A soil detection method, system, medium and device based on nutrient zoning

Abstract

The present application belongs to the field of image processing, and particularly relates to a soil detection method, system, medium and equipment based on nutrient partitioning. According to the variability of the data source, the optimal partition number is automatically determined, and the partitioning does not depend on the collection of ground data. The partitioning can ensure that the actual soil nutrient and other physicochemical property data and production potential in different partitions have significant differences. The present application can call satellite, unmanned aerial vehicle remote sensing growth, yield map and satellite remote sensing yield estimation data of different data sources for farmland management partitioning. The present application designs a reasonable soil sampling point number and position, and can ensure that the collected soil information is representative with the least samples.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of image processing, and particularly relates to a soil detection method, system, medium and equipment based on nutrient partitioning. BACKGROUND

[0002] Traditional farmland management is extensive, that is, each field is regarded as a homogeneous region, and the same input and management module is used regardless of the differences in soil nutrient content, topography, land use mode, etc. This management method belongs to an extensive management mode. Since 2005, the soil testing and formula fertilization technology has been promoted. Through the testing and analysis of regional soil samples, the fertilization scheme of elements such as nitrogen, phosphorus and potassium with regional adaptability is formulated by referring to the basic parameters such as crop fertilizer requirement law and soil fertilizer supply capacity. However, long-term large-scale soil testing and formula fertilization has high investment and low timeliness.

[0003] The soil testing and formula fertilization in the form of soil investigation and sampling is difficult to meet the requirements of precision fertilization and digital soil mapping due to the factors such as large quantity, high cost, insufficient typicality of distribution or lack of accurate geographic reference.

[0004] The Kriging interpolation technology is used to map the spatial continuous soil information by taking the geographic information system as the main technical method, and the model of regional soil nutrient distribution can be established, but the prerequisite is to collect ground information.

[0005] The Kmeans mean clustering or fuzzy C mean clustering algorithm can be used for differentiated management of farmland partitioning by taking remote sensing monitoring as the main technical method, but the Kmeans mean clustering or fuzzy C mean clustering algorithm needs to define the number of partitions, calculate the loss function value under different partition numbers according to the elbow rule, and constantly debug the parameters of the number of partitions, so the calculation time is long and it is difficult to apply on a large scale. SUMMARY

[0006] The technical problem to be solved by the application is to provide a soil detection method, system, medium and equipment based on nutrient partitioning.

[0007] The technical solution for solving the above technical problem is as follows: a soil detection method based on nutrient partitioning, comprising:

[0008] Step 1, acquiring a remote sensing image shot for soil, and converting the remote sensing image into an array corresponding thereto;

[0009] Step 2, calculating the variability of the array, and determining the number N of center points of the array based on the variability;

[0010] Step 3, determining the raster file of the array based on the number N of center points;

[0011] Step 4, converting the grid file to obtain a vector file, and determining a vector range of the soil sampling point according to the vector file, and performing soil detection in an actual soil range corresponding to the vector range.

[0012] The application has the advantages that the optimal number of partitions is automatically determined according to the variability of the data source, and the partitioning can ensure that the actual soil nutrient and physical and chemical property data and production potential of different partitions have significant differences; the application can call satellite, unmanned aerial vehicle remote sensing vigor, yield map and satellite remote sensing yield estimation data of different data sources for farmland management partitioning; the application designs reasonable soil sampling point quantity and position, and can ensure that the collected soil information is representative with the least samples.

[0013] Based on the above technical solution, the application can be further improved as follows.

[0014] Further, the remote sensing image comprises at least one of:

[0015] a satellite remote sensing vigor image of vegetation, a satellite remote sensing yield estimation image of vegetation, a yield map of agricultural machinery and an unmanned aerial vehicle vigor map.

[0016] Further, the process of determining the number of center points of each array based on the variability comprises:

[0017] searching for the number of center points corresponding to the variability in a preset center point corresponding table, and the preset center point corresponding table represents the corresponding relationship between variability and the number of center points.

[0018] Further, the process of determining the grid file of the array based on the number of center points comprises:

[0019] randomly selecting N centers in the array, calculating the distance between each sample data in the array and any one center point, and classifying the sample data to the center cluster corresponding to the minimum distance; until the classification of all sample data in the array is completed, N center clusters corresponding to the array are obtained, the first center point of each center cluster is calculated, and whether each first center point converges is determined; if yes, a grid file is generated according to the N center clusters, and if not, the N centers are reselected until each first center point converges.

