Land space intelligent planning system and method based on multi-source data fusion

Abstract

The invention relates to the technical field of land space planning, and discloses an intelligent land space planning system and method based on multi-source data fusion, and the system comprises an analysis unit construction module which constructs an analysis unit; the effective capillary rising height calculation module is used for calculating the effective capillary rising height; the image date set construction module is used for constructing an image date set; the salt surface response index calculation module is used for calculating the scaled salt surface response index; the self-adaptive threshold calculation module is used for calculating a self-adaptive threshold; and the land parcel category division module outputs land parcel categories. According to the method, the effective capillary rising height is calculated through the soil texture and the underground water salinity, the underground physical condition and the remote sensing image are associated, the subjectivity of manually setting the threshold value is avoided through the self-adaptive threshold value, the territorial space attribute graph with the attributes capable of being queried and the area capable of being traced is constructed, and the precision and reliability of cultivated land category division are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of land spatial planning technology, and more specifically, to a land spatial intelligent planning system and method based on multi-source data fusion. Background Technology

[0002] In the field of land spatial planning and farmland protection, salinization is the core factor restricting the availability of farmland in arid and semi-arid regions. Accurately determining the degree of farmland salinization and classifying farmland into usable, unusable and need-to-improve categories is a key requirement for optimizing farmland layout.

[0003] Traditional methods for identifying farmland salinization largely rely solely on spectral indices from remote sensing imagery, such as the ratio of shortwave infrared to near-infrared, or only roughly assess the risk of salt rise based on groundwater depth. However, they fail to link underground physical conditions, such as capillary action capacity determined by soil texture and groundwater salinity, with remote sensing spectral characteristics, such as equivalent capillary radius and effective capillary reach. Actual salinization requires the physical process of underground salts reaching the surface and accumulating through capillary action; therefore, relying on a single data point is insufficient to distinguish true farmland salinization and is prone to misjudgment. Furthermore, traditional farmland salinization identification methods mostly use manually set salinization spectral index thresholds, without considering the different regional characteristics. Differences in soil texture and groundwater conditions, such as the effective capillary reach height being much greater in sandy soil than in clay, mean that the same threshold may misclassify saline-free farmland as saline-alkali farmland in sandy soil areas, while it may miss mildly saline-alkali farmland in clay areas. This makes the classification results difficult to reuse in different regions and has extremely poor transferability. In addition, traditional remote sensing image selection mostly relies on experience and does not consider the impact of groundwater depth fluctuations on capillary action. If images are selected during periods of temporary groundwater rise, capillary action may produce false salinity, while if images are selected during periods of excessive groundwater drop, they may mask true salinity, which cannot guarantee the reliability of the image data and further exacerbates the classification error.

[0004] Therefore, the above schemes are difficult to accurately and reliably classify farmland salinization, resulting in an inability to meet the requirements of intelligent land space planning for accuracy, reliability, and cross-regional applicability. Summary of the Invention

[0005] This invention provides a land space intelligent planning system and method based on multi-source data fusion, which solves the technical problems in the background art mentioned above.

[0006] This invention provides a land spatial intelligent planning system based on multi-source data fusion, comprising: The analysis unit construction module is used to map multi-source data of gridded plots to construct analysis units according to spatial location and observation date. The multi-source data includes: groundwater depth and salinity of observation wells, soil texture of soil samples, and shortwave infrared and near-infrared reflectance of remote sensing images; The effective capillary reach height calculation module is used to determine the equivalent capillary radius based on soil texture and calculate the effective capillary reach height in combination with groundwater salinity. The image date set construction module is used to calculate the capillary reach margin based on the effective capillary reach height and the groundwater depth, determine the image selection weight by the ratio of the capillary reach margin to the effective capillary reach height, and construct the image date set based on the image selection weight. The salt surface response index calculation module is used to calculate the salt surface response index within the image date set and scale it to the scaled salt surface response index according to the image selection weight. An adaptive threshold calculation module is used to filter the time marker with the highest image selection weight from the image date set of each analysis unit, and calculate the adaptive threshold based on the capillary margin under that time marker. The land parcel classification module is used to output the land parcel classification of the analysis unit based on the capillary surfacing margin, the scaled salt surface response index and the adaptive threshold, and to construct a land space attribute map by calculating the land parcel area. The land parcel categories include: usable arable land, unusable arable land, and arable land requiring improvement.

[0007] Furthermore, the analysis units are defined based on the coordinate system used by the base map, and each analysis unit has a unique spatial boundary and identifier. A time identifier is established to be associated with the analysis units, with the time identifier being the Gregorian calendar date and the date precision being the day. An observation data catalog is established to be associated with the analysis units. The observation data catalog consists of a list of observation wells, a list of soil sample points, and a list of remote sensing images. The list of observation wells includes the well location coordinates of each observation well, the list of soil sample points includes the sample point coordinates of each soil sample point, and the list of remote sensing images includes the shortwave infrared reflectance, near-infrared reflectance, and pixel center coordinates of each pixel in the remote sensing images. Moreover, the well location coordinates, sample point coordinates, and pixel center coordinates are all consistent with the coordinate system used by the base map.

