Intelligent processing method and system for soil multi-disc sampling data

Abstract

The application provides a soil multi-disc sampling data intelligent processing method and system, relates to the intelligent technical field, obtains unique identification information of each sampling disc, uses a three-dimensional space difference identification method to judge whether the sampling disc is an abnormal sampling disc or a normal sampling disc, generates a candidate source unit and an independent point, generates main soil and non-main soil according to each candidate source unit, judges whether the non-main soil is candidate special soil, uses an embedded rule method to obtain a patch type of the candidate special soil, constructs a survey area range, uses an evidence strength grid connected region extraction method to obtain an evidence missing area and an information weak area in the survey area range, generates a recommended supplementary evidence point, and visualizes the sampling disc, the evidence missing area, the information weak area, the recommended supplementary evidence point and the survey area range, thereby significantly improving the automation level and sampling efficiency of soil investigation.

Description

Technical Field

[0001] This invention relates to the field of intelligent technology, specifically to an intelligent processing method and system for multi-pane soil sampling data. Background Technology

[0002] With the development of intelligent processing technology for multi-pane soil sampling, the manpower and time costs of field sampling have been greatly reduced, subjective judgment errors have been reduced, and scientific and traceable data support has been provided for farmland quality protection, pollution control and land remediation. This has important practical significance for promoting precision agriculture and sustainable management of soil resources. However, in the existing multi-pan soil sampling process, each sampling pan is usually managed and analyzed as an independent data unit, which can only obtain single-point physicochemical index results. It lacks the ability to systematically identify the spatial correlation, source consistency, and abnormal distribution patterns between sampling pans. This makes it difficult to automatically detect abnormal samples, identify the boundaries of special soil patches, assess the integrity of the investigation evidence, and guide subsequent supplementary sampling. As a result, the soil survey results have problems such as low analysis efficiency, difficulty in finding blind spots, and insufficient reliability of the survey results. Summary of the Invention

[0003] To achieve the above objectives, the present invention provides the following technical solution: an intelligent processing method for multi-pane soil sampling data, the method comprising: Obtain the unique identification information of each sampling disk, and use the three-dimensional spatial difference recognition method to determine whether the sampling disk is an abnormal sampling disk or a normal sampling disk based on the unique identification information of each sampling disk; Based on the normal sampling disk, candidate source cells and independent points are generated; For each candidate source unit, its spatial extent is obtained. The candidate source unit with the largest spatial extent is taken as the main soil, and the remaining candidate source units are taken as non-main soil. It is then determined whether the non-main soil is a candidate special soil. For the candidate special soil, the built-in rule method is used to obtain the patch type of the candidate special soil. Based on the main soil and non-main soil, the survey area is constructed, and the evidence strength grid connected region extraction method is used to obtain the evidence-deficient areas and information-weak areas in the survey area. For the evidence-deficient areas, the optimal sequential supplementation method is used to generate recommended supplementation points. For the information-weak areas, the radial extension method is used to generate recommended supplementation points. Supplementation sequence is generated based on the recommended supplementation points. The sampling disk, areas with missing evidence, areas with weak information, and recommended supplementary evidence points are visualized.

[0004] Furthermore, obtaining the unique identifier information for each sampling disk includes: For each sampling point, basic information and spatial structure correlation data of the sampling point are obtained; and the soil collected from the sampling point is put into a sampling tray to obtain rapid on-site physicochemical data of the sampling tray. The basic information includes temperature, humidity, and wind speed data at the time of sampling; spatial structure correlation data includes the location, depth, and soil texture of the sampling point; rapid on-site physicochemical data includes organic matter content, pH, salinity, characteristic element content, and characteristic element content ratio data. By constructing a unique identifier from basic information, spatial structure association data, and rapid on-site physicochemical data, and binding it with the sampling disk, the unique identifier of the sampling disk is obtained.

[0005] Furthermore, the method of using a three-dimensional spatial difference recognition to determine whether a sampling disk is an abnormal or normal sampling disk, based on the unique identifier information of each sampling disk, includes: Step ss1: Traverse each sampling disk, set the current sampling disk as the target sampling disk, set the search radius, and record the sampling disks within the search radius as feature sampling disks; if there are no feature sampling disks within the search radius, skip them; if they exist, check whether the target sampling disk meets any of the conditions in the horizontal spatial neighborhood comparison conditions; if it does, it means that the target sampling disk is an abnormal sampling disk. Step ss2: For sampling disks at different depths in the same borehole, sort them in ascending order of depth. Define the sampling disk with the smaller depth as a shallow sampling disk and the sampling disk with the larger depth as a deep sampling disk. Compare the shallow sampling disk and the deep sampling disk in a pair of adjacent depths. If any of the conditions in the vertical spatial neighborhood comparison conditions are met, the deep sampling disk is considered an abnormal sampling disk. Step ss3: For the rapid field physicochemical data in all sampling disks, obtain the mean and standard deviation of the rapid field physicochemical data, and generate a virtual disk based on the mean of the rapid field physicochemical data; Each sampling disk is compared with a virtual disk; if the difference between any data in the rapid on-site physicochemical data of the sampling disk and the corresponding data in the virtual disk exceeds the baseline standard deviation, the sampling disk is considered an abnormal sampling disk.

[0006] Furthermore, the step of generating candidate source cells and independent points based on the normal sampling disk includes: For each normal sampling disc, the distance between each pair of normal sampling discs and the difference in organic matter content were obtained; The distance between each pair of normal sampling discs is divided into H intervals, and the median of the difference in organic matter content within each interval is obtained to construct a list of the median differences in organic matter content among the H intervals. And obtain the upper limit of the homogeneous spatial distance based on the distance of each interval; and retrieve the median of the largest organic matter content difference from the list of median differences in organic matter content. For the distance between each pair of normal sampling discs, if the distance is less than the upper limit of the distance between homologous spaces and the difference in organic matter content is less than the median of the largest difference in organic matter content, then the pair of normal sampling discs is considered to be homologous candidates. Transitively connect the homologous candidates to generate K candidate source units; Obtain the representative fingerprint of each candidate source unit, and check whether the difference in organic matter content between the normal sampling disk in each candidate source unit and the representative fingerprint is greater than the median of the largest organic matter content difference. If it is greater, the corresponding normal sampling disk is separated from the candidate source unit and used as an independent point.

[0007] Furthermore, for candidate special soils, the method of obtaining patch types of candidate special soils using built-in rules includes: For the main soil sample, rapid on-site physicochemical data were retrieved from each normal sampling tray in the main soil sample. Calculate the direction of change of rapid field physicochemical data between each pair of normal sampling discs to construct a group of rapid field physicochemical data change directions of the main soil sample. For candidate special soils, we obtained the group of rapid on-site physicochemical data changes in candidate special soils. If the similarity between the rapid field physicochemical data change direction group in the candidate special soil and the rapid field physicochemical data change direction group in the main soil is greater than or equal to the basic similarity, it indicates that the candidate special soil has undergone natural variation rather than human disturbance. If the similarity between the rapid field physicochemical data change direction group of the candidate special soil and the rapid field physicochemical data change direction group of the main soil is less than the basic similarity, it indicates that the source of the candidate special soil is different from that of the main soil. In this case, the multi-index deviation decision method is used to determine the patch type of the candidate special soil.

