A method for estimating crop yield in smart agriculture

Abstract

The present invention provides a method for estimating crop yields in smart agricultural planting, and relates to the technical field of agricultural management. The present invention comprehensively analyzes environmental data in historical planting data, conducts in-depth analysis on the trend and seasonality of these data, generates environmental data residuals and random fluctuation indexes, screens historical planting data based on the random fluctuation indexes, constructs a more reliable historical reference data set, removes historical data affected by extreme climatic conditions, constructs an environmental approximation index by comprehensively considering the trends and seasonal differences in temperature, precipitation and light intensity of each historical data in the historical reference data set, and uses the environmental approximation index to select two groups of historical planting data, uses historical data with similar environmental conditions, and combines the relationship between planting density and yield to estimate the yield of crops using the average value method.

Description

Technical Field

[0001] The present invention relates to the field of agricultural management technology, and in particular to a method for estimating crop yields in smart agriculture. Background Art

[0002] With the continuous growth of population and the increasing shortage of agricultural resources, improving crop yields and production efficiency has become an important issue in the agricultural field. Therefore, crop yield estimation is particularly critical. It is not only related to the effective utilization of agricultural resources, but also directly affects the decision-making and planning of agricultural production. Smart agriculture, as an innovative agricultural model that integrates modern information technology, aims to optimize agricultural production processes and improve the sustainability and economic benefits of agricultural production through precise data analysis and intelligent decision-making support.

[0003] Among existing smart agricultural technologies, crop yield estimation is a key link, which relies on the analysis of historical planting data and model building. These technologies usually include using sensors to collect environmental data, applying statistical models to analyze crop growth trends, and using machine learning algorithms to predict yields. Although these technologies have improved the accuracy of yield estimation to a certain extent, they often ignore the trend and seasonal changes of environmental data, as well as the comprehensive impact of these changes on crop growth.

[0004] However, existing crop yield estimation methods have obvious shortcomings. First, existing methods often rely on historical data sets for prediction, but crop yields are greatly affected by natural factors, such as non-seasonal climate change and extreme weather events. The impact of these factors on yields is very accidental. If they are not screened, the accuracy of the prediction will be affected. Second, existing methods fail to fully consider the dynamic changes in environmental and management data, especially the differences in environmental conditions and management measures between different planting cycles. Therefore, existing estimation methods lack sufficient flexibility and adaptability, resulting in unsatisfactory results in practical applications.

[0005] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0006] The purpose of the present invention is to provide a method for estimating crop yield in smart agriculture to solve the problems raised in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A method for estimating crop yields in smart agriculture, the specific steps comprising:

[0009] Step 1: For the crop type whose yield is to be estimated, obtain the same type of crops Historical planting data of a historical planting cycle, wherein the historical planting data includes monthly environmental data, crop management data and final yield data during the crop growth cycle;

[0010] Step 2: Based on the historical planting data of each historical planting cycle, the environmental data between two adjacent months in the planting cycle are subjected to trend and seasonal analysis, and the residual of the environmental data between the two months is generated according to the results of the trend and seasonal analysis. Based on the residual data of each historical planting cycle, the stochastic fluctuation index of each historical planting cycle is generated;

[0011] Step 3: Based on the Stochastic Volatility Index Filter the historical planting data, remove the historical planting data whose random fluctuation index is greater than the fluctuation threshold, and use the filtered historical planting data to form a historical reference data set;

[0012] Step 4: Collect the monthly environmental data and crop management data of the crops whose yield is to be estimated up to the current growth cycle to form an evaluation data set, traverse the trend and seasonal analysis results of each historical planting data in the historical reference data set, and combine them with the environmental data in the evaluation data set to generate an environmental approximation index;

[0013] Step 5: Based on the environmental approximation index, two sets of historical planting data are selected, and the yield of the crops to be estimated is predicted based on the management data and final yield data of the crops in the selected historical planting data, combined with the management data of the crops in the evaluation data set.

[0014] Furthermore, the environmental data includes temperature, precipitation and light intensity data of the crop planting environment. The logic for performing trend and seasonal analysis based on the environmental data between two adjacent months is:

[0015] Number each month of the crop growth cycle in chronological order, and collect the average temperature, average precipitation, and average light intensity of each month in the crop growth cycle;

[0016] Calculate the trend index and seasonal index of temperature, precipitation and light intensity data between two adjacent months based on the average temperature, average precipitation and average light intensity of each month;

[0017] Based on the trend index and seasonal index, the residuals of temperature, precipitation and light intensity between two adjacent months are calculated.

