A wheat and rice interplanting self-adaptive sowing width regulation method based on crop and soil parameter identification

Abstract

The application relates to the field of crop seeding technology, and particularly discloses a wheat-corn interplanting self-adaptive seeding width regulation method based on crop and soil parameter identification, which comprises the following steps: normalizing soil body penetration resistance data to generate soil body penetration resistance score values and dividing seeding condition grades; jointly analyzing wheat grain moisture content data and soil slope data to construct a slope moisture content correlation matrix and perform dimension expansion to obtain interference characteristic factors; inputting wheat row spacing data, the seeding condition grades and the interference characteristic factors into a decision rule library to generate seeding width values and seeding density values; converting the seeding width values into channel boundary coordinate instructions of seeding machines, and driving the seeding machines to adjust the operation width of corn seeding according to the channel boundary coordinate instructions; and synchronously regulating the single irrigation amount of wheat medium-shallow buried drip irrigation. The application can improve the seeding precision and resource utilization efficiency of wheat-corn interplanting operation.

Description

Technical Field

[0001] This invention relates to the field of crop sowing technology, and in particular to an adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification. Background Technology

[0002] Wheat-maize intercropping is an important planting model for improving land productivity in the oasis irrigation areas of Northwest China. Its core lies in optimizing the spatial configuration of wheat and maize to achieve efficient utilization of light, heat, water, and fertilizer resources. In wheat-maize intercropping, the setting of maize sowing width and density directly determines the spatial distribution structure of the intercropping population and the interspecific resource competition pattern. In traditional sowing operations, maize sowing width and density are usually preset based on fixed agronomic parameters, and the sowing machine performs a single-mode sowing operation according to the preset row and plant spacing to ensure the basic spatial distribution of the intercropping population.

[0003] In existing technologies, wheat row spacing in wheat-maize intercropping fields exhibits significant spatial variability due to factors such as sowing errors, field management compaction, and soil subsidence. Fixed-width sowing methods struggle to adapt to the dynamic changes in wheat row spacing, leading to maize sowing channel deviation and uneven population density. Furthermore, current sowing decision-making methods lack multi-dimensional parameter integration of soil physicochemical conditions and wheat growth status, making it difficult to quantify the constraining effect of soil conditions on sowing quality and the degree to which wheat growth encroaches on maize sowing space. This results in a single basis for setting sowing width and density, failing to achieve dynamic coupling between sowing width and density. In addition, existing sowing machinery execution mechanisms and decision-making systems lack a precise spatial coordinate-based linkage mechanism; sowing width settings and machinery operation trajectories are independent, making it difficult to form a closed-loop control of decision-making, execution, and feedback. These deficiencies result in poor population structure coordination in wheat-maize intercropping fields, hindering the improvement of inter-species resource utilization efficiency and restricting further breakthroughs in the overall productivity of the intercropping system. Therefore, there is an urgent need to develop an adaptive sowing width control method that can identify crop and soil parameters in real time, dynamically couple sowing width and density decisions, and achieve precise execution by machinery, in order to solve the problems of fixed sowing width, blind density, and disconnected execution in existing technologies, and improve the coordination of population structure and resource utilization efficiency of wheat-maize intercropping. Summary of the Invention

[0004] This invention provides an adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, this invention provides an adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification, comprising:

[0006] Data on wheat row spacing, soil slope, wheat grain moisture content, and soil penetration resistance were collected.

[0007] The soil penetration resistance data is normalized and mapped to generate a soil penetration resistance score, and the sowing condition level is classified according to the soil penetration resistance score.

[0008] The wheat grain moisture content data and the soil slope data are jointly analyzed to construct a slope moisture content correlation matrix, and the dimensions of the slope moisture content correlation matrix are expanded to obtain interference characteristic factors for corn sowing.

[0009] The wheat row spacing data, the sowing condition level, and the interference feature factor are input into a preset decision rule base to generate the sowing width value and sowing density value for corn sowing.

[0010] The seeding width value is converted into channel boundary coordinate instructions for the seeding machine, and the channel boundary coordinate instructions are analyzed by Beidou navigation, and the seeding machine is driven to adjust the working width of corn seeding according to the channel boundary coordinate instructions;

[0011] Based on the sowing width value of the corn, the single irrigation amount of shallow buried drip irrigation in wheat is simultaneously adjusted.

[0012] In a preferred embodiment, the step of normalizing and mapping the soil penetration resistance data to generate a soil penetration resistance score, and classifying the sowing condition level based on the soil penetration resistance score, includes:

[0013] Based on the maximum and minimum values ​​of soil penetration resistance data, a normalized mapping interval is constructed.

[0014] Based on the normalized mapping interval, the soil penetration resistance data is mapped to a scoring interval to generate a soil penetration resistance score value.

[0015] The soil penetration resistance scores were arithmetically averaged, and the sowing conditions were classified according to the average results of the soil penetration resistance scores.

[0016] In a preferred embodiment, classifying the sowing condition level based on the soil penetration resistance score includes:

[0017] A first scoring threshold and a second scoring threshold are preset, wherein the first scoring threshold is greater than the second scoring threshold;

[0018] When the soil penetration resistance score is greater than the first score threshold, the sowing condition level is classified as Level 1.

[0019] When the soil penetration resistance score is between the second score threshold and the first score threshold, the sowing condition level is classified as Level II.

[0020] When the soil penetration resistance score is less than the second score threshold, the sowing condition level is classified as Level 3.

[0021] In a preferred embodiment, the joint analysis of the wheat grain moisture content data and the soil slope data is performed to construct a slope moisture content correlation matrix, and the dimensionality of the slope moisture content correlation matrix is ​​expanded to obtain interference feature factors for corn sowing, including:

[0022] The wheat grain moisture content data and the soil slope data were normalized respectively to obtain normalized moisture content values ​​and normalized slope values.

[0023] Perform a Cartesian product operation on the normalized moisture content value and the normalized slope value to construct a slope-moisture content correlation matrix indexed by the moisture content dimension and the slope dimension;

[0024] Using a preset spatial interpolation algorithm, the slope moisture content correlation matrix is ​​expanded into a three-dimensional tensor containing spatial location information, and the three-dimensional tensor is used as an interference feature factor for corn planting.

[0025] In a preferred embodiment, the step of expanding the dimension of the slope moisture content correlation matrix using a preset spatial interpolation algorithm further includes:

[0026] Obtain the BeiDou coordinate information of sampling points in the field, and establish a spatial mapping relationship between the BeiDou coordinate information and the elements in the slope moisture content correlation matrix;

[0027] Using the BeiDou coordinate information as a spatial reference, spatial interpolation is performed on the slope moisture content correlation matrix to generate a three-dimensional interference feature tensor corresponding to the grid coordinates in the field.

[0028] In a preferred embodiment, the step of inputting the wheat row spacing data, the sowing condition level, and the interference feature factor into a preset decision rule base to generate the sowing width value and sowing density value for corn sowing includes:

[0029] The pre-defined decision rule base includes row spacing hierarchical rules and span-density mapping tables;

[0030] The row spacing data is input into the row spacing classification rules to obtain the row spacing level value;

[0031] The row spacing level value, the sowing condition level, and the interference feature factor are concatenated to form a decision feature vector;

[0032] The decision feature vector is matched and retrieved with the sowing width-density mapping table to obtain the initial value of the sowing width and the initial value of the sowing density corresponding to the decision feature vector;

[0033] The initial values ​​of the sowing width and sowing density are smoothed and filtered to generate the sowing width and sowing density values ​​for corn sowing.

