detection station and the No. 3 detection station, and the first regional annual average ground temperature data, linear regression analysis of the fitting equation of the local annual average temperature data and the regional annual average ground temperature data (1991-2018): T s =0.8722T c +3.2551, where the fitting accuracy is 0.99.
[0169] In this embodiment, the local annual average temperature data of the No. 2 and No. 3 detection stations from 1981 to 1990 are calculated and then substituted into T s =0.8722T c +3.2551, we can get the average annual ground temperature data for the whole area from 1981 to 1990.
[0170] In this embodiment, the annual average geothermal data of the whole area from 1981 to 1990 is spliced with the annual average geothermal data of the whole area from 1991 to 2018 to obtain the annual average geothermal sequence from 1981 to 2018.
[0171] (III) Crop growth period prediction model training
[0172] In this embodiment, a crop growth period prediction model is trained through the annual average ground temperature sequence from 1981 to 2018 and the various growth period nodes of crops from 1981 to 2018, wherein the input is the annual average ground temperature of the prediction year, and the output is the various growth period nodes of crops.
[0173] (IV) Forecasting of crop growth period nodes
[0174] In this embodiment, the relevant technology for predicting temperature data is relatively mature. For example, if the growth period nodes of crops in 2025 are predicted, a linear fitting curve of the annual average ground temperature data relative to time is constructed based on the annual average ground temperature sequence to obtain the corresponding linear regression equation; then 2025 is inserted into the linear regression equation to predict the annual average ground temperature data T in 2025. 2025c Then, T 2025c The crop growth period prediction model can predict the various growth period nodes of crops in 2025.
[0175] The above describes the crop growth period prediction method based on ground temperature in the embodiment of the present invention. The following describes the crop growth period prediction device based on ground temperature in the embodiment of the present invention. Figure 5 In one embodiment of the present invention, a device for predicting the growth period of crops based on ground temperature includes:
[0176] An acquisition module 501 is used to acquire N consecutive years of temperature data and M non-continuous years of ground temperature data from multiple adjacent observation stations in different regions, where N is greater than M and the same base year is used;
[0177] A construction module 502 is used to construct an annual average ground temperature sequence for the entire region for N consecutive years based on the air temperature data and the ground temperature data;
[0178] A training module 503 is used to use the annual average ground temperature sequence of N consecutive years and the growing period nodes of crops as training samples, and adopt a preset machine learning algorithm for training to generate a crop growing period prediction model;
[0179] An estimation module 504 is used to estimate the annual average ground temperature corresponding to the year to be predicted based on the annual average ground temperature sequence;
[0180] The prediction module 505 is used to input the annual average ground temperature into the crop growth period prediction model for processing, and output the growth period node of the crop in the year to be predicted.
[0181] In the embodiment of the present invention, firstly, multiple linear regression is performed through non-continuous M-year ground temperature data, missing ground temperature data is interpolated, and ground temperature data of continuous M years is obtained; then, ground temperature data of NM years is predicted through continuous N-year air temperature data and continuous M-year ground temperature data, and ground temperature data of continuous N years is obtained; then, ground temperature data of continuous N years and related crop growth period nodes are trained to generate a crop growth period prediction model; finally, the average annual ground temperature is predicted, and then the crop growth period prediction model is input to predict the growth period of crops. The present invention realizes the prediction of crop generation period nodes through ground temperature, which can effectively control the growth of crops.
[0182] See also Figure 6 Another embodiment of the crop growth period prediction device based on ground temperature in the embodiment of the present invention includes:
[0183] An acquisition module 501 is used to acquire N consecutive years of temperature data and M non-continuous years of ground temperature data from multiple adjacent observation stations in different regions, where N is greater than M and the same base year is used;
[0184] A construction module 502 is used to construct an annual average ground temperature sequence for the entire region for N consecutive years based on the air temperature data and the ground temperature data;
[0185] A training module 503 is used to use the annual average ground temperature sequence of N consecutive years and the growing period nodes of crops as training samples, and adopt a preset machine learning algorithm for training to generate a crop growing period prediction model;
[0186] An estimation module 504 is used to estimate the annual average ground temperature corresponding to the year to be predicted based on the annual average ground temperature sequence;
[0187] The prediction module 505 is used to input the annual average ground temperature into the crop growth period prediction model for processing, and output the growth period node of the crop in the year to be predicted.