[0018] Furthermore, the formula for calculating the trend index of temperature, precipitation and light intensity between two adjacent months is: in, Respectively The historical planting data Month and Trend index of temperature, precipitation and light intensity between months, Respectively represent The historical planting data The average temperature, average precipitation and average light intensity for each month. Respectively represent The historical planting data Average temperature, average precipitation and average light intensity for each month;

[0019] The formula for calculating the seasonal index of temperature, precipitation and light intensity between two consecutive months is: in, Respectively represent The historical planting data Month and Seasonality indices of temperature, precipitation, and light intensity between months;

[0020] The formulas for calculating the residuals of temperature, precipitation and light intensity between two consecutive months are: in, Respectively represent The historical planting data Month and The residuals of temperature, precipitation and light intensity between months, Indicates the number of historical planting data. is a positive integer, , Indicates the number of the month in the planting cycle.

[0021] Furthermore, the logic for generating the stochastic volatility index for each historical planting cycle is:

[0022] According to the residuals of temperature, precipitation and light intensity between two adjacent months in each historical planting data, the mean and variance of temperature residuals, the mean and variance of precipitation residuals, and the mean and variance of light intensity residuals of each historical planting cycle are generated respectively. The specific formula is as follows: in, Respectively represent The mean and variance of the temperature residuals for the historical planting data, Indicates The number of months in the crop planting cycle in the historical planting data;

[0023] in, Respectively represent The mean and variance of the precipitation residuals for the historical planting data, Respectively represent The mean and variance of the light intensity residuals for historical planting data.

[0024] Furthermore, the formula for generating the stochastic volatility index for each historical planting cycle is: in, Indicates Stochastic volatility index of historical planting data, denote the first weight coefficient and the second weight coefficient respectively, where ,and ;

[0025] According to the stochastic volatility index The method for filtering historical planting data is:

[0026] like , then the The historical planting data are filtered into the historical reference data set. , then the The historical planting data were eliminated.

[0027] Furthermore, the method for generating the environmental similarity index between each historical planting data in the evaluation dataset and the historical reference dataset is as follows:

[0028] Based on the environmental data in the evaluation data set, the trend index and seasonal index of temperature, precipitation and light intensity between two adjacent months in the current growth cycle of the crop whose yield is to be estimated are generated respectively;

[0029] The trend index and seasonal index of temperature, precipitation and light intensity between two adjacent months corresponding to the evaluation data set are matched with the trend index and seasonal index of temperature, precipitation and light intensity between two adjacent months corresponding to each historical planting data in the historical reference data set in chronological order, and the months are matched one by one to generate the environmental approximation index between the two.

[0030] The logic behind generating the environmental proximity index is: in, Represents the evaluation dataset and the historical reference dataset. The trend similarity index, seasonal similarity index and environmental similarity index between historical planting data are Indicates the number of the growing months of the crop whose yield is to be estimated as of now. That is, the crop whose yield is to be estimated is calculated from the planting month, and the current time is Month, They represent the evaluation data set Month and Trend index of temperature, precipitation and light intensity between months, They represent the evaluation data set Month and Seasonality indices of temperature, precipitation, and light intensity between months;

[0031] Respectively represent the The historical planting data Month and Trend index of temperature, precipitation and light intensity between months, Respectively represent the The historical planting data Month and Seasonality index of temperature, precipitation and light intensity between months, Indicates the number of the historical planting data in the historical reference data set. Indicates the number of the month within the time period up to the current growth cycle of the crop for which the estimated yield is generated.

[0032] Furthermore, the logic for generating the trend index of temperature, precipitation and light intensity between two adjacent months in the evaluation data set is as follows: The logic behind generating the seasonal index of temperature, precipitation, and light intensity between two adjacent months in the evaluation dataset is as follows: in, They are respectively The average temperature, average precipitation and average light intensity for each month. Respectively represent The average temperature, average precipitation and average sunshine intensity for each month.

[0033] Furthermore, when selecting two sets of historical planting data according to the environmental similarity index, two historical planting data with the smallest environmental similarity index between the historical reference data set and the evaluation data set are selected, and the management data of the crops corresponding to the two historical planting data are obtained, and the management data of the crops include planting density, and the generation logic of the planting density is: in, represents the planting density, Indicates the weight of the seeds at the time of sowing. The area of ​​farmland in the table.