[0008] Furthermore, similar observations from different locations within the same analysis unit are merged using a distance-inverse weighted method. First, the centroid of the analysis unit is used as a reference point. The planar distance from each point to the corresponding reference point of the analysis unit is calculated. A very small positive number is added to this planar distance, and the result is squared and the reciprocal is taken to obtain the weight corresponding to that point. Then, the groundwater depth of each observation well is multiplied by the corresponding weight, the sum is divided by the sum of all weights, and the groundwater depth merged value of the analysis unit under the time marker is obtained. The groundwater salinity merged value, shortwave infrared reflectance merged value, and near-infrared reflectance merged value are calculated in the same way. Finally, the soil texture is stored in the corresponding analysis unit in the form of a classification code.

[0009] Furthermore, the equivalent capillary radius is first determined based on the soil texture using the soil water characteristic curve. Then, the surface tension and density of the saline water are determined based on the salinity and temperature of the groundwater. The surface tension of the saline water is multiplied by 2, and then multiplied by the cosine of the contact angle between the groundwater and soil particles as the numerator. The product of the saline water density, gravitational acceleration, and equivalent capillary radius is used as the denominator. The effective capillary height is obtained by dividing the numerator by the denominator.

[0010] Furthermore, the difference between the effective capillary reach height and the groundwater depth of the analysis unit under the same time marker is first calculated to obtain the capillary reach margin. Then, the ratio between the absolute value of the capillary reach margin and the effective capillary reach height is calculated. The intermediate result is obtained by subtracting the ratio from 1. The value range of the intermediate result is controlled between 0 and 1 by the clipping function to obtain the image selection weight.

[0011] Furthermore, the analysis unit selects all time markers for images with a weight greater than or equal to a preset weight threshold at each time marker, and then aggregates these time markers to construct the image date set corresponding to the analysis unit, where the preset weight threshold is a custom parameter.

[0012] Furthermore, the ratio of short-wave infrared reflectance to near-infrared reflectance at each time marker is calculated and analyzed as the salt surface response index. If the near-infrared reflectance is less than a preset reflectance threshold, the salt surface response index at that time marker is marked as the missing salt surface response index. The preset reflectance threshold is a user-defined parameter. The product between the image selection weight and the non-missing salt surface response index is calculated and used as the scaled salt surface response index.

[0013] Furthermore, from the image date set of each analysis unit, the time identifier with the highest image selection weight is selected, and this time identifier is determined as the representative time. The capillary reach margin and effective capillary reach height of the analysis unit at the representative time are obtained simultaneously as the representative values ​​of capillary reach margin and effective capillary reach height. All analysis units with a capillary reach margin representative value greater than or equal to zero are selected, and these analysis units are aggregated to construct a sample set. A histogram distribution is established based on the scaled salt surface response index corresponding to each analysis unit in the sample set. The inter-class variance corresponding to different candidate thresholds in the histogram distribution is calculated. The inter-class variance is the product of the proportion of samples on both sides of the candidate threshold and the square of the difference between the means of samples on both sides of the candidate threshold. The candidate threshold with the largest inter-class variance is selected as the adaptive threshold.

[0014] Furthermore, within the time range corresponding to the image date set, the median of the scaled salt surface response index is taken. If the representative value of capillary reach margin is greater than or equal to 0 and the median of the scaled salt surface response index is greater than or equal to the adaptive threshold, the corresponding analysis unit is determined to be unusable arable land. If the representative value of capillary reach margin is less than or equal to -30% of the representative value of effective capillary reach height and the median of the scaled salt surface response index is less than the adaptive threshold, the corresponding analysis unit is determined to be usable arable land. In other cases, the corresponding analysis unit is determined to be arable land that needs improvement.

[0015] This invention provides a land spatial intelligent planning method based on multi-source data fusion, comprising the following steps: Step S201: Construct analysis units by mapping the multi-source data of the gridded plots according to spatial location and observation date; Step S202: Determine the equivalent capillary radius based on soil texture, and calculate the effective capillary height by combining it with groundwater salinity; Step S203: Calculate the capillary reach margin based on the effective capillary reach height and groundwater depth, determine the image selection weight by the ratio of the capillary reach margin to the effective capillary reach height, and construct an image date set based on the image selection weight. Step S204: Calculate the salt surface response index within the image date set, and scale it to the scaled salt surface response index according to the image selection weight. Step S205: Select the time marker with the highest image selection weight from the image date set of each analysis unit, and calculate the adaptive threshold based on the capillary margin under the time marker. Step S206: Output the land parcel category of the analysis unit based on the capillary surface margin, the scaled salt surface response index and the adaptive threshold, and construct a land space attribute map by calculating the land parcel area.