[0188] Specifically, the crop growth period prediction method and device based on ground temperature also includes:
[0189] A test module 506 is used to perform a uniformity test on the temperature data of each observation station for M consecutive years, and retain the temperature data that passes the uniformity test;
[0190] The analysis module 507 is used to use the temperature data of multiple adjacent observation stations that have passed the homogeneity check for M consecutive years to perform stepwise multivariate linear regression analysis on the ground temperature data of a single observation station for M non-continuous years, and establish a regression equation between the ground temperature data of a single observation station and the temperature data of multiple adjacent observation stations;
[0191] The interpolation module 508 is used to use the temperature data of multiple adjacent observation stations that have passed the uniformity check for M consecutive years as reference values, and based on the regression equation, interpolate the non-continuous ground temperature data of the single observation station for M consecutive years to obtain the ground temperature data of each observation station for M consecutive years.
[0192] Specifically, the construction module 502 includes:
[0193] The first mean value calculation unit 5021 is used to select a single or multiple observation stations as local stations, obtain the monthly temperature data of the local station for M consecutive years and perform mean calculation year by year to obtain the first local annual average temperature data for M consecutive years; and to obtain the monthly ground temperature data of each observation station for M consecutive years and perform mean calculation year by year to obtain the first annual average ground temperature data of the whole area for M consecutive years;
[0194] An analysis unit 5022 is used to perform a linear regression analysis on the first local annual average temperature data and the first regional annual average ground temperature data for M consecutive years, and establish a linear regression equation between the local annual average temperature data and the regional annual average ground temperature data;
[0195] The second mean value calculation unit 5023 is used to obtain the monthly temperature data of the local station for consecutive (NM) years and perform mean value calculation year by year to obtain the second local annual mean temperature data for consecutive (NM) years, wherein the consecutive (NM) years correspond to the years with temperature data but without ground temperature data;
[0196] The ground temperature data calculation unit 5024 is used to calculate the second annual average ground temperature data of the whole area for consecutive (NM) years according to the linear regression equation and the second local annual average temperature data;
[0197] The splicing unit 5025 is used to splice the first average annual geothermal data of the whole area for M consecutive years with the second average annual geothermal data of the whole area for (NM) consecutive years to obtain the average annual geothermal sequence of the whole area for N consecutive years.
[0198] Specifically, the inspection module 506 includes:
[0199] The verification unit 5061 is used to respectively use the preset first mutation year verification algorithm and the second mutation year verification algorithm to verify the temperature data of each observation station for M consecutive years, and obtain the first predicted mutation year and the second predicted mutation year accordingly;
[0200] A first determining unit 5062 is used to determine whether the first predicted mutation year and the second predicted mutation year are the same;
[0201] A first screening unit 5063 is configured to remove the temperature data of the corresponding year if the first predicted mutation year and the second predicted mutation year are the same;
[0202] A second determining unit 5064 is configured to determine whether the first predicted mutation year or the second predicted mutation year is marked with a preset mutation label if the first predicted mutation year and the second predicted mutation year are different;
[0203] The second screening unit 5065 is used to remove the temperature data of the corresponding year if it is marked with a preset mutation label, and to ensure that the temperature data that has not been removed passes the homogeneity test and is retained.
[0204] In an embodiment of the present invention, a regression equation related to the existing continuous air temperature data and discontinuous ground temperature data is fitted to interpolate the missing ground temperature data and obtain continuous ground temperature data; then, a linear regression equation related to the continuous local annual average air temperature data and the continuous regional annual average ground temperature data is fitted to achieve the prediction and completion of ground temperature data for uncollected years; wherein, for ground temperature data and air temperature data, the abnormally collected air temperature data is eliminated through uniformity judgment, and the air temperature data and ground temperature data with higher uniformity are used for fitting subsequent regression equations and linear regression equations, and the goodness of fit is higher, and the prediction of missing and uncollected ground temperature data is more accurate.
[0205] above Figure 5 and Figure 6 The crop growth period prediction device based on ground temperature in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The crop growth period prediction device based on ground temperature in the embodiment of the present invention is described in detail from the perspective of hardware processing.
[0206] Figure 71 is a schematic diagram of the structure of a device for predicting the growth period of crops based on ground temperature provided by an embodiment of the present invention. The device for predicting the growth period of crops based on ground temperature 700 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 710 (for example, one or more processors) and a memory 720, and one or more storage media 730 (for example, one or more mass storage devices) storing application programs 733 or data 732. Among them, the memory 720 and the storage medium 730 may be temporary storage or permanent storage. The program stored in the storage medium 730 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the device for predicting the growth period of crops based on ground temperature 700. Furthermore, the processor 710 may be configured to communicate with the storage medium 730 to execute a series of instruction operations in the storage medium 730 on the device for predicting the growth period of crops based on ground temperature 700.
[0207] The crop growth period prediction device 700 based on ground temperature may also include one or more power supplies 740, one or more wired or wireless network interfaces 750, one or more input and output interfaces 760, and / or one or more operating systems 731, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. It can be understood by those skilled in the art that Figure 7 The structure of the device for predicting the growth period of crops based on ground temperature shown does not constitute a limitation of the device for predicting the growth period of crops based on ground temperature, and may include more or less components than shown in the figure, or a combination of certain components, or a different arrangement of components.