[0034] Furthermore, the method for predicting the yield of the crop to be estimated is: in, It indicates the estimated yield of the crop to be estimated. represents the planting density of crops to be estimated in the evaluation data set. They represent the final yield data of the two historical planting data with the smallest environmental approximation index. and They respectively represent the planting density of crops in the two historical planting data with the smallest environmental approximation index.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] The present invention comprehensively analyzes the environmental data in the historical planting data, conducts in-depth analysis of the trend and seasonality of these data, generates environmental data residuals and random fluctuation indexes, screens historical planting data based on the random fluctuation indexes, constructs a more reliable historical reference data set, removes historical data affected by extreme climate conditions, and thus improves the representativeness of the historical reference data set and the accuracy of the estimate;

[0037] The present invention constructs an environmental approximation index by comprehensively considering the trends and seasonal differences of temperature, precipitation and light intensity of each historical data in the historical reference data set, and uses the environmental approximation index to select two groups of historical planting data. The smallest environmental approximation index means that the environmental conditions of the two historical data are closest to the environmental conditions of the crops to be estimated. The historical data with similar environmental conditions are combined with the relationship between planting density and yield, and the average value method is used to estimate the yield of the crops. The present invention integrates the historical reference data set and real-time data, makes an estimate based on the historical data most relevant to the environmental conditions, and combines it with management data, thereby effectively improving the accuracy of the estimate. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a schematic diagram of the overall method flow of the present invention. DETAILED DESCRIPTION

[0039] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0040] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0041] Example

[0042] See also Figure 1 , the present invention provides a technical solution:

[0043] A method for estimating crop yields in smart agriculture, the specific steps comprising:

[0044] Step 1: For the crop type whose yield is to be estimated, obtain the same type of crops The historical planting data includes monthly environmental data, crop management data and final yield data during the crop growth cycle.

[0045] Crops of the same type refer to crops that belong to the same species in biological classification and have the same planting and harvesting cycle. For example, if we want to estimate wheat yield, then "crops of the same type" refers to all wheat crops of the same variety.

[0046] The monthly data within the growth cycle means that the month of crop planting is the first month and the month of harvest is the last month. For example, if we want to estimate the yield of corn, the corn is planted on May 12 and harvested on October 23, then May is the first month, June is the second month, and so on, October is the sixth month. The average temperature, average precipitation and average light intensity of each month are only calculated for that month. For example, the average temperature in May is the average temperature from May 12 to May 31, and the average rainfall in October is the average rainfall from October 1 to October 23.

[0047] In this embodiment, the environmental data includes temperature, precipitation and light intensity data of the crop planting environment. The logic for performing trend and seasonal analysis based on the environmental data between two adjacent months is:

[0048] Number each month of the crop growth cycle in chronological order, and collect the average temperature, average precipitation, and average light intensity of each month in the crop growth cycle;

[0049] Calculate the trend index and seasonal index of temperature, precipitation and light intensity data between two adjacent months based on the average temperature, average precipitation and average light intensity of each month;

[0050] Based on the trend index and seasonal index, the residuals of temperature, precipitation and light intensity between two adjacent months are calculated.

[0051] Collect daily temperature, precipitation, and light intensity data from weather stations, agricultural monitoring systems, or online meteorological databases. For temperature, record the maximum and minimum temperatures of each day of each month, and then calculate the average maximum and average minimum temperatures of each month. Calculate the average of the average maximum and average minimum temperatures of each month to obtain the monthly average temperature. For precipitation, add up the daily precipitation data of each month to obtain the total precipitation of the month. For light intensity, record the light intensity of each hour to calculate the total light intensity of each day. Add the total light intensity of each day of each month to obtain the total light intensity of the month. Then, divide the total light intensity of the month by the number of days in the month to obtain the monthly average light intensity.

[0052] By decomposing the trend and seasonal components of environmental data, we can more clearly understand the long-term trends and cyclical fluctuations of environmental changes. Calculating residuals helps identify outliers in environmental data, which may be caused by non-seasonal climate changes, extreme weather events or other external factors. It can also determine whether there are abnormal climate and weather changes in historical data.

[0053] Step 2: Based on the historical planting data of each historical planting cycle, trend and seasonal analysis is performed on the environmental data between two adjacent months in the planting cycle. Based on the results of trend and seasonal analysis, the residual of the environmental data between the two months is generated. Based on the residual data of each historical planting cycle, the stochastic fluctuation index of each historical planting cycle is generated.

[0054] In this embodiment, the formula for calculating the trend index of temperature, precipitation and light intensity between two adjacent months is: in, Respectively The historical planting data Month and Trend index of temperature, precipitation and light intensity between months, Respectively represent The historical planting data The average temperature, average precipitation and average light intensity for each month. Respectively represent The historical planting data The average temperature, average precipitation and average sunshine intensity for each month.

[0055] The formula for calculating the seasonal index of temperature, precipitation and light intensity between two consecutive months is: in, Respectively represent The historical planting data Month and Seasonality index of temperature, precipitation, and light intensity between months.

[0056] The trend index and seasonal index are two important indicators used to analyze the changes of environmental data over time. They reflect the long-term trend and fluctuation of environmental conditions respectively. The trend index is obtained by calculating the average value of two adjacent months. It reflects the average change trend of environmental conditions such as temperature, precipitation and light intensity over a period of time, and can effectively capture the changing trend of environmental conditions. The seasonal index is calculated by comparing the actual environmental data of each month with the trend index. It reflects the fluctuation of environmental conditions relative to the trend, which helps to identify the seasonal change pattern of environmental conditions.