[0020] Further, the process of performing soil detection in the actual soil range corresponding to the vector range comprises:

[0021] The area proportion of the vector range in the remote sensing image corresponding to the vector range is calculated, the number of sampling points in the vector range is determined based on a preset area corresponding table, the position of each sampling point in the vector range is determined based on a preset minimum interval, the position is mapped in the remote sensing image to determine the actual position of each sampling point, and soil detection is completed.

[0022] Another technical solution of the present application to solve the above technical problems is as follows: a soil detection system based on nutrient zoning, comprising:

[0023] The acquisition module is configured to acquire a remote sensing image taken for soil, and convert the remote sensing image into an array corresponding thereto.

[0024] The calculation module is configured to calculate the variability of the array, and determine the number N of center points of the array based on the variability.

[0025] The determination module is configured to determine a raster file of the array based on the number N of center points.

[0026] The repetition module is configured to convert the raster file to obtain a vector file, and determine a vector range of a soil sampling point according to the vector file, and perform soil detection in an actual soil range corresponding to the vector range.

[0027] The present application has the following advantages: the optimal number of partitions is automatically determined according to the variability of the data source, and the collection of ground data is not dependent, and the partitions can ensure that the actual soil nutrient and other physical and chemical property data and production potential have significant differences between different partitions; the present application can call satellite, unmanned aerial vehicle remote sensing vigor, yield map, and satellite remote sensing yield estimation data different data sources for farmland management zoning; the present application designs a reasonable number and position of soil sampling points, which can ensure that the collected soil information is representative with the least samples.

[0028] Further, the remote sensing image comprises:

[0029] At least one of a satellite remote sensing vigor image of vegetation, a satellite remote sensing yield estimation image of vegetation, a yield map of agricultural machinery, and an unmanned aerial vehicle vigor map.

[0030] Further, the process of determining the number of center points of each array based on the variability is as follows:

[0031] The number of center points corresponding to the variability is searched in a preset center point corresponding table, and the preset center point corresponding table represents the correspondence between variability and the number of center points.

[0032] Further, the process of determining the raster file of the array based on the number of center points is as follows:

[0033] Randomly select N centers in the array, calculate the distance between each sample data in the array and any center point, and classify the sample data to the center cluster corresponding to the minimum distance; until the classification of all sample data in the array is completed, the N center clusters corresponding to the array are obtained, the first center point of each center cluster is calculated, and whether each first center point converges is judged; if yes, a raster file is generated according to the N center clusters, and if not, the N centers are reselected until each first center point converges.

[0034] Further, the soil detection process in the actual soil range corresponding to the vector range is as follows:

[0035] The area proportion of the vector range in the remote sensing image corresponding to the vector range is calculated, the number of sampling points in the vector range is determined based on a preset area corresponding table, the position of each sampling point in the vector range is determined based on a preset minimum interval, the position is mapped in the remote sensing image, the actual position of each sampling point is determined, and the soil detection is completed.

[0036] Another technical solution of the present application to solve the above technical problems is as follows: a storage medium, the storage medium stores instructions, when the computer reads the instructions, the computer executes the method as claimed in any one of the above.

[0037] The beneficial effects of the present application are: the optimal number of partitions is automatically determined according to the variability of the data source, and does not depend on the collection of ground data, the partition can ensure that the actual soil nutrient and other physical and chemical property data and production potential have significant differences between different partitions; the present application can call satellite, unmanned aerial vehicle remote sensing vigor, yield map and satellite remote sensing yield estimation data different data sources for farmland management partitioning; the present application designs a reasonable number and position of soil sampling points, which can ensure that the collected soil information is representative with the least samples.

[0038] Another technical solution of the present application to solve the above technical problems is as follows: an electronic device, comprising the above storage medium and a processor executing the instructions in the above storage medium.

[0039] The beneficial effects of the present application are: the optimal number of partitions is automatically determined according to the variability of the data source, and does not depend on the collection of ground data, the partition can ensure that the actual soil nutrient and other physical and chemical property data and production potential have significant differences between different partitions; the present application can call satellite, unmanned aerial vehicle remote sensing vigor, yield map and satellite remote sensing yield estimation data different data sources for farmland management partitioning; the present application designs a reasonable number and position of soil sampling points, which can ensure that the collected soil information is representative with the least samples. BRIEF DESCRIPTION OF DRAWINGS

[0040] Figure 1A flowchart diagram is provided for an embodiment of the soil detection method based on nutrient zoning of the present application.

