A soil property map inference mapping method and system
Through machine learning algorithms and statistical principles, a spatial prediction model is established, a spatial distribution map of soil attributes is generated, and a humidity hot spot area is identified. The problems of low accuracy and high cost of soil attribute map mapping in traditional methods are solved, and a detailed irrigation plan is achieved. and hot spot identification.
Patent Information
- Application Number
- CN202411198966.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-29
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-08-29
AI Technical Summary
The traditional soil attribute map speculative mapping method relies on experience and manual drawing, with high cost, long periods and subjective judgment influences, and the importance weight of environmental variables is fixed, making it difficult to effectively identify hot spots in soil moisture and conductivity properties.
Using machine learning algorithms and statistical principles, by obtaining soil attributes and environmental variable data, preprocessing and feature extraction, establishing spatial prediction models, performing spatial interpolation, generating spatial distribution maps of soil attributes, identifying humidity hot spots and formulating detailed irrigation plan.
It improves the accuracy of soil attribute mapping, provides a detailed soil water replenishment irrigation plan, reduces costs and cycles, reduces the impact of subjective judgment, and can effectively identify hot spots in soil moisture and conductivity properties.
Smart Images

Figure CN118710841B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of data prediction and mapping, and in particular to a soil property map inference mapping method and system. Background Art
[0002] Digital soil attribute mapping methods are now widely used in the field of soil attribute mapping. Digital soil attribute mapping methods are based on the soil information of known points, and use statistics, spatial analysis, machine learning and other means to infer the soil characteristics of other points to generate soil attribute spatial distribution maps. Traditional soil attribute map inference mapping methods mainly rely on the experience and manual drawing of soil experts. They are usually based on field sampling, laboratory analysis and historical soil data, and then use manual drawing to produce soil attribute distribution maps. Traditional soil attribute map inference mapping methods have high costs, long cycles and subjective judgment. In contrast, digital soil mapping uses spatial statistical theory and advanced spatial information processing technology to infer the spatial distribution of soil attributes through machine learning and other models. It is a more modern, sophisticated and efficient method. In existing digital soil attribute mapping methods, the importance weights of environmental variables are usually fixed. In large areas, the effects of environmental variables on soil attributes vary. This paper conducts regional hotspot analysis on the changes in soil moisture attributes and soil conductivity attributes. According to the inference mapping of soil attribute maps, the hotspot areas of soil moisture attributes are inferred to provide detailed irrigation plans for agricultural production. Summary of the invention
[0003] (1) Technical issues to be resolved
[0004] The purpose of the present invention is to provide a method and system for inferential mapping of soil property maps, which uses a machine learning algorithm to improve the mapping accuracy of soil property maps, and uses statistical principles to solve the identification of hot spots of soil moisture properties and soil conductivity properties, providing a detailed plan for soil water replenishment irrigation.
[0005] (2) Technical solution
[0006] To achieve the above object, on the one hand, the present invention provides a soil property map estimation method, the method comprising:
[0007] Step S1, obtaining soil attribute data and environmental variable data of a target area as raw data; preprocessing the raw data to obtain standardized data, and obtaining soil conductivity attribute data and soil moisture attribute data of the target area by feature extraction.
[0008] Step S2, establishing a first spatial prediction model, inputting the soil electrical conductivity attribute data and the soil moisture attribute data into the first spatial prediction model, and obtaining soil moisture attribute spatial distribution data and soil electrical conductivity attribute spatial distribution data.
[0009] Step S3, performing spatial interpolation on the soil moisture attribute spatial distribution data and the soil conductivity attribute spatial distribution data to obtain the spatial distribution data of the soil attributes, and generating a soil attribute spatial distribution map.
[0010] Step S4, by analyzing the spatial distribution map of soil properties, the correlation between soil moisture properties and soil electrical conductivity properties is obtained, the humidity hotspot areas are identified, and a detailed soil water replenishment irrigation plan for the study area is formulated accordingly.