[0057] This embodiment not only considers the changing trend of environmental data, but also seasonal fluctuations, thereby providing more comprehensive information on environmental changes. In addition, this calculation method can also identify outliers in environmental data, which may be caused by non-seasonal climate change, extreme weather events or other external factors. Timely detection of these anomalies can help growers take preventive measures and reduce losses.

[0058] The formulas for calculating the residuals of temperature, precipitation and light intensity between two consecutive months are: in, Respectively represent The historical planting data Month and The residuals of temperature, precipitation and light intensity between months, Indicates the number of historical planting data. is a positive integer, , Indicates the number of the month in the planting cycle.

[0059] In this embodiment, the residual is used to evaluate the accuracy of the trend and seasonal analysis of the environmental data. The residual reflects the difference between the actually observed environmental data and the trend index and the seasonal index. If the residual is close to zero, it means that the data changes more regularly. If the residual is large, it means that non-seasonal climate change, extreme weather events or other external factors may have occurred, because under normal planting conditions, temperature, precipitation and light intensity have a certain trend of change. For example, from winter to summer, the temperature will gradually increase without sudden changes. If the residual is large, it means that it does not conform to the normal trend. At this time, non-seasonal climate change, extreme weather events or other external factors may have occurred. The impact of non-seasonal climate change, extreme weather events or other external factors on crop yield is negative and exceeds the suitable range of crop growth and development. For example, non-seasonal climate change may lead to the interruption of the crop growth cycle. Extreme weather events, such as heat waves, droughts, floods or hail, will bring physiological stress to crops. Non-seasonal climate change and extreme weather events may change the occurrence of pests and diseases, leading to the outbreak of pests and diseases.

[0060] Non-seasonal climate change and extreme weather events increase the uncertainty of environmental conditions. These events are often difficult to predict, and their timing, intensity and scope of impact may exceed conventional climate patterns and historical data, making it difficult for forecasting models based on historical data to accurately predict future production. Therefore, non-seasonal climate change, extreme weather events or other external factors will cause inaccurate production forecasts.

[0061] Furthermore, the logic for generating the stochastic volatility index for each historical planting cycle is:

[0062] According to the residuals of temperature, precipitation and light intensity between two adjacent months in each historical planting data, the mean and variance of temperature residuals, the mean and variance of precipitation residuals, and the mean and variance of light intensity residuals of each historical planting cycle are generated respectively. The specific formula is as follows: in, Respectively represent The mean and variance of the temperature residuals for the historical planting data, Indicates The number of months in the crop planting cycle in the historical planting data;

[0063] in, Respectively represent The mean and variance of the precipitation residuals for the historical planting data, Respectively represent The mean and variance of the light intensity residuals for historical planting data.

[0064] In this embodiment, the mean and variance of the residuals respectively reflect the fluctuation of the temperature, precipitation and light intensity relative to their average levels in the historical planting cycle. The mean of the residuals reflects the average deviation of the temperature, precipitation and light intensity in the historical planting cycle. If the mean of the residuals is close to zero, it means that these climate factors are generally close to their average levels in the historical planting cycle; if the mean of the residuals is large or small, it means that these climate factors generally deviate from their average levels in the historical planting cycle. The variance of the residuals reflects the degree of fluctuation of the temperature, precipitation and light intensity from their average levels in the historical planting cycle. The larger the variance, the greater the volatility of these climate factors in the historical planting cycle, that is, the larger the variation range. The smaller the variance, the smaller the volatility of these climate factors in the historical planting cycle, that is, the smaller the variation range.

[0065] The purpose of calculating the mean and variance of the residuals is to quantify the fluctuations of climate factors in historical planting cycles, so as to provide a reference for yield prediction. The mean and variance of the residuals can be used to evaluate the stability of climate factors in historical planting cycles. Stable climate conditions are usually conducive to crop growth and yield prediction.

[0066] In agricultural production, the volatility of climate factors has an important impact on crop yields. Large climate fluctuations may lead to instability in the crop growth environment, increasing the risk and uncertainty of agricultural production. Therefore, using the stochastic volatility index to assess the volatility of climate factors in historical planting cycles can help agricultural producers better understand and respond to the impact of climate change on crop yields, thereby taking corresponding management measures to reduce risks and increase crop yields.

[0067] If the mean of the residuals is small and the variance is small, this indicates that the changing trends of temperature, precipitation and light intensity in the historical planting cycle are relatively stable, and the prediction of crop yield is more accurate because the crop growth conditions are more consistent. If the mean of the residuals is large or the variance is large, this indicates that the climate conditions in the historical planting cycle fluctuated greatly. This instability will lead to increased uncertainty in crop yield predictions because crop growth may be affected by extreme climate events.