[0041] Figure 2 A structural framework diagram is provided for an embodiment of the soil detection system based on nutrient zoning of the present application.

[0042] Figure 3 A satellite remote sensing yield estimation raw data diagram is provided for an embodiment of the soil detection method based on nutrient zoning of the present application.

[0043] Figure 4 A satellite remote sensing yield estimation clustering zoning diagram is provided for an embodiment of the soil detection method based on nutrient zoning of the present application.

[0044] Figure 5 A satellite remote sensing yield estimation zoning vector diagram is provided for an embodiment of the soil detection method based on nutrient zoning of the present application.

[0045] Figure 6 A satellite remote sensing yield estimation zoning boundary buffer diagram is provided for an embodiment of the soil detection method based on nutrient zoning of the present application.

[0046] Figure 7 A satellite remote sensing yield estimation data space data calculation result diagram is provided for an embodiment of the soil detection method based on nutrient zoning of the present application.

[0047] Figure 8 A satellite remote sensing yield estimation management zoning and soil sampling point result diagram is provided for an embodiment of the soil detection method based on nutrient zoning of the present application. DETAILED DESCRIPTION

[0048] The principles and features of the present application are described below, and the examples are only used to explain the present application, and are not used to limit the scope of the present application.

[0049] As shown in Figure 1 A soil detection method based on nutrient zoning, comprising:

[0050] Step 1, acquiring a remote sensing image taken for soil, and converting each remote sensing image into an array corresponding thereto;

[0051] Step 2, calculating the variability of any one array, and determining the number N of center points of each array based on the variability;

[0052] Step 3, determining the grid file of the array based on the number N of center points;

[0053] Step 4: Convert the raster file into a vector file, determine the vector range of the soil sampling points based on the vector file, and perform soil detection within the actual soil range corresponding to the vector range; repeat steps 2 to 4 until soil detection for all remote sensing images is completed.

[0054] In some possible implementations, the optimal number of partitions is automatically determined based on the variability of the data source, without relying on the collection of ground data. The partitioning can ensure that there are significant differences in the actual soil nutrient and other physicochemical properties and production potential between different partitions. The present invention can call different data sources such as satellite and UAV remote sensing of growth, yield maps and satellite remote sensing yield estimation data as needed for farmland management partitioning. The present invention designs a reasonable number and location of soil sampling points, which can ensure that the collected soil information is representative with the fewest samples.

[0055] S1, the specific process of acquiring remote sensing images of the soil and converting each remote sensing image into a corresponding array includes:

[0056] S11, such as Figure 3 As shown, at least one of the following is acquired: satellite remote sensing crop image, satellite remote sensing yield estimation image, agricultural machinery output map, and UAV crop image;

[0057] Among them, satellite remote sensing vegetation images refer to raster files of vegetation indices obtained by processing satellite images through band operations.

[0058] Satellite remote sensing yield estimation images refer to raster files of yield obtained by inputting satellite images into a yield estimation model.

[0059] Agricultural machinery production map refers to a raster file obtained by raster interpolation of agricultural machinery production measurement points.

[0060] A drone vegetation map refers to a raster file of vegetation indices obtained by processing drone images through band calculations.

[0061] S12 uses the open-source library NumPy to convert any image and obtain an array. The values ​​of the array represent the grayscale value of each pixel in the image. It is a set of two-dimensional numbers representing the grayscale value of the image, with values ​​ranging from 0 to 255.

[0062] S2, as Figure 4 As shown, the process of calculating the variability of any array and determining the number N of center points for each array based on the variability is as follows:

[0063] S21, calculate the variability of any array using the first formula, which is:

[0064] C·V = (Standard Deviation SD / Mean) × 100%

[0065] S22, determining the number of partitions according to the variability, when C·V≥12%, N1(partition number)=5, when 12%>C·V≥10%, N1(partition number)=4, when 10%>C·V≥8%, N1(partition number)=3, when 8%>C·V≥6%, N1(partition number)=2, when 6%>C·V, N1(partition number)=1.

[0066] S23, each partition corresponds to contain a center point.

[0067] S3, as shown in Figure 4 and Figure 5 , the process of determining the grid file of the array based on the number of center points N is:

[0068] S31, by kmeans clustering algorithm, randomly select N center points in the two-dimensional array, respectively: μ1... μ N ;

[0069] S32, calculate the distance from each sample data x i in the two-dimensional array to all center points, and configure the sample data to the center cluster corresponding to the nearest center point;

[0070] c i =arg min||x i -μ j || 2 ;

[0071] Wherein, c i indicates the center cluster to which the sample data x i belongs, and μ j indicates the center point of the center cluster.