[0208] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the method for predicting the growth period of crops based on ground temperature.
[0209] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.
[0210] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.
[0211] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for predicting the growth period of crops based on ground temperature, characterized in that: The crop growth period prediction method based on ground temperature includes: Obtain N consecutive years of temperature data and M discontinuous years of ground temperature data from multiple adjacent and different regional observation stations, and perform local ground temperature data interpolation and global ground temperature data correction processing on the discontinuous M-year ground temperature data based on the M consecutive years of temperature data in the N consecutive years of temperature data, to obtain M consecutive years of ground temperature data, where N is greater than M and the same base year is used; Based on N consecutive years of air temperature data and M consecutive years of ground temperature data, an annual average ground temperature sequence for the entire region for N consecutive years is constructed, wherein the annual average ground temperature sequence is a linear regression equation obtained based on the periodic ground temperature average temperature recorded in the ground temperature data of a specified year; The annual average ground temperature sequence and the growing period nodes of crops for N consecutive years are used as training samples, and a preset machine learning algorithm is used for training to generate a crop growing period prediction model; Inputting the year to be predicted into the linear regression equation to obtain the annual average ground temperature corresponding to the year to be predicted; The annual average ground temperature is input into the crop growth period prediction model for processing, and the growth period nodes of the crops in the year to be predicted are output.
2. The method for predicting crop growth period based on ground temperature according to claim 1, characterized in that: The method of performing local geothermal data interpolation and global geothermal data correction processing on the non-continuous geothermal data of M years to obtain the geothermal data of M years in a continuous manner includes: Conduct homogeneity tests on the temperature data of each observation station for M consecutive years, and retain the temperature data that pass the homogeneity test; Using the temperature data of multiple adjacent observation stations that have passed the homogeneity check for M consecutive years, a stepwise multivariate linear regression analysis is performed on the non-continuous ground temperature data of a single observation station for M consecutive years, and a regression equation is established between the ground temperature data of a single observation station and the temperature data of multiple adjacent observation stations; Taking the M-year continuous air temperature data of the plurality of adjacent observation stations that have passed the homogeneity check as reference values, the M-year discontinuous ground temperature data of the single observation station are interpolated based on the regression equation to obtain the M-year continuous ground temperature data of each observation station.
3. The method for predicting crop growth period based on ground temperature according to claim 2, characterized in that: The method of constructing the average annual ground temperature sequence for the entire region for N consecutive years based on the air temperature data for N consecutive years and the ground temperature data for M consecutive years includes: Select one or more observation stations as local stations, obtain monthly temperature data of the local stations for M consecutive years and calculate the average value year by year to obtain the first local annual average temperature data for M consecutive years; Obtain the monthly ground temperature data of each observation station for M consecutive years and calculate the average value year by year to obtain the first annual average ground temperature data of the whole area for M consecutive years; Conduct linear regression analysis on the first local annual average temperature data and the first regional annual average ground temperature data for M consecutive years, and establish a linear regression equation between the local annual average temperature data and the regional annual average ground temperature data; Obtain the monthly temperature data of the local station for consecutive (NM) years and calculate the average year by year to obtain the second local annual average temperature data for consecutive (NM) years, wherein the consecutive (NM) years correspond to the years with temperature data but without ground temperature data; Calculate the second regional average annual ground temperature data for consecutive (NM) years according to the linear regression equation and the second local average annual temperature data; The first average annual geothermal data of the whole region for M consecutive years is spliced with the second average annual geothermal data of the whole region for (NM) consecutive years to obtain the average annual geothermal sequence of the whole region for N consecutive years.
4. The method for predicting crop growth period based on ground temperature according to claim 2, characterized in that: The temperature data of each observation station for M consecutive years are subjected to a homogeneity test, and the temperature data that pass the homogeneity test are retained, including: The preset first mutation year verification algorithm and the second mutation year verification algorithm are respectively used to verify the temperature data of each observation station for M consecutive years, and the first predicted mutation year and the second predicted mutation year are obtained accordingly; Determining whether the first predicted mutation year and the second predicted mutation year are the same; If they are the same, the temperature data of the corresponding year will be eliminated; If different, determining whether the first predicted mutation year or the second predicted mutation year is marked with a preset mutation label; If it is marked with a preset mutation label, the temperature data of the corresponding year will be eliminated, and the temperature data that have not been eliminated will be determined to pass the homogeneity test and be retained.
