Corn yield prediction and planting income evaluation system based on big data

Abstract

The application discloses a corn yield prediction and planting income evaluation system based on big data and belongs to the technical field of agricultural big data analysis. The system comprises the following steps: obtaining historical meteorological sequences and dividing the sequences into stage sub-sequences according to growth stages, and mapping soil attribute sequences into soil attribute vectors; constructing a spatial grid according to the stage sub-sequences and the soil attribute vectors, dividing a planting area into multiple evaluation units and associating corresponding segments; calculating yield contribution coefficients by using the stage sub-sequences and the soil attribute vectors, so as to represent the contribution degree to the final yield in the growth stage; performing yield grade sorting according to the contribution coefficients and the soil attributes to generate a yield prediction interval; obtaining production cost and market price sequences, aligning the prediction interval with the cost and price time periods to generate an income evaluation result. The application realizes comprehensive evaluation of corn yield prediction and planting income.

Description

Technical Field

[0001] This invention relates to the field of agricultural big data analysis technology, specifically a big data-based corn yield prediction and planting income assessment system. Background Technology

[0002] Current maize yield forecasting and planting benefit assessments typically focus on the entire planting area or a large-scale grid, using regional average meteorological observations, soil sampling averages, historical yield statistical models, or remote sensing vegetation indices to estimate yield. This is then combined with average production costs and market prices to calculate planting benefits. Meteorological and soil property sequences are often processed using spatial interpolation or regional averages, homogenizing the periodic differences in meteorological conditions and soil properties across different spatial locations within the planting area. The predicted results are often single yield values ​​or wide yield ranges, failing to represent the differences in yield contribution from each spatial unit within the planting area. The combined effects of soil organic matter, pH, moisture content, and precipitation and sunshine during the grain-filling period at the spatial grid scale are not quantified on a unit-by-unit basis. The determination of the upper and lower limits of the yield prediction range lacks adjustment for soil moisture differences between high-contribution and low-contribution units, leading to inconsistencies between the predicted range and the actual spatial distribution of yield in the planting area.

[0003] In the benefit assessment stage, the alignment of yield forecasts with production cost and market price sequences is rather coarse. In particular, the matching between the current harvest price and cost inputs and the yield forecast range is insufficient, making it difficult to form a benefit assessment range that reflects the differences within the planting area. Reflecting spatial differences within the planting area requires constructing a spatial grid within the area, associating meteorological subsequences and soil attribute vectors at different growth stages with each assessment unit, calculating the yield contribution coefficient for each unit, and determining the upper and lower limits of the yield forecast range based on the yield contribution ranking and the difference in soil moisture content between high and low contribution units. This range is then aligned with the current planting season's cost price sequence to generate the planting benefit assessment result. Summary of the Invention

[0004] This invention provides a big data-based corn yield prediction and planting income assessment system. It is used to divide the planting area into spatial assessment units, associate the stage meteorological subsequence and soil attribute vector with each assessment unit, quantify the yield contribution of each unit, adjust the upper and lower limits of the yield prediction interval by ranking the yield contribution and the difference in soil moisture content, and generate planting income assessment results by aligning the current planting season production cost sequence and market price sequence. This enables the yield prediction and income assessment to reflect the spatial differences and time-series cost and price changes within the planting area.

[0005] To achieve the above objectives, this invention provides the following technical solution: This invention provides a big data-based maize yield prediction and planting benefit assessment system, comprising a data acquisition module, a grid division module, a contribution calculation module, a ranking module, and a benefit assessment module. The data acquisition module acquires historical meteorological sequences and soil attribute sequences of the planting area, divides the historical meteorological sequences into stage subsequences according to growth stages, and maps the soil attribute sequences into soil attribute vectors. The grid division module constructs a spatial grid based on the stage subsequences and soil attribute vectors, dividing the planting area into multiple assessment units. Each assessment unit is associated with a corresponding stage subsequence segment and soil attribute vector segment. The contribution calculation module calculates the yield contribution coefficient for each assessment unit using the stage subsequence segment and soil attribute vector segment. This yield contribution coefficient characterizes the degree of contribution of the assessment unit to the final yield during the growth stage. The ranking module ranks the multiple assessment units by yield level based on the yield contribution coefficient and soil attribute vector segment of each assessment unit, and generates a yield prediction interval for the planting area. The profit assessment module obtains the production cost series and market price series for the planting area, aligns the yield forecast interval with the production cost series and market price series over time, and generates the planting profit assessment results. This improves the spatial granularity of yield forecasting and the accuracy of profit assessment by ranking the contributions of spatial grid assessment units.

[0006] In the data acquisition module, historical meteorological sequences are arranged by timestamp as daily meteorological records. Each daily meteorological record includes temperature, precipitation, and sunshine duration. The daily meteorological records are further divided into seedling stage, jointing stage, tasseling stage, and grain-filling stage subsequences according to the number of days in each maize growth stage. The organic matter content, pH, and moisture content values ​​of each sampling point in the soil attribute sequence are combined into a soil attribute vector. These stage subsequences and soil attribute vectors provide input data matching the maize growth stages for subsequent spatial grid construction and yield contribution calculation.

[0007] Preferably, the grid division module divides the planting area into initial grids at latitude and longitude intervals. Soil sampling points are selected within each initial grid, and the soil attribute vectors of these sampling points are used as the soil attribute vector segments of the initial grid. For each initial grid, a meteorological observation station is searched for using the geometric center point of the initial grid as the center and a preset search radius. The stage subsequences of the searched meteorological observation stations are then merged by distance weighting into the stage subsequence segments of that initial grid, and the initial grids that have completed the association are marked as evaluation units. Through spatial grid subdivision and distance-weighted merging, the impact of local soil sampling errors and uneven distribution of meteorological observation stations on data representativeness is reduced.