[0034] In a preferred embodiment, the spread-density mapping table includes:

[0035] To obtain data on the optimal sowing width and optimal sowing density under different wheat row spacing, different sowing condition levels, and different combinations of interference characteristic factors in historical field trials in the field.

[0036] A decision tree classification model is constructed using the wheat row spacing data, sowing condition level, and interference feature factors as input variables, and the optimal sowing width and optimal sowing density as output variables.

[0037] Extract the decision paths of the leaf nodes of the decision tree classification model, and store the correspondence between the input variables and output variables in the decision tree classification model into the spread-density mapping table.

[0038] In a preferred embodiment, the step of converting the sowing width value into channel boundary coordinates of the seeder, parsing the channel boundary coordinates using BeiDou navigation, and driving the seeder to adjust the working width of corn sowing according to the channel boundary coordinates includes:

[0039] Using the current BeiDou positioning coordinates of the seeding machine as the reference point and the working direction of the seeding machine as the axis, half of the sowing width of the corn seeding is used as the unilateral offset. The unilateral offset is applied to the left and right sides perpendicular to the working direction, and the left boundary coordinates and right boundary coordinates are calculated by superimposing the offsets.

[0040] The left and right boundary coordinates are serialized and encapsulated to generate a channel boundary coordinate instruction containing coordinate fields and verification fields;

[0041] The channel boundary coordinate command is sent to the Beidou navigation terminal of the seeding machine. The Beidou navigation terminal extracts the left boundary coordinate and right boundary coordinate from the channel boundary coordinate command, and calculates the lateral deviation and heading deviation between the machine centerline and the left and right boundary coordinates based on the real-time Beidou positioning data of the seeding machine.

[0042] Based on the lateral deviation value and the heading deviation value, the steering correction amount of the automatic navigation controller in the seeder is calculated, and the working posture of the seeder is dynamically adjusted according to the steering correction amount, so that the working width of the seeder is dynamically adapted to the channel range defined by the left boundary coordinate and the right boundary coordinate.

[0043] In a preferred embodiment, the step of simultaneously controlling a single irrigation of wheat using shallow-buried drip irrigation based on the sowing width value of the corn seed includes:

[0044] Based on the sowing width value and the sowing width reference value of the corn, the sowing width change rate is calculated;

[0045] The rate of change of the sowing width is input into a preset irrigation volume control function to generate an irrigation volume adjustment coefficient;

[0046] Multiply the irrigation volume adjustment coefficient by the preset standard single irrigation volume to obtain the target single irrigation volume;

[0047] Based on the target single irrigation volume, the irrigation volume adjustment command for shallow-buried drip irrigation is executed.

[0048] In a preferred embodiment, the formula for the irrigation volume control function is:

[0049]

[0050] in, The target irrigation volume per irrigation session, This is the standard single irrigation volume. The sensitivity coefficient for broadcast width. This is the seeding width value for the currently generated corn seed. This is the baseline value for the broadcast width. As a sowing condition level adjustment factor, This represents the sowing condition level value.

[0051] Compared with the prior art, the present invention has the following beneficial effects:

[0052] 1. This invention achieves dynamic and precise matching of sowing width and sowing density by constructing a multi-source parameter fusion sowing width decision mechanism. In existing technologies, the wheat-maize intercropping sowing method lacks real-time perception and fusion capabilities for multi-dimensional parameters such as wheat row spacing spatial variation, soil penetration resistance, soil slope, and wheat grain moisture content, leading to a disconnect between sowing width settings and actual field conditions, and poor coordination of the plant population structure. By collecting wheat row spacing data, soil slope data, wheat grain moisture content data, and soil penetration resistance data, the soil penetration resistance data is normalized and mapped, and sowing condition levels are classified. Wheat grain moisture content data and soil slope data are embedded in the same feature space and expanded into three-dimensional interference feature factors. Then, the wheat row spacing data, sowing condition levels, and interference feature factors are input into a preset decision rule base to dynamically generate the sowing width and sowing density values ​​for maize sowing. This mechanism couples multi-source information, such as soil physical properties, crop growth status, and field spatial structure, to the sowing decision-making process, solving the technical problems of fixed sowing width and blind density in existing technologies, and significantly improving the coordination of intercropping population structure and the efficiency of inter-species resource utilization.

[0053] 2. This invention constructs a closed-loop operation system of decision-making, execution, and feedback by precisely linking BeiDou navigation with the sowing implement and combining synchronous regulation of irrigation volume with the sowing width. The generated sowing width value is converted into channel boundary coordinate commands containing left and right boundary coordinates. The BeiDou navigation terminal calculates the lateral and heading deviations between the implement's centerline and the boundary coordinates, driving the PID control unit of the automatic navigation controller to adjust the steering actuator in real time. This dynamically adapts the sowing implement's operating width to the range defined by the channel boundary coordinates, achieving a precise closed loop between sowing decisions and implement execution. Simultaneously, based on the sowing width value for corn, a preset irrigation volume control function synchronously regulates the single irrigation volume of shallow-buried drip irrigation in wheat, transmitting the demand for population density due to changes in sowing width to the water and fertilizer supply stage, forming a synergistic coupling mechanism between sowing width and irrigation volume. This design solves the technical problems of decision-making and execution disconnect and system silos in existing technologies, improving the overall accuracy and system efficiency of wheat-corn intercropping operations. Attached Figure Description

[0054] Figure 1 A flowchart illustrating an adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification, provided in an embodiment of the present invention.

[0055] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0056] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0057] This application provides an adaptive sowing amplitude control method for wheat-maize intercropping based on crop and soil parameter identification. The executing entity of this method includes, but is not limited to, at least one electronic device that can be configured to execute the method provided in this application, such as a server or a terminal. In other words, the adaptive sowing amplitude control method for wheat-maize intercropping based on crop and soil parameter identification can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0058] Reference Figure 1 The diagram shown is a flowchart illustrating an adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification, according to an embodiment of the present invention. In this embodiment, the adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification includes:

[0059] Data on wheat row spacing, soil slope, wheat grain moisture content, and soil penetration resistance were collected.

[0060] In this embodiment of the invention, before the corn sowing operation begins in the wheat-corn intercropping field, a row spacing sensor array, an inclination sensor, a near-infrared spectral moisture meter, and a soil penetration resistance sensor are fixedly installed on the front crossbeam of the sowing machine. The row spacing sensor array consists of multiple sets of lidar. Each set of lidar emits a laser beam towards the wheat plant canopy. After the laser beam contacts the wheat stem, it is reflected back to the sensor. The sensor calculates the horizontal distance between the sensor and the wheat plant based on the time difference between emission and reception. After continuous scanning, the center line positions of two adjacent rows of wheat are identified, the horizontal distance between the two row center lines is measured, and the wheat row spacing data is output in real time. The inclination sensor uses a microelectromechanical system inclination meter, which contains a miniature mass block and a capacitance detection circuit. When the sowing machine moves in the field, the inclination sensor detects the tilt angle of the machine's chassis relative to the horizontal plane. The change in capacitance is converted into a voltage signal, and the soil slope data of the current machine location is output. The near-infrared spectroscopy moisture analyzer continuously emits a near-infrared beam towards the wheat ears during the machine's movement. The beam penetrates the surface of the wheat grains and reflects back to the analyzer. The analyzer analyzes the absorption characteristics of the reflected spectrum at a specific wavelength, compares it with the internally stored moisture content characteristic spectral lines, and outputs the wheat grain moisture content data for the current sampling point. The soil penetration resistance sensor is installed on the support in front of the furrow opener of the seeder. A conical probe is installed at the front end of the sensor. During the machine's movement, the probe continuously penetrates vertically into the soil surface. The force-sensitive element inside the sensor detects the resistance encountered by the probe during penetration, converting the mechanical force into an electrical signal, and outputting the soil penetration resistance data. These four sensors synchronously collect data at a sampling frequency of fifty times per second and transmit the collected data in real time to the memory of the on-board control terminal via the vehicle bus.