5. A device for predicting crop growth period based on ground temperature, characterized in that: The crop growth period prediction method and device based on ground temperature comprises: An acquisition module is used to acquire N consecutive years of temperature data and M discontinuous years of ground temperature data from multiple adjacent and different regional observation stations, and to perform local ground temperature data interpolation and global ground temperature data correction processing on the discontinuous M-year ground temperature data based on the M consecutive years of temperature data in the N consecutive years of temperature data, so as to obtain the ground temperature data for M consecutive years, wherein N is greater than M and the same base year is used; A construction module is used to construct an annual average geothermal sequence for the entire region for N consecutive years based on N consecutive years of air temperature data and M consecutive years of geothermal data, wherein the annual average geothermal sequence is a linear regression equation obtained based on the periodic geothermal average temperature recorded in the geothermal data of a specified year; A training module, for using the annual average ground temperature sequence of N consecutive years and the growing period nodes of crops as training samples, using a preset machine learning algorithm for training, and generating a crop growing period prediction model; An estimation module, used for inputting the year to be predicted into the linear regression equation to obtain the annual average ground temperature corresponding to the year to be predicted; The prediction module is used to input the annual average ground temperature into the crop growth period prediction model for processing, and output the growth period node of the crop in the year to be predicted.
6. The crop growth period prediction device based on ground temperature according to claim 5, characterized in that: The crop growth period prediction method device based on ground temperature also includes: The test module is used to perform a homogeneity test on the temperature data of each observation station for M consecutive years, and retain the temperature data that pass the homogeneity test; An analysis module is used to use the temperature data of multiple adjacent observation stations that have passed the homogeneity check for M consecutive years to perform stepwise multivariate linear regression analysis on the non-continuous ground temperature data of a single observation station for M consecutive years, and to establish a regression equation between the ground temperature data of a single observation station and the temperature data of multiple adjacent observation stations; The interpolation module is used to use the temperature data of multiple adjacent observation stations that have passed the uniformity check for M consecutive years as reference values, and based on the regression equation, interpolate the discontinuous ground temperature data of the single observation station for M consecutive years to obtain the ground temperature data of each observation station for M consecutive years.
7. The crop growth period prediction device based on ground temperature according to claim 6, characterized in that: The building blocks include: The first mean value calculation unit is used to select a single or multiple observation stations as local stations, obtain the monthly temperature data of the local station for M consecutive years and perform mean calculation year by year to obtain the first local annual average temperature data for M consecutive years; and to obtain the monthly ground temperature data of each observation station for M consecutive years and perform mean calculation year by year to obtain the first annual average ground temperature data of the whole area for M consecutive years; An analysis unit is used to perform a linear regression analysis on the first local annual average temperature data and the first regional annual average ground temperature data for M consecutive years, and establish a linear regression equation between the local annual average temperature data and the regional annual average ground temperature data; A second mean value calculation unit is used to obtain the monthly temperature data of the local station for consecutive (NM) years and perform mean calculation year by year to obtain the second local annual mean temperature data for consecutive (NM) years, wherein the consecutive (NM) years correspond to years with temperature data but without ground temperature data; A ground temperature data calculation unit, used for calculating the second regional average annual ground temperature data for consecutive (NM) years according to the linear regression equation and the second local average annual air temperature data; The splicing unit is used to splice the first average annual geothermal data of the whole area for M consecutive years with the second average annual geothermal data of the whole area for (NM) consecutive years to obtain the average annual geothermal sequence of the whole area for N consecutive years.
8. The crop growth period prediction device based on ground temperature according to claim 6, characterized in that: The inspection module comprises: A verification unit is used to respectively use a preset first mutation year verification algorithm and a preset second mutation year verification algorithm to verify the temperature data of each observation station for M consecutive years, and obtain a first predicted mutation year and a second predicted mutation year accordingly; A first judging unit, used for judging whether the first predicted mutation year and the second predicted mutation year are the same; A first screening unit, configured to remove temperature data of a corresponding year if the first predicted mutation year and the second predicted mutation year are the same; a second determination unit, configured to determine whether the first predicted mutation year or the second predicted mutation year is marked with a preset mutation label if the first predicted mutation year and the second predicted mutation year are different; The second screening unit is used to eliminate the temperature data of the corresponding year if it is marked with a preset mutation label, and to ensure that the temperature data that has not been eliminated passes the homogeneity test and is retained.
9. A device for predicting crop growth period based on ground temperature, characterized in that: The crop growth period prediction device based on ground temperature includes: a memory and at least one processor; The at least one processor calls the instructions in the memory to enable the ground temperature-based crop growth period prediction device to execute the ground temperature-based crop growth period prediction method according to any one of claims 1-4.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting the growth period of crops based on ground temperature as described in any one of claims 1 to 4 is implemented.