[0011] Furthermore, the original data includes: soil property data and environmental variable data of the target area obtained through stratified random sampling and satellite remote sensing data; the soil property data includes at least: texture, pH value, organic matter content, conductivity, humidity, and the environmental variable data includes: slope and weather data.
[0012] The preprocessing of the raw data to obtain standardized data specifically includes: calculating the Z value of each data point by the Z-score method, and when the Z value is greater than a preset threshold, marking the data as abnormal data and deleting it.
[0013] The stratified random sampling data and the satellite remote sensing data are merged to obtain fused data, and feature extraction is performed based on the fused data to obtain soil conductivity attribute data and soil moisture attribute data.
[0014] Furthermore, the method for obtaining the spatial distribution data of soil moisture attributes and the spatial distribution data of soil electrical conductivity attributes includes:
[0015] The soil moisture attribute data and the soil conductivity attribute data are used to obtain a hyperplane through the first spatial prediction model. The hyperplane is:
[0016] ;
[0017] in, is the normal vector of the hyperplane, is the input data, b is the intercept of the hyperplane, express The transpose of .
[0018] The hyperplane that maximizes the interval between soil moisture attribute data and soil conductivity attribute data to the hyperplane is obtained through the first spatial prediction model:
[0019] ;
[0020] ;
[0021] in, yes The square of the Euclidean norm of It is The labels of the sample points, belong , It is The feature vector of the sample points, is a constraint condition, which ensures that all sample points are correctly classified.
[0022] The soil moisture attribute data and the soil conductivity attribute data are used to train the first spatial prediction model by introducing the second kernel function, where the second kernel function is:
[0023] ;
[0024] in, is the soil moisture attribute data, is the soil conductivity property data, are the parameters of the kernel function.
[0025] By introducing the second kernel function, the parameters are obtained , and obtain the spatial distribution data of soil moisture attributes and the spatial distribution data of soil electrical conductivity attributes.
[0026] Furthermore, the spatial distribution data of soil properties is obtained, and the method for generating the spatial distribution map of soil properties includes:
[0027] The soil moisture attribute data and soil conductivity attribute data are spatially interpolated. Assuming there are N sample points, the position of each sample point i1 is , the corresponding soil attribute value is , for any unknown position , the predicted value of its soil property value Calculated by the formula:
[0028] ;
[0029] in, is the weight of the i1th sample point, Defined as:
[0030] ;
[0031] in, The point to be calculated With the i1th sample point The distance between Is 2 or 3.
[0032] Furthermore, the method of performing spatial interpolation on the spatial distribution data of the soil moisture attribute and the spatial distribution data of the soil conductivity attribute includes:
[0033] The spatial interpolation method is:
[0034] ;
[0035] in, is the attribute value of the position to be calculated, is the attribute value of the known point, is the weight value, and the calculation of the weight depends on the semivariance function:
[0036] ;
[0037] in, is the semivariance at distance h, is the number of position points with a distance h, is the attribute value of the known point, is the attribute value of the point at distance h.
[0038] The weight value The following system equations must be satisfied:
[0039] ;
[0040] ;
[0041] in, For point and Point The semivariance between is the Lagrange multiplier used to ensure that the weight sum is 1, is the attribute value of the known point, x is the attribute value of the point to be calculated, and the weight value is obtained by solving the above system equation , the predicted point is calculated by the weight value The value of .
[0042] The spatial autocorrelation of regionalized variables is used to estimate the values of unknown points through semivariogram functions and system equations to obtain the interpolation results of soil property data.
[0043] Furthermore, the spatial distribution data of the soil moisture attribute and the spatial distribution data of the soil conductivity attribute are spatially interpolated to obtain the spatial distribution data of the soil attribute, and a spatial distribution map of the soil attribute is generated:
[0044] The soil moisture attribute data and soil conductivity attribute data after the spatial prediction model analysis are spatially interpolated to obtain the soil attribute data at the unknown point;
[0045] The soil property data generated by the spatial interpolation method are used to generate a soil property spatial distribution map to obtain the distribution of soil properties in the entire area.