[0068] Furthermore, the formula for generating the stochastic volatility index for each historical planting cycle is: in, Indicates Stochastic volatility index of historical planting data, denote the first weight coefficient and the second weight coefficient respectively, where ,and .

[0069] The stochastic fluctuation index reflects the fluctuations in temperature, precipitation and light intensity during the historical planting cycle. The larger the stochastic fluctuation index, the more abnormal the environmental data and climate conditions during the planting process of the historical crops. The stochastic fluctuation index comprehensively considers the residual mean and variance of multiple temperatures, precipitation and light intensity, and can more comprehensively evaluate the climate fluctuations. The stochastic fluctuation index considers not only the residual mean of temperature, precipitation and light intensity, but also their variance. This allows the index to comprehensively reflect the fluctuations of these climate factors, rather than just the influence of a single factor.

[0070] The reason why this implementation uses a logarithmic function to fit the temperature and light intensity is that the logarithmic function can amplify smaller values ​​and reduce larger values. The impact of temperature and light intensity fluctuations on crop yield presents a nonlinear relationship. The logarithmic function can better capture this nonlinear relationship, thereby more accurately reflecting the impact of these factors on crop yield, because when the fluctuations in temperature and light intensity are small, they have a significant impact on crop yield. As the temperature and light intensity gradually increase, this impact will gradually decrease.

[0071] The reason why this implementation uses an exponential function to fit rainfall is that the logarithmic function can amplify larger values ​​and shrink smaller values. In agricultural production, precipitation fluctuations may more often manifest as extreme events (such as floods or droughts). The impact of these extreme events on crop yields is often sudden and significant. The exponential function can better capture this sudden and significant impact, thereby more accurately reflecting the impact of precipitation fluctuations on crop yields.

[0072] In the stochastic volatility index, the application of logarithmic functions and exponential functions is based on their respective characteristics and applicable scenarios. The logarithmic function is suitable for processing temperature and light intensity data with nonlinear relationships, and can amplify the impact of smaller fluctuations; while the exponential function is suitable for processing precipitation data with extreme values, and can amplify the impact of extreme events. Through this design, the stochastic volatility index can more comprehensively and accurately reflect the impact of different climate factors on crop yields. The larger the value of the stochastic volatility index, the more abnormal the climate conditions in the historical planting cycle. This anomaly may be caused by extreme climate events (such as extreme high temperature, extreme precipitation, or extreme light), which may have an adverse effect on crop growth.

[0073] Temperature and light intensity are the basic physiological requirements for crop growth, and have a direct impact on crop photosynthesis, respiration, and growth and development. Therefore, fluctuations in temperature and light intensity have a more significant impact on crop yield, while fluctuations in precipitation have a relatively small impact, so the first weight coefficient is greater than the second weight coefficient.

[0074] When the stochastic volatility index is large, it means that the climate factors in the historical planting cycle fluctuated greatly, which may include more extreme climate events or abnormal climate conditions. These extreme or abnormal climate conditions may have a greater impact on crop yields, resulting in larger fluctuations in crop yields. Therefore, the larger the stochastic volatility index is, the higher the uncertainty and risk of climate factors in the historical planting cycle.

[0075] Step 3: Based on the Stochastic Volatility Index The historical planting data are screened, and the historical planting data whose random fluctuation index is greater than the fluctuation threshold are removed, and the screened historical planting data constitute a historical reference data set.

[0076] In this embodiment, according to the random fluctuation index The method for filtering historical planting data is:

[0077] like , then the The historical planting data are filtered into the historical reference data set. , then the The historical planting data were eliminated.

[0078] In this embodiment, historical planting data are screened based on the random fluctuation index to construct a more stable and reliable historical reference data set. Specifically, historical planting data with a random fluctuation index less than or equal to the threshold are retained, while historical planting data with a random fluctuation index greater than the threshold are eliminated. The purpose of this is to remove historical data affected by extreme climatic conditions, thereby improving the representativeness of the historical reference data set and the accuracy of the prediction model. This embodiment improves the quality of the data set and the readiness for subsequent predictions through sophisticated data screening.

[0079] Step 4: Collect monthly environmental data and crop management data of the crops whose yield is to be estimated up to the current growth cycle to form an evaluation data set, traverse the trend and seasonal analysis results of each historical planting data in the historical reference data set, and combine them with the environmental data in the evaluation data set to generate an environmental approximation index.