[0008] In one embodiment of the present invention, three soil sampling points are uniformly arranged diagonally within each initial grid. The organic matter content, pH value, and moisture content of the three soil sampling points are collected respectively, forming three sampling attribute vectors. The three sampling attribute vectors are then weighted and averaged according to their position coordinates to generate a soil attribute vector fragment for the initial grid. Sampling attribute vectors with position coordinates closer to the center point of the initial grid are given higher weights, while those with position coordinates farther from the center point are given lower weights. A preset search radius is set to twice the length of the diagonal of the initial grid. The reciprocal of the distance between each meteorological observation station searched within the search radius and its geometric center point is used as the weight coefficient for each meteorological observation station. Subsequences of the same growth stage within the stage subsequences of each meteorological observation station are weighted and merged according to their weight coefficients to generate a stage subsequence fragment for the initial grid.

[0009] The contribution calculation module extracts the cumulative precipitation and sunshine duration values ​​of the grain-filling period subsequence from the stage subsequence fragments, and the organic matter content and pH values ​​from the soil attribute vector fragments. The ratio of the cumulative precipitation value of the grain-filling period subsequence of the evaluation unit to the median of the cumulative precipitation values ​​of the grain-filling period subsequences of all evaluation units in the planting area is used as the precipitation contribution factor for that evaluation unit; the ratio of the cumulative sunshine duration value of the grain-filling period subsequence of the evaluation unit to the median of the cumulative sunshine duration of the grain-filling period subsequences of all evaluation units in the planting area is used as the sunshine contribution factor for that evaluation unit; the product of the organic matter content value and pH value in the soil attribute vector fragment of the evaluation unit is used as the soil basic factor for that evaluation unit. The precipitation contribution factor, sunshine contribution factor, and soil basic factor are then superimposed according to the growth stage weights to generate the yield contribution coefficient for that evaluation unit. By highlighting the impact of precipitation and sunshine during the grain-filling period on yield formation, the contribution of the evaluation unit to the final yield can be more accurately reflected.

[0010] In one embodiment of the present invention, the maize growth stages are divided into seedling stage, jointing stage, tasseling stage, and grain-filling stage, and seedling option weights, jointing option weights, tasseling option weights, and grain-filling option weights are assigned to each stage respectively, with the grain-filling option weight being greater than the seedling option weight and the jointing option weight. The yield contribution coefficient of the assessment unit is generated by multiplying the precipitation contribution factor and sunshine contribution factor of the assessment unit during the grain-filling stage by the grain-filling option weight, and adding the value of the soil basic factor during the grain-filling stage multiplied by the grain-filling option weight.

[0011] The ranking module arranges all assessment units from high to low according to their yield contribution coefficients to form a yield contribution sequence. Starting from the beginning of the yield contribution sequence, a first preset proportion of assessment units are selected as the high contribution unit set, and starting from the end of the yield contribution sequence, a second preset proportion of assessment units are selected as the low contribution unit set. The ratio of the average water content value of the soil attribute vector fragments in the high contribution unit set to the average water content value of the soil attribute vector fragments in the low contribution unit set is calculated as the water content adjustment coefficient. The average yield contribution coefficient of the high contribution unit set is multiplied by the water content adjustment coefficient to obtain the upper limit of the yield prediction interval, and the average yield contribution coefficient of the low contribution unit set is multiplied by the water content adjustment coefficient to obtain the lower limit of the yield prediction interval.

[0012] In one embodiment of the present invention, when the moisture adjustment coefficient is greater than a first adjustment threshold, the upper limit of the yield prediction interval is adjusted to the average yield contribution coefficient of the high-contribution unit set multiplied by the moisture adjustment coefficient and then multiplied by an upward correction factor; when the moisture adjustment coefficient is less than a second adjustment threshold, the lower limit of the yield prediction interval is adjusted to the average yield contribution coefficient of the low-contribution unit set multiplied by the moisture adjustment coefficient and then multiplied by a downward correction factor; wherein the first adjustment threshold is greater than the second adjustment threshold, the upward correction factor is greater than 1, and the downward correction factor is less than 1. Adjusting the yield interval based on the moisture difference between high-contribution and low-contribution units makes the yield prediction interval closer to the actual yield fluctuation range.

[0013] The revenue assessment module maps the yield forecast interval to a predicted upper limit and a predicted lower limit. The total cost is calculated as the sum of seed, fertilizer, and pesticide costs for the current planting season within the production cost series. The selling price is calculated as the harvest price for the current planting season within the market price series. The upper limit of the revenue assessment is calculated by subtracting the total cost from the product of the predicted upper limit and the selling price, and the lower limit is calculated by subtracting the total cost from the product of the predicted lower limit and the selling price. Combining these upper and lower limits generates the planting revenue assessment interval. This process aligns the yield forecast interval with production costs and market prices over time, making the planting revenue assessment results more reflective of actual revenue fluctuations in the current planting season.

[0014] The technical effects and advantages provided by the present invention in the above technical solution are as follows: In the spatial grid division and evaluation unit association, an initial grid is divided according to latitude and longitude intervals. Soil sampling points are arranged within each initial grid, and soil attribute vector fragments are generated. Meteorological observation stations are searched using a preset search radius, with the geometric center of the initial grid as the center. The stage subsequences of the searched meteorological observation stations are then weighted by distance and merged into stage subsequence fragments, enabling each evaluation unit to simultaneously associate local soil attribute vector fragments and local meteorological stage subsequence fragments. This processing method preserves the differences in soil organic matter, pH, moisture content, and meteorological elements at different spatial locations within the planting area during the seedling, jointing, tasseling, and grain-filling stages. This ensures that subsequent yield contribution calculations no longer rely solely on regional averages as input, but instead reflect the impact of spatial combinations of precipitation, sunshine, and soil organic matter and pH on yield at the evaluation unit scale, reducing contribution bias caused by spatial homogenization.