[0046] Furthermore, the method for obtaining a detailed soil water replenishment irrigation plan for the study area according to the humidity hotspot area includes:
[0047] Hot spot analysis is performed by acquiring soil property data of the target area. The hot spot analysis method is as follows:
[0048] ;
[0049] in, This is the hot spot analysis at position i3 Statistics, is the spatial weight between positions i3 and j3, is the attribute value data at position j3, is the mean value of the attribute, S is the standard deviation of the attribute, and n is the total number of samples.
[0050] Introducing the key statistics of spatial clustering significance, Z-score and p-value, into hot spot analysis In the statistical formula, the Z score and p-value are used to determine whether the attribute values of the predicted location form significant hot spots and cold spots. A positive Z score indicates the aggregation of high-value features, and a negative value indicates the aggregation of low-value features. Usually, a Z score greater than 1.96 or less than -1.96 is considered significant, and a p-value less than 0.05 is considered significant. Specifically, a high Z score and a small p-value indicate that the location is a significant hot spot, a low Z score and a small p-value indicate that the location is a significant cold spot, and a Z score close to zero and a large p-value indicate that the location has no significant spatial clustering. By calculating the Z score and p-value of each location, the hot spots and cold spots in the data can be identified, thereby performing spatial clustering analysis.
[0051] The spatial distribution relationship between soil moisture attributes and soil conductivity attributes is analyzed through hotspot distribution analysis of soil attribute data and soil attribute distribution map. The change relationship between soil conductivity attributes and soil moisture attributes is positively correlated. The hotspot areas with high and low humidity are identified through the soil conductivity attribute distribution map, and a detailed plan for regional soil water replenishment irrigation is obtained.
[0052] Based on the same inventive concept, the present invention provides a soil property map speculation mapping system, the system comprising: a first data processing module, a second data processing module, a key speculation mapping module, a hot spot analysis and decision module, the modules are connected in sequence;
[0053] The first data processing module is used to obtain soil property data and environmental variable data of the target area as raw data; preprocess the raw data to obtain standardized data, and obtain soil conductivity property data and soil moisture property data of the target area through feature extraction.
[0054] The second data processing module is used to establish a first spatial prediction model, input the soil conductivity attribute data and the soil moisture attribute data into the first spatial prediction model, and obtain soil moisture attribute spatial distribution data and soil conductivity attribute spatial distribution data.
[0055] The key inference mapping module is used to perform spatial interpolation on the soil moisture attribute spatial distribution data and the soil conductivity attribute spatial distribution data to obtain the spatial distribution data of the soil attributes and generate a soil attribute spatial distribution map.
[0056] The hot spot analysis and decision-making module is used to obtain the correlation between soil moisture attributes and soil conductivity attributes by analyzing the spatial distribution map of soil attributes, identify moisture hot spots, and formulate a detailed soil water replenishment irrigation plan for the study area based on this.
[0057] (3) Beneficial effects
[0058] Compared with the prior art, the present invention has the following beneficial effects: raw data is obtained by collecting soil attribute data of a target area, soil conductivity attribute data and soil moisture attribute data of the target area are obtained by data processing and feature extraction, a first spatial prediction model is established, soil moisture attribute spatial distribution data and soil conductivity attribute spatial distribution data are obtained, a spatial interpolation method is used to obtain the spatial distribution of soil attributes in an unknown area, a soil attribute spatial distribution map is generated, the correlation between soil moisture attributes and soil conductivity attributes is obtained through the soil attribute spatial distribution map, and a humidity hotspot area is obtained, and a detailed soil water replenishment irrigation plan for the study area is obtained according to the humidity hotspot area. BRIEF DESCRIPTION OF THE DRAWINGS
[0059] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.
[0060] Figure 1 A flowchart of a soil property map estimation method according to Embodiment 1 of the present invention;
[0061] Figure 2This is a schematic diagram of the module composition of a soil property map inference mapping system according to Example 2 of the present invention. DETAILED DESCRIPTION
[0062] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0063] Example 1: Figure 1 As shown, this embodiment provides a soil property map inference mapping method, the method comprising:
[0064] S10, obtaining soil property data and environmental variable data of the target area as raw data; preprocessing the raw data to obtain standardized data, and obtaining soil conductivity property data and soil moisture property data of the target area by feature extraction.