[0080] In this embodiment, the method for generating the environmental similarity index between the evaluation data set and each historical planting data in the historical reference data set is:

[0081] Based on the environmental data in the evaluation data set, the trend index and seasonal index of temperature, precipitation and light intensity between two adjacent months in the current growth cycle of the crop whose yield is to be estimated are generated respectively;

[0082] The trend index and seasonal index of temperature, precipitation and light intensity between two adjacent months corresponding to the evaluation data set are matched with the trend index and seasonal index of temperature, precipitation and light intensity between two adjacent months corresponding to each historical planting data in the historical reference data set in chronological order, and the months are matched one by one to generate the environmental approximation index between the two.

[0083] The logic behind generating the environmental proximity index is: in, Represents the evaluation dataset and the historical reference dataset. The trend similarity index, seasonal similarity index and environmental similarity index between historical planting data are Indicates the number of the growing months of the crop whose yield is to be estimated as of now. That is, the crop whose yield is to be estimated is calculated from the planting month, and the current time is Month, They represent the evaluation data set Month and Trend index of temperature, precipitation and light intensity between months, They represent the evaluation data set Month and Seasonality indices of temperature, precipitation, and light intensity between months;

[0084] Respectively represent the The historical planting data Month and Trend index of temperature, precipitation and light intensity between months, Respectively represent the The historical planting data Month and Seasonality index of temperature, precipitation and light intensity between months, Indicates the number of the historical planting data in the historical reference data set. Indicates the number of the month within the time period up to the current growth cycle of the crop for which the estimated yield is generated.

[0085] The environmental similarity index reflects the degree of similarity of environmental conditions between the evaluation dataset and the historical planting data in the historical reference dataset. The trend similarity index and seasonal similarity index are obtained by calculating the sum of the square differences of the trend index and the seasonal index between adjacent months in the evaluation dataset and the historical reference dataset, and then taking the average value. The environmental similarity index is the sum of the square roots of the trend similarity index and the seasonal similarity index. The environmental similarity index comprehensively considers the trend and seasonal differences in temperature, precipitation and light intensity, thereby reflecting the degree of similarity of environmental conditions between the evaluation dataset and the historical planting data in the historical reference dataset.

[0086] By comprehensively considering the trends and seasonal differences in temperature, precipitation and light intensity, the environmental similarity index can comprehensively reflect the similarity of environmental conditions between the evaluation dataset and the historical reference dataset. By calculating the sum of the square differences between the trend index and the seasonal index, the environmental similarity index can quantify the differences in environmental conditions between the evaluation dataset and the historical reference dataset, thereby providing an objective similarity measure for subsequent yield estimation.

[0087] The larger the environmental similarity index is, the greater the difference in environmental conditions between the evaluation dataset and the historical planting data in the historical reference dataset. When the environmental similarity index is large, it means that there are large differences in environmental conditions between the evaluation dataset and the historical reference dataset. Specifically, the trends and seasonal patterns of temperature, precipitation, and light intensity are significantly different between the evaluation dataset and the historical reference dataset, reflecting that the current environmental conditions deviate significantly from the environmental conditions in the historical reference dataset.

[0088] The larger the environmental approximation index is, the greater the difference between the current environmental conditions and the environmental conditions in the historical reference data set. This difference may reduce the applicability of historical data in estimating crop yields, because crop yields are greatly affected by environmental conditions. If the current environmental conditions are significantly different from those in the historical data, then the yield may not be accurately predicted based on historical data.

[0089] Furthermore, the logic for generating the trend index of temperature, precipitation and light intensity between two adjacent months in the evaluation data set is as follows: The logic behind generating the seasonal index of temperature, precipitation, and light intensity between two adjacent months in the evaluation dataset is as follows: in, They are respectively The average temperature, average precipitation and average light intensity for each month. Respectively represent The average temperature, average precipitation and average sunshine intensity for each month.

[0090] The generation method of the trend index and seasonal index of temperature, precipitation and light intensity between two adjacent months in the evaluation dataset is the same as that in the historical data. This is mainly to ensure that when calculating the environmental approximation index, the comparison between the evaluation dataset and the historical dataset is based on consistent standards and logic. Using the same generation method can ensure that the trend index and seasonal index of the evaluation dataset and the historical dataset are calculated under the same logic and standards. In this way, the comparison between the two is fair and meaningful, avoiding comparison deviations caused by different calculation methods. The consistent generation method makes the trend index and seasonal index of the evaluation dataset and the historical dataset comparable. In this way, by comparing the trend index and seasonal index of the two, the similarity between the current environmental conditions and the historical environmental conditions can be more accurately evaluated.

[0091] Step 5: Based on the environmental approximation index, two sets of historical planting data are selected, and the yield of the crops to be estimated is predicted based on the management data and final yield data of the crops in the selected historical planting data, combined with the management data of the crops in the evaluation data set.