[0016] The beneficial effects of this invention are as follows: This invention constructs analysis units based on spatial location and observation date, employs distance-inverse weighted merging of multi-source data to ensure the accuracy of basic data; it calculates effective capillary reach height by combining soil texture and groundwater salinity, correlates underground physical conditions with remote sensing images, and determines image selection weight through capillary reach margin to screen physically reliable image dates; it calculates adaptive thresholds based on the sample set and maximum inter-class variance, avoiding the subjectivity of manually setting thresholds; finally, it classifies usable arable land, unusable arable land, and arable land requiring improvement into categories, and constructs a land spatial attribute map with queryable attributes and traceable area, significantly improving the accuracy and reliability of arable land classification, while enhancing cross-regional portability, providing accurate and stable decision support for intelligent land spatial planning. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of a land spatial intelligent planning system based on multi-source data fusion according to the present invention; Figure 2 This is a flowchart of a land spatial intelligent planning method based on multi-source data fusion according to the present invention.

[0018] In the figure: Analysis unit construction module 101, effective capillary height calculation module 102, image date set construction module 103, salt surface response index calculation module 104, adaptive threshold calculation module 105, and land parcel classification module 106. Detailed Implementation

[0019] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0020] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of the present invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in one or more embodiments of the present invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" indicate that the element or object preceding the term encompasses the elements or objects listed following the term and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0021] like Figures 1-2 As shown, a land spatial intelligent planning system based on multi-source data fusion includes: Analysis unit construction module 101 is used to map multi-source data of gridded plots to construct analysis units according to spatial location and observation date; The multi-source data includes: groundwater depth and salinity of observation wells, soil texture of soil samples, and shortwave infrared and near-infrared reflectance of remote sensing images; The effective capillary reach height calculation module 102 is used to determine the equivalent capillary radius based on soil texture and calculate the effective capillary reach height in combination with groundwater salinity. The image date set construction module 103 is used to calculate the capillary reach margin based on the effective capillary reach height and the groundwater burial depth, determine the image selection weight based on the ratio of the capillary reach height to the effective capillary reach height, and construct the image date set based on the image selection weight. Salt surface response index calculation module 104 is used to calculate the salt surface response index within the image date set and scale it to the scaled salt surface response index according to the image selection weight. The adaptive threshold calculation module 105 is used to filter the time identifier with the highest image selection weight from the image date set of each analysis unit, and calculate the adaptive threshold based on the capillary margin under the time identifier. The land parcel classification module 106 is used to output the land parcel classification of the analysis unit based on the capillary expression margin, the scaled salt surface response index and the adaptive threshold, and to construct a land space attribute map by calculating the land parcel area. The land parcel categories include: usable arable land, unusable arable land, and arable land requiring improvement.

[0022] In one embodiment of the present invention, analysis units are defined based on the coordinate system used by a map base plate, and each analysis unit has a unique spatial boundary and identifier; a time identifier is established and associated with the analysis unit, the time identifier being a Gregorian calendar date with a date precision of day; an observation data catalog is established and associated with the analysis unit, the observation data catalog consisting of an observation well list, a soil sample point list, and a remote sensing image list, wherein the observation well list includes the well location coordinates of each observation well, the soil sample point list includes the sample point coordinates of each soil sample point, and the remote sensing image list includes the shortwave infrared reflectance, near-infrared reflectance, and pixel center coordinates of each pixel in the remote sensing image, and the well location coordinates, sample point coordinates, and pixel center coordinates are all consistent with the coordinate system used by the map base plate.

[0023] It should be noted that a single base map refers to a benchmark map used in fields such as land spatial planning and natural resource management, serving as a unified carrier of spatial data. It contains a unified geographical framework, administrative divisions, topography, and other basic spatial information, acting as the foundation for integrating various professional data. The coordinate system used in this base map is the national legal plane rectangular coordinate system, typically the CGCS2000, ensuring that all spatial data are overlaid and correlated within the same coordinate framework. Analysis units can be obtained manually. If analyzing at the parcel level, the boundaries of already-established parcels on the current land use map (e.g., contracted land parcels) are used. For land parcels and cultivated land parcels, if a unified scale is required, the unit boundaries are determined manually based on administrative or natural zoning maps (such as watershed zoning, soil type zoning, etc.) to ensure that the boundaries match the actual geographical features. In addition, analysis units can also be automatically obtained through grid division. For example, the grid size can be set according to the density of observation data (for example, when the average spacing between observation wells is 50 meters, the grid side length can be set to 50 meters so that each grid contains at least one observation point), and an equally spaced rectangular grid can be generated on a map base. The unique identifier of the analysis unit can be generated by the snowflake algorithm or UUID, which will not be elaborated here.

[0024] It should be noted that, for a given time marker, the groundwater depth and salinity of each observation well in the observation well list, the soil texture of each soil sample point in the soil sample point list, and the shortwave infrared reflectance and near-infrared reflectance of each pixel in the remote sensing image list under the time marker are mapped to the corresponding analysis units according to the nearest neighbor principle or the distance weight principle.

[0025] In one embodiment of the present invention, similar observations from different points within the same analysis unit are merged. The merging adopts a distance-inverse weighting method. First, the centroid of the analysis unit is used as a reference point. The planar distance from each point to the corresponding reference point of the analysis unit is calculated. A very small positive number is added to the planar distance. The result is squared and the reciprocal is taken to obtain the weight corresponding to that point. Then, the groundwater depth of each observation well is multiplied by the corresponding weight, the sum is divided by the sum of all weights, and the groundwater depth merged value of the analysis unit under the time identifier is obtained. The groundwater salinity merged value, shortwave infrared reflectance merged value, and near-infrared reflectance merged value are calculated in the same way. Finally, the soil texture is stored in the corresponding analysis unit in the form of classification code.