[0072] S33, the center point μ j corresponding to each partition is recalculated by the mean method, and the calculation formula is:

[0073]

[0074] S34, repeat S31 to S33 until the center point μ j no longer changes, that is, converges, and generate a grid file according to the current center point μ j .

[0075] S4, convert the grid file to obtain a vector file, and determine the vector range of the soil sampling point according to the vector file, and perform soil detection in the actual soil range corresponding to the vector range:

[0076] S41, as shown in Figure 6 and Figure 8As shown, the raster file is converted using the open source library shapely and rasterio to obtain a raster file for each partition, and a corresponding vector file is generated for each partition. The vector boundary of each partition is extracted, and the vector boundary is buffered inward by a predetermined distance to obtain a region as the vector range of the partition.

[0077] The vector boundary refers to the size and range of each level partition (each level partition contains multiple cells).

[0078] S42, calculate the area ratio of the vector range, and determine the number of sampling points of the vector range according to the area ratio;

[0079] The area ratio of the vector range can be calculated in the following way: the area ratio of each cell in the vector range partition to the total area of the partition level

[0080] When the area ratio is greater than or equal to 90%, the number of sampling points is 5;

[0081] When the area ratio is 70%-90%, the number of sampling points is 4;

[0082] When the area ratio is 50%-70%, the number of sampling points is 3;

[0083] When the area ratio is 30%-50%, the number of sampling points is 2;

[0084] When the area ratio is 10%-30%, the number of sampling points is 1;

[0085] The remaining area ratio is set to 0 according to the area size until the total number of sampling points in the partition is at most 5.

[0086] S43, using the Create Random Point tool of ArcGis, inputting the vector range, the number of sampling points in each range and the minimum interval between sampling points, to generate the vector random sampling points of the partition;

[0087] The minimum interval is obtained by experience or artificial definition, and usually one quarter of the square root of the area of the partition is taken as the minimum interval of the sampling points;

[0088] The process of generating the vector random sampling points of the partition is as follows:

[0089] S431, using a standard polygon partition algorithm to divide the polygon into polygons of different sizes, and each polygon is filled with triangles.

[0090] S432, placing the first point in the partitioned polygon, a triangle in the whole polygon will be randomly selected, the larger the triangle, the higher the probability of being selected. Two sides of the triangle will become the two axes on which the random point is placed. The next unused value will be obtained from the random stream and transformed into a "uniform" distribution, with the start of the triangle axis as the minimum value and the end of the edge as the maximum value. A value is randomly selected on the axis. The same operation is performed on the other axis or edge of the triangle. The two random values obtained will be used to place the point. The point will fall within the parallelogram created by the two axes of the triangle.

[0091] S433, if the point falls on the third side of the triangle used to define the triangle, the point is selected. If the point falls outside the triangle, it will fall within the other half of the parallelogram. Since this half of the parallelogram is a mirror image of the defined triangle, the point will be mirrored back into the defined triangle. This process will be repeated until the specified number of points are placed within the polygon that manages the partition, and this process is repeated for each management partition.

[0092] S44, mapping the sampling points to the remote sensing image, determining the actual position corresponding to the sampling points, and performing soil sampling processing according to the actual position.

[0093] In addition, as Figure 7 shown, the figure mainly explains the area proportion of each subzone, the number of sampling points calculated, wherein gridcode is a subzone, such as FID (number) 4, 5, 21 gridcode is 1, representing the same subzone, the buffer distance of the subzone is 5 meters, the total area is 4655 square meters, the area of each subzone is 2866 square meters, 774 square meters, and 1015 square meters, the proportion PROP is 0.615682, 0.166273, and 0.218045, respectively, the number of points num_poi is 3, 1, and 1, respectively, and the distance SQRT between the points in each subzone is 13.3838 meters, 6.95521 meters, and 7.96477 meters, respectively.

[0094] Preferably, in any of the above embodiments, the remote sensing image comprises:

[0095] At least one of the satellite remote sensing vigor image of vegetation, the satellite remote sensing yield estimation image of vegetation, the yield map of agricultural machinery, and the unmanned aerial vehicle vigor map.