[0008] Furthermore, the method of using multiple indicators to deviate from the decision-making criteria to determine the patch type of candidate special soil includes: Extract the maximum and minimum values ​​of each data point from all normal sampling pans in the main soil sample; Retrieve representative fingerprints from candidate special soils, i.e., representative fingerprints of the corresponding candidate source units; For each data point, its maximum value, minimum value, and the corresponding data in the fingerprint are compared. If the fingerprint of a certain data point is greater than the maximum value of the corresponding data point, it indicates that the data point is significantly high. If the representative fingerprint of a certain data is less than the minimum value of the corresponding data, it indicates that the data is significantly low. If the fingerprint representing a certain data point is between the minimum and maximum values ​​of the corresponding data, then the data is considered normal. When the organic matter content data of a candidate special soil is significantly low and the ratio of characteristic element content is significantly deviated (significant deviation includes significantly high or significantly low), the patch type of the candidate special soil is a backfill area. When both pH and salinity data are significantly high, the plaque type is salinization. If the organic matter content is significantly higher and the depth is significantly lower, then the patch type is a sedimentary area. If at least one characteristic element in the characteristic element content data significantly exceeds the corresponding risk screening value, then the patch type is a pollution disturbance area. When the organic matter content, pH, and salinity data are all normal, but the ratio of characteristic element content deviates significantly, the patch type is an area where exogenous soil has been introduced. If none of the above conditions are met, the plaque type is set to uncertain and marked as "to be verified".

[0009] Furthermore, based on the main soil and non-main soil, the survey area is constructed, and the evidence strength grid connected region extraction method is used to obtain the evidence-deficient areas and information-weak areas within the survey area, including: The survey area is constructed by using the main soil and non-main soil, and D grids are generated for the survey area. For each grid, check whether there is a normal sampling disk in the grid. If not, the grid is defined as a grid without evidence. If it exists, check whether the normal sampling disk in the grid belongs to the candidate source unit. If it does, the grid is defined as a grid with strong evidence. If it does not belong to the candidate source unit, i.e., an independent point, the grid is defined as a grid with weak evidence. When a grid is in a candidate special soil, the grid is upgraded to obtain an upgraded grid; among them, the grid with no evidence is upgraded to the grid with weak evidence, the grid with weak evidence is upgraded to the grid with strong evidence, and the grid with strong evidence remains the grid with strong evidence. For each grid, including both upgraded and unupgraded grids, they are merged according to the four-connectivity rule to obtain regions with missing evidence and weak information.

[0010] Furthermore, for regions lacking evidence, the optimal sequential supplementation method is used to generate recommended supplementary evidence points, including: For areas with missing evidence, retrieve the number of grids in each area, calculate the minimum distance from the center point of each grid in the area to all normal sampling disks in the survey area, and select the grid center point with the largest minimum distance as the first recommended supplementary evidence point; if this point happens to coincide with a normal sampling disk, select the grid center point with the second largest minimum distance as the first recommended supplementary evidence point. For the center points of the remaining grid, obtain the minimum distance to each center point and the distance to the first recommended supplementary proof point; The comprehensive score of the center point of each remaining grid is obtained by weighting the minimum distance to the center point of the remaining grid and the distance to the first recommended supplementary proof point. Select the center with the highest overall score as the second recommended supplementary certification point; Remove the grids of the first and second recommended supplementary evidence points. For the remaining grids, calculate their comprehensive scores and select the center point with the highest comprehensive score as the third recommended supplementary evidence point. Stop generating recommended supplementary evidence points when the minimum distance is less than the standard distance. For the recommended supplementary evidence points, a list of recommended supplementary evidence point locations for areas with missing evidence is constructed.

[0011] Furthermore, the method of generating recommended supplementary evidence points using a radial extension approach for areas with weak information includes: Calculate the difference intensity of all independent points in the information-weak region, and arrange the difference intensity in descending order to obtain the difference intensity sequence; Using the intensity of difference as the weight, all independent points are weighted to obtain a weighted position. This weighted position is then used as the first recommended supplementary evidence point. If the weighted position falls exactly on a normal sampling point, the weighted position is moved one grid step in any direction and a new weighted position is obtained in the information-weak region. This new weighted position is then used as the first recommended supplementary evidence point. For each independent point and the first recommended supplementary evidence point, determine the direction and distance of extension; Starting from each independent point, according to the extension direction and extension distance, a group of candidate recommended supplementary evidence points is obtained; If a candidate recommended supplementary evidence point in the candidate recommended supplementary evidence point group overlaps with an existing sampling point or is not located in an area with weak information, then that candidate recommended supplementary evidence point is discarded. The remaining candidate recommended supplementary evidence points and the first recommended supplementary evidence point are used as recommended supplementary evidence points for the area with weak information, and a list of recommended supplementary evidence point locations for the area with weak information is constructed.

[0012] A soil multi-pane sampling data intelligent processing system, the system comprising: Anomaly detection module: Obtain the unique identification information of each sampling disk, and use the three-dimensional spatial difference recognition method to determine whether the sampling disk is an abnormal sampling disk or a normal sampling disk based on the unique identification information of each sampling disk; Homogeneous clustering module: Generates candidate source cells and independent points based on the normal sampling disk; Subject identification module: Based on each candidate source unit, obtain its spatial range, take the candidate source unit with the largest spatial range as the subject soil, and the remaining candidate source units as non-subject soil, and determine whether the non-subject soil is a candidate special soil; for candidate special soil, use the built-in rule method to obtain the patch type of the candidate special soil; The evidence supplementation planning module: Based on the main soil and non-main soil, the survey area is constructed, and the evidence strength grid connected region extraction method is used to obtain the evidence-deficient areas and information-weak areas in the survey area; for the evidence-deficient areas, the optimal sequential evidence supplementation method is used to generate recommended evidence supplementation points; for the information-weak areas, the radial extension method is used to generate recommended evidence supplementation points; and a evidence supplementation sequence is generated based on the recommended evidence supplementation points. Visualization module: Visualizes the sampling disk, areas with missing evidence, areas with weak information, recommended supplementary evidence points, and the scope of the survey area.