[0065] S20, establishing a first spatial prediction model, inputting the soil electrical conductivity attribute data and the soil moisture attribute data into the first spatial prediction model, and obtaining soil moisture attribute spatial distribution data and soil electrical conductivity attribute spatial distribution data.
[0066] S30, performing spatial interpolation on the spatial distribution data of soil moisture attributes and the spatial distribution data of soil electrical conductivity attributes to obtain spatial distribution data of soil attributes, and generating a spatial distribution map of soil attributes.
[0067] S40. By analyzing the spatial distribution map of soil properties, the correlation between soil moisture properties and soil electrical conductivity properties is obtained, the moisture hotspot areas are identified, and a detailed soil water replenishment irrigation plan for the study area is formulated accordingly.
[0068] It should be noted that the original data include: soil property data and environmental variable data of the target area obtained through stratified random sampling and satellite remote sensing data; the soil property data include at least: texture, pH value, organic matter content, conductivity, humidity, and the environmental variable data include: slope and weather data.
[0069] S11, the preprocessing of the original data to obtain standardized data is specifically: calculating the Z value of each data point by the Z-score method, and when the Z value is greater than a preset threshold, marking the data as abnormal data and deleting it.
[0070] S12, merging the stratified random sampling data and the satellite remote sensing data to obtain fused data, and performing feature extraction based on the fused data to obtain soil conductivity attribute data and soil moisture attribute data.
[0071] It should be noted that the method for obtaining the spatial distribution data of soil moisture attributes and the spatial distribution data of soil conductivity attributes includes:
[0072] S21. Using the soil moisture attribute data and the soil conductivity attribute data through the first spatial prediction model to obtain a hyperplane, wherein the hyperplane is:
[0073] ;
[0074] in, is the normal vector of the hyperplane, is the input data, b is the intercept of the hyperplane, express The transpose of .
[0075] S22. The hyperplane that maximizes the interval between the soil moisture attribute data and the soil conductivity attribute data and the hyperplane is obtained by the first spatial prediction model:
[0076] ;
[0077] ;
[0078] in, yes The square of the Euclidean norm of It is The labels of the sample points, belong , It is The feature vector of the sample points, is a constraint condition, which ensures that all sample points are correctly classified.
[0079] S23, training the first spatial prediction model by introducing a second kernel function into the soil moisture attribute data and the soil conductivity attribute data, wherein the second kernel function is:
[0080] ;
[0081] in, is the soil moisture attribute data, is the soil conductivity property data, are the parameters of the kernel function;
[0082] By introducing the second kernel function, the parameters are obtained , and obtain the spatial distribution data of soil moisture attributes and the spatial distribution data of soil electrical conductivity attributes.
[0083] For example: Randomly select soil conductivity data of 4 sample points from the soil attribute sampling point data, sample point 1: X coordinate is 100, Y coordinate is 200, conductivity is 2.5mS / m, sample point 2: X coordinate is 150, Y coordinate is 180, conductivity is 3mS / m, sample point 3: X coordinate is 120, Y coordinate is 220, conductivity is 2.7mS / m. =20, predict the soil conductivity at the position (130,190), calculate the second kernel function values between this position and the sample point as 0.8, 0.6, 0.7 respectively, and then perform weighted average to get the soil conductivity at (130,190) as 2.71.
[0084] It should be noted that the method of performing spatial interpolation on the spatial distribution data of soil moisture attributes and the spatial distribution data of soil conductivity attributes to obtain the spatial distribution of soil attributes and generate a spatial distribution map of soil attributes includes:
[0085] S31, perform spatial interpolation on the spatial distribution data of soil moisture attributes and the spatial distribution data of soil conductivity attributes. Assuming there are N sample points, the position of each sample point i1 is , the corresponding soil attribute value is , for any unknown position , the predicted value of its soil property value Calculated by the formula:
[0086] ;
[0087] in, is the weight of the i1th sample point, Defined as:
[0088] ;
[0089] in, The point to be calculated With the i1th sample point The distance between is a tunable parameter, usually 2 or 3, = 2 for Euclidean distance, = 3 for Manhattan distance.