[0026] Specifically, the formula for calculating the aggregated value of groundwater depth for the u-th analysis unit at time t is as follows: , ,in This indicates the groundwater depth of the j-th observation well at time t. This represents the planar distance from the j-th observation well to the u-th analysis unit reference point. It represents a very small positive number.

[0027] It should be noted that there may be multiple observation points within the spatial range of an analysis unit. These points will generate multiple similar observations. Therefore, it is necessary to merge similar observations from different points within the same analysis unit to integrate these discrete data into the representative value of the analysis unit, avoiding data redundancy or conflicts. The setting of the minimum positive number should be much smaller than the minimum planar distance that may occur in actual observation. Its default value is 0.01 meters. The main purpose is to prevent the denominator from being 0 when the point happens to fall on the reference point of the analysis unit, which would result in a planar distance of 0. Soil texture classification coding refers to the type of soil texture, such as sandy soil, loam, clay, sandy loam, loam-clay, etc. The corresponding classification codes can adopt the coding rules in national standards or industry specifications. For example, the value 1 represents sandy soil, the value 2 represents loam, etc., which will not be elaborated here.

[0028] In one embodiment of the present invention, the equivalent capillary radius is first determined based on the soil texture using the soil water characteristic curve. Then, the surface tension and density of the saline water are determined based on the salinity and temperature of the groundwater. The surface tension of the saline water is multiplied by 2, and then multiplied by the cosine of the contact angle between the groundwater and the soil particles as the numerator. The product of the saline water density, gravitational acceleration, and equivalent capillary radius is used as the denominator. The effective capillary height is obtained by dividing the numerator by the denominator.

[0029] Specifically, the height reached by the u-th analysis unit at time t is calculated using soil texture. The calculation formula is as follows: ,in This represents the groundwater salinity in the u-th analysis unit at time t, where T represents the temperature. Indicates the surface tension of salt water. The value represents the density of the salt water, and g represents the acceleration due to gravity. Let represent the equivalent capillary radius of the u-th analysis unit. This indicates the contact angle between groundwater and soil particles.

[0030] It should be noted that the soil water characteristic curve describes the relationship between soil moisture content and soil water suction. Different soil textures correspond to different curve shapes. This curve can be used to obtain relevant information about soil pores, such as the equivalent capillary radius. In addition, the equivalent capillary radius can be obtained by consulting an industry reference table, which is a table that directly corresponds soil texture to the equivalent capillary radius. The equivalent capillary radius approximates the complex and diverse pore distribution in the soil as the radius of a uniform capillary, which simplifies the calculation of capillary rise height.

[0031] For example, the equivalent capillary radius values ​​for sandy soil, sandy loam, loam, loam-clay, and clay are 150–300 micrometers, 50–150 micrometers, 30–50 micrometers, 10–30 micrometers, and 5–10 micrometers, respectively. In practical applications, these values ​​can be fine-tuned based on local soil measurement data, with the default value being the median of the corresponding range.

[0032] It should be noted that the surface tension and density of saline solution can be obtained by consulting manuals or through laboratory measurements, such as measuring surface tension with a surface tension meter and density with a densitometer. The contact angle between groundwater and soil particles represents the angle between the edge of the droplet formed by groundwater on the surface of the soil particles and the surface of the soil particles. This angle reflects the wetting ability of groundwater on soil particles; the smaller the angle, the stronger the wetting ability, and the easier it is for water to diffuse in the soil pores. Conversely, the larger the angle, the weaker the wetting ability, and the more difficult it is for water to diffuse in the pores. The default value of the cosine of the groundwater is 1. At this time, the groundwater can completely wet the soil particles and the contact angle is close to zero degrees. If there is severe salinization that causes the soil particle surface to crust, the cosine of the contact angle between the groundwater and the soil particles is artificially reduced. The effective capillary height represents the maximum vertical height that groundwater can be transported upward by capillary action under soil pore conditions and saline physical properties. It directly reflects whether groundwater can reach the vicinity of the surface through capillary action. If this height is greater than the groundwater burial depth, it means that the groundwater can reach the vicinity of the surface. Otherwise, it means that the groundwater cannot reach the surface.

[0033] In one embodiment of the present invention, if there is bubble infiltration pressure in the analysis unit, the effective capillary height is calculated based on the bubble infiltration pressure. First, the surface tension and density of the saline water are determined based on the salinity and temperature of the groundwater. The ratio of the surface tension of the saline water to the density of the saline water is calculated. Then, the ratio of the surface tension of the clean water to the density of the clean water is calculated. Finally, the bubble infiltration pressure is multiplied by the ratio of the surface tension of the saline water to the density of the saline water and divided by the ratio of the surface tension of the clean water to the density of the clean water to obtain the effective capillary height.

[0034] Specifically, the height reached by the u-th analysis unit on the effective capillary at time t is calculated using the bubble infiltration pressure. The calculation formula is as follows: ,in This represents the bubble infiltration pressure in the u-th analysis unit. Indicates the surface tension of clear water. This indicates the density of pure water.