[0013] This invention provides an intelligent processing method and system for multi-pane soil sampling data. It has the following beneficial effects: By integrating multi-pane sampling data with intelligent processing workflows, the entire chain of automated processing has been achieved, from anomaly identification of sampling points, division of soil source units, identification of candidate special soil patches to grid analysis of areas with missing evidence and weak information, and adaptive generation of supplementary points. This significantly reduces the blindness and repetitive workload of field sampling, improves the pertinence and sampling efficiency of soil surveys, and provides scientific and efficient technical support for practical engineering applications such as farmland quality assessment, pollution site identification, and backfill area determination. At the same time, the generated visualization results facilitate on-site decision-making and long-term traceability. The three-dimensional spatial difference identification method can accurately distinguish between abnormal sampling disks and normal sampling disks; the transitive clustering and decomposition mechanism based on distance and organic matter content differences can automatically generate homologous candidate source units and independent points, avoiding the subjectivity of manual division; the evidence strength grid connected region extraction method is used to quantitatively identify information-weak areas and evidence-deficient areas, and the radial extension method and the optimal sequential supplementation method are used for different areas to generate recommended supplementation points, realizing the adaptability and spatial rationality of the supplementation strategy. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating the steps of the intelligent processing method for multi-pane soil sampling data of the present invention. Figure 2This is a flowchart of the three-dimensional spatial difference recognition method of the present invention; Figure 3 This is an architecture diagram of the intelligent processing system for multi-disc soil sampling data of the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] like Figures 1 to 2 As shown, an intelligent processing method for multi-pane soil sampling data is described, and the method includes: Obtain the unique identification information of each sampling disk, and use the three-dimensional spatial difference recognition method to determine whether the sampling disk is an abnormal sampling disk or a normal sampling disk based on the unique identification information of each sampling disk; Based on the normal sampling disk, candidate source cells and independent points are generated; For each candidate source unit, its spatial extent is obtained. The candidate source unit with the largest spatial extent is taken as the main soil, and the remaining candidate source units are taken as non-main soil. It is then determined whether the non-main soil is a candidate special soil. For the candidate special soil, the built-in rule method is used to obtain the patch type of the candidate special soil. Based on the main soil and non-main soil, the survey area is constructed, and the evidence strength grid connected region extraction method is used to obtain the evidence-deficient areas and information-weak areas in the survey area. For the evidence-deficient areas, the optimal sequential supplementation method is used to generate recommended supplementation points. For the information-weak areas, the radial extension method is used to generate recommended supplementation points. Supplementation sequence is generated based on the recommended supplementation points. The sampling trays, areas with missing evidence, areas with weak information, recommended supplementary evidence points, and the scope of the survey area are visualized.

[0017] Obtain the unique identifier information for each sampling disk, including: For each sampling point, basic information and spatial structure correlation data of the sampling point are obtained; and the soil collected from the sampling point is put into a sampling tray to obtain rapid on-site physicochemical data of the sampling tray. By using mobile devices, basic information, spatial structure association data, and rapid on-site physical and chemical data are combined to form unique identification information, which is then bound to the sampling disk. For example, the sampling personnel use the APP to generate a barcode, print it out and stick it on the sampling tray containing the soil at the sampling point; create a blank record for the barcode generated by the APP; store the basic information collected at the sampling point, the spatial structured correlation data, and the rapid on-site physicochemical data of the sampling tray into the blank record and save it, thus completing the binding. The basic information includes temperature data, humidity data, and wind speed data at the time of sampling, which are obtained through a handheld environmental monitoring device. Spatial structure association data includes sampling point location, sampling depth, and soil texture. Sampling point location is obtained using a handheld RTK-GPS sensor, and sampling depth is obtained using a depth encoder. Rapid on-site physicochemical data includes organic matter content data, pH data, salinity data, characteristic element content data, and characteristic element content ratio data. Soil texture and organic matter content are obtained using a portable near-infrared spectrometer, pH and salinity are obtained using a portable pH meter, and characteristic element content and characteristic element content ratio are obtained using a portable X-ray fluorescence spectrometer.

[0018] For each sampling disk's unique identifier, a three-dimensional spatial difference recognition method is used to determine whether the sampling disk is an abnormal or normal sampling disk, including: Step ss1: Traverse each sampling disk, set the current sampling disk as the target sampling disk, set the search radius, and record the sampling disks within the search radius as feature sampling disks; if there are no feature sampling disks within the search radius, skip to step ss2; if there are, check whether the target sampling disk satisfies any of the conditions in the horizontal spatial neighborhood comparison conditions; if it does, it means that the target sampling disk is an abnormal sampling disk. Comparison conditions for horizontal spatial domains include: Condition 1: The soil textures of the target sampling disk and the characteristic sampling disk are different; Condition 2: The organic matter content data of the target sampling disk is not within the upper and lower limits of the organic matter content data of the feature sampling disk; the upper limit of the organic matter content data of the feature sampling disk is 1.2 times the organic matter content, and the lower limit is 0.8 times the organic matter content data. If there are multiple feature sampling disks, the upper limit of the organic matter content data is 1.2 times the average of the organic matter content data of the multiple feature sampling disks; the lower limit is 0.8 times the average of the organic matter content data of the multiple sampling disks. Condition 3: The characteristic element content ratio data of the target sampling disk is not within the upper and lower limits of the characteristic element content ratio data of the characteristic sampling disk; the upper limit of the characteristic element content ratio data is 1.2 times the average value of the characteristic element content data, and the lower limit of the characteristic element content ratio data is 0.8 times the average value of the characteristic element content data; the upper and lower limits are 1.2 times and 0.8 times the average value of the corresponding data, respectively, which are set according to the engineering inspection method.

[0019] Step ss2: For sampling disks at different depths in the same borehole, sort them in ascending order of depth. Define the sampling disk with the smaller depth as a shallow sampling disk and the sampling disk with the larger depth as a deep sampling disk. Compare the shallow sampling disk and the deep sampling disk in a pair of adjacent depths. If any of the conditions in the vertical spatial neighborhood comparison conditions are met, the deep sampling disk is considered an abnormal sampling disk. Vertical spatial neighborhood comparison conditions include: Condition 1: The organic matter content data of the deep sampling disk is greater than or equal to the organic matter content data of the shallow sampling disk, that is, the organic matter content does not decrease with depth, which violates the normal vertical distribution law. Condition 2: The difference in pH between the deep sampling disk and the shallow sampling disk is less than or equal to the standard pH difference, which is 0.5 and is set according to engineering experience. Condition 3: The difference between the characteristic element content ratio data between the deep disk and the shallow disk is less than or equal to the preset characteristic element content ratio threshold, which is 0.2 and is set according to engineering experience. Step ss3: For the rapid field physicochemical data in all sampling disks, obtain the mean and standard deviation of the rapid field physicochemical data, and generate a virtual disk based on the mean of the rapid field physicochemical data; Each sampling disk is compared with the virtual disk; if the difference between any data in the rapid on-site physicochemical data of the sampling disk and the corresponding data in the virtual disk exceeds the baseline standard deviation (the baseline standard deviation is 2.5 times the standard deviation, set according to engineering experience), then the sampling disk is considered an abnormal sampling disk. The search radius is obtained by calculating the distance between each sampling disk and the target sampling disk (using the Euclidean method to obtain the distance), calculating the average distance, and using the average distance as the search radius; Step SS1 is executed first, based on the first law of geography that the closer the distance, the more similar the soil properties should be. By comparing the target sampling disk with the feature sampling disks within a certain radius around it, it can sensitively capture anomalies in the local space. If an anomaly is judged in this step, there is no need to proceed to the subsequent steps. Step SS2 is executed next. For sampling trays at different depths in the same borehole, the pattern of decreasing organic matter content with depth, and the gradual change of pH and characteristic element content ratios in normal soil profiles is used to identify whether there are vertical anomalies in the deep sampling trays, such as an increase in organic matter content instead of a decrease, or a sharp jump in pH. If an anomaly is detected in this step, there is no need to proceed to the next step. Step SS3 is executed last, using the average value of all sampled disks to construct a "virtual disk" as a global background; by comparing the difference between each sampled disk and the virtual disk, sampled disks that are statistically outliers can be identified.