[0096] Preferably, in any of the above embodiments, the process of determining the number of center points of each array based on the variability is:

[0097] searching a preset center point corresponding table corresponding to the variability, the preset center point corresponding table representing the corresponding relationship between variability and the number of center points.

[0098] Preferably, in any of the above embodiments, the process of determining the raster file of the array based on the number of center points is:

[0099] In the array, N centers are randomly selected, the distance between each sample data in the array and any one center point is calculated, and the sample data is classified into the center cluster corresponding to the minimum distance; until the classification of all sample data in the array is completed, N center clusters corresponding to the array are obtained, the first center point of each center cluster is calculated, and whether each first center point converges is determined; if yes, a raster file is generated according to the N center clusters, and if not, N centers are reselected until each first center point converges.

[0100] Preferably, in any of the above embodiments, the process of soil detection in the actual soil range corresponding to the vector range is:

[0101] The area proportion of the vector range in the remote sensing image corresponding to the vector range is calculated, and the number of sampling points in the vector range is determined based on a preset area corresponding table. The position of each sampling point in the vector range is determined based on a preset minimum interval, the position is mapped in the remote sensing image, the actual position of each sampling point is determined, and soil detection is completed.

[0102] As shown in Figure 2 a soil detection system based on nutrient zoning, comprising:

[0103] The acquisition module 100 is configured to acquire remote sensing images taken for soil, and convert each remote sensing image into an array corresponding thereto.

[0104] The calculation module 200 is configured to calculate the variability of any one array, and determine the number N of center points of each array based on the variability.

[0105] The determination module 300 is configured to determine a raster file of the array based on the number N of center points.

[0106] The repetition module 400 is configured to convert the raster file to obtain a vector file, determine a vector range of soil sampling points according to the vector file, and perform soil detection in the actual soil range corresponding to the vector range. The calculation module, the determination module, and the repetition module are repeatedly executed until soil detection corresponding to all remote sensing images is completed.

[0107] In some possible embodiments, the optimal number of partitions is automatically determined according to variability of a data source, and is independent of acquisition of ground data, and the partitions can ensure that actual soil nutrient and other physical and chemical property data and production potential in different partitions have significant differences; the application can call satellite, unmanned aerial vehicle remote sensing vigor, yield map and satellite remote sensing yield estimation data of different data sources to perform farmland management partitioning; and the application designs reasonable soil sampling point number and position, and can ensure that collected soil information is representative with the least samples.

[0108] Preferably, in any of the above embodiments, the remote sensing image comprises:

[0109] At least one of a satellite remote sensing vigor image of vegetation, a satellite remote sensing yield estimation image of vegetation, a yield map of agricultural machinery and an unmanned aerial vehicle vigor map.

[0110] Preferably, in any of the above embodiments, the process of determining the number of center points of each array based on the variability comprises:

[0111] Searching for the number of center points corresponding to the variability in a preset center point corresponding table, wherein the preset center point corresponding table represents a corresponding relationship between variability and the number of center points.

[0112] Preferably, in any of the above embodiments, the process of determining the raster file of the array based on the number of center points comprises:

[0113] Randomly selecting N centers in the array, calculating the distance between each sample data in the array and any one of the center points, and classifying the sample data to the center cluster corresponding to the minimum distance; until the classification of all sample data in the array is completed, N center clusters corresponding to the array are obtained, the first center point of each center cluster is calculated, and whether each first center point converges is determined; if yes, a raster file is generated according to the N center clusters, and if not, the N centers are reselected until each first center point converges.

[0114] Preferably, in any of the above embodiments, the process of performing soil detection in the actual soil range corresponding to the vector range comprises:

[0115] Calculating the area proportion of the vector range in the remote sensing image corresponding to the vector range, determining the number of sampling points in the vector range based on a preset area corresponding table, determining the position of each sampling point in the vector range based on a preset minimum interval, mapping the position in the remote sensing image, determining the actual position of each sampling point, and completing soil detection.

[0116] Another technical solution of the application to solve the above technical problems is as follows: a storage medium, wherein the storage medium stores instructions, and when a computer reads the instructions, the computer executes the method according to any of the above embodiments.

[0117] In some possible implementation manners, the optimal number of partitions is automatically determined according to variability of data sources, and the partitions can ensure that actual soil nutrient and other physicochemical property data and production potential in different partitions have significant differences, without relying on ground data collection; the application can call satellite, unmanned aerial vehicle satellite remote sensing growth, yield map and satellite remote sensing yield estimation data of different data sources to perform farmland management partitioning; and the application designs reasonable soil sampling point number and position, and can ensure that collected soil information is representative with the least samples.