[0061] The beneficial effects are as follows: By simultaneously collecting data on wheat row spacing, soil slope, wheat grain moisture content, and soil penetration resistance, a multi-dimensional field information perception system covering crop spatial structure, surface geometry, crop physiological state, and soil physical properties is constructed. The row spacing sensor array acquires the horizontal distance between wheat rows in real time, providing a spatial benchmark for the precise division of corn planting channels. The tilt sensor continuously detects the field surface slope, enabling subsequent sowing width decisions to adapt to sowing needs under different terrain conditions. The near-infrared spectroscopy moisture meter measures wheat grain moisture content in real time during machine movement, incorporating the degree of wheat growth's encroachment on corn sowing space into the decision-making process. The soil penetration resistance sensor continuously penetrates the soil to acquire resistance values, quantifying the constraining effect of soil compaction on sowing quality. The four sensors simultaneously collect data at high frequency and transmit it in real time to the vehicle-mounted control terminal, providing complete, real-time, and accurate input parameters for the dynamic decision-making of subsequent sowing width and density, solving the problems of single decision-making basis and insufficient perception dimensions in existing technologies.

[0062] The soil penetration resistance data is normalized and mapped to generate a soil penetration resistance score, and the sowing condition level is classified according to the soil penetration resistance score.

[0063] In this embodiment of the invention, the soil penetration resistance data is normalized and mapped to generate a soil penetration resistance score, and the sowing condition level is classified according to the soil penetration resistance score, including:

[0064] Based on the maximum and minimum values ​​of soil penetration resistance data, a normalized mapping interval is constructed.

[0065] Based on the normalized mapping interval, the soil penetration resistance data is mapped to a scoring interval to generate a soil penetration resistance score value.

[0066] The soil penetration resistance scores were arithmetically averaged, and the sowing conditions were classified according to the average results of the soil penetration resistance scores.

[0067] The classification of sowing conditions based on soil penetration resistance scores includes:

[0068] A first scoring threshold and a second scoring threshold are preset, wherein the first scoring threshold is greater than the second scoring threshold;

[0069] When the soil penetration resistance score is greater than the first score threshold, the sowing condition level is classified as Level 1.

[0070] When the soil penetration resistance score is between the second score threshold and the first score threshold, the sowing condition level is classified as Level II.

[0071] When the soil penetration resistance score is less than the second score threshold, the sowing condition level is classified as Level 3.

[0072] After the vehicle-mounted control terminal reads all the sampling point data collected by the soil penetration resistance sensor from the memory, it iterates through all the soil penetration resistance data and extracts the maximum and minimum values. The terminal uses the maximum value as the upper limit of the normalized mapping interval and the minimum value as the lower limit of the normalized mapping interval, mapping the soil penetration resistance data of each sampling point to a scoring interval of zero to one according to a linear proportional relationship. During mapping, the original value of each sampling point is subtracted from the minimum value and then divided by the difference between the maximum and minimum values ​​to obtain the soil penetration resistance score value corresponding to that sampling point.

[0073] The terminal sums the soil penetration resistance scores of all sampling points within the same sowing channel and divides the sum by the total number of sampling points to obtain the average soil penetration resistance score for that sowing channel. The vehicle-mounted control terminal pre-stores a first scoring threshold and a second scoring threshold, with the first threshold set to 0.7 and the second threshold set to 0.3.

[0074] The terminal compares the average soil penetration resistance score with the first and second scoring thresholds: when the average score is greater than the first scoring threshold, the terminal classifies the sowing condition level of the sowing channel as Level 1, indicating that the soil conditions are good and suitable for sowing; when the average score is between the second and first scoring thresholds, the terminal classifies it as Level 2, indicating that the soil conditions are moderate and there is some sowing resistance; when the average score is less than the second scoring threshold, the terminal classifies it as Level 3, indicating that the soil conditions are poor and the sowing resistance is high.

[0075] The beneficial effects are as follows: the above steps convert the original soil penetration resistance data into dimensionless score values ​​through normalization mapping, eliminating the influence of differences in soil background values ​​between different fields on decision-making, and providing a unified comparison benchmark for the classification of sowing condition levels. Arithmetic averaging eliminates the interference of measurement errors from individual sampling points on the level classification, ensuring that the sowing condition levels accurately reflect the overall soil condition of the entire sowing channel. Based on a preset threshold, the continuously changing soil penetration resistance is discretized into three levels of sowing condition levels, providing clear and comparable soil condition input parameters for subsequent sowing width decisions, and solving the problem in existing technologies where soil physical properties are difficult to quantify and integrate into sowing decisions.

[0076] The wheat grain moisture content data and the soil slope data are jointly analyzed to construct a slope moisture content correlation matrix, and the dimensions of the slope moisture content correlation matrix are expanded to obtain interference characteristic factors for corn sowing.

[0077] In this embodiment of the invention, the wheat grain moisture content data and the soil slope data are jointly analyzed to construct a slope moisture content correlation matrix. The slope moisture content correlation matrix is ​​then expanded in dimension to obtain interference feature factors for corn sowing, including:

[0078] The wheat grain moisture content data and the soil slope data were normalized respectively to obtain normalized moisture content values ​​and normalized slope values.

[0079] Perform a Cartesian product operation on the normalized moisture content value and the normalized slope value to construct a slope-moisture content correlation matrix indexed by the moisture content dimension and the slope dimension;

[0080] Using a preset spatial interpolation algorithm, the slope moisture content correlation matrix is ​​expanded into a three-dimensional tensor containing spatial location information, and the three-dimensional tensor is used as an interference feature factor for corn planting.

[0081] The slope moisture content correlation matrix is ​​expanded in dimensionality using a preset spatial interpolation algorithm, and the method further includes:

[0082] Obtain the BeiDou coordinate information of sampling points in the field, and establish a spatial mapping relationship between the BeiDou coordinate information and the elements in the slope moisture content correlation matrix;

[0083] Using the BeiDou coordinate information as a spatial reference, spatial interpolation is performed on the slope moisture content correlation matrix to generate a three-dimensional interference feature tensor corresponding to the grid coordinates in the field.

[0084] The vehicle-mounted control terminal reads wheat grain moisture content data from its memory. This data consists of moisture content values ​​collected by a near-infrared spectroscopy moisture analyzer at various sampling points in the field, with each value corresponding to the geographical location of a sampling point. The terminal also reads soil slope data, which consists of the surface tilt angle values ​​collected by an inclination sensor at the same sampling point. The terminal iterates through all the collected moisture content values, comparing each value with the currently recorded maximum and minimum values. When a larger value is encountered, the maximum value record is updated; when a smaller value is encountered, the minimum value record is updated. After the iteration is complete, the maximum and minimum values ​​are determined. The terminal then iterates through all the collected moisture content values ​​again, subtracting the minimum value from the moisture content value of each sampling point, and then dividing the result by the difference between the maximum and minimum values ​​to obtain the normalized moisture content value for that sampling point. The terminal processes soil slope data in the same way, iterates through all slope collection values, compares each slope value with the currently recorded maximum and minimum values ​​during the iteration, determines the maximum and minimum values ​​among all slope values, it iterates through all slope collection values ​​again, subtracts the minimum value from the slope value of each sampling point, and then divides the subtraction result by the difference between the maximum and minimum values ​​to obtain the normalized slope value of that sampling point.