[0090] For example: Location Note that p1 = (1,2), the corresponding attribute value = 5, position Note that p2 = (3,4), the corresponding attribute value is = 8, position Note that p3 = (5,6) corresponds to the attribute value = 6, now predict the soil property value at position p = (4,3), calculate , using the parameter = 2 Calculate weight , , , then calculate the soil property value at p = (4,3) , that is, the predicted value of the soil property value at position p = (4,3) is 8.27.
[0091] It should be noted that the method for performing spatial interpolation on the spatial distribution data of the soil moisture attribute and the spatial distribution data of the soil conductivity attribute includes:
[0092] S311, the spatial interpolation method is:
[0093] ;
[0094] in, is the attribute value of the position to be calculated, is the attribute value of the known point, is the weight value, and the calculation of the weight depends on the semivariance function:
[0095] ;
[0096] in, is the semivariance at distance h, is the number of position points with a distance h, is the attribute value of the known point, is the attribute value of the point at distance h.
[0097] S312: The weight value The following system equations must be satisfied:
[0098] ;
[0099] ;
[0100] in, For point and Point The semivariance between is the Lagrange multiplier used to ensure that the weight sum is 1, is the attribute value of the known point, x is the attribute value of the point to be calculated, and the weight value is obtained by solving the above system equation , the predicted point is calculated by the weight value The value of .
[0101] S313. Utilize the spatial autocorrelation of regionalized variables, estimate the values of unknown points through semivariogram functions and system equations, and obtain soil property data interpolation results.
[0102] For example: the position of sample point 1 is (0,0), the attribute value is 10, the position of sample point 2 is (1,0), the attribute value is 12, the position of sample point 3 is (0,1), the attribute value is 14, the position of sample point 4 is (1,1), the attribute value is 16, and the attribute value at the predicted position (0.5,0.5) is calculated. Now calculate the distance between each sample point. The distance between (0,0) and (1,0) is 1, the distance between (0,0) and (0,1) is 1, and the distance between (0,0) and (1,1) is , the distance between (1,0) and (0,1) is , the distance (1,0) and (1,1) is 1, the distance (0,1) and (1,1) is 1, and the weights are obtained by solving the semivariance model, that is, the system equation =0.25, =0.25, =0.25, = 0.25, substituting the weight and the known point attribute value, we get , then the estimated attribute value at position (0.5,0.5) is 13.
[0103] It should be noted that the method of performing spatial interpolation on the soil moisture attribute spatial distribution data and the soil conductivity attribute spatial distribution data to obtain the spatial distribution data of the soil attribute and generate the soil attribute spatial distribution map includes:
[0104] S3141. Perform spatial interpolation on the soil moisture attribute data and soil electrical conductivity attribute data analyzed by the spatial prediction model to obtain soil attribute data at unknown points.
[0105] S3142. Generate a soil property spatial distribution map using the soil property data generated by the spatial interpolation method to obtain the distribution of soil properties in the entire area.
[0106] It should be noted that the method for obtaining a detailed soil water replenishment irrigation plan for the study area according to the humidity hotspot area includes:
[0107] S41, performing hot spot analysis by acquiring soil property data of the target area, wherein the hot spot analysis method is:
[0108] ;
[0109] in, This is the hot spot analysis at position i3 Statistics, is the spatial weight between positions i3 and j3, is the attribute value data at position j3, is the mean value of the attribute, S is the standard deviation of the attribute, and n is the total number of samples.
[0110] For example, if the attribute value of sample point 1 at position (0,0) is 10, the attribute value of sample point 2 at position (1,0) is 12, the attribute value of sample point 3 at position (0,1) is 14, and the attribute value of sample point 4 at position (1,1) is 16, then if the spatial weight is 1, the average value is , the standard deviation is , calculate the position (0,0) value , since all points have equal weights and attribute values are evenly distributed, A value of 0 indicates no significant hot or cold spots.