[0020] Based on the normal sampling disk, candidate source cells and independent points are generated, including: For each normal sampling pan, the distance between each pair of normal sampling pans was obtained (the location of the corresponding soil sampling point in each sampling pan was known, and the distance was obtained using the Euclidean formula) and the difference in organic matter content (i.e., the difference in organic matter content); for example, the distance between normal sampling pan 1 and normal sampling pan 2 is 2.8m, and the difference in organic matter content is 1g / kg; the distance between normal sampling pan 2 and normal sampling pan 3 is 5.6m, and the difference in organic matter content is 1g / kg. The distance between each pair of normal sampling trays is divided into H intervals (H ≥ 2), and the median of the organic matter content difference within each interval is obtained to construct a list of median organic matter content differences for the H intervals. For example, if the minimum distance is 0m and the maximum distance is 50m, then the intervals can be divided into 10 intervals of 5m each. The interval from 0m to 5m includes normal sampling tray 1 and normal sampling tray 2, with a distance of 2.8m and an organic matter content difference of 1g / kg; normal sampling tray 4 and normal sampling tray 5, with a distance of 2.8m and an organic matter content difference of 4g / kg; and normal sampling tray 4 and normal sampling tray 7, with a distance of 2.9m and an organic matter content difference of 1g / kg. The median of the organic matter content difference within this interval is 1g / kg. And obtain the upper limit of the distance between homogeneous spaces based on the distance between each interval; The specific steps to obtain the upper limit of the distance between similar spaces are as follows: Calculate the rate of change between the differences in organic matter content in each adjacent interval, and select the distance between the two adjacent intervals with the largest rate of change as the upper limit of the spatial distance between the same source; for example, if the difference in organic matter content between 20 and 25m is 2g / kg, and the difference in organic matter content between 25 and 30m is 8g / kg, and the rate of change is 3 (i.e., the ratio of the difference in organic matter content between 25 and 30m to the difference in organic matter content between 20 and 25m, to the difference in organic matter content between 20 and 25m), and it is the largest rate of change, then 25m is the upper limit of the spatial distance between the same source; And retrieve the median of the largest difference in organic matter content from the list of median differences in organic matter content; For the distance between each pair of normal sampling discs, if the distance is less than the upper limit of the same-source spatial distance and the difference in organic matter content is less than the median of the largest difference in organic matter content, then the pair of normal sampling discs is considered to be same-source candidates. For example, if the distance between normal sampling disc 1 and normal sampling disc 3 is 5.6, which is less than the upper limit of the same-source spatial distance and the difference in organic matter content is 1, which is less than the largest difference in organic matter content, then normal sampling disc 1 and normal sampling disc 3 are considered to be same-source candidates. By transitively connecting the same source candidates, K candidate source units are generated, where K is greater than or equal to 1. For example, if normal sampling disk 1 and normal sampling disk 3 are the same source candidates, normal sampling disk 3 and normal sampling disk 9 are the same source candidates, and normal sampling disk 9 and normal sampling disk 11 are the same source candidates, then by transitively connecting the same source candidates, we get normal sampling disk 1 → normal sampling disk 3 → normal sampling disk 9 → normal sampling disk 11. Then, candidate unit 1 is: {normal sampling disk 1, normal sampling disk 3, normal sampling disk 9, normal sampling disk 11}. Obtain the representative fingerprint of each candidate source unit, which is the average organic matter content of all normal sampling disks in the candidate source unit. Check whether the difference in organic matter content of normal sampling disks in each candidate source unit is greater than the median of the largest organic matter content difference. If it is greater, the corresponding normal sampling disk is separated from the candidate source unit and used as an independent point.

[0021] Spatially adjacent main soil sampling points with similar organic matter content are grouped into the same candidate source unit as a reliable background reference for subsequent identification of special soils (such as backfill areas); while unclassifiable points are treated as independent points to avoid contaminating the background benchmark.

[0022] For each candidate source unit, its spatial extent is obtained, and the candidate source unit with the largest spatial extent is taken as the main soil, while the other candidate source units are taken as non-main soil. Traverse the non-main soil and calculate the distance between each pair of normal sampling disks in the non-main soil. If there are all pairs of normal sampling disks with a distance less than a preset distance threshold and the number of normal sampling disks is greater than or equal to the number of reference sampling disks, then mark the non-main soil as a candidate special soil. Based on candidate special soils, the patch types of candidate special soils are inferred using the built-in rule method.

[0023] The preset distance threshold is the average distance between each pair of normal sampling discs within the main soil layer; The number of reference sampling disks is set to 3 because only three points can form a surface in a space; If the distance between any two points is less than a preset distance threshold, it indicates that the non-main soil forms a highly compact cluster in space and has basic geometric morphology and statistical reliability, and is therefore identified as a candidate special soil. For each candidate source unit, its spatial range is obtained using the minimum convex hull method or the outer matrix method. Both the minimum convex hull method and the outer matrix method are existing technologies. The specific steps for obtaining the spatial range of candidate source cells using the minimum convex hull method are as follows: Retrieve the positions of all normal sampling disks in the candidate source unit, such as normal sampling disk 1 (1,1,2), normal sampling disk 2 (4,2,3), normal sampling disk 3 (3,5,1), and normal sampling disk 4 (6,4,4). Find the extreme points in the X, Y, and Z directions. The minimum extreme point in the X direction is normal sampling disk 1 (1,1,2), and the maximum extreme point is normal sampling disk 4 (6,4,4). The minimum extreme point in the Y direction is normal sampling disk 1 (1,1,2), and the maximum extreme point is normal sampling disk 3 (3,5,1). The minimum extreme point in the Z direction is normal sampling disk 3 (3,5,1), and the maximum extreme point is normal sampling disk 4 (6,4,4). Since the positions of the four normal sampling disks are not coplanar, the positions of these four normal sampling disks are taken as vertices, and the volume of the tetrahedron is obtained according to the tetrahedron volume calculation formula. If all positive sampling disks have a Z-axis of 2, they are coplanar.