[0092] When selecting two sets of historical planting data according to the environmental similarity index, two historical planting data with the smallest environmental similarity index between the historical reference data set and the evaluation data set are selected, and the management data of the crops corresponding to the two historical planting data are obtained.

[0093] The historical planting data with the smallest environmental similarity index means that the historical environmental conditions corresponding to these data are closest to the environmental conditions in the current evaluation data set. Selecting such historical data can ensure that the environmental conditions of the selected data have a high similarity with the current environmental conditions. Crop growth and yield are affected by environmental factors. When environmental factors remain stable or similar within a certain range, the conditions for crop growth are also relatively stable, resulting in relatively stable or similar yields. The genetic and physiological characteristics of crops determine how they respond to environmental conditions. Under similar environmental conditions, crops of the same type will exhibit similar growth patterns and yield levels.

[0094] In this embodiment, the management data of the crops includes planting density, and the generation logic of the planting density is: in, represents the planting density, Indicates the weight of the seeds at the time of sowing. The area of ​​farmland in the table.

[0095] In this embodiment, the planting density is defined as the number of seeds per unit area. By dividing the total seed weight by the farmland area, the seed weight per unit area can be obtained, thereby indirectly reflecting the planting density. The crop types applicable to this embodiment are crops with relatively uniform seed size and weight and uniform sowing, such as wheat, corn, etc.

[0096] Furthermore, the method for predicting the yield of the crop to be estimated is: in, It indicates the estimated yield of the crop to be estimated. represents the planting density of crops to be estimated in the evaluation data set. They represent the final yield data of the two historical planting data with the smallest environmental approximation index. and They respectively represent the planting density of crops in the two historical planting data with the smallest environmental approximation index.

[0097] When predicting the yield of the crops to be estimated, the historical planting data with the smallest two environmental approximation indexes and their corresponding planting densities are used. The smallest environmental approximation index means that the environmental conditions of the two historical data are closest to the environmental conditions of the crops to be estimated. By referring to these historical data, the yield of the crops to be estimated under similar environmental conditions can be better estimated.

[0098] As the planting density increases, the number of individuals per unit area increases, and resource competition intensifies. When resources are sufficient and reasonably distributed, each individual can still obtain sufficient resources, and the yield increases linearly with the increase in planting density. Therefore, there is a certain linear relationship between planting density and yield. By dividing the yield of historical data by its corresponding planting density, the yield under unit planting density can be obtained. Then, take the average of the yields under these two unit planting densities and multiply it by the planting density of the crop to be estimated, so as to obtain the estimated yield of the crop to be estimated. By calculating the average of the yields under the unit planting density of the two historical data, the influence of the two historical data can be balanced to ensure that the prediction results are more robust. In this way, the impact of possible deviations or outliers in a single historical data on the prediction results can be reduced.

[0099] This embodiment uses historical data with similar environmental conditions, combines the relationship between planting density and yield, and adopts the average value method to predict the yield of the crop to be estimated. Based on the relationship between planting density and yield of historical data and the application of the average value, a relatively accurate and robust yield prediction method is provided.