[0111] S42. Introducing the key statistics of spatial clustering significance, Z-score and p-value, into hot spot analysis In the statistical formula, the Z score and p-value are used to determine whether the attribute values of the predicted location form significant hot spots and cold spots. A positive Z score indicates the aggregation of high-value features, and a negative value indicates the aggregation of low-value features. Usually, a Z score greater than 1.96 or less than -1.96 is considered significant, and a p-value less than 0.05 is considered significant. Specifically, a high Z score and a small p-value indicate that the location is a significant hot spot, a low Z score and a small p-value indicate that the location is a significant cold spot, and a Z score close to zero and a large p-value indicate that the location has no significant spatial clustering. By calculating the Z score and p-value of each location, the hot spots and cold spots in the data can be identified, thereby performing spatial clustering analysis.
[0112] S43. The spatial distribution relationship between soil moisture attributes and soil electrical conductivity attributes is analyzed through hotspot distribution analysis of soil attribute data and soil attribute distribution map. The change relationship between soil electrical conductivity attributes and soil moisture attributes is positively correlated. The hotspot areas with high and low humidity are identified through the soil electrical conductivity attribute distribution map, and a detailed plan for regional soil water replenishment irrigation is obtained.
[0113] For example, the total area of the study area is 260 square kilometers. The terrain is mainly plains and hills. The soil types mainly include loam and sandy loam. The climate of this area belongs to the temperate monsoon climate. The annual precipitation is about 150 mm. The precipitation distribution is uneven and the seasonal drought is obvious. Collect soil attribute data of the area, including texture, pH value, organic matter content, conductivity, humidity, etc. and environmental variable data, such as slope, weather data, etc. Data analysis is performed through spatial interpolation methods and machine learning models to obtain the spatial distribution data of soil moisture and conductivity, and generate a spatial distribution map of soil attributes. Then perform hot spot area analysis to conclude that areas with higher conductivity overlap with humidity hot spot areas, and the soil salinity in these areas is higher. Areas with higher humidity are concentrated in the north and east of the farmland, usually in low-lying areas or near rivers. The soil humidity is higher, which is conducive to crop growth, but there is also a risk of over-wetting. According to the humidity hot spot area, we can formulate the following detailed irrigation plan: A. Divide the area into several irrigation areas, and each area is subdivided according to the soil moisture and conductivity properties to ensure the accuracy and effectiveness of irrigation. B. In humidity hotspots, irrigate once a week to maintain soil moisture, use water-saving irrigation methods such as drip irrigation or micro-sprinkler irrigation to reduce water evaporation and waste, improve water resource utilization efficiency, and avoid crop diseases caused by excessive soil moisture. C. Use irrigation methods such as sprinkler irrigation or furrow irrigation in areas with low humidity to ensure sufficient soil moisture and promote crop growth. D. According to the conductivity data, control the amount of water for each irrigation to avoid salt accumulation and nutrient loss caused by excessive irrigation. Regularly maintain and inspect the irrigation system to ensure the normal operation of the irrigation equipment and the irrigation effect. E. After irrigation, regularly monitor soil moisture and crop growth to evaluate the irrigation effect. According to the evaluation results, adjust the irrigation plan in time to ensure healthy crop growth and increased yield.
[0114] Embodiment 2: Based on the same inventive concept, Figure 2 As shown, this embodiment provides a soil property map inference mapping system, the system includes: a first data processing module, a second data processing module, a key inference mapping module, a hot spot analysis and decision module, and the modules are connected in sequence.
[0115] The first data processing module is used to obtain soil property data and environmental variable data of the target area as raw data; preprocess the raw data to obtain standardized data, and obtain soil conductivity property data and soil moisture property data of the target area through feature extraction.
[0116] The second data processing module is used to establish a first spatial prediction model, input the soil conductivity attribute data and the soil moisture attribute data into the first spatial prediction model, and obtain soil moisture attribute spatial distribution data and soil conductivity attribute spatial distribution data.