[0024] For candidate special soils, the built-in rule method is used to obtain the patch types of the candidate special soils, including: For the main soil sample, rapid on-site physicochemical data were retrieved from each normal sampling tray in the main soil sample. Calculate the direction of change of rapid field physicochemical data between each pair of normal sampling discs to construct a group of rapid field physicochemical data change directions for the main soil sample; that is, a "+" direction indicates a positive direction, a "-" direction indicates a negative direction, and if there is no change, the pair is ignored. For all rapid field physicochemical data in the main soil sample tray, the number of times each data point appears in the positive and negative directions is considered. If the number of times a data point appears in the positive direction is more than the number of times it appears in the negative direction, then the direction of change of that data point is positive, and vice versa. For candidate special soils, we obtained the group of rapid on-site physicochemical data changes in candidate special soils. If the similarity between the rapid field physicochemical data change direction group in the candidate special soil and the rapid field physicochemical data change direction group in the main soil is greater than or equal to the basic similarity, it indicates that the candidate special soil has undergone natural variation rather than human disturbance. If the similarity between the rapid field physicochemical data change direction group of the candidate special soil and the rapid field physicochemical data change direction group of the main soil is less than the basic similarity, it indicates that the source of the candidate special soil is different from that of the main soil. In this case, the multi-index deviation decision method is used to determine the patch type of the candidate special soil.

[0025] The similarity calculation uses the matching coefficient method. For example, if the direction change group of the main soil rapid field physicochemical data is (+, +, +, -, +, -, +), and the direction change group of the candidate special soil rapid field physicochemical data is (-, +, +, +, +, -, +), then each direction is compared one by one to obtain (different, same, same, different, same, same, same). The number of same is 5, and the total number is 7, so the similarity is 71%. The basic similarity was set to 80%, based on engineering experience. By comparing the change directions of candidate special soils and the main soil in multiple rapid field physicochemical indicators, it can be determined whether the candidate special soil belongs to natural variation. If the change directions of the two are highly consistent, it is determined to be natural variation; otherwise, it indicates that the source or cause is different, and further identification of the specific patch type is required.

[0026] A multi-index deviation decision method is used to determine the patch type of candidate special soils, including: Extract the maximum and minimum values ​​of each data point from all normal sampling pans in the main soil sample; Retrieve representative fingerprints from candidate special soils, i.e., representative fingerprints of the corresponding candidate source units; For each data point, its maximum value, minimum value, and the corresponding data in the fingerprint are compared. If the fingerprint of a certain data point is greater than the maximum value of the corresponding data point, it indicates that the data point is significantly high. If the representative fingerprint of a certain data is less than the minimum value of the corresponding data, it indicates that the data is significantly low. If the fingerprint representing a certain data point is between the minimum and maximum values ​​of the corresponding data, then the data is considered normal. When the organic matter content data of a candidate special soil is significantly low and the ratio of characteristic element content is significantly deviated (significant deviation includes significantly high or significantly low), the patch type of the candidate special soil is a backfill area. When both pH and salinity data are significantly high, the plaque type is salinization. If the organic matter content is significantly higher and the depth is significantly lower, then the patch type is a sedimentary area. If at least one characteristic element in the characteristic element content data significantly exceeds the corresponding risk screening value, then the patch type is a pollution disturbance area. The specific risk screening value for each characteristic element content data is obtained according to the national standard GB15618. For example, when the pH data is greater than 7.5, the risk screening value for lead in the characteristic element content data is 170 mg / kg. When the organic matter content, pH, and salinity data are all normal, but the ratio of characteristic element content deviates significantly, the patch type is an area where exogenous soil has been introduced. If none of the above conditions are met, such as only the condition that the organic matter content data is significantly low, then the plaque type is set to uncertain and marked as "to be verified". If multiple conditions are met simultaneously, the selection should be made in the following order: pollution disturbance zone > backfill zone > saline zone > sedimentation zone > foreign soil introduction zone.

[0027] By comparing the representative fingerprints of candidate special soils with the maximum / minimum values ​​of the main soil, significant deviations of each indicator are identified. Different combinations of indicator deviations are then mapped to specific patch types, and multiple condition conflicts are handled with priority. Cases that cannot be matched are marked as "to be verified". This enables rapid, transparent, and interpretable inference of patch types of candidate special soils, serving soil quality assessment and decision-making.

[0028] Based on the main soil and non-main soil, the survey area was constructed, and the evidence strength grid connected region extraction method was used to obtain the areas with missing evidence and weak information within the survey area, including: The survey area is constructed using the main soil and non-main soil samples (the survey area is generated using the minimum convex hull method based on the coordinates of the normal sampling disks within it); for the survey area, D grids are generated, and the grids are user-defined, such as 5m*5m and 10m*10m; For each grid, check whether there is a normal sampling disk in the grid. If not, the grid is defined as a grid without evidence. If it exists, check whether the normal sampling disk in the grid belongs to the candidate source unit. If it does, the grid is defined as a grid with strong evidence. If it does not belong to the candidate source unit, i.e., an independent point, the grid is defined as a grid with weak evidence. If a grid contains multiple sampling disks, and at least one normal sampling disk belongs to a candidate source cell, then the grid is defined as a strong evidence grid; if all normal sampling disks in the grid belong to independent points, then the grid is defined as a weak evidence grid. When a grid is in a candidate special soil, the grid is upgraded to obtain an upgraded grid; among them, the grid with no evidence is upgraded to the grid with weak evidence, the grid with weak evidence is upgraded to the grid with strong evidence, and the grid with strong evidence remains the grid with strong evidence. For each grid, including both upgraded and unupgraded grids, they are merged according to the four-connectivity rule to obtain regions with missing evidence and weak information.

[0029] Since candidate special soils (such as backfill areas) are already known specific patch types, there may be sparse or even no sampling points within them. If conventional grid evidence judgment rules are followed, these grids will be marked as "no evidence" or "weak evidence," and subsequently merged into "evidence-deficient areas" or "information-weak areas" by connected regions. This would cause the candidate special soil to be misjudged as a blank area that needs supplementary sampling, thus "disappearing" spatially. Therefore, it is necessary to enforce evidence upgrades on all grids within the candidate special soil range, so that they at least reach the level of "weak evidence" or even "strong evidence." This would preserve the integrity and special identity of these candidate special soils in subsequent processing and avoid misjudgments caused by gridding.

[0030] For each grid, including both upgraded and unupgraded grids, they are merged according to the four-connectivity rule to obtain regions with missing evidence and weak information. In the upgraded grid and the non-upgraded grid, the grids without evidence are merged into connected regions according to the four-connectivity rule. The connected regions are called regions with missing evidence. If weak evidence grids are merged into connected regions according to the four-connectivity rule, then the connected regions are called information-weak regions.

[0031] The four-connectivity rule is an existing technology, and the specific steps are as follows: For example, set up a 3x3 grid, mark the grid with no evidence as 1, and mark the other grids as 0, and now we want to obtain a region where evidence is missing; The 3x3 grid is as follows: Using the four-connection rule, the 1 in the top left corner is adjacent to the 1 on the right, so they are connected, but the 1 is not connected to the 0 below; the 1 in the middle is connected to the 1 above and the 1 above to the left, but not connected to the 0 on the right and below. Therefore, the three 1s in the top left, top middle, and middle constitute a region of missing evidence. For areas with missing evidence, the optimal sequential supplementation method is used to generate recommended supplementary evidence points; for areas with weak information, the radial extension method is used to generate recommended supplementary evidence points; a supplementary evidence sequence is generated based on the recommended supplementary evidence points; the current location of the sampling team is obtained, such as by using the location function built into the mobile device, and the distance between the recommended supplementary evidence points in the supplementary evidence sequence and the current location of the sampling team is collected in a smooth manner from near to far, and the soil of the recommended supplementary evidence points is collected.