[0085] After normalization, the terminal performs a Cartesian product operation on the normalized moisture content and normalized slope values. The terminal divides the interval from zero to one into multiple equally wide sub-intervals as the moisture content dimension and the interval from zero to one into multiple equally wide sub-intervals as the slope dimension. The terminal creates a two-dimensional matrix where the number of rows equals the number of sub-intervals for the moisture content dimension, and the number of columns equals the number of sub-intervals for the slope dimension. The row index corresponds to the sequence number of the moisture content sub-interval, and the column index corresponds to the sequence number of the slope sub-interval. The terminal iterates through all sampling points. For each sampling point, it determines the sequence number of the moisture content sub-interval to which its normalized moisture content value falls, using this as the row index, and determines the sequence number of the slope sub-interval to which its normalized slope value falls, using this as the column index. The counter value at that position in the matrix is ​​then incremented by one. After the terminal completes the statistics of all sampling points, it traverses each position in the matrix, divides the count value of that position by the total number of all sampling points, and obtains the correlation characterization value of that position. This two-dimensional matrix is ​​the slope moisture content correlation matrix. Each element in the matrix represents the density of the distribution of sampling points under the combination of the corresponding moisture content interval and slope interval.

[0086] The terminal acquires BeiDou coordinate information from multiple sampling points in the field. Each sampling point simultaneously possesses BeiDou coordinate information and its corresponding element position in the slope-moisture content correlation matrix. The terminal establishes a spatial mapping relationship between these BeiDou coordinate information and the elements in the slope-moisture content correlation matrix; that is, the BeiDou coordinates of each sampling point uniquely correspond to an element in the slope-moisture content correlation matrix, and the value of this element is the correlation characterization value of that sampling point. Using the BeiDou coordinate information of all sampling points as a spatial reference, the terminal divides the field into grids, sets a fixed grid spacing, and generates a uniform grid point array covering the entire field. Each grid point has independent north and east coordinates.

[0087] The terminal acquires BeiDou coordinate information from multiple sampling points in the field. Using these coordinate points as a spatial reference, the field is interpolated in a gridded manner using the inverse distance weighted interpolation method. During the interpolation process, its power exponent is set. The value is 2, which represents the number of neighboring points searched. There are 12 sampling points, and no variogram is used. The specific interpolation calculation method is as follows: For each grid point in the field, the terminal traverses all known sampling points, calculates the Euclidean distance between the BeiDou coordinates of each known sampling point and the coordinates of the grid point, and selects the 12 sampling points with the smallest Euclidean distance as neighboring sampling points; for each selected neighboring sampling point, the terminal calculates the reciprocal of the square of the Euclidean distance between the sampling point and the grid point as the weight value, multiplies the element value of the slope and moisture content correlation matrix corresponding to each neighboring sampling point with its weight value, and obtains the weighted contribution value of the sampling point; the terminal sums the weighted contribution values ​​of all neighboring sampling points and divides them by the sum of all weight values ​​to obtain the correlation characterization value at the grid point. Finally, a three-dimensional tensor is generated, whose first and second dimensions correspond to the north and east coordinates of the grid point, and the third dimension stores the correlation characterization value of the grid point. This three-dimensional tensor is the interference feature factor of corn planting.

[0088] The beneficial effects are as follows: by embedding two heterogeneous parameters, wheat grain moisture content and soil slope, into the same feature space and constructing a correlation matrix, a joint representation of crop physiological state and surface geometry is achieved, enabling unified quantification of interspecific competition interference and terrain constraint interference affecting maize planting space. By introducing BeiDou coordinate information for spatial interpolation of the correlation matrix, the correlation representation values ​​of discrete sampling points are expanded into a continuous three-dimensional tensor covering the entire field, giving the interference characteristic factors spatial distribution characteristics and accurately reflecting the spatial heterogeneity of interference intensity at different field locations. This three-dimensional tensor serves as the input parameter for subsequent sowing width decisions, providing an accurate spatial benchmark for spatially differentiated decisions on sowing width and density, solving the problem of difficulty in spatially quantifying interference factors in existing technologies.

[0089] The wheat row spacing data, the sowing condition level, and the interference feature factor are input into a preset decision rule base to generate the sowing width value and sowing density value for corn sowing.

[0090] In this embodiment of the invention, the wheat row spacing data, the sowing condition level, and the interference feature factor are input into a preset decision rule base to generate the sowing width value and sowing density value for corn sowing, including:

[0091] The pre-defined decision rule base includes row spacing hierarchical rules and span-density mapping tables;

[0092] The row spacing data is input into the row spacing classification rules to obtain the row spacing level value;

[0093] The row spacing level value, the sowing condition level, and the interference feature factor are concatenated to form a decision feature vector;

[0094] The decision feature vector is matched and retrieved with the sowing width-density mapping table to obtain the initial value of the sowing width and the initial value of the sowing density corresponding to the decision feature vector;

[0095] The initial values ​​of the sowing width and sowing density are smoothed and filtered to generate the sowing width and sowing density values ​​for corn sowing.

[0096] The broadcast width-density mapping table includes:

[0097] To obtain data on the optimal sowing width and optimal sowing density under different wheat row spacing, different sowing condition levels, and different combinations of interference characteristic factors in historical field trials in the field.

[0098] A decision tree classification model is constructed using the wheat row spacing data, sowing condition level, and interference feature factors as input variables, and the optimal sowing width and optimal sowing density as output variables.

[0099] Extract the decision paths of the leaf nodes of the decision tree classification model, and store the correspondence between the input variables and output variables in the decision tree classification model into the spread-density mapping table.

[0100] The vehicle-mounted control terminal reads wheat row spacing data from its memory. This data represents the horizontal distance between adjacent rows of wheat collected by the row spacing sensor array at various sampling points in the field. The terminal has a pre-set row spacing grading rule that divides the wheat row spacing values ​​into three grade intervals: the first grade interval corresponds to wheat row spacing values ​​greater than 45 cm, the second grade interval corresponds to wheat row spacing values ​​between 35 cm and 45 cm, and the third grade interval corresponds to wheat row spacing values ​​less than 35 cm. The terminal inputs the wheat row spacing data into the row spacing grading rule, determines which grade interval the current sampling point's wheat row spacing value falls into, and outputs the grade value corresponding to that grade interval as the row spacing grade value for that sampling point.

[0101] The terminal simultaneously reads the sowing condition level and interference feature factor generated in the previous steps from the memory. The sowing condition level is a discrete value, taking values ​​of level one, two, or three. The interference feature factor is a three-dimensional tensor containing spatial location information, which is composed of the correlation representation values ​​between grid point coordinates and corresponding grid points. The terminal performs feature concatenation of the row spacing level value, sowing condition level, and interference feature factor. Specifically, the row spacing level value is placed at the beginning of the feature vector, the sowing condition level value is placed at the second position, and the correlation representation values ​​of each grid point in the three-dimensional tensor of the interference feature factor are expanded sequentially according to the grid point's north coordinates in ascending order and east coordinates in ascending order, forming a continuous feature sequence, which is placed at the end of the feature vector, combining to form a complete decision feature vector.