[0117] The key inference mapping module is used to perform spatial interpolation on the soil moisture attribute spatial distribution data and the soil conductivity attribute spatial distribution data to obtain the spatial distribution data of the soil attributes and generate a soil attribute spatial distribution map.
[0118] The hot spot analysis and decision-making module is used to obtain the correlation between soil moisture attributes and soil conductivity attributes by analyzing the spatial distribution map of soil attributes, identify moisture hot spots, and formulate a detailed soil water replenishment irrigation plan for the study area based on this.
[0119] It should be noted that, regarding the system in the above embodiment, the specific manner in which each module performs operations has been described in detail in Embodiment 1 of the method, and will not be elaborated here.
[0120] Finally, it should be noted that: Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for inferential mapping of soil property maps, It is characterized in that The method comprises: Step S1, obtaining soil attribute data and environmental variable data of a target area as raw data; preprocessing the raw data to obtain standardized data, and obtaining soil conductivity attribute data and soil moisture attribute data of the target area by feature extraction; Step S2, establishing a first spatial prediction model, inputting the soil electrical conductivity attribute data and the soil moisture attribute data into the first spatial prediction model, and obtaining soil moisture attribute spatial distribution data and soil electrical conductivity attribute spatial distribution data; Step S3, performing spatial interpolation on the soil moisture attribute spatial distribution data and the soil conductivity attribute spatial distribution data to obtain the spatial distribution data of the soil attributes, and generating a soil attribute spatial distribution map; Step S4, by analyzing the spatial distribution map of soil properties, obtaining the correlation between soil moisture properties and soil conductivity properties, identifying humidity hotspots, and formulating a detailed soil water replenishment irrigation plan for the study area accordingly; The soil moisture attribute data and the soil conductivity attribute data are used to obtain a hyperplane through the first spatial prediction model. The hyperplane is: ; in, is the normal vector of the hyperplane, is the input data, b is the intercept of the hyperplane, express The transpose of The hyperplane that maximizes the interval between soil moisture attribute data and soil conductivity attribute data to the hyperplane is obtained through the first spatial prediction model: ; ; in, yes The square of the Euclidean norm of It is The labels of the sample points, belong , It is The feature vector of the sample points, is a constraint condition, and the constraint condition is used to ensure that all sample points are correctly classified; The soil moisture attribute data and the soil conductivity attribute data are used to train the first spatial prediction model by introducing the second kernel function, where the second kernel function is: ; in, is the soil moisture attribute data, is the soil conductivity property data, are the parameters of the kernel function; By introducing the second kernel function, the parameters are obtained , and obtain the spatial distribution of soil moisture attribute data and the spatial distribution of soil conductivity attribute data.
2. The method for inferring and mapping a soil property map according to claim 1, It is characterized in that The raw data include: soil property data and environmental variable data of the target area obtained through stratified random sampling and satellite remote sensing data; the soil property data include at least: texture, pH value, organic matter content, conductivity, humidity, and the environmental variable data include: slope and weather data; The preprocessing of the raw data to obtain standardized data specifically includes: calculating the Z value of each data point by the Z-score method, and when the Z value is greater than a preset threshold, marking the data as abnormal data and deleting it; The stratified random sampling data and the satellite remote sensing data are merged to obtain fused data, and feature extraction is performed based on the fused data to obtain soil conductivity attribute data and soil moisture attribute data.
3. The method for inferring and mapping a soil property map according to claim 1, It is characterized in that The method for generating the soil property spatial distribution map comprises: The soil moisture attribute data and soil conductivity attribute data are spatially interpolated. Assuming there are N sample points, each sample point The location is , the corresponding soil attribute value is , for any unknown position , the predicted value of its soil property value Calculated by the formula: ; in, It is The weight of the sample points, Defined as: ; in, The point to be calculated With Sample points The distance between Is 2 or 3.