[0032] For areas with missing evidence, the optimal sequential evidence completion method is used to generate recommended supplementary evidence points, including: For areas with missing evidence, retrieve the number of grids in each area, calculate the minimum distance from the center point of each grid in the area to all normal sampling disks in the survey area, and select the grid center point with the largest minimum distance as the first recommended supplementary evidence point; if this point happens to coincide with a normal sampling disk (distance is 0), then select the grid center point with the second largest minimum distance as the first recommended supplementary evidence point. For the center points of the remaining grid, obtain the minimum distance to each center point and the distance to the first recommended supplementary proof point; The comprehensive score of each remaining grid center point is obtained by weighting the minimum distance to the center point of the remaining grid and the distance to the first recommended supplementary point. Specifically, the minimum distance to the center point and the distance to the first recommended supplementary point are first normalized, and then the weighted calculation is performed again with the weights set to be the same to obtain the comprehensive score of the center point. Select the center with the highest overall score as the second recommended supplementary certification point; Remove the grids of the first and second recommended supplementary evidence points. For the remaining grids, calculate their comprehensive scores and select the center point with the highest comprehensive score as the third recommended supplementary evidence point. Stop generating recommended supplementary evidence points when the minimum distance is less than the standard distance. For the recommended supplementary evidence points, a list of recommended supplementary evidence point locations for areas with missing evidence is constructed; The standard distance is 0.5 times the average distance between all normal sampling pans within the survey area. The average distance between normal sampling pans reflects the typical scale of spatial variability of soil properties under the current sampling density. When the minimum distance from the center point of a grid in an area lacking evidence to the nearest normal sampling pan is less than half of the average distance, it indicates that the uncertainty at that location is lower than the average interval between normal sampling pans. Continuing to add recommended supplementary evidence points may result in them being too close to existing sampling points, wasting sampling resources. Therefore, using 0.5 times the average distance as the stopping threshold can effectively control the number of supplementary points while ensuring the integrity of spatial information coverage, achieving a reasonable balance between sampling efficiency and information acquisition.

[0033] Within areas lacking evidence, the system maximizes spatial information coverage with the fewest recommended supplementary evidence points. First, the grid center point with the largest minimum distance from all normal sampling points is selected as the first recommended supplementary evidence point, ensuring priority coverage of the least information-scarce areas. Then, by weighted balancing the information deficiency of the remaining grid center points with their distance from the first recommended supplementary evidence point, subsequent recommended supplementary evidence points that best fill gaps and avoid clustering are selected. When the minimum distance between the remaining grid center points is less than half the average distance of normal sampling points, it indicates that the uncertainty in the area has decreased to an acceptable level, and the generation of recommended supplementary evidence points ceases. This approach enables efficient and economical supplementation of spatial information in areas lacking evidence, even with limited sampling resources.

[0034] For areas with weak information, a radial extension method is used to generate recommended supplementary evidence points, including: Calculate the difference intensity of all independent points in the information-weak region, and arrange the difference intensity in descending order to obtain the difference intensity sequence; Using the intensity of difference as the weight, all independent points are weighted to obtain a weighted position. This weighted position is then used as the first recommended supplementary evidence point. If the weighted position falls exactly on a normal sampling point, the weighted position is moved one grid step in any direction and a new weighted position is obtained in the information-weak region. This new weighted position is then used as the first recommended supplementary evidence point. For each independent point and the first recommended supplementary evidence point, determine the direction and distance of extension; Starting from each independent point, according to the extension direction and extension distance, a group of candidate recommended supplementary evidence points is obtained; If a candidate recommended supplementary evidence point in the candidate recommended supplementary evidence point group overlaps with an existing sampling point or is not in the information-weak area, then the candidate recommended supplementary evidence point is discarded, and the remaining candidate recommended supplementary evidence points and the first recommended supplementary evidence point are used as recommended supplementary evidence points for the information-weak area, and a list of recommended supplementary evidence point locations for the information-weak area is constructed. The direction of extension is the direction from the independent point to the first recommended supplementary evidence point; The extension distance is the product of the distance between the independent point and the first recommended supplementary point and the difference strength of that independent point.

[0035] The steps to obtain the difference intensity of independent points are as follows: retrieve the rapid field physicochemical data of each normal sampling plate in the main soil, and obtain the mean and standard deviation of the rapid field physicochemical data in the main soil; The deviation of the rapid field physicochemical data is obtained by retrieving the rapid field physicochemical data of independent points and comparing it with the mean and standard deviation of the rapid field physicochemical data of the main soil. The rapid field physicochemical data includes organic matter content data, pH data, salinity data, characteristic element content data, and characteristic element content ratio data. The deviation of each data point is the ratio of the absolute difference between the mean and the standard deviation of the corresponding data in the rapid field physicochemical data of the main soil to the standard deviation of the corresponding data. The deviation of the obtained rapid on-site physicochemical data is averaged to obtain the average anomaly intensity of the independent point. The average anomaly intensity is then normalized to obtain the difference intensity of the independent point.

[0036] Specifically, the weighted position is obtained by weighting all independent points with the intensity of difference as the weight; the weight is set to be the same, that is, 1 divided by the number of independent points. In information-weak regions, the difference intensity of independent points is used as weights. First, a weighted centroid is determined as the most likely location to represent the anomaly center. Then, adaptive extension is performed along the direction of each independent point pointing to the centroid to explore whether the independent points expand towards the center or have continuity. This allows for efficient verification of the true boundary and internal structure of information-weak regions with the fewest recommended supplementary evidence points.

[0037] Visualize the sampling trays, areas with missing evidence, areas with weak information, recommended supplementary evidence points, and the scope of the survey area; First, visualize the survey area. Then, overlay sampling panels, areas with missing evidence, areas with weak information, and recommended supplementary evidence points on the survey area. The sampling panels include normal sampling panels and abnormal sampling panels.

[0038] like Figure 3 As shown, the intelligent processing system for multi-pane soil sampling data includes: Anomaly detection module: Obtain the unique identification information of each sampling disk, and use the three-dimensional spatial difference recognition method to determine whether the sampling disk is an abnormal sampling disk or a normal sampling disk based on the unique identification information of each sampling disk; Homogeneous clustering module: Generates candidate source cells and independent points based on the normal sampling disk; Subject identification module: Based on each candidate source unit, obtain its spatial range, take the candidate source unit with the largest spatial range as the subject soil, and the remaining candidate source units as non-subject soil, and determine whether the non-subject soil is a candidate special soil; for candidate special soil, use the built-in rule method to obtain the patch type of the candidate special soil; The evidence supplementation planning module: Based on the main soil and non-main soil, the survey area is constructed, and the evidence strength grid connected region extraction method is used to obtain the evidence-deficient areas and information-weak areas in the survey area; for the evidence-deficient areas, the optimal sequential evidence supplementation method is used to generate recommended evidence supplementation points; for the information-weak areas, the radial extension method is used to generate recommended evidence supplementation points; and a evidence supplementation sequence is generated based on the recommended evidence supplementation points. Visualization module: Visualizes the sampling disk, areas with missing evidence, areas with weak information, recommended supplementary evidence points, and the scope of the survey area.