[0035] It should be noted that bubble infiltration pressure represents the minimum pressure required for water to begin entering the soil's air-phase pores and forming a continuous flow. Determining whether bubble infiltration pressure exists in an analytical unit involves verifying the existence of measured soil infiltration test data for that unit. Bubble infiltration pressure can be obtained using the pressure membrane method. This involves selecting sampling points within the analytical unit using a five-point sampling method, collecting undisturbed soil samples using a ring sampler (a cylindrical metal container with a fixed volume) to avoid damaging the soil structure, and collecting three parallel samples from each sampling point. The collected undisturbed soil samples are then soaked in clean water for 24 hours to allow for complete osmosis. Saturation means the pores are filled with water and no air remains. The saturated soil sample is then placed in the sample chamber of the pressure membrane apparatus. The bottom of the sample chamber is a porous membrane (allowing only water to pass through and preventing air from passing through). A pressure source is connected, and the pressure is gradually increased from a low pressure (for example, an initial pressure of 0.01 MPa, increasing by 0.01 MPa each time, and stabilizing each pressure level for 2 hours). After each pressure level stabilizes, the weight of the soil sample is weighed (to calculate the change in water content). When the pressure increases to the point where the soil sample begins to lose water significantly, the corresponding pressure value is the bubble infiltration pressure. The bubble infiltration pressure can be obtained by periodic field testing, which will not be elaborated here.

[0036] It should be noted that the surface tension and density of clean water can also be obtained by consulting manuals or by laboratory measurement. When the analysis unit has bubble infiltration pressure, using this parameter to calculate the effective capillary height is more accurate than calculating it based on soil texture. Bubble infiltration pressure is a measured parameter that directly reflects the connectivity and pore size of soil pores, and can accurately capture the pore differences within the same soil texture. In contrast, the equivalent capillary radius obtained based on soil texture is an average approximation of similar soils, and the calculation deviation may be large. However, the cost of calculating the effective capillary height based on bubble infiltration pressure is higher. To ensure the accuracy of the calculation, this method can be given priority.

[0037] In one embodiment of the present invention, the difference between the effective capillary reach height and the groundwater depth under the same time identifier of the analysis unit is first calculated to obtain the capillary reach margin. Then, the ratio between the absolute value of the capillary reach margin and the effective capillary reach height is calculated. The intermediate result is obtained by subtracting the ratio from 1. The value range of the intermediate result is controlled between 0 and 1 by the clipping function to obtain the image selection weight.

[0038] Specifically, the image selection weight of the u-th analysis unit at time t is... The calculation formula is as follows: ,in and represents the capillary reach margin and effective capillary reach height of the u-th analysis unit at time t, respectively, and clip represents the clipping function.

[0039] In one embodiment of the present invention, the filtering analysis unit selects all time markers whose image selection weight is greater than or equal to a preset weight threshold under each time marker, and summarizes these time markers to construct the image date set corresponding to the analysis unit. The preset weight threshold is a custom parameter, and the default value of the preset weight threshold is 0.8.

[0040] It should be noted that the closer the image selection weight is to 1, the smaller the relative deviation between the capillary reach margin and the effective capillary reach height, and the more stable the matching state between groundwater level and capillary action. At this time, the image can more accurately reflect the relevant characteristics of surface salinization. Therefore, the subsequent selection of image selection weights to construct the image date set can avoid the interference of factors such as groundwater fluctuations and capillary instability, thereby improving the accuracy of subsequent land parcel classification.

[0041] In one embodiment of the present invention, the ratio of short-wave infrared reflectance to near-infrared reflectance at each time marker is calculated and analyzed as the salt surface response index. If the near-infrared reflectance is less than a preset reflectance threshold, the salt surface response index at that time marker is marked as the missing salt surface response index. The preset reflectance threshold is a user-defined parameter with a default value of 0.01. The product between the image selection weight and the non-missing salt surface response index is calculated and used as the scaled salt surface response index.

[0042] It should be noted that salt crusts and salt frosts, among other salinized materials, have a strong scattering effect on the short-wave infrared band, significantly increasing short-wave infrared reflectance. However, they have a relatively strong absorption effect on the near-infrared band, keeping near-infrared reflectance at a low level. The ratio of these two factors can amplify the spectral differences between salinized and non-salinized surfaces; a larger ratio indicates more obvious salt traces on the surface. The atmospherically corrected remote sensing reflectance ranges from 0 to 1. When the near-infrared reflectance is less than the preset reflectance threshold, it indicates that the area is mostly shadowed water or has data anomalies. In this case, a small denominator will cause the salt surface response index to increase sharply, resulting in anomalies. Therefore, threshold judgment is used to exclude anomalies and ensure the accuracy and stability of subsequent calculations.

[0043] It should be noted that the salt surface response index is based solely on remote sensing spectral characteristics and only reflects the spectral information visible on the surface. It does not take into account whether the physical state of the underground soil and groundwater supports the true manifestation of salt traces. For example, a high salt surface response index on a certain date may indicate temporary overburden disturbance rather than true salinization. However, the image selection weight can quantify the reliability of the underground physical conditions on that date. That is, the closer the image selection weight is to 1, the more stable the underground conditions are, which is conducive to the true manifestation of salt traces. Therefore, the scaled salt surface response index obtained by multiplying the two can more closely approximate the true salinization state.