[0015] In the calculation of yield contribution coefficient and generation of yield prediction interval, the cumulative values ​​of precipitation and sunshine duration during the grain-filling period are extracted from the stage subsequence fragments, and the values ​​of organic matter content and pH are extracted from the soil attribute vector fragments, forming precipitation contribution factor, sunshine contribution factor, and soil basic factor, respectively. These are then superimposed according to the weight of the growth stage to form the yield contribution coefficient. All evaluation units are then sorted according to their yield contribution coefficients, and high-contribution unit sets and low-contribution unit sets are extracted. The ratio of the average soil moisture content of the two sets is used as the water adjustment coefficient. The upper limit of the yield prediction interval is obtained by multiplying the average yield contribution coefficient of the high-contribution unit set by the water adjustment coefficient, and the lower limit is obtained by multiplying the average yield contribution coefficient of the low-contribution unit set by the water adjustment coefficient. This processing method distinguishes high-contribution and low-contribution spatial units by yield contribution sorting and adjusts the upper and lower limits of the prediction interval using differences in soil moisture content. This ensures that the yield prediction interval reflects the differences in moisture conditions between high-contribution and low-contribution areas within the planting area, avoiding deviations caused by determining the upper and lower limits of the interval solely by average values ​​or fixed proportions. By aligning the yield forecast interval with the current planting season's production cost sequence and harvest period price sequence, the planting income assessment results can be consistent with spatial differences and temporal cost price changes within the planting area. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0017] Figure 1 This is a flowchart of a big data-based corn yield prediction and planting benefit assessment system. Figure 2 This is a flowchart for generating planting benefit assessment ranges; Figure 3 This is a schematic diagram showing the ranking of corn yield contribution coefficients and the yield prediction range; Figure 4 It is a graph showing the production costs, market prices, and planting profits during the planting season. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] See Figure 1 This invention provides a big data-based corn yield prediction and planting benefit assessment system, comprising: The data acquisition module acquires historical meteorological sequences and soil attribute sequences of the planting area, divides the historical meteorological sequences into stage subsequences according to the growth stages, and maps the soil attribute sequences into soil attribute vectors. The grid division module constructs a spatial grid based on the stage subsequence and soil attribute vector. The planting area is divided into multiple evaluation units through the spatial grid, and each evaluation unit is associated with the corresponding stage subsequence segment and soil attribute vector segment. The contribution calculation module calculates the yield contribution coefficient for each evaluation unit using stage subsequence fragments and soil attribute vector fragments. The yield contribution coefficient is used to characterize the degree of contribution of the evaluation unit to the final yield during the growth stage. The ranking module ranks multiple assessment units by yield contribution coefficient and soil attribute vector fragments, and generates a yield prediction range for the planting area based on the ranking results. The profit assessment module obtains the production cost series and market price series of the planting area, aligns the yield forecast interval with the production cost series and market price series over time, and generates the planting profit assessment results.

[0020] In practical implementation, the data acquisition process for the big data-based corn yield prediction and planting benefit assessment system is as follows: Acquire historical meteorological and soil property sequences for the planting area. The historical meteorological sequence contains daily meteorological data recorded by meteorological stations within the planting area over a historical period. The soil property sequence contains soil property data recorded by soil sampling points within the planting area over a historical period.

[0021] The historical meteorological series is processed. The historical meteorological series is then arranged into daily meteorological records by timestamp. Each daily meteorological record includes temperature, precipitation, and sunshine duration. The timestamp for each daily meteorological record is the date.

[0022] Daily meteorological records were divided into seedling stage, jointing stage, tasseling stage, and grain-filling stage subsequences based on the number of days between each growth stage. Specifically, the seedling stage subsequence is the set of daily meteorological records from sowing to the end of the seedling emergence period; the jointing stage subsequence is the set of daily meteorological records from the beginning to the end of the jointing stage; the tasseling stage subsequence is the set of daily meteorological records from the beginning to the end of the tasseling stage; and the grain-filling stage subsequence is the set of daily meteorological records from the beginning to the end of the grain-filling stage. The number of days between each growth stage was set according to the growth cycle characteristics of the maize variety.

[0023] The soil property sequence is processed. The organic matter content, pH, and moisture content values ​​of each sampling point in the soil property sequence are combined into a soil property vector. The soil property vector is represented as follows: in, Indicates the first Soil attribute vectors from each soil sampling point; Indicates the first Organic matter content values ​​at each soil sampling point; Indicates the first pH values ​​at each soil sampling point; Indicates the first Moisture content values ​​at each soil sampling point; This indicates the sequence number of the soil sampling point. The organic matter content, pH, and moisture content values ​​are all extracted from the soil property sequence, which is obtained by testing and analyzing soil samples collected from the soil sampling points.

[0024] In practice, a spatial grid is constructed based on the stage subsequence and soil attribute vectors. The planting area is divided into initial grids at latitude and longitude intervals. The latitude and longitude intervals are determined based on the spatial extent of the planting area and the resolution requirements of the evaluation units. The spatial extent of the planting area is determined by the boundary coordinates of the planting area, and the resolution requirements of the evaluation units are set according to the spatial accuracy of maize yield prediction. Each initial grid has a geometric center point.

[0025] Soil sampling points were selected within each initial grid. Three soil sampling points were evenly distributed along the diagonal of each initial grid. The first soil sampling point was located at the first equidistant point of the first diagonal of the initial grid; the second soil sampling point was located at the intersection of the first and second diagonals of the initial grid; and the third soil sampling point was located at the second equidistant point of the second diagonal of the initial grid. The first and third soil sampling points were symmetrical about the geometric center of the initial grid. The organic matter content, pH value, and moisture content of the three soil sampling points were collected respectively, forming three sampling attribute vectors.

[0026] The three sampled attribute vectors are weighted and averaged according to their location coordinates to generate a soil attribute vector fragment for the initial grid. Sampling attribute vectors with location coordinates closer to the center point of the initial grid are assigned higher weights, while those with location coordinates farther from the center point are assigned lower weights. The location weight coefficients are determined as follows: the Euclidean distance between each soil sampling point and the geometric center point of the initial grid is calculated, and the reciprocal of this Euclidean distance is used as the initial location weight. The initial location weights of the three soil sampling points are then normalized to obtain the location weight coefficient for each soil sampling point. The Euclidean distance is calculated based on the planar coordinates of the soil sampling point within the initial grid and the planar coordinates of the geometric center point of the initial grid. The planar coordinates of the soil sampling points are obtained by converting the latitude and longitude range of the initial grid.

