A multi-scale remote sensing estimation method for spatially variable soil organic matter

Abstract

The present invention relates to the field of remote sensing estimation technology, specifically to a spatially variable multi-scale soil organic matter remote sensing estimation method, comprising the following steps: by collecting soil samples at differentiated locations, testing the soil organic matter content and extracting environmental factors, performing spatial difference analysis, analyzing the spatial variation relationship between environmental factors and soil organic matter, generating spatial variation patterns, optimizing the spatial weighted relationship between remote sensing indexes and environmental factors, and obtaining spatial weighted results of environmental factors. In the present invention, by analyzing the spatial variation relationship between environmental factors and soil organic matter, optimizing the weighted results, improving the estimation accuracy, bandwidth optimization and soil organic matter weight adjustment list correct local outliers, enhancing robustness, spatial heterogeneity adjustment and error correction improve the reliability of results, dynamic optimization and real-time updating overcome the limitations of static models, ensure the timeliness and accuracy of estimation results, and provide more accurate decision support.

Description

Technical Field

[0001] The present invention relates to the field of remote sensing estimation technology, and in particular to a spatially variable multi-scale soil organic matter remote sensing estimation method. Background Art

[0002] The field of remote sensing estimation technology is the technology of collecting surface or atmospheric data at a long distance using sensors carried by spacecraft or aircraft. The data is used to monitor and evaluate various characteristics of the earth's surface, including vegetation cover, soil conditions, water conditions, and urban development. Remote sensing technology has the advantages of wide coverage, high update frequency and relatively low cost, which makes it play an important role in environmental monitoring, agriculture, forestry, urban planning and natural disaster management. Including in the monitoring of soil organic matter, remote sensing technology can quickly evaluate soil fertility and carbon reserves over a large area by analyzing the spectral reflectance characteristics of the surface, which is of great significance to agricultural production and environmental protection.

[0003] Among them, the multi-scale soil organic matter remote sensing estimation method is a technology that uses remote sensing data to estimate the soil organic matter (SOM) content. This method takes into account the spatial variability of the relationship between remote sensing parameters and soil organic matter, and attempts to reveal this variability through multi-scale analysis, thereby optimizing the estimation accuracy of soil organic matter. This technology uses environmental covariates such as elevation, slope, and aspect to construct a model, which can more effectively reflect the uneven distribution of soil organic matter at different spatial scales. The development of the technology not only helps to improve the accuracy of soil quality monitoring, but also provides a scientific basis for the formulation of agricultural management and environmental protection strategies.

[0004] Although existing technologies can use remote sensing data to monitor soil organic matter over a large area, there are still several shortcomings in practical applications. Since existing technologies mainly rely on the weighted relationship between fixed remote sensing indices and environmental factors, it is difficult to accurately reflect the complex spatial variability between different regions and different environmental factors, resulting in low accuracy of the estimation results in spatial distribution. Especially in areas where soil organic matter concentration varies greatly, existing methods are difficult to capture the differences in local areas and ignore the special effects of local environmental factors. For example, in areas with large slopes or humid areas, existing methods fail to fully consider the nonlinear relationship between environmental factors and soil organic matter. Therefore, the estimation accuracy of soil organic matter concentration in special areas is poor. Existing technologies have poor adaptability to dynamic changes in soil organic matter concentration and fail to adjust the parameters in the estimation model in a timely manner, which limits the stability and predictive ability of the model. Due to the lack of a flexible local optimization mechanism, the error correction and spatial heterogeneity adjustment of existing methods are not fine enough, which affects the accuracy and scope of application of the monitoring results. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a spatially variable multi-scale soil organic matter remote sensing estimation method.

[0006] In order to achieve the above object, the present invention adopts the following technical scheme: a spatially variable multi-scale soil organic matter remote sensing estimation method, comprising the following steps:

[0007] S1: By collecting soil samples at different locations, testing the soil organic matter content and extracting environmental factors, conducting spatial difference analysis, analyzing the spatial variation relationship between environmental factors and soil organic matter, generating spatial variation patterns, optimizing the spatial weighted relationship between remote sensing index and environmental factors, and obtaining the spatial weighted results of environmental factors;

[0008] S2: Based on the spatial weighted results of the environmental factors, the spatial weighted values ​​of the environmental factors are adjusted through bandwidth optimization, the spatial regression parameters of the environmental factors and the soil organic matter concentration are analyzed, the weighted regression parameter matrix of the environmental factors is obtained, the soil organic matter concentration of the differentiated regions is estimated, and the predicted soil organic matter concentration value is obtained;

[0009] S3: Based on the predicted soil organic matter concentration value, the spatial scale is adjusted, the weights of environmental factors in differentiated areas are optimized, the local effects of environmental factors on soil organic matter concentration are identified, a soil organic matter weight adjustment list is generated, and the soil organic matter concentration of the region is updated in combination with the spatial change pattern to obtain an optimized soil organic matter concentration dataset;

[0010] S4: Based on the optimized soil organic matter concentration dataset, spatial heterogeneity adjustment is performed, the soil organic matter concentration of the region is dynamically optimized and error correction is performed, and the feature subset is extracted and adjusted and updated according to the characteristics of the current environmental factors to generate dynamic soil organic matter estimation results.

[0011] As a further solution of the present invention, the step of