[0044] In one embodiment of the present invention, the time identifier with the highest image selection weight is selected from the image date set of each analysis unit, and the time identifier is determined as the representative time. The capillary reach margin and effective capillary reach height of the analysis unit at the representative time are obtained simultaneously as the representative values ​​of capillary reach margin and effective capillary reach height. All analysis units with a capillary reach margin representative value greater than or equal to zero are selected, and these analysis units are aggregated to construct a sample set.

[0045] In one embodiment of the present invention, a histogram distribution is established based on the scaled salt surface response index corresponding to each analysis unit in the sample set. The inter-class variance corresponding to different candidate thresholds in the histogram distribution is calculated. The inter-class variance is the product of the proportion of samples on both sides of the candidate threshold and the square of the difference between the means of samples on both sides of the candidate threshold. The candidate threshold with the largest inter-class variance is selected as the adaptive threshold.

[0046] Specifically, adaptive threshold The calculation formula is as follows: ,in and These represent the candidate thresholds as follows: The sample proportions on the left and right sides, and These represent the candidate thresholds as follows: The sample means on the left and right sides, This indicates that the candidate threshold with the largest inter-class variance is selected.

[0047] It should be noted that, firstly, the range of values ​​for all scaled salt surface response indices in the sample set is determined. Then, intervals are divided according to a preset fixed class interval. For example, if the value range is between 0.5 and 1.5, dividing the range with a preset fixed class interval of 0.2 will yield five consecutive intervals: 0.5–0.7, 0.7–0.9, 0.9–1.1, 1.1–1.3, and 1.3–1.5. Next, the number of samples in each interval is counted, and a histogram is constructed with the interval as the horizontal axis and the number of samples as the vertical axis. The candidate threshold is a key value for covering the histogram, and is usually taken as the actual sample value at the interval boundary or midpoint. The target value, i.e., the scaled salt surface response index in the sample set, is used to determine whether the value is less than the candidate threshold. Samples with values ​​less than or equal to the candidate threshold are considered to be on the left, and samples with values ​​greater than or equal to the candidate threshold are considered to be on the right. In the classification scenario of farmland salinization, high values ​​of the scaled salt surface response index usually correspond to areas with high salinization, while low values ​​correspond to areas with low or no salinization. Selecting the candidate threshold with the largest inter-class variance as the adaptive threshold can avoid the problem of classification errors caused by subjectively setting a fixed threshold. It adapts to the distribution characteristics of sample data in different regions and at different times, without the need for manual adjustment of the threshold, thereby ensuring the objectivity of the threshold setting and the reliability of the classification results.

[0048] In one embodiment of the present invention, within the time range corresponding to the image date set, the median of the scaled salt surface response index is taken. If the representative value of capillary reach margin is greater than or equal to 0 and the median of the scaled salt surface response index is greater than or equal to the adaptive threshold, the corresponding analysis unit is determined to be unusable arable land. If the representative value of capillary reach margin is less than or equal to -30% of the representative value of effective capillary reach height and the median of the scaled salt surface response index is less than the adaptive threshold, the corresponding analysis unit is determined to be usable arable land. In other cases, the corresponding analysis unit is determined to be arable land to be improved.

[0049] It should be noted that a capillary reach margin value greater than or equal to 0 indicates that the effective capillary reach height is greater than or equal to the groundwater depth, meaning that the maximum transport height of capillary action can cover the distance from groundwater to the surface. Groundwater containing dissolved salts can be continuously transported to the surface through capillary ascent. Furthermore, a median salt surface response index after scaling greater than or equal to the adaptive threshold indicates that obvious salt crusts, salt frost, and other salinization substances have formed on the surface. The effect of these substances on the short-wave infrared reflectance causes the spectral index to exceed the adaptive threshold. The combination of these two factors indicates that the underground salt transport path of the analysis unit is unobstructed, surface salinization has become a given fact, and the soil salinity exceeds the tolerance range of crops, making normal cultivation impossible. Therefore, the land is determined to be unusable for cultivation.

[0050] It should be noted that if the representative value of capillary reach margin is less than or equal to -30% of the representative value of effective capillary reach height, it means that the groundwater depth is more than 30% deeper than the effective capillary reach height. In other words, the maximum transport height of capillary action cannot cover the distance from groundwater to the surface, and there is a 30% margin. Salt cannot be transported to the surface through capillary rise. The median of the scaled salt surface response index is less than the adaptive threshold, indicating that there are no obvious salinization substances on the surface, the soil salinity is within the range that crops can tolerate, and there is no threat of salt damage. The combination of these two factors indicates that the analysis unit has neither the possibility of underground salt transport nor any signs of surface salinization. The soil conditions meet the needs of crop growth, and therefore it is determined to be usable arable land.