[0027] The weighted average of soil attribute vector fragments is expressed as: in, Indicates the first A fragment of soil property vector from an initial grid; Indicates the initial grid number; Indicates the first A sampling attribute vector of each soil sampling point; This indicates the sequence number of the soil sampling point within the grid, with values ​​of 1, 2, and 3. Indicates the first Location weighting coefficients for each soil sampling point. The range of values ​​is arrive And the sum of the location weight coefficients of the three soil sampling points equals The location weighting coefficient is determined based on the distance between the soil sampling point and the geometric center of the initial grid; the smaller the distance, the larger the location weighting coefficient.

[0028] For each initial grid, a search for meteorological stations is conducted with the geometric center of the initial grid as the center, within a preset search radius. The preset search radius is set to twice the diagonal length of the initial grid. The diagonal length of the initial grid is calculated based on the longitude and latitude span of the initial grid. The reciprocal of the distance between each meteorological station found within the search radius and the geometric center is used as the weight coefficient for each meteorological station. The distance between the meteorological station and the geometric center of the initial grid is calculated based on the latitude and longitude coordinates of the meteorological station and the geometric center of the initial grid.

[0029] Subsequences of the same growth stage from each meteorological observation station are weighted and merged according to weight coefficients to generate stage subsequence fragments for the initial grid. These stage subsequence fragments include seedling stage, jointing stage, tasseling stage, and grain-filling stage subsequence fragments. The weighted merging process is as follows: for subsequences of the same growth stage, the temperature values, precipitation values, and sunshine duration values ​​for the same date are weighted and averaged according to the weight coefficients of each meteorological observation station, respectively, to obtain the meteorological element values ​​for the corresponding dates in the stage subsequence fragments. The larger the weight coefficient of the meteorological observation station, the higher the proportion of the meteorological element values ​​of that station in the weighted merging result.

[0030] The initial grid after association is marked as the evaluation unit. The evaluation unit is associated with the corresponding stage subsequence fragment and soil property vector fragment.

[0031] In practice, yield contribution coefficients are calculated for each assessment unit using stage subsequence fragments and soil attribute vector fragments. Cumulative precipitation and cumulative sunshine hours are extracted from the grain-filling period subsequence fragments, while organic matter content and pH values ​​are extracted from the soil attribute vector fragments.

[0032] For each assessment unit, a grouting period subsequence segment is obtained from the associated stage subsequence segment. This segment contains precipitation and sunshine duration values ​​for multiple dates. The precipitation values ​​for all dates within the grouting period subsequence segment are summed to obtain the cumulative precipitation value for the grouting period subsequence of the assessment unit. Similarly, the sunshine duration values ​​for all dates within the grouting period subsequence segment are summed to obtain the cumulative sunshine duration value for the grouting period subsequence of the assessment unit. The cumulative precipitation values ​​for the grouting period subsequence of all assessment units within the planting area are collected, sorted by value, and the median value is taken as the median of the cumulative precipitation values ​​for the grouting period subsequence of all assessment units within the planting area. If the number of assessment units within the planting area is even, the average of the two middle values ​​after sorting is taken as the median. The cumulative precipitation value for the grouting period subsequence of the assessment unit is divided by the median of the cumulative precipitation values ​​for the grouting period subsequence of all assessment units within the planting area to obtain the precipitation contribution factor for the assessment unit. Collect the cumulative sunshine hours for the grain-filling period subsequence of all assessment units within the planting area, sort them by value, and take the median value as the median of the cumulative sunshine hours for the grain-filling period subsequence of all assessment units within the planting area. If the number of assessment units within the planting area is even, take the average of the two middle values ​​after sorting as the median value. Divide the cumulative sunshine hours for the grain-filling period subsequence of each assessment unit by the median value of the cumulative sunshine hours for the grain-filling period subsequence of all assessment units within the planting area to obtain the sunshine contribution factor of each assessment unit.

[0033] Organic matter content and pH values ​​are obtained from the soil attribute vector fragments associated with the assessment unit. The organic matter content value and pH value in the soil attribute vector fragment of the assessment unit are multiplied together to obtain the basic soil factors of the assessment unit.

[0034] The maize growth stages are divided into seedling stage, jointing stage, tasseling stage, and grain-filling stage, with corresponding option weights assigned to each stage. The seedling option weight is set at 0.15, the jointing option weight at 0.20, the tasseling option weight at 0.25, and the grain-filling option weight at 0.40. The grain-filling option weight is greater than the seedling option weight, which is greater than the jointing option weight, and the grain-filling option weight is greater than the tasseling option weight. The weighting values ​​are set as follows: Based on the agronomical characteristics of maize growth stages, the grain-filling stage directly determines the degree of grain filling and has the greatest contribution to yield formation; therefore, the weighting value for the grain-filling option is set at 0.40. The tasseling stage determines pollination quality and has a significant impact on the number of grains per ear; therefore, the weighting value for the tasseling option is set at 0.25. The jointing stage determines plant growth rate and ear differentiation, and has a moderate impact on yield; therefore, the weighting value for the jointing option is set at 0.20. The seedling stage determines emergence quality and early vegetative growth, and has a relatively small impact on final yield; therefore, the weighting value for the seedling option is set at 0.15. The sum of the weighting values ​​for each growth stage equals 1.

[0035] For each assessment unit, the yield contribution coefficient is generated by multiplying the product of the precipitation contribution factor and the sunshine contribution factor during the grouting period by the grouting option weight, and then adding the value of the soil baseline factor during the grouting period multiplied by the grouting option weight. The yield contribution coefficient is expressed as: in, Indicates the first The output contribution coefficient of each evaluation unit. For the evaluation unit number, The total number of assessment units within the planting area is obtained from 1; Indicates the first Precipitation contribution factors for each assessment unit; Indicates the first Sunlight contribution factor of each assessment unit; Indicates the first The relative soil fundamental factors of the assessment unit, through the first assessment unit The product of the organic matter content and pH value in the soil attribute vector segment of each assessment unit is divided by the median value of the soil basic factors of all assessment units in the planting area. This indicates the weight of the grouting option. The value range is from 0 to 1, and in this embodiment it is set to 0.40. This setting is based on the fact that the contribution of the grain-filling stage to maize yield formation is higher than that of the seedling stage, jointing stage, and tasseling stage. Precipitation contribution factor. Through the first The cumulative precipitation value of the grain-filling period subsequence of each assessment unit is divided by the median of the cumulative precipitation values ​​of the grain-filling period subsequences of all assessment units in the planting area. Sunshine contribution factor. Through the first The cumulative sunshine duration of the grain-filling period subsequence in each assessment unit is obtained by dividing the cumulative sunshine duration of the grain-filling period subsequence in all assessment units of the planting area by the median value. Soil basic factors. Through the first The organic matter content value and pH value are obtained by multiplying the soil attribute vector fragment of each assessment unit.