4. The method for inferring and mapping a soil property map according to claim 3, It is characterized in that The method for performing spatial interpolation on the spatial distribution data of the soil moisture attribute and the spatial distribution data of the soil conductivity attribute includes: The spatial interpolation method is: ; in, is the attribute value of the position to be calculated, is the attribute value of the known point, is the weight value, and the calculation of the weight depends on the semivariance function: ; in, is the semivariance at distance h, is the number of position points with a distance h, is the attribute value of the known point, is the attribute value of the point at distance h; The weight value The following system equations must be satisfied: ; ; in, For point and Point The semivariance between is the Lagrange multiplier used to ensure that the weight sum is 1, is the attribute value of the known point, x is the attribute value of the point to be calculated, and the weight value is obtained by solving the above system equation , the predicted point is calculated by the weight value The value of The spatial autocorrelation of regionalized variables is used to estimate the values of unknown points through semivariogram functions and system equations to obtain the interpolation results of soil property data.
5. A soil property map estimation method according to claim 4, It is characterized in that The method of performing spatial interpolation on the spatial distribution data of soil moisture attributes and the spatial distribution data of soil conductivity attributes to obtain the spatial distribution data of soil attributes and generate a spatial distribution map of soil attributes includes: The soil moisture attribute data and soil conductivity attribute data after the spatial prediction model analysis are spatially interpolated to obtain the soil attribute data at the unknown point; The soil property data generated by the spatial interpolation method are used to generate a soil property spatial distribution map to obtain the distribution of soil properties in the entire area.
6. The method for inferring and mapping a soil property map according to claim 1, It is characterized in that The method for obtaining a detailed soil water replenishment irrigation plan for the study area according to the humidity hotspot area includes: Hot spot analysis is performed by acquiring soil property data of the target area. The hot spot analysis method is as follows: ; in, It's location Hot spot analysis Statistics, is the spatial weight between positions i3 and j3, It's location Attribute value data at is the mean value of the attribute, S is the standard deviation of the attribute, and n is the total number of samples; Introducing the key statistics of spatial clustering significance, Z-score and p-value, into hot spot analysis In the statistical formula, the Z score and p-value are used to determine whether the attribute values of the predicted location form significant hot spots and cold spots. A positive Z score indicates the aggregation of high-value features, and a negative value indicates the aggregation of low-value features. A Z score greater than 1.96 or less than -1.96 is considered significant, and a p-value less than 0.05 is considered significant. Specifically, a high Z score and a small p-value indicate that the location is a significant hot spot, a low Z score and a small p-value indicate that the location is a significant cold spot, and a Z score close to zero and a large p-value indicate that the location has no significant spatial clustering. By calculating the Z score and p-value of each location, the hot spots and cold spots in the data can be identified, thereby performing spatial clustering analysis. The spatial distribution relationship between soil moisture attributes and soil conductivity attributes is analyzed through hotspot distribution analysis of soil attribute data and soil attribute distribution map. The change relationship between soil conductivity attributes and soil moisture attributes is positively correlated. The hotspot areas with high and low humidity are identified through the soil conductivity attribute distribution map, and a detailed plan for regional soil water replenishment irrigation is obtained.
7. A soil property map speculative mapping system, used to execute the method according to any one of claims 1 to 6, It is characterized in that The system comprises: a first data processing module, a second data processing module, a key inference mapping module, and a hot spot analysis and decision module, wherein the modules are connected in sequence; The first data processing module is used to obtain soil property data and environmental variable data of the target area as raw data; preprocess the raw data to obtain standardized data, and obtain soil conductivity property data and soil moisture property data of the target area through feature extraction; The second data processing module is used to establish a first spatial prediction model, input the soil conductivity attribute data and the soil moisture attribute data into the first spatial prediction model, and obtain soil moisture attribute spatial distribution data and soil conductivity attribute spatial distribution data; The key inference mapping module is used to perform spatial interpolation on the soil moisture attribute spatial distribution data and the soil conductivity attribute spatial distribution data to obtain the spatial distribution data of the soil attributes and generate a soil attribute spatial distribution map; The hot spot analysis and decision-making module is used to obtain the correlation between soil moisture attributes and soil conductivity attributes by analyzing the spatial distribution map of soil attributes, identify moisture hot spots, and formulate a detailed soil water replenishment irrigation plan for the study area based on this.