[0102] The terminal has a pre-set sowing width-density mapping table, which is constructed using a decision tree classification model. The construction process of the decision tree classification model is as follows: The terminal acquires historical field trial data from the field, which includes the optimal sowing width and optimal sowing density verified in the field for multiple experimental points under different wheat row spacing data, different sowing condition levels, and different combinations of interference feature factors. The terminal uses wheat row spacing data, sowing condition levels, and interference feature factors as input variables, and the optimal sowing width and optimal sowing density as output variables to construct the decision tree classification model. The construction of the decision tree classification model adopts a recursive splitting method. The terminal traverses all input variables, selects the variable that minimizes the uncertainty of the output variable as the splitting node, divides the corresponding numerical range of the variable into multiple branches, and continues the splitting process on each branch until the output variables on each branch tend to be consistent or reach a preset stopping condition. After the decision tree classification model is constructed, the terminal extracts the decision paths of the leaf nodes of the model. Each leaf node decision path corresponds to a set of input variable value range combinations and the corresponding output variable values. The terminal stores the correspondence between the value ranges of these input variables and the values ​​of the output variables in the amplitude-density mapping table, forming query entries.

[0103] The terminal matches the generated decision feature vector with the sowing width-density mapping table. During the matching retrieval, the terminal traverses each query entry in the sowing width-density mapping table, sequentially comparing whether the row spacing level value in the decision feature vector falls within the row spacing level interval corresponding to the entry, whether the sowing condition level value in the decision feature vector is equal to the sowing condition level value corresponding to the entry, and whether the correlation representation values ​​of each grid point of the interference feature factor in the decision feature vector fall within the value interval of the interference feature factor corresponding to the entry. The terminal calculates the matching degree between the decision feature vector and each query entry, selects the query entry with the highest matching degree, and extracts the initial value of the sowing width and the initial value of the sowing density corresponding to that entry.

[0104] For the current planting row, the terminal calculates a weighted average of the initial seed width of the current row and the seed width of the previous row, with the initial seed width of the current row weighted at 70% and the seed width of the previous row weighted at 30%. The result of this weighted average is used as the seed width value of the current row. Similarly, the terminal calculates a weighted average of the initial seed density of the current row and the seed density of the previous row, with the initial seed density of the current row weighted at 70% and the seed density of the previous row weighted at 30%. The result of this weighted average is used as the seed density value of the current row. The terminal outputs the calculated seed width and seed density values ​​as the seed width and seed density values ​​for corn planting.

[0105] The beneficial effects are as follows: By converting continuous wheat row spacing data into discrete row spacing level values ​​through row spacing grading rules, the decision rule base can handle the spatially variable wheat row distribution characteristics. By concatenating row spacing level values, sowing condition levels, and interference feature factors, three heterogeneous types of information—soil physical properties, crop spatial structure, and crop growth status—are integrated into a unified decision feature vector, providing multi-dimensional input for joint decision-making on sowing width and density. The sowing width-density mapping table constructed through a decision tree classification model transforms the optimal sowing width and density experience verified in historical field trials into searchable decision knowledge, enabling the decision-making process to be matched based on real agronomic experience. This eliminates abrupt changes in sowing width and density between adjacent sowing rows, ensuring the continuity of sowing operations and the gradual change in population structure, and solving the problem of existing technologies having a single basis for sowing width and density decisions and being unable to dynamically adapt to field spatial heterogeneity.

[0106] The seeding width value is converted into channel boundary coordinate instructions for the seeding machine, and the channel boundary coordinate instructions are analyzed by Beidou navigation, and the seeding machine is driven to adjust the working width of corn seeding according to the channel boundary coordinate instructions;

[0107] In this embodiment of the invention, the sowing width value is converted into channel boundary coordinate instructions for the sowing machine, and the channel boundary coordinate instructions are parsed using BeiDou navigation. The sowing machine is then driven to adjust the working width of the corn sowing according to the channel boundary coordinate instructions, including:

[0108] Using the current BeiDou positioning coordinates of the seeding machine as the reference point and the working direction of the seeding machine as the axis, half of the sowing width of the corn seeding is used as the unilateral offset. The unilateral offset is applied to the left and right sides perpendicular to the working direction, and the left boundary coordinates and right boundary coordinates are calculated by superimposing the offsets.

[0109] The left and right boundary coordinates are serialized and encapsulated to generate a channel boundary coordinate instruction containing coordinate fields and verification fields;

[0110] The channel boundary coordinate command is sent to the Beidou navigation terminal of the seeding machine. The Beidou navigation terminal extracts the left boundary coordinate and right boundary coordinate from the channel boundary coordinate command, and calculates the lateral deviation and heading deviation between the machine centerline and the left and right boundary coordinates based on the real-time Beidou positioning data of the seeding machine.

[0111] Based on the lateral deviation value and the heading deviation value, the steering correction amount of the automatic navigation controller in the seeder is calculated, and the working posture of the seeder is dynamically adjusted according to the steering correction amount, so that the working width of the seeder is dynamically adapted to the channel range defined by the left boundary coordinate and the right boundary coordinate.

[0112] The vehicle-mounted control terminal obtains the BeiDou positioning coordinates of the current seeding machine as a reference point. These BeiDou positioning coordinates are the north and east coordinates of the machine's center point in the geodetic coordinate system, calculated in real time by the BeiDou navigation terminal from satellite signals received.

[0113] The terminal simultaneously acquires the operating direction of the seeding implement as its axis. This operating direction is the angle between the implement's forward direction, calculated by the BeiDou navigation terminal based on the implement's positioning coordinates over multiple consecutive moments, and true north. The terminal divides the corn sowing width value generated in the previous steps by two to obtain the unilateral offset. Using the reference point as the starting point and the operating direction as the axis, the terminal determines the left and right directions perpendicular to the operating direction.

[0114] The terminal adds the northward component of the left offset perpendicular to the direction of operation to the northward coordinates of the reference point to obtain the northward coordinates of the left boundary. It adds the eastward component of the left offset perpendicular to the direction of operation to the eastward coordinates of the reference point to obtain the eastward coordinates of the left boundary. The northward coordinates of the left boundary and the eastward coordinates of the left boundary are combined to form the left boundary coordinates.

[0115] The terminal subtracts the north component of the right offset perpendicular to the working direction from the north coordinate of the reference point to obtain the north coordinate of the right boundary. It also subtracts the east component of the right offset perpendicular to the working direction from the east coordinate of the reference point to obtain the east coordinate of the right boundary. The north coordinate of the right boundary and the east coordinate of the right boundary are combined to form the right boundary coordinate.

[0116] The terminal serializes and encapsulates the left and right boundary coordinates. The terminal constructs instruction data according to a preset instruction frame format, which consists of a frame header identifier, a left boundary north coordinate field, a left boundary east coordinate field, a right boundary north coordinate field, a right boundary east coordinate field, a check field, and a frame tail identifier. The terminal converts the left boundary north coordinate to a hexadecimal byte sequence and writes it into the left boundary north coordinate field; it also converts the left boundary east coordinate to a hexadecimal byte sequence and writes it into the left boundary east coordinate field; similarly, it converts the right boundary north coordinate to a hexadecimal byte sequence and writes it into the right boundary north coordinate field; and finally, it converts the right boundary east coordinate to a hexadecimal byte sequence and writes it into the right boundary east coordinate field.

[0117] The terminal performs a summation operation on all bytes after the frame header identifier and before the check field, writes the sum as the check value into the check field, and finally writes it into the frame tail identifier to generate a complete channel boundary coordinate command. The terminal sends this command to the Beidou navigation terminal of the seeding machine via the serial communication interface.

[0118] After receiving the channel boundary coordinate command, the BeiDou navigation terminal extracts the left and right boundary coordinates from the command frame. The BeiDou navigation terminal parses the command frame, locating the left boundary north coordinate field, the left boundary east coordinate field, the right boundary north coordinate field, and the right boundary east coordinate field. It converts the hexadecimal byte sequence of each field into north and east coordinate values, thus recovering the left and right boundary coordinates. Simultaneously, the BeiDou navigation terminal acquires the real-time BeiDou positioning data of the seeding machine, which represents the current north and east coordinates of the machine's center point.