[0118] Another technical solution of the application to solve the above technical problems is as follows: an electronic device, comprising the above storage medium and a processor for executing instructions in the storage medium.

[0119] In some possible implementation manners, the optimal number of partitions is automatically determined according to variability of data sources, and the partitions can ensure that actual soil nutrient and other physicochemical property data and production potential in different partitions have significant differences, without relying on ground data collection; the application can call satellite, unmanned aerial vehicle satellite remote sensing growth, yield map and satellite remote sensing yield estimation data of different data sources to perform farmland management partitioning; and the application designs reasonable soil sampling point number and position, and can ensure that collected soil information is representative with the least samples.

[0120] It should be understood by the reader that, in the description of the specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the application. In the specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, the skilled person in the art can combine and combine the different embodiments or examples described in the specification and the features of the different embodiments or examples without contradiction.

[0121] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the above-described method embodiments are only illustrative, for example, the division of steps is only a logical function division, and actual implementation can have another division manner, for example, multiple steps can be combined or integrated into another step, or some features can be ignored or not executed.

[0122] If the above method is realized in the form of a software function unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or all or part of the technical solutions that essentially contribute to the prior art can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0123] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present application, and these modifications or replacements should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A soil testing method based on nutrient zoning, characterized in that, include: Step 1: Acquire remote sensing images of the soil and convert the images into corresponding arrays; Step 2: Calculate the variability of the array and determine the number N of the array's center points based on the variability; Step 3: Determine the raster file of the array based on the number N of center points; Step 4: Convert the raster file into a vector file, determine the vector range of the soil sampling points based on the vector file, and perform soil testing within the actual soil range corresponding to the vector range; The process of determining the raster file of the array based on the number of center points is as follows: Randomly select N centers in the array, calculate the distance between each sample data in the array and any center point, and classify the sample data into the center cluster of the center point corresponding to the minimum distance; The process continues until all sample data in the array are classified, resulting in N central clusters. The first center point of each central cluster is calculated, and it is determined whether each first center point has converged. If it has, a raster file is generated based on the N central clusters. If not, N new centers are selected until each first center point has converged. The process of conducting soil testing within the actual soil area corresponding to the vector range is as follows: Calculate the area ratio of the vector range in the remote sensing image corresponding to the vector range, and determine the number of sampling points within the vector range based on a preset area correspondence table. Determine the position of each sampling point within the vector range based on a preset minimum interval, map the position onto the remote sensing image, determine the actual position of each sampling point, and complete the soil detection.

2. The soil testing method based on nutrient zoning according to claim 1, characterized in that, Remote sensing images include: At least one of the following: satellite remote sensing images of vegetation growth, satellite remote sensing images of vegetation yield estimation, agricultural machinery yield maps, and drone growth maps.

3. The soil testing method based on nutrient zoning according to claim 1, characterized in that, The process of determining the number of center points of the array based on the variability is as follows: Search a preset center point correspondence table for the number of center points corresponding to the variability. The preset center point correspondence table represents the correspondence between variability and the number of center points.

4. A soil testing system based on nutrient zoning, employing the soil testing method based on nutrient zoning as described in claim 1, characterized in that, include: The acquisition module is used to: acquire remote sensing images of the soil and convert the remote sensing images into corresponding arrays; The calculation module is used to: calculate the variability of the array, and determine the number N of the array's center points based on the variability; The determination module is used to: determine the raster file of the array based on the number N of center points; The repeat module is used to: convert the raster file into a vector file, determine the vector range of the soil sampling points based on the vector file, and perform soil testing within the actual soil range corresponding to the vector range.

5. A soil testing system based on nutrient zoning according to claim 4, characterized in that, Remote sensing images include: At least one of the following: satellite remote sensing images of vegetation growth, satellite remote sensing images of vegetation yield estimation, agricultural machinery yield maps, and drone growth maps.

6. A soil testing system based on nutrient zoning according to claim 4, characterized in that, The process of determining the number of center points of the array based on the variability is as follows: Search a preset center point correspondence table for the number of center points corresponding to the variability. The preset center point correspondence table represents the correspondence between variability and the number of center points.

7. A storage medium, characterized in that, The medium stores instructions that, when read by a computer, cause the computer to execute the method as described in any one of claims 1 to 3.

8. An electronic device, characterized in that, Includes the storage medium of claim 7 and a processor that executes instructions within the storage medium.

Images