[0036] In practice, multiple assessment units are ranked by yield level based on their yield contribution coefficients and soil attribute vector fragments. All assessment units are then arranged from highest to lowest yield contribution coefficient, forming a yield contribution sequence. The yield contribution coefficient is derived from the output of the contribution calculation module, with each assessment unit corresponding to one coefficient. In the yield contribution sequence, the assessment unit at the beginning has the highest yield contribution coefficient, and the assessment unit at the end has the lowest yield contribution coefficient.

[0037] The first set of evaluation units, representing a predetermined percentage from the beginning of the yield contribution sequence, is designated as the high-contribution unit set. This predetermined percentage is set to 20% of the total number of evaluation units in the planting area. The selection method is as follows: all evaluation units ranking in the top 20% of the yield contribution sequence are included in the high-contribution unit set. The second set of evaluation units, representing a predetermined percentage from the end of the yield contribution sequence, is designated as the low-contribution unit set. This second predetermined percentage is set to 20% of the total number of evaluation units in the planting area. The selection method is as follows: all evaluation units ranking in the bottom 20% of the yield contribution sequence are included in the low-contribution unit set. The basis for setting the first predetermined percentage to 20% is that the top 20% of evaluation units with the highest yield contribution coefficients cover evaluation units whose yield contribution in the planting area is significantly higher than the regional average, while the bottom 20% of evaluation units with the lowest yield contribution coefficients cover evaluation units whose yield contribution in the planting area is significantly lower than the regional average. Both sets contain a sufficient number of evaluation units to ensure the stability of subsequent mean calculations. The basis for setting the second predetermined percentage to 20% is the same as that for setting the first predetermined percentage to 20%.

[0038] Obtain the water content value from the soil attribute vector fragment of each evaluation unit within the high-contribution unit set. Sum the water content values ​​from the soil attribute vector fragments of all evaluation units within the high-contribution unit set, and then divide the sum by the number of evaluation units in the high-contribution unit set to obtain the average water content value of the soil attribute vector fragments for the high-contribution unit set. Obtain the water content value from the soil attribute vector fragment of each evaluation unit within the low-contribution unit set. Sum the water content values ​​from the soil attribute vector fragments of all evaluation units within the low-contribution unit set, and then divide the sum by the number of evaluation units in the low-contribution unit set to obtain the average water content value of the soil attribute vector fragments for the low-contribution unit set.

[0039] The ratio of the mean water content of soil attribute vector fragments from high-contribution unit sets to the mean water content of soil attribute vector fragments from low-contribution unit sets is calculated and used as the water content adjustment coefficient. The water content adjustment coefficient is expressed as: in, Indicates the moisture adjustment coefficient; The mean water content value of the soil attribute vector fragment representing the high-contribution unit set; This represents the mean water content value of the soil attribute vector fragment of the low contribution unit set. The result is obtained by summing the water content values ​​in the soil attribute vector fragments of all evaluation units within the high contribution unit set and then dividing by the number of evaluation units within the high contribution unit set. The moisture adjustment coefficient is obtained by summing the water content values ​​in the soil attribute vector fragments of all evaluation units within the low contribution unit set and dividing by the number of evaluation units in the low contribution unit set. The value is greater than 0. When the moisture adjustment coefficient A value greater than 1 indicates that the average soil moisture content of the high-contribution unit set is higher than that of the low-contribution unit set; when the moisture adjustment coefficient... When the value is less than 1, it indicates that the average soil moisture content of the high contribution unit set is lower than the average soil moisture content of the low contribution unit set.

[0040] Obtain the output contribution coefficient of each evaluation unit within the high-contribution unit set. Sum the output contribution coefficients of all evaluation units within the high-contribution unit set, and then divide by the number of evaluation units in the high-contribution unit set to obtain the average output contribution coefficient of the high-contribution unit set. Obtain the output contribution coefficient of each evaluation unit within the low-contribution unit set. Sum the output contribution coefficients of all evaluation units within the low-contribution unit set, and then divide by the number of evaluation units in the low-contribution unit set to obtain the average output contribution coefficient of the low-contribution unit set.

[0041] The upper limit of the yield prediction interval is calculated by multiplying the average yield contribution coefficient of the high-contribution unit set by the moisture adjustment coefficient. The lower limit of the yield prediction interval is calculated by multiplying the average yield contribution coefficient of the low-contribution unit set by the moisture adjustment coefficient.

[0042] When the moisture adjustment coefficient exceeds the first adjustment threshold, the upper limit of the yield prediction interval is adjusted to the average yield contribution coefficient of the high-contribution unit set multiplied by the moisture adjustment coefficient, and then multiplied by the upward correction factor. The first adjustment threshold is set to 1.2. The basis for setting the first adjustment threshold to 1.2 is that when the average soil moisture content of the high-contribution unit set is more than 20% higher than that of the low-contribution unit set, soil moisture conditions have an additional boosting effect on the yield upper limit of the high-contribution unit set, requiring a further upward adjustment of the upper limit of the yield prediction interval. The upward correction factor is set to 1.1. The basis for setting the upward correction factor to 1.1 is that when the moisture adjustment coefficient is greater than 1.2, the soil moisture advantage of the high-contribution unit set is increased by 10% based on the calculated upper limit value to reflect the additional yield-increasing effect of soil moisture advantage on high-yield areas.