[0039] In this embodiment, by integrating multi-pane sampling data with intelligent processing procedures, the entire chain of automated processing is realized, from sampling point anomaly identification, soil source unit division, candidate special soil patch identification to grid analysis of areas with missing evidence and weak information, and adaptive point supplementation generation. This significantly reduces the blindness and repetitive workload of field sampling, improves the pertinence and sampling efficiency of soil surveys, and provides scientific and efficient technical support for practical engineering applications such as farmland quality evaluation, contaminated site identification, and backfill area determination. At the same time, the generated visualization results facilitate on-site decision-making and long-term traceability. The three-dimensional spatial difference identification method can accurately distinguish between abnormal sampling disks and normal sampling disks; the transitive clustering and decomposition mechanism based on distance and organic matter content differences can automatically generate homologous candidate source units and independent points, avoiding the subjectivity of manual division; the evidence strength grid connected region extraction method is used to quantitatively identify information-weak areas and evidence-deficient areas, and the radial extension method and the optimal sequential supplementation method are used for different areas to generate recommended supplementation points, realizing the adaptability and spatial rationality of the supplementation strategy.

[0040] This application also provides an electronic device. The electronic device may include one or more processors and one or more memories. The memories store computer-readable code, which, when executed by the one or more processors, can perform the intelligent processing method and system for multi-disc soil sampling data as described above.

[0041] The methods and systems according to the embodiments of this application can also be implemented using the architecture of the electronic device shown in this application. The electronic device may include a bus, one or more CPUs, ROM, RAM, a communication port connected to a network, input / output, a hard disk, etc. The storage device in the electronic device, such as a ROM or hard disk, may store the intelligent processing method and system for multi-disc soil sampling data provided in this application. Furthermore, the electronic device may also include a user interface. Of course, the architecture shown in this application is merely exemplary; when implementing different devices, one or more components of the electronic device shown in this application may be omitted according to actual needs.

[0042] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising a reference structure" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.

[0043] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for intelligent processing of soil multi-pane sampling data, characterized in that, The method includes: Step S1: Obtain the unique identification information of each sampling plate. For the unique identification information of each sampling plate, use the three-dimensional spatial difference recognition method to determine whether the sampling plate is an abnormal sampling plate or a normal sampling plate. Step S2: Generate candidate source cells and independent points based on the normal sampling disk; Step S3: Based on each candidate source unit, obtain its spatial range, take the candidate source unit with the largest spatial range as the main soil, and the remaining candidate source units as non-main soil, and determine whether the non-main soil is a candidate special soil; for the candidate special soil, use the built-in rule method to obtain the patch type of the candidate special soil. Step S4: Based on the main soil and non-main soil, construct the survey area and use the evidence strength grid connected region extraction method to obtain the evidence-deficient areas and information-weak areas in the survey area; for the evidence-deficient areas, use the optimal sequential supplementation method to generate recommended supplementary evidence points; for the information-weak areas, use the radial extension method to generate recommended supplementary evidence points; generate a supplementary evidence sequence based on the recommended supplementary evidence points. Step S5: Visualize the sampling disk, areas with missing evidence, areas with weak information, recommended supplementary evidence points, and the scope of the investigation area.

2. The intelligent processing method for multi-disc soil sampling data according to claim 1, characterized in that, The process of obtaining the unique identifier information for each sampling disk includes: For each sampling point, basic information and spatial structure correlation data of the sampling point are obtained; and the soil collected from the sampling point is put into a sampling tray to obtain rapid on-site physicochemical data of the sampling tray. The basic information includes temperature, humidity, and wind speed data at the time of sampling; spatial structure correlation data includes the location, depth, and soil texture of the sampling point; rapid on-site physicochemical data includes organic matter content, pH, salinity, characteristic element content, and characteristic element content ratio data. By constructing a unique identifier from basic information, spatial structure association data, and rapid on-site physicochemical data, and binding it with the sampling disk, the unique identifier of the sampling disk is obtained.

3. The intelligent processing method for multi-pane soil sampling data according to claim 2, characterized in that, The method of using a three-dimensional spatial difference recognition to determine whether a sampling disk is an abnormal or normal sampling disk, based on the unique identifier information of each sampling disk, includes: Step ss1: Traverse each sampling disk, set the current sampling disk as the target sampling disk, set the search radius, and record the sampling disks within the search radius as feature sampling disks; check whether the target sampling disk satisfies any of the conditions in the horizontal spatial neighborhood comparison conditions. If it does, it is considered an abnormal sampling disk. Step ss2: For sampling disks at different depths in the same borehole, sort them in ascending order of depth. Define the sampling disk with the smaller depth as a shallow sampling disk and the sampling disk with the larger depth as a deep sampling disk. Compare the shallow sampling disk and the deep sampling disk in a pair of adjacent depths. If any of the conditions in the vertical spatial neighborhood comparison conditions are met, the deep sampling disk is considered an abnormal sampling disk. Step ss3: For the rapid field physicochemical data in all sampling disks, obtain the mean and standard deviation of the rapid field physicochemical data, and generate a virtual disk based on the mean; Each sampling disk is compared with a virtual disk; if the difference between any data point in the rapid on-site physicochemical data of the sampling disk and the corresponding data point in the virtual disk exceeds the baseline standard deviation, it is considered an abnormal sampling disk.

4. The intelligent processing method for multi-pane soil sampling data according to claim 3, characterized in that, The step of generating candidate source cells and independent points based on the normal sampling disk includes: For each normal sampling disc, the distance between each pair of normal sampling discs and the difference in organic matter content were obtained; The distance between each pair of normal sampling discs is divided into H intervals, and the median of the difference in organic matter content within each interval is obtained to construct a list of the median differences in organic matter content among the H intervals. And obtain the upper limit of the homogeneous spatial distance based on the distance of each interval; and retrieve the median of the largest organic matter content difference from the list of median differences in organic matter content. For the distance between each pair of normal sampling discs, if the distance is less than the upper limit of the distance between homologous spaces and the difference in organic matter content is less than the median of the largest difference in organic matter content, then the pair of normal sampling discs is considered to be homologous candidates. Transitively connect the homologous candidates to generate K candidate source units; Obtain the representative fingerprint of each candidate source unit, and check whether the difference in organic matter content between the normal sampling disk in each candidate source unit and the representative fingerprint is greater than the median of the largest organic matter content difference. If it is greater, the corresponding normal sampling disk is separated from the candidate source unit and used as an independent point.