[0051] It should be noted that, based on the above, other situations may indicate that the risk of salinization has not been eliminated or the degree of salinization is relatively mild, requiring intervention before cultivation. For example, if the representative value of capillary margin is greater than or equal to 0, but the median of the scaled salt surface response index is less than the adaptive threshold, it means that underground capillary action can transport salt to the surface, but due to insufficient salt accumulation time or recent rainfall leaching, obvious salinization substances have not yet formed on the surface. For farmland to be improved, soil or groundwater conditions can be improved through drainage to reduce waterlogging, deep plowing and loosening the soil, increasing the application of organic fertilizer, and planting salt-tolerant crops, thereby eliminating the risk of salinization.

[0052] In one embodiment of the present invention, the spatial boundaries of each analysis unit and the corresponding land parcel category are first obtained. The area of ​​each analysis unit is calculated using geographic information system tools. The analysis units are grouped and summarized according to land parcel category. The areas of all analysis units in the same category are added together to obtain the total area of ​​the three types of arable land, and the area values ​​of each category are recorded. The spatial boundaries of the analysis units are associated with the area data of the classification results, as well as the representative value of capillary reach margin, effective capillary reach height, and adaptive threshold. Using a base map as the base map, the above dataset is superimposed on the base map, and different visual labels are assigned to different land parcel categories. For example, usable arable land is marked in green, unusable arable land is marked in red, and arable land to be improved is marked in yellow. The category and area information of each analysis unit are marked, and finally a land spatial attribute map is generated to ensure that the attributes of each unit in the map are queryable and the area statistics are traceable. This will not be elaborated here.

[0053] In one embodiment of the present invention, such as Figure 2 As shown, a land spatial intelligent planning method based on multi-source data fusion includes the following steps: Step S201: Construct analysis units by mapping the multi-source data of the gridded plots according to spatial location and observation date; Step S202: Determine the equivalent capillary radius based on soil texture, and calculate the effective capillary height by combining it with groundwater salinity; Step S203: Calculate the capillary reach margin based on the effective capillary reach height and groundwater depth, determine the image selection weight by the ratio of the capillary reach margin to the effective capillary reach height, and construct an image date set based on the image selection weight. Step S204: Calculate the salt surface response index within the image date set, and scale it to the scaled salt surface response index according to the image selection weight. Step S205: Select the time marker with the highest image selection weight from the image date set of each analysis unit, and calculate the adaptive threshold based on the capillary margin under the time marker. Step S206: Output the land parcel category of the analysis unit based on the capillary surface margin, the scaled salt surface response index and the adaptive threshold, and construct a land space attribute map by calculating the land parcel area.

[0054] It should be noted that the interval and threshold sizes are set for ease of comparison. The size of the threshold depends on the amount of sample data and the base number set by those skilled in the art for each set of sample data, as long as it does not affect the proportional relationship between the parameter and the quantized value. Furthermore, the above formulas are all dimensionless calculations, and the formulas are derived from software simulations using a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0055] The embodiments of this example have been described above. However, this example is not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms based on the guidance of this example, and all of them are within the protection scope of this example.