[0119] The BeiDou navigation terminal calculates the lateral deviation and heading deviation between the equipment's centerline and the coordinates of its left and right boundaries. The lateral deviation is calculated as follows: the BeiDou navigation terminal calculates the perpendicular distance from the equipment's center point coordinates to the line connecting the left and right boundary coordinates, determines whether the equipment's center point is to the left or right of this line, and obtains the lateral deviation value with a directional sign. The heading deviation is calculated as follows: the BeiDou navigation terminal calculates the angle between the direction of the line connecting the left and right boundary coordinates and true north as the target heading angle, and subtracts the equipment's current heading angle from the target heading angle to obtain the heading deviation value.

[0120] The BeiDou navigation terminal inputs the lateral deviation and heading deviation values ​​into the PID control unit of the automatic navigation controller of the seeder. The proportional element of the PID control unit multiplies the lateral deviation value by a proportional coefficient to generate a lateral proportional correction, and multiplies the heading deviation value by a proportional coefficient to generate a heading proportional correction. The integral element of the PID control unit accumulates and integrates the lateral deviation values, then multiplies them by an integral coefficient to generate a lateral integral correction, and accumulates and integrates the heading deviation values, then multiplies them by an integral coefficient to generate a heading integral correction.

[0121] The derivative element of the PID control unit differentiates the rate of change of the lateral deviation value and multiplies it by a derivative coefficient to generate the lateral derivative correction. It also differentiates the rate of change of the heading deviation value and multiplies it by a derivative coefficient to generate the heading derivative correction. The PID control unit then sums the lateral proportional correction, lateral integral correction, and lateral derivative correction to obtain the lateral correction, and sums the heading proportional correction, heading integral correction, and heading derivative correction to obtain the heading correction. Finally, the lateral and heading corrections are combined according to preset weights to obtain the final steering correction.

[0122] The PID control unit converts the steering correction amount into the duty cycle of a pulse width modulation signal and outputs it to the drive circuit of the steering actuator motor. After receiving the drive signal, the steering actuator motor rotates, which drives the steering tie rod to move. The steering tie rod pushes the front wheel steering knuckle to rotate, changing the deflection angle of the front wheel, so that the working width of the seeder dynamically adapts to the channel range defined by the left and right boundary coordinates.

[0123] The beneficial effects are as follows: This step establishes a spatial coordinate relationship between sowing decisions and implement execution by converting the sowing width value into channel boundary coordinate commands recognizable by the BeiDou Navigation Satellite System, enabling the sowing width decision result to be accurately mapped to the actual travel path of the implement. By calculating the lateral and heading deviation values ​​through the BeiDou navigation terminal, the deviation between the actual trajectory of the implement and the target channel is quantified in real time, providing precise feedback input for closed-loop control. The PID control unit calculates the steering correction based on the deviation value and drives the steering motor, achieving real-time dynamic adjustment of the implement's working posture. This ensures that the sowing implement always operates within the channel range defined by the left and right boundary coordinates, solving the problem of independent sowing width settings and implement working trajectories, and the disconnect between decision-making and execution in existing technologies. This improves the width control accuracy and spatial consistency of the corn sowing population.

[0124] Based on the sowing width value of the corn, the single irrigation amount of shallow buried drip irrigation in wheat is simultaneously adjusted.

[0125] In this embodiment of the invention, based on the sowing width value of the corn seeding, the single irrigation amount of shallow-buried drip irrigation in wheat is simultaneously controlled, including:

[0126] Based on the sowing width value and the sowing width reference value of the corn, the sowing width change rate is calculated;

[0127] The rate of change of the sowing width is input into a preset irrigation volume control function to generate an irrigation volume adjustment coefficient;

[0128] Multiply the irrigation volume adjustment coefficient by the preset standard single irrigation volume to obtain the target single irrigation volume;

[0129] Based on the target single irrigation volume, the irrigation volume adjustment command for shallow-buried drip irrigation is executed.

[0130] The formula for the irrigation volume regulation function is as follows:

[0131]

[0132] in, The target irrigation volume per irrigation session, This is the standard single irrigation volume. The sensitivity coefficient for broadcast width. This is the seeding width value for the currently generated corn seed. This is the baseline value for the broadcast width. As a sowing condition level adjustment factor, This represents the sowing condition level value.

[0133] The vehicle-mounted control terminal reads the currently generated corn sowing width value from the memory. This sowing width value is the width of the corn sowing row in the current row, output after processing with a moving average filter. The window size of the moving average filter is set to 3, meaning the current sowing width value is the arithmetic mean of the initial sowing width values ​​of the current row, the previous row, and the next row. Simultaneously, the terminal reads a pre-stored sowing width reference value, which is the standard sowing width determined through field trials under standard intercropping conditions. The terminal subtracts the sowing width reference value from the current sowing width value to obtain the sowing width difference. Dividing this difference by the reference value yields the sowing width change rate.

[0134] The sowing width change rate is input into a preset irrigation volume control function. This function uses the sowing width change rate as its input parameter and internally processes it logarithmically. When the sowing width change rate is positive, the function outputs an irrigation volume adjustment coefficient greater than one, which increases as the sowing width change rate increases. Conversely, when the sowing width change rate is negative, the function outputs an irrigation volume adjustment coefficient less than one, which decreases as the sowing width change rate decreases. The irrigation volume control function also receives the sowing condition level value generated in the previous step as an input parameter. The higher the sowing condition level value, the greater the deviation of the irrigation volume adjustment coefficient from one. Specifically, the adjustment range of the irrigation volume adjustment coefficient is largest when the sowing condition level is level three, and smallest when the sowing condition level is level one. The terminal substitutes the sowing width change rate and the sowing condition level value into the irrigation volume control function, performs internal calculations, and outputs the irrigation volume adjustment coefficient.

[0135] The system reads the preset standard single irrigation volume, which is the baseline water usage for each irrigation determined based on the crop's water requirement under the standard intercropping model. The terminal multiplies the standard single irrigation volume by the irrigation volume adjustment coefficient to obtain the target single irrigation volume. When the irrigation volume adjustment coefficient is greater than one, the target single irrigation volume is greater than the standard single irrigation volume; when the irrigation volume adjustment coefficient is less than one, the target single irrigation volume is less than the standard single irrigation volume.

[0136] The target irrigation volume for a single application is transmitted to the controller of the shallow-buried drip irrigation system via a wireless communication module. Upon receiving the target volume, the controller reads the feedback value from the current irrigation volume sensor to obtain the current actual irrigation volume. The controller compares the target volume with the current actual volume and calculates the irrigation volume deviation. Based on the deviation, the controller uses an internal adjustment algorithm to calculate the valve opening adjustment, outputs a control signal to the electric valve, drives it to rotate, changes the valve opening angle, and regulates the water flow through the valve. Simultaneously, the controller adjusts the water supply duration according to the target volume; when the target volume increases, the water supply duration is extended; when the target volume decreases, the water supply duration is shortened. The controller continuously monitors the feedback value from the irrigation volume sensor until the actual irrigation volume reaches the target volume, completing the execution of the irrigation volume adjustment command.

[0137] The sowing width value of the corn seeding currently generated by the vehicle control terminal is derived from the sowing width value generated by the decision rule base and processed by the moving average filter in the previous step. The window size of the moving average filter is 3.

[0138] The seeding width baseline value is a pre-stored baseline parameter, which is the standard seeding width determined through field trials in the standard intercropping pattern.