[0043] When the moisture adjustment coefficient is less than the second adjustment threshold, the lower limit of the yield prediction interval is adjusted to the average yield contribution coefficient of the low contribution unit set multiplied by the moisture adjustment coefficient, and then multiplied by the downward correction factor. The second adjustment threshold is set to 0.8. The basis for setting the second adjustment threshold to 0.8 is that when the average soil moisture content of the high contribution unit set is more than 20% lower than the average soil moisture content of the low contribution unit set, the soil moisture conditions have an additional suppressive effect on the yield lower limit of the low contribution unit set, and the lower limit of the yield prediction interval needs to be further lowered. The downward correction factor is set to 0.9. The basis for setting the downward correction factor to 0.9 is that when the moisture adjustment coefficient is less than 0.8, the soil moisture deficiency of the low contribution unit set is reduced by 10% based on the calculated lower limit value, to reflect the additional yield reduction effect of soil moisture deficiency on low-yield areas. The first adjustment threshold is greater than the second adjustment threshold, the upward correction factor is greater than 1, and the downward correction factor is less than 1.

[0044] Based on the sorting results and the corrected upper and lower limits, a yield prediction interval for the planting area is generated. If the moisture adjustment coefficient is less than or equal to the first adjustment threshold and greater than or equal to the second adjustment threshold, the upper limit of the yield prediction interval remains the average yield contribution coefficient of the high-contribution unit set multiplied by the moisture adjustment coefficient, and the lower limit remains the average yield contribution coefficient of the low-contribution unit set multiplied by the moisture adjustment coefficient. If the moisture adjustment coefficient is greater than the first adjustment threshold, the upper limit of the yield prediction interval is updated to the corrected upper limit, while the lower limit remains unchanged. If the moisture adjustment coefficient is less than the second adjustment threshold, the lower limit of the yield prediction interval is updated to the corrected lower limit, while the upper limit remains unchanged.

[0045] See Figure 3 In the figure, the horizontal axis represents the ranking number of the evaluation units, indicating the unit numbers arranged from highest to lowest according to the output contribution coefficient, ranging from 1 to approximately 500. The vertical axis represents the dimensionless value of the output contribution coefficient. The black solid line represents the output contribution sequence of all evaluation units. The curve gradually decreases from a high value of approximately 1.55 on the left to a low value of approximately 0.5 on the right, reflecting the decreasing trend of the output contribution coefficient of the evaluation units.

[0046] In the figure, the red dashed line represents the upper limit of the yield prediction interval, with a value of approximately 1.48; the blue dashed line represents the lower limit of the yield prediction interval, with a value of approximately 0.86. The upper and lower limits constitute the boundaries of the yield prediction interval, reflecting the predicted range of maize yield in the planting area after combining the ranking of evaluation units and the moisture adjustment coefficient.

[0047] The red shaded area in the figure marks the top 20% of high-contribution units (approximately the first 100 evaluation units). The output contribution coefficient of this area is significantly higher than that of subsequent units, reflecting the large contribution of high-contribution units to the overall output. The blue shaded area marks the bottom 20% of low-contribution units (approximately the last 100 evaluation units). Their output contribution coefficient is significantly lower, representing evaluation units with weaker output contribution.

[0048] According to the above embodiment, the yield contribution sequence and the division of the evaluation units into 20% before and after the curve, as shown in the figure, are used to calculate the average soil moisture content and average yield contribution coefficient of the high contribution unit set and the low contribution unit set, and then calculate the moisture adjustment coefficient to adjust the upper and lower limits of the yield prediction interval. The upper and lower limits of the yield prediction interval in the figure are the yield prediction results after adjustment by the moisture adjustment coefficient, reflecting the moderating effect of soil moisture on the yield prediction interval.

[0049] In specific implementation, please refer to Figure 2 The production cost series and market price series for the planting area are obtained. The production cost series contains production cost records for multiple planting seasons, and each production cost record includes a planting season identifier, seed cost value, fertilizer cost value, and pesticide cost value. The market price series contains market price records for multiple planting seasons, and each market price record includes a planting season identifier and harvest price value. The yield forecast interval is aligned with the production cost series and market price series by time period. The time period alignment method is as follows: using the current planting season as the time base, production cost records with the planting season identifier of the current planting season are selected from the production cost series, and market price records with the planting season identifier of the current planting season are selected from the market price series. The current planting season is the planting season for which the current assessment is being conducted.

[0050] The production forecast interval is mapped to the upper and lower limits of the predicted production. The production forecast interval is generated by the ranking module. The upper limit of the production forecast interval is directly mapped to the upper limit of the predicted production, and the lower limit of the production forecast interval is directly mapped to the lower limit of the predicted production.

[0051] In some embodiments, seed cost values, fertilizer cost values, and pesticide cost values ​​are extracted from the production cost records of the current planting season in the production cost sequence. The seed cost values, fertilizer cost values, and pesticide cost values ​​are added together to obtain the total cost value. The total cost value represents the sum of inputs into the three cost items of seeds, fertilizer, and pesticides for the current planting season.

[0052] In some embodiments, the harvest price value is extracted from the market price record of the current planting season in the market price series. The harvest price value is the market selling price value of corn harvested during the current planting season. The harvest price value is used as the selling price value.

[0053] Calculate the upper limit of revenue assessment. The upper limit of revenue assessment is calculated by subtracting the total cost from the product of the projected upper limit of production and the selling price. The formula for calculating the upper limit of revenue assessment is as follows: in, This indicates the upper limit of the revenue assessment; This represents the upper limit of the predicted output, obtained by mapping the upper limit of the output prediction interval; The selling price value is obtained by extracting the harvest price value from the market price record of the current planting season in the market price series. This represents the total cost value, obtained by adding the seed cost, fertilizer cost, and pesticide cost values ​​for the current planting season in the production cost sequence. (Upper limit for revenue assessment) The units for the sales price and total cost are consistent with those for the projected production ceiling. This represents the upper limit of the yield forecast range, indicating the highest predicted yield per unit area for the planting region in the current planting season. (Sales price value) This represents the selling price per unit weight of corn at harvest time during the current planting season. Total cost value. This represents the total cost of seeds, fertilizers, and pesticides per unit area during the current planting season.