5. The intelligent processing method for multi-disc soil sampling data according to claim 4, characterized in that, The process of obtaining patch types for candidate special soils using a built-in rule-based method includes: For the main soil sample, rapid on-site physicochemical data were retrieved from each normal sampling tray in the main soil sample. Calculate the direction of change of rapid field physicochemical data between each pair of normal sampling discs to construct a group of rapid field physicochemical data change directions of the main soil sample. For candidate special soils, we obtained the group of rapid on-site physicochemical data changes in candidate special soils. If the similarity between the rapid field physicochemical data change direction group in the candidate special soil and the rapid field physicochemical data change direction group in the main soil is greater than or equal to the basic similarity, it indicates that the candidate special soil has undergone natural variation. If the similarity between the rapid field physicochemical data change direction group of the candidate special soil and the rapid field physicochemical data change direction group of the main soil is less than the basic similarity, it indicates that the source of the candidate special soil is different from that of the main soil. The multi-index deviation decision method is used to determine the patch type of the candidate special soil.

6. The intelligent processing method for multi-disc soil sampling data according to claim 5, characterized in that, The method of using multiple index deviation decision-making to determine the patch type of candidate special soil includes: Extract the maximum and minimum values ​​of each data point from all normal sampling pans in the main soil sample; Retrieve representative fingerprints from candidate special soils; For each data point, its maximum and minimum values ​​are compared with the corresponding data in the fingerprint. If the fingerprint of a certain data point is greater than the maximum value of the corresponding data point, it indicates that the data point is significantly high; if it is less than the minimum value of the corresponding data point, it indicates that the data point is significantly low. When the organic matter content data of the candidate special soil is significantly low and the characteristic element content ratio data is significantly deviated (significant deviation includes significantly high or significantly low), it is a backfill area. When both pH and salinity data are significantly high, it indicates a salinized area. If the organic matter content is significantly higher and the depth is significantly lower, it indicates a sedimentary zone. If the content of at least one characteristic element in the characteristic element content data significantly exceeds the corresponding risk screening value, then it is a pollution disturbance area; When the data on organic matter content, pH, and salinity are all normal, but the ratio of characteristic elements deviates significantly, it indicates an area where foreign soil has been introduced. If none of the above conditions are met, the plaque type is uncertain and marked as "to be verified".

7. The intelligent processing method for multi-disc soil sampling data according to claim 6, characterized in that, The survey area is constructed based on the main soil and non-main soil, and the evidence strength grid connected region extraction method is used to obtain the evidence-deficient areas and information-weak areas within the survey area, including: The survey area is constructed by using the main soil and non-main soil, and D grids are generated for the survey area. For each grid, check whether there is a normal sampling disk in the grid. If not, the grid is defined as a grid without evidence. If it exists, check whether the normal sampling disk in the grid belongs to the candidate source unit. If it does, the grid is defined as a grid with strong evidence. If it does not belong to the candidate source unit, i.e., an independent point, the grid is defined as a grid with weak evidence. When a grid is in a candidate special soil, the grid is upgraded to obtain an upgraded grid; among them, the grid with no evidence is upgraded to the grid with weak evidence, the grid with weak evidence is upgraded to the grid with strong evidence, and the grid with strong evidence remains the grid with strong evidence. For each grid, including both upgraded and unupgraded grids, they are merged according to the four-connectivity rule to obtain regions with missing evidence and weak information.

8. The intelligent processing method for multi-pane soil sampling data according to claim 7, characterized in that, The method of generating recommended supplementary evidence points using the optimal sequential supplementation method for regions with missing evidence includes: For areas with missing evidence, retrieve the number of grids in each area, calculate the minimum distance from the center point of each grid in the area to all normal sampling disks in the survey area, and select the grid center point with the largest minimum distance as the first recommended supplementary evidence point; if this point happens to coincide with a normal sampling disk, select the grid center point with the second largest minimum distance as the first recommended supplementary evidence point. For the center points of the remaining grid, obtain the minimum distance to each center point and the distance to the first recommended supplementary proof point; The comprehensive score of the center point of each remaining grid is obtained by weighting the minimum distance to the center point of the remaining grid and the distance to the first recommended supplementary proof point. Select the center with the highest overall score as the second recommended supplementary certification point; Remove the grids of the first and second recommended supplementary evidence points. For the remaining grids, calculate their comprehensive scores and select the center point with the highest comprehensive score as the third recommended supplementary evidence point. Stop generating recommended supplementary evidence points when the minimum distance is less than the standard distance. For the recommended supplementary evidence points, a list of recommended supplementary evidence point locations for areas with missing evidence is constructed.

9. The intelligent processing method for multi-disc soil sampling data according to claim 8, characterized in that, The method of generating recommended supplementary evidence points using a radial extension approach for areas with weak information includes: Calculate the difference intensity of all independent points in the information-weak region, and arrange the difference intensity in descending order to obtain the difference intensity sequence; Using the intensity of difference as the weight, all independent points are weighted to obtain a weighted position. This weighted position is then used as the first recommended supplementary evidence point. If the weighted position falls exactly on a normal sampling point, the weighted position is moved one grid step in any direction and a new weighted position is obtained in the information-weak region. This new weighted position is then used as the first recommended supplementary evidence point. For each independent point and the first recommended supplementary evidence point, determine the direction and distance of extension; Starting from each independent point, according to the extension direction and extension distance, a group of candidate recommended supplementary evidence points is obtained; If a candidate recommended supplementary evidence point in the candidate recommended supplementary evidence point group overlaps with an existing sampling point or is not located in an area with weak information, then that candidate recommended supplementary evidence point is discarded. The remaining candidate recommended supplementary evidence points and the first recommended supplementary evidence point are used as recommended supplementary evidence points for the area with weak information, and a list of recommended supplementary evidence point locations for the area with weak information is constructed.

10. A soil multi-pan sampling data intelligent processing system, applied to the soil multi-pan sampling data intelligent processing method according to any one of claims 1 to 9, characterized in that, include: Anomaly detection module: Obtain the unique identification information of each sampling disk, and use the three-dimensional spatial difference recognition method to determine whether the sampling disk is an abnormal sampling disk or a normal sampling disk based on the unique identification information of each sampling disk; Homogeneous clustering module: Generates candidate source cells and independent points based on the normal sampling disk; Subject identification module: Based on each candidate source unit, obtain its spatial range, take the candidate source unit with the largest spatial range as the subject soil, and the remaining candidate source units as non-subject soil, and determine whether the non-subject soil is a candidate special soil; for candidate special soil, use the built-in rule method to obtain the patch type of the candidate special soil; The evidence supplementation planning module: Based on the main soil and non-main soil, the survey area is constructed, and the evidence strength grid connected region extraction method is used to obtain the evidence-deficient areas and information-weak areas in the survey area; for the evidence-deficient areas, the optimal sequential evidence supplementation method is used to generate recommended evidence supplementation points; for the information-weak areas, the radial extension method is used to generate recommended evidence supplementation points; and a evidence supplementation sequence is generated based on the recommended evidence supplementation points. Visualization module: Visualizes the sampling disk, areas with missing evidence, areas with weak information, recommended supplementary evidence points, and the scope of the survey area.

Images