[0139] The seeding width sensitivity coefficient is a pre-stored empirical coefficient, which is a constant determined by fitting multiple sets of field test data. The value range of this parameter is 0.1~0.3.

[0140] The sowing condition level adjustment factor read by the terminal is a pre-stored empirical constant. This empirical constant is a constant determined through irrigation response experiments under different soil conditions, and the value range of this parameter is 1.1~1.3.

[0141] The sowing condition level is a level value obtained by averaging the soil penetration resistance scores from the previous steps, and this level value can be classified as Level 1, Level 2, or Level 3. The standard single irrigation amount read by the terminal is a pre-stored baseline irrigation parameter, which is a standard irrigation amount determined through crop water requirement experiments.

[0142] The irrigation volume control function expresses the mathematical relationship between the target single irrigation volume, the standard single irrigation volume, the sowing width change rate, and the sowing condition level. The function first calculates the sowing width change rate, which is the difference between the currently generated corn sowing width value and the sowing width benchmark value, divided by the sowing width benchmark value. This change rate reflects the degree of deviation of the current sowing width from the standard sowing width.

[0143] The target single irrigation volume is calculated by the terminal. The data is transmitted to the controller of the shallow-buried drip irrigation system via a wireless communication module. The controller adjusts the opening of the electric valve and the water supply duration according to the target irrigation volume, executing the irrigation volume adjustment command. It should be noted that in this embodiment, the seeding width reference value... The value is 30cm, which is the current width of the seeding area. The value range is 20cm~35cm, therefore the seeding width variation rate The value range is -0.33 to 0.17, which satisfies... Ensure the logarithmic function Valid within the defined domain. According to the experimental data of this embodiment, this irrigation regulation strategy increases the amount of irrigation per irrigation by 3%-5% for every 5cm increase in seeding width, achieving coordinated adaptation between water and fertilizer supply and intercropping population structure.

[0144] The function performs a logarithmic operation on the result of adding the sowing width change rate to the given value. The logarithmic result is then multiplied by the sowing width sensitivity coefficient, and finally by the sowing condition level adjustment factor raised to the power of its value, yielding the adjustment term. The function adds this adjustment term to the given value to obtain the overall adjustment coefficient. Finally, the standard single irrigation amount is multiplied by this overall adjustment coefficient to obtain the target single irrigation amount.

[0145] In the irrigation volume control function, when the currently generated corn seeding width value equals the baseline seeding width value, the seeding width change rate is zero. Adding one to the seeding width change rate equals one, the logarithmic result is zero, the adjustment term is zero, the overall adjustment coefficient is one, and the target single irrigation volume equals the standard single irrigation volume. When the currently generated corn seeding width value is greater than the baseline seeding width value, the seeding width change rate is positive. Adding one to the seeding width change rate is greater than one, the logarithmic result is positive, the adjustment term is positive, the overall adjustment coefficient is greater than one, the target single irrigation volume is greater than the standard single irrigation volume, and the larger the seeding width change rate, the larger the logarithmic result, the larger the overall adjustment coefficient, and the larger the target single irrigation volume. When the current generated corn sowing width value is less than the sowing width benchmark value, the sowing width change rate is negative. When one is added to the sowing width change rate, which is less than one, the logarithmic result is negative, the adjustment term is negative, the overall adjustment coefficient is less than one, the target single irrigation amount is less than the standard single irrigation amount, and the smaller the sowing width change rate, the smaller the logarithmic result, the smaller the overall adjustment coefficient, and the smaller the target single irrigation amount.

[0146] When the sowing condition level adjustment factor is greater than one, the sowing condition level value acts as an index on this adjustment factor. When the sowing condition level is level one, the exponential result of the adjustment factor is the smallest, the exponential factor in the adjustment term is the smallest, and the overall adjustment coefficient deviates from one the smallest. When the sowing condition level is level three, the exponential result of the adjustment factor is the largest, the exponential factor in the adjustment term is the largest, and the overall adjustment coefficient deviates from one the largest. Therefore, the higher the sowing condition level value, i.e., the worse the soil conditions, the stronger the adjustment effect of the sowing width change rate on the target single irrigation amount, and the greater the deviation of the target single irrigation amount from the standard single irrigation amount. Conversely, the lower the sowing condition level value, i.e., the better the soil conditions, the weaker the adjustment effect of the sowing width change rate on the target single irrigation amount, and the smaller the deviation of the target single irrigation amount from the standard single irrigation amount.

[0147] The beneficial effects are as follows: This step establishes a quantitative correlation between sowing and irrigation decisions by comparing the sowing width value of maize with a benchmark sowing width value to calculate the rate of change in sowing width. This allows irrigation adjustments to respond to the impact of sowing width changes on plant density. By introducing sowing condition levels as adjustment factors into the irrigation volume control function, the irrigation volume adjustment range can be differentiated according to soil conditions. The adjustment range increases when soil conditions are poor and decreases when soil conditions are good, achieving coordinated adaptation between irrigation volume and overall field conditions. By sending the target single irrigation volume to the shallow-buried drip irrigation system controller for adjustment, a real-time linkage between sowing decisions and irrigation execution is formed. This solves the problem of lack of coordinated water and fertilizer supply response after sowing width adjustment in existing technologies and improves the matching accuracy of water and fertilizer resources with intercropping plant structure.

[0148] In this embodiment, a core demonstration area in the Northwest Oasis Irrigation District was used as the experimental field. This field is a typical wheat-maize intercropping area, covering an area of ​​approximately 870 mu (about 58 hectares). Before maize sowing, a row spacing sensor array was fixedly installed on the front crossbeam of the sowing machine. The laser radar had a sampling frequency of 50Hz, the tilt sensor had an accuracy of 0.1°, the near-infrared spectroscopy moisture analyzer had a wavelength range of 900nm-1700nm, and the soil penetration resistance sensor had a cone angle of 30°. The sensors transmitted the collected data to the vehicle control terminal in real time via the vehicle's CAN bus.

[0149] The terminal, based on collected data on wheat row spacing, soil slope, wheat grain moisture content, and soil penetration resistance, classifies sowing conditions into levels through normalized mapping. It then jointly analyzes the moisture content and slope data to construct a slope-moisture content correlation matrix, employing an inverse distance weighted interpolation method, where the power exponent... Number of neighboring points The parameters are expanded into a three-dimensional disturbance feature tensor. The above parameters are input into a preset decision rule base, and after being filtered by a moving average with a window size of 3, the sowing width and sowing density values ​​for corn are generated. The sowing width value is converted into channel boundary coordinate commands through Beidou navigation, driving the sowing machine to dynamically adjust the operating width. At the same time, the single irrigation volume of shallow-buried drip irrigation is synchronously controlled based on the sowing width value. In the irrigation volume control function, the sowing width benchmark value is 30cm, the standard single irrigation volume is 30m³ / mu, the sowing width sensitivity coefficient is 0.2, and the sowing condition level adjustment factor is 1.2.

[0150] Field verification showed that in this embodiment, the accuracy rate of sowing width control reached 95.3%, the accuracy rate of corn sowing channel identification was 96.2%, and the target sowing density control error was 4.8%. The average response time of the irrigation regulation system from command generation to irrigation volume reaching the target was 18.6 seconds, and the maximum response time was 25.3 seconds. Field yield measurements showed that the annual yield of wheat and maize in the core demonstration area reached 1044.74 kg / mu, including 272.13 kg of wheat and 772.61 kg of maize. Compared with the control area, the wheat yield per mu increased by 27.33%, and the maize yield per mu increased by 25.08%. Maize seedling uniformity improved by 18.6%, the lodging rate decreased by 12.3%, and the light and heat resource utilization rate increased by 16.8%. Simultaneously, through the coordinated control of sowing width and irrigation, the demonstration area saved 16.7% of irrigation water, reduced chemical nitrogen fertilizer use by 15%, and increased soil organic matter content by 10.43%–34.23%, achieving significant yield increases, efficiency improvements, and soil improvement effects.