[0100] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0101] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0102] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0103] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for estimating crop yields in smart agriculture, characterized in that: The specific steps include: Step 1: For the crop type whose yield is to be estimated, obtain the same type of crops Historical planting data of a historical planting cycle, wherein the historical planting data includes monthly environmental data, crop management data and final yield data during the crop growth cycle; Step 2: Based on the historical planting data of each historical planting cycle, the environmental data between two adjacent months in the planting cycle are subjected to trend and seasonal analysis, and the residual of the environmental data between the two months is generated according to the results of the trend and seasonal analysis. Based on the residual data of each historical planting cycle, the stochastic fluctuation index of each historical planting cycle is generated; Step 3: Based on the Stochastic Volatility Index Filter the historical planting data, remove the historical planting data whose random fluctuation index is greater than the fluctuation threshold, and use the filtered historical planting data to form a historical reference data set; Step 4: Collect the monthly environmental data and crop management data of the crops whose yield is to be estimated up to the current growth cycle to form an evaluation data set, traverse the trend and seasonal analysis results of each historical planting data in the historical reference data set, and combine them with the environmental data in the evaluation data set to generate an environmental approximation index; Step 5: Select two sets of historical planting data based on the environmental approximation index, and predict the yield of the crop to be estimated based on the management data and final yield data of the crops in the selected historical planting data and the management data of the crops in the evaluation data set; The environmental data include the temperature, precipitation and light intensity data of the crop planting environment. The logic for trend and seasonal analysis based on the environmental data between two adjacent months is as follows: Number each month of the crop growth cycle in chronological order, and collect the average temperature, average precipitation, and average light intensity of each month in the crop growth cycle; Calculate the trend index and seasonal index of temperature, precipitation and light intensity data between two adjacent months based on the average temperature, average precipitation and average light intensity of each month; Based on the trend index and seasonal index, the residuals of temperature, precipitation and light intensity between two adjacent months are calculated; The formula for calculating the trend index of temperature, precipitation and light intensity between two consecutive months is: in, Respectively The historical planting data Month and Trend index of temperature, precipitation and light intensity between months, Respectively represent The historical planting data The average temperature, average precipitation and average light intensity for each month. Respectively represent The historical planting data Average temperature, average precipitation and average light intensity for each month; The formula for calculating the seasonal index of temperature, precipitation and light intensity between two consecutive months is: in, Respectively represent The historical planting data Month and Seasonality indices of temperature, precipitation, and light intensity between months; The formulas for calculating the residuals of temperature, precipitation and light intensity between two consecutive months are: in, Respectively represent The historical planting data Month and The residuals of temperature, precipitation and light intensity between months, Indicates the number of historical planting data. is a positive integer, , Indicates the number of the month in the planting cycle; The logic behind generating the stochastic volatility index for each historical planting cycle is: According to the residuals of temperature, precipitation and light intensity between two adjacent months in each historical planting data, the mean and variance of temperature residuals, the mean and variance of precipitation residuals, and the mean and variance of light intensity residuals of each historical planting cycle are generated respectively. The specific formula is as follows: in, Respectively represent The mean and variance of the temperature residuals for the historical planting data, Indicates The number of months in the crop planting cycle in the historical planting data; in, Respectively represent The mean and variance of the precipitation residuals for the historical planting data, Respectively represent The mean and variance of the light intensity residuals of the historical planting data; The formula used to generate the Stochastic Volatility Index for each historical planting cycle is: in, Indicates Stochastic volatility index of historical planting data, denote the first weight coefficient and the second weight coefficient respectively, where ,and ; According to the stochastic volatility index The method for filtering historical planting data is: like , then the The historical planting data are filtered into the historical reference data set. , then the The historical planting data were eliminated; The method for generating the environmental similarity index between each historical planting data in the evaluation dataset and the historical reference dataset is: Based on the environmental data in the evaluation data set, the trend index and seasonal index of temperature, precipitation and light intensity between two adjacent months in the current growth cycle of the crop whose yield is to be estimated are generated respectively; The trend index and seasonal index of temperature, precipitation and light intensity between two adjacent months corresponding to the evaluation data set are matched with the trend index and seasonal index of temperature, precipitation and light intensity between two adjacent months corresponding to each historical planting data in the historical reference data set in chronological order, and the months are matched one by one to generate the environmental approximation index between the two. The logic behind generating the environmental proximity index is: in, Represents the evaluation dataset and the historical reference dataset. The trend similarity index, seasonal similarity index and environmental similarity index between historical planting data are Indicates the number of the growing months of the crop whose yield is to be estimated as of now. That is, the crop whose yield is to be estimated is calculated from the planting month, and the current time is Month, They represent the evaluation data set Month and Trend index of temperature, precipitation and light intensity between months, They represent the evaluation data set Month and Seasonality indices of temperature, precipitation, and light intensity between months; Respectively represent the The historical planting data Month and Trend index of temperature, precipitation and light intensity between months, Respectively represent the The historical planting data Month and Seasonality index of temperature, precipitation and light intensity between months, Indicates the number of the historical planting data in the historical reference data set. Indicates the number of the month within the time period up to the current growth cycle of the crop for which the estimated yield is generated; The logic behind generating the trend index of temperature, precipitation, and light intensity between two adjacent months in the evaluation dataset is as follows: The logic behind generating the seasonal index of temperature, precipitation, and light intensity between two adjacent months in the evaluation dataset is as follows: in, They are respectively The average temperature, average precipitation and average light intensity for each month. Respectively represent The average temperature, average precipitation and average sunshine intensity for each month.

2. The method for estimating crop yields in smart agriculture according to claim 1, characterized in that: When selecting two sets of historical planting data according to the environmental similarity index, two historical planting data with the smallest environmental similarity index between the historical reference data set and the evaluation data set are selected, and the management data of the crops corresponding to the two historical planting data are obtained. The management data of the crops includes planting density, and the generation logic of the planting density is: in, represents the planting density, Indicates the weight of the seeds at the time of sowing. The area of ​​farmland in the table.

3. The method for estimating crop yields in smart agriculture according to claim 2, characterized in that: The method for predicting the yield of the crop to be estimated is: in, It indicates the estimated yield of the crop to be estimated. represents the planting density of crops to be estimated in the evaluation data set. They represent the final yield data of the two historical planting data with the smallest environmental approximation index. and They respectively represent the planting density of crops in the two historical planting data with the smallest environmental approximation index.

Images