Claims

1. A land space intelligent planning system based on multi-source data fusion, characterized in that, The method comprises the following steps: An analysis unit construction module is used to map and construct analysis units according to spatial positions and observation dates of multi-source data of the grid land; The multi-source data comprises groundwater depth and groundwater salinity of observation wells, soil texture of soil sample points, short-wave infrared reflectivity and near-infrared reflectivity of remote sensing images; An effective capillary rise height calculation module is used to determine an equivalent capillary radius according to the soil texture, and to calculate the effective capillary rise height in combination with the groundwater salinity; An image date set construction module is used to calculate a capillary rise height according to the effective capillary rise height and the groundwater depth, to determine an image selection date weight according to the ratio of the capillary rise height to the effective capillary rise height, and to construct an image date set according to the image selection date weight; A salt surface response index calculation module is used to calculate a salt surface response index in the image date set, and to scale the salt surface response index into a scaled salt surface response index according to the image selection date weight; An adaptive threshold calculation module is used to select a time identifier with the largest image selection date weight from the image date set of each analysis unit, and to calculate an adaptive threshold according to the capillary rise height at the time identifier; A land class division module is used to output a land class of the analysis unit according to the capillary rise height, the scaled salt surface response index and the adaptive threshold, and to construct a national land space attribute map by counting land areas; The land class comprises available farmland, unavailable farmland and farmland to be improved. 2.The land space intelligent planning system based on multi-source data fusion according to claim 1, characterized in that, The analysis unit is defined with reference to a coordinate system adopted by a base map, and each analysis unit has a unique spatial boundary and identifier; A time identifier is associated with the analysis unit, and the time identifier is a Gregorian calendar date with a date precision of one day; 3.The land space intelligent planning system based on multi-source data fusion according to claim 1, characterized in that, An observation data directory is associated with the analysis unit, and the observation data directory comprises a well list, a soil sample list and a remote sensing image list, wherein the well list comprises well coordinates of each observation well, the soil sample list comprises sample coordinates of each soil sample, and the remote sensing image list comprises short-wave infrared reflectivity, near-infrared reflectivity and pixel center coordinates of each pixel in the remote sensing image, and the well coordinates, the sample coordinates and the pixel center coordinates are consistent with the coordinate system adopted by the base map. The same type of observation data from different points in the same analysis unit are merged by using a distance inverse weighting method, the center of gravity of the analysis unit is taken as a reference point, the plane distance from each point to the reference point of the corresponding analysis unit is calculated, a small positive number is added to the plane distance, the obtained result is squared and then inverted to obtain the weight of the point, the groundwater depth of each observation well is multiplied by the corresponding weight, the sum is divided by the sum of all weights to obtain the merged value of the groundwater depth of the analysis unit at the time identifier, the groundwater salinity merged value, the short-wave infrared reflectivity merged value and the near-infrared reflectivity merged value are obtained by the same method, and finally the soil texture is stored in the corresponding analysis unit in the form of classification code. 4.The land space intelligent planning system based on multi-source data fusion according to claim 1, wherein, Firstly, the equivalent capillary radius is determined according to the soil texture through the soil water characteristic curve, and then the salt water surface tension and the salt water density are determined according to the underground water salinity and temperature, the salt water surface tension is multiplied by 2, and then multiplied by the cosine value of the contact angle between the underground water and the soil particles as the numerator, and the product of the salt water density, the gravity acceleration and the equivalent capillary radius as the denominator, and the numerator is divided by the denominator to obtain the effective capillary rise height. 5.The land space intelligent planning system based on multi-source data fusion according to claim 1, wherein, Firstly, the difference between the effective capillary rise height and the underground water depth of the analysis unit under the same time identifier is calculated to obtain the capillary rise margin, and then the ratio between the absolute value of the capillary rise margin and the effective capillary rise height is calculated, the intermediate result is obtained by subtracting the ratio from 1, the value range of the intermediate result is controlled between 0 and 1 through the clipping function to obtain the image timing weight. 6.The land space intelligent planning system based on multi-source data fusion according to claim 1, wherein, All time identifiers of the analysis unit whose image timing weight is greater than or equal to the preset weight threshold value under each time identifier are screened and analyzed, and these time identifiers are collected to construct the image date set corresponding to the analysis unit, wherein the preset weight threshold value is a custom parameter. 7.The land space intelligent planning system based on multi-source data fusion of claim 1, wherein, The ratio between the short-wave infrared reflectivity and the near-infrared reflectivity of the analysis unit under each time identifier is calculated as the salt surface response index, and it is judged that the near-infrared reflectivity is less than the preset reflectivity threshold value, then the salt surface response index under the time identifier is marked as the missing salt surface response index, wherein the preset reflectivity threshold value is a custom parameter; the product between the image timing weight and the non-missing salt surface response index is calculated as the scaled salt surface response index. 8.The land space intelligent planning system based on multi-source data fusion of claim 1, wherein, The time identifier with the maximum image timing weight is selected from the image date set of each analysis unit, the capillary rise margin and the effective capillary rise height of the analysis unit at the representative time are obtained as the capillary rise margin representative value and the effective capillary rise height representative value, and the analysis unit whose capillary rise margin representative value is greater than or equal to zero is screened to collect and construct a sample set. A histogram distribution is established according to the scaled salt surface response index corresponding to each analysis unit in the sample set, the inter-class variance corresponding to different candidate thresholds in the histogram distribution is calculated, the inter-class variance is the product of the difference between the sample proportion on both sides of the candidate threshold and the square of the difference between the sample mean on both sides of the candidate threshold, and the candidate threshold with the maximum inter-class variance is selected as the adaptive threshold. 9.The land space intelligent planning system based on multi-source data fusion of claim 8, wherein, Within the time range corresponding to the image date set, the median of the scaled salt surface response index is taken, when the capillary rise margin representative value is greater than or equal to 0 and the median of the scaled salt surface response index is greater than or equal to the adaptive threshold, the corresponding analysis unit is determined as unusable farmland; when the capillary rise margin representative value is less than or equal to 30% of the effective capillary rise height representative value and the median of the scaled salt surface response index is less than the adaptive threshold, the corresponding analysis unit is determined as usable farmland; otherwise, the corresponding analysis unit is determined as the farmland to be improved.

10. A land space intelligent planning method based on multi-source data fusion, characterized in that, The land space intelligent planning system based on multi-source data fusion is executed, and the system comprises the following steps: Step S201, mapping and constructing analysis units of the multi-source data of the grid block according to the spatial position and the observation date; In step S202, the equivalent capillary radius is determined according to the soil texture, and the effective capillary rise height is calculated in combination with the groundwater salinity; In step S203, the capillary rise degree is calculated according to the effective capillary rise height and the groundwater depth, the image selection weight is determined according to the proportion of the capillary rise degree to the effective capillary rise height, and the image date set is constructed according to the image selection weight; In step S204, the salt surface response index is calculated in the image date set, and the scaled salt surface response index is obtained by scaling the salt surface response index according to the image selection weight; In step S205, the time identifier with the maximum image selection weight is screened from the image date set of each analysis unit, and the adaptive threshold is calculated according to the capillary rise degree corresponding to the time identifier; In step S206, the plot category of the analysis unit is output according to the capillary rise degree, the scaled salt surface response index and the adaptive threshold, and the plot area is counted to construct the land space attribute map.

Images