[0151] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0152] The embodiments of this application can acquire and process relevant data based on an artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0153] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for adaptive sowing width control in wheat-maize intercropping based on crop and soil parameter identification, characterized in that, The method includes: Data on wheat row spacing, soil slope, wheat grain moisture content, and soil penetration resistance were collected. The soil penetration resistance data is normalized and mapped to generate a soil penetration resistance score, and the sowing condition level is classified according to the soil penetration resistance score. The wheat grain moisture content data and the soil slope data are jointly analyzed to construct a slope moisture content correlation matrix, and the dimensions of the slope moisture content correlation matrix are expanded to obtain interference characteristic factors of corn sowing. The wheat row spacing data, the sowing condition level, and the interference feature factor are input into a preset decision rule base to generate the sowing width value and sowing density value for corn sowing. The seeding width value is converted into channel boundary coordinate instructions for the seeding machine, and the channel boundary coordinate instructions are analyzed by Beidou navigation, and the seeding machine is driven to adjust the working width of corn seeding according to the channel boundary coordinate instructions; Based on the sowing width value of the corn, the single irrigation amount of shallow buried drip irrigation in wheat is simultaneously adjusted.

2. The adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification as described in claim 1, characterized in that, The process of normalizing and mapping the soil penetration resistance data to generate a soil penetration resistance score, and classifying the sowing condition level based on the soil penetration resistance score, includes: Based on the maximum and minimum values ​​of soil penetration resistance data, a normalized mapping interval is constructed. Based on the normalized mapping interval, the soil penetration resistance data is mapped to a scoring interval to generate a soil penetration resistance score value. The soil penetration resistance scores were arithmetically averaged, and the sowing conditions were classified according to the average results of the soil penetration resistance scores.

3. The adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification as described in claim 2, characterized in that, The classification of sowing conditions based on soil penetration resistance scores includes: A first scoring threshold and a second scoring threshold are preset, wherein the first scoring threshold is greater than the second scoring threshold; When the soil penetration resistance score is greater than the first score threshold, the sowing condition level is classified as Level 1. When the soil penetration resistance score is between the second score threshold and the first score threshold, the sowing condition level is classified as Level II. When the soil penetration resistance score is less than the second score threshold, the sowing condition level is classified as Level 3.

4. The adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification as described in claim 1, characterized in that, The joint analysis of the wheat grain moisture content data and the soil slope data is performed to construct a slope moisture content correlation matrix. The dimensions of this correlation matrix are then expanded to obtain interference characteristic factors for corn sowing, including: The wheat grain moisture content data and the soil slope data were normalized respectively to obtain normalized moisture content values ​​and normalized slope values. Perform a Cartesian product operation on the normalized moisture content value and the normalized slope value to construct a slope-moisture content correlation matrix indexed by the moisture content dimension and the slope dimension; Using a preset spatial interpolation algorithm, the slope moisture content correlation matrix is ​​expanded into a three-dimensional tensor containing spatial location information, and the three-dimensional tensor is used as an interference feature factor for corn planting.

5. The adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification as described in claim 4, characterized in that, The step of expanding the dimensionality of the slope moisture content correlation matrix using a preset spatial interpolation algorithm also includes: Obtain the BeiDou coordinate information of sampling points in the field, and establish a spatial mapping relationship between the BeiDou coordinate information and the elements in the slope moisture content correlation matrix; Using the BeiDou coordinate information as a spatial reference, spatial interpolation is performed on the slope moisture content correlation matrix to generate a three-dimensional interference feature tensor corresponding to the grid coordinates in the field.

6. The adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification as described in claim 1, characterized in that, The step of inputting the wheat row spacing data, the sowing condition level, and the interference feature factor into a preset decision rule base to generate the sowing width and sowing density values ​​for corn sowing includes: The pre-defined decision rule base includes row spacing hierarchical rules and span-density mapping tables; The row spacing data is input into the row spacing classification rules to obtain the row spacing level value; The row spacing level value, the sowing condition level, and the interference feature factor are concatenated to form a decision feature vector; The decision feature vector is matched and retrieved with the sowing width-density mapping table to obtain the initial value of the sowing width and the initial value of the sowing density corresponding to the decision feature vector; The initial values ​​of the sowing width and sowing density are smoothed and filtered to generate the sowing width and sowing density values ​​for corn sowing.

7. The adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification as described in claim 6, characterized in that, The broadcast width-density mapping table includes: To obtain data on the optimal sowing width and optimal sowing density under different wheat row spacing, different sowing condition levels, and different combinations of interference characteristic factors in historical field trials in the field. A decision tree classification model is constructed using the wheat row spacing data, sowing condition level, and interference feature factors as input variables, and the optimal sowing width and optimal sowing density as output variables. Extract the decision paths of the leaf nodes of the decision tree classification model, and store the correspondence between the input variables and output variables in the decision tree classification model into the spread-density mapping table.

8. The adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification as described in claim 6, characterized in that, The process of converting the sowing width value into channel boundary coordinate commands for the seeder, parsing the channel boundary coordinate commands via BeiDou navigation, and driving the seeder to adjust the corn sowing width according to the channel boundary coordinate commands includes: Using the current BeiDou positioning coordinates of the seeding machine as the reference point and the working direction of the seeding machine as the axis, half of the sowing width of the corn seeding is used as the unilateral offset. The unilateral offset is applied to the left and right sides perpendicular to the working direction, and the left boundary coordinates and right boundary coordinates are calculated by superimposing the offsets. The left and right boundary coordinates are serialized and encapsulated to generate a channel boundary coordinate instruction containing coordinate fields and verification fields; The channel boundary coordinate command is sent to the Beidou navigation terminal of the seeding machine. The Beidou navigation terminal extracts the left boundary coordinate and right boundary coordinate from the channel boundary coordinate command, and calculates the lateral deviation and heading deviation between the machine centerline and the left and right boundary coordinates based on the real-time Beidou positioning data of the seeding machine. Based on the lateral deviation value and the heading deviation value, the steering correction amount of the automatic navigation controller in the seeder is calculated, and the working posture of the seeder is dynamically adjusted according to the steering correction amount, so that the working width of the seeder is dynamically adapted to the channel range defined by the left boundary coordinate and the right boundary coordinate.

9. The adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification as described in claim 1, characterized in that, The method of simultaneously controlling the single irrigation volume of shallow-buried drip irrigation in wheat based on the sowing width value of the corn seeding includes: Based on the sowing width value of the corn sowing and the sowing width reference value, calculate the sowing width change rate; The rate of change of the sowing width is input into a preset irrigation volume control function to generate an irrigation volume adjustment coefficient; Multiply the irrigation volume adjustment coefficient by the preset standard single irrigation volume to obtain the target single irrigation volume; Based on the target single irrigation volume, the irrigation volume adjustment command for shallow-buried drip irrigation is executed.

10. The adaptive sowing width control method for wheat-maize intercropping based on crop and soil parameter identification as described in claim 9, characterized in that, The formula for the irrigation volume regulation function is as follows: in, The target irrigation volume per irrigation session, This is the standard single irrigation volume. The sensitivity coefficient for broadcast width. This is the seeding width value for the currently generated corn seed. This is the baseline value for the broadcast width. As a sowing condition level adjustment factor, This represents the sowing condition level value.

Images