[0054] Calculate the lower limit of the revenue assessment. The lower limit of the revenue assessment is calculated by subtracting the total cost from the product of the predicted lower limit of the yield and the selling price. The calculation method for the lower limit of the revenue assessment is the same as that for the upper limit of the revenue assessment, except that the predicted upper limit of the yield in the formula for the upper limit of the revenue assessment is replaced by the predicted lower limit of the yield. The predicted lower limit of the yield is the lower limit of the yield forecast interval, representing the lowest predicted yield per unit area for the planting area in the current planting season.

[0055] The upper and lower limits of the combined benefit assessment are used to generate the planting benefit assessment range. The planting benefit assessment range is represented as a closed interval formed by the lower and upper limits of the benefit assessment.

[0056] See Figure 4 The graph uses the planting season as the horizontal axis, covering multiple planting seasons without specifying the unit of measurement, representing different planting season sequences. The left vertical axis represents the cost (in yuan / mu), used to depict the cost of seeds, fertilizers, pesticides, and the upper and lower limits of profit assessment; the right vertical axis represents the price (in yuan / kg), used to depict the price at harvest. The legend shows six curves, corresponding to the cost of seeds (blue dashed line), the cost of fertilizers (orange dashed line), the cost of pesticides (green dotted line), the price at harvest (black solid line), the upper limit of profit assessment (red solid line), and the lower limit of profit assessment (purple dashed line).

[0057] The cost curves for seeds, fertilizers, and pesticides remained relatively stable overall. Fertilizer costs were the highest, fluctuating around 100 yuan / mu, followed by seed costs, with pesticide costs being the lowest. None of these showed significant fluctuations, indicating that the input of each cost item was relatively stable across different planting seasons. The price curve at harvest time fluctuated significantly, with prices ranging from approximately 1.4 yuan / kg to 2.7 yuan / kg, reflecting the strong volatility of market prices due to multiple factors.

[0058] The upper and lower limits of the profit assessment are represented by solid red lines and dashed purple lines, respectively. Both curves exhibit overall volatility but are significantly higher than the cost curve, especially the fertilizer cost curve, indicating that after considering the predicted yield range and market prices, the planting profit is in a positive range. The upper limit of the profit assessment is generally higher than the lower limit, and there is a certain gap between the two, reflecting the impact of the uncertainty of the upper and lower limits of the yield prediction range on planting profit.

[0059] The profit assessment curve shows peaks and troughs across multiple planting seasons, roughly synchronizing with the price fluctuations during harvest, indicating that market prices have a direct driving effect on profit assessment. The amplitude and fluctuation range of the upper and lower limits of profit assessment are significantly greater than those of cost items, highlighting the dominant role of yield and market prices in profit fluctuations.

[0060] The curves in the figure illustrate the specific implementation of calculating the upper and lower limits of the revenue assessment by aligning the production forecast interval with the production cost and market price series over time. The fluctuating upper and lower limits of the revenue assessment show the combined impact of the upper and lower limits of the predicted production and the market price and total cost of the corresponding planting season, verifying the dynamic assessment capability of the revenue assessment module in this embodiment.

[0061] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A big data-based corn yield prediction and planting benefit assessment system, characterized in that, include: The data acquisition module acquires historical meteorological sequences and soil attribute sequences of the planting area, divides the historical meteorological sequences into stage subsequences according to the growth stages, and maps the soil attribute sequences into soil attribute vectors. The grid division module constructs a spatial grid based on the stage subsequence and soil attribute vector. The planting area is divided into multiple evaluation units through the spatial grid, and each evaluation unit is associated with the corresponding stage subsequence segment and soil attribute vector segment. The contribution calculation module calculates the yield contribution coefficient for each evaluation unit using stage subsequence fragments and soil attribute vector fragments. The yield contribution coefficient is used to characterize the degree of contribution of the evaluation unit to the final yield during the growth stage. The ranking module ranks multiple assessment units by yield contribution coefficient and soil attribute vector fragments, and generates a yield prediction range for the planting area based on the ranking results. The profit assessment module obtains the production cost series and market price series of the planting area, aligns the yield forecast interval with the production cost series and market price series over time, and generates the planting profit assessment results.

2. The big data-based corn yield prediction and planting income assessment system according to claim 1, characterized in that, In the steps of obtaining the historical meteorological sequence and soil property sequence of the planting area, dividing the historical meteorological sequence into stage sub-sequences according to the growth stage, and mapping the soil property sequence into a soil property vector: Historical meteorological sequences are arranged by timestamp into daily meteorological records, each of which includes temperature, precipitation, and sunshine duration. The daily meteorological records were divided into seedling stage subsequence, jointing stage subsequence, tasseling stage subsequence, and grain filling stage subsequence according to the number of days in the corn growth stage. The organic matter content, pH value, and water content value of each sampling point in the soil attribute sequence are combined into a soil attribute vector.

3. The big data-based corn yield prediction and planting income assessment system according to claim 2, characterized in that, In the step of constructing a spatial grid based on stage subsequences and soil attribute vectors, dividing the planting area into multiple evaluation units using the spatial grid, and associating each evaluation unit with a corresponding stage subsequence fragment and soil attribute vector fragment: The planting area is divided into initial grids according to latitude and longitude intervals. Soil sampling points are selected in each initial grid, and the soil attribute vector of the soil sampling point is used as the soil attribute vector fragment of the initial grid. For each initial grid, meteorological observation stations are searched with the geometric center point of the initial grid as the center and the search radius is set. The stage subsequences of the searched meteorological observation stations are then merged into a stage subsequence fragment of the initial grid by distance weighting. Mark the initial mesh after association as the evaluation unit.

4. The big data-based corn yield prediction and planting income assessment system according to claim 3, characterized in that, In the step of calculating the yield contribution coefficient for each evaluation unit using stage subsequence fragments and soil attribute vector fragments, the yield contribution coefficient is used to characterize the degree of contribution of the evaluation unit to the final yield during the growth stage: The cumulative precipitation and cumulative sunshine duration of the grouting period subsequence were extracted from the stage subsequence fragments, and the organic matter content and pH values ​​were extracted from the soil property vector fragments. The ratio of the cumulative precipitation value of the grain-filling period subsequence of the assessment unit to the median value of the cumulative precipitation value of the grain-filling period subsequence of all assessment units in the planting area is used as the precipitation contribution factor of the assessment unit. The ratio of the cumulative sunshine hours of the grain-filling period subsequence of the evaluation unit to the median of the cumulative sunshine hours of the grain-filling period subsequence of all evaluation units in the planting area is used as the sunshine contribution factor of the evaluation unit. The product of the organic matter content and pH value in the soil attribute vector segment of the assessment unit is used as the basic soil factor of the assessment unit. The yield contribution coefficient of this assessment unit is generated by superimposing the precipitation contribution factor, sunshine contribution factor, and soil basic factor according to the weight of growth stage.

5. The big data-based corn yield prediction and planting income assessment system according to claim 4, characterized in that, In the step of ranking multiple assessment units by yield contribution coefficient and soil attribute vector fragment based on each assessment unit, and generating a yield prediction interval for the planting area based on the ranking results: All evaluation units are arranged from high to low according to their output contribution coefficients to form an output contribution sequence. The first preset proportion of evaluation units are selected from the beginning of the output contribution sequence as the high contribution unit set, and the second preset proportion of evaluation units are selected from the end of the output contribution sequence as the low contribution unit set. The ratio of the mean water content of the soil attribute vector fragments of the high contribution unit set to the mean water content of the soil attribute vector fragments of the low contribution unit set is calculated and used as the water adjustment coefficient. The upper limit of the yield prediction interval is the average yield contribution coefficient of the high-contribution unit set multiplied by the moisture adjustment coefficient, and the lower limit of the yield prediction interval is the average yield contribution coefficient of the low-contribution unit set multiplied by the moisture adjustment coefficient.

6. The big data-based corn yield prediction and planting income assessment system according to claim 5, characterized in that, In the step of obtaining the production cost series and market price series of the planting area, aligning the yield prediction interval with the production cost series and market price series over time, and generating the planting income assessment result: Map the production forecast interval to a predicted upper limit and a predicted lower limit; The total cost value is the sum of the seed cost, fertilizer cost, and pesticide cost values ​​for the current planting season in the production cost sequence. The selling price is determined by using the harvest price value of the current planting season in the market price series. The upper limit of the profit assessment is calculated by subtracting the total cost from the product of the predicted upper limit of the yield and the selling price, and the lower limit of the profit assessment is calculated by subtracting the total cost from the product of the predicted lower limit of the yield and the selling price. The upper and lower limits of the profit assessment are combined to generate the planting profit assessment range.

7. The big data-based corn yield prediction and planting income assessment system according to claim 6, characterized in that, In the step of dividing the planting area into initial grids according to latitude and longitude intervals, selecting soil sampling points within each initial grid, and using the soil attribute vector of the soil sampling points as the soil attribute vector fragment of the initial grid: Three soil sampling points are evenly arranged diagonally within each initial grid. The organic matter content, pH value, and water content of the three soil sampling points are collected respectively to form three sampling attribute vectors. The three sampled attribute vectors are weighted and averaged according to their location coordinates to generate soil attribute vector fragments for the initial grid. The sampled attribute vectors whose position coordinates are closer to the center point of the initial grid are given higher weights, while the sampled attribute vectors whose position coordinates are farther from the center point of the initial grid are given lower weights.

8. The big data-based corn yield prediction and planting income assessment system according to claim 7, characterized in that, In the step of searching for meteorological observation stations with a preset search radius around the geometric center point of each initial grid, and merging the stage subsequences of the searched meteorological observation stations into a stage subsequence fragment of the initial grid by distance weighting: The preset search radius is set to twice the length of the initial grid diagonal. The reciprocal of the distance between each meteorological observation station found within the search radius and the geometric center point is used as the weight coefficient of each meteorological observation station. The subsequences of the same growth stage in the stage subsequences of each meteorological observation station are weighted and merged according to the weight coefficient to generate the stage subsequence fragments of the initial grid.

9. The big data-based corn yield prediction and planting income assessment system according to claim 8, characterized in that, In the step of superimposing the precipitation contribution factor, sunshine contribution factor, and basic soil factors according to the weight of growth stage to generate the yield contribution coefficient of the evaluation unit: The growth stages of maize are divided into seedling stage, jointing stage, tasseling stage and grain filling stage, and seedling option weights are assigned to seedling option weights, jointing option weights, tasseling option weights and grain filling option weights respectively. The weight of the grouting option is greater than the weight of the seedling option, and the weight of the grouting option is greater than the weight of the jointing option; The yield contribution coefficient of the assessment unit is generated by multiplying the precipitation contribution factor and sunshine contribution factor of the assessment unit during the grouting period by the grouting option weight, and adding the value of the soil basic factor of the assessment unit during the grouting period by the grouting option weight.

10. The big data-based corn yield prediction and planting income assessment system according to claim 9, characterized in that, In the step of using the average yield contribution coefficient of the high-contribution unit set multiplied by the moisture adjustment coefficient as the upper limit of the yield prediction interval, and using the average yield contribution coefficient of the low-contribution unit set multiplied by the moisture adjustment coefficient as the lower limit of the yield prediction interval: When the moisture adjustment coefficient is greater than the first adjustment threshold, the upper limit of the yield prediction interval is adjusted to the average yield contribution coefficient of the high contribution unit set multiplied by the moisture adjustment coefficient and then multiplied by the upward correction factor. When the moisture adjustment coefficient is less than the second adjustment threshold, the lower limit of the yield prediction interval is adjusted to the average yield contribution coefficient of the low contribution unit set multiplied by the moisture adjustment coefficient and then multiplied by the downward correction factor. The first adjustment threshold is greater than the second adjustment threshold, the upward correction factor is greater than 1, and the downward correction factor is less than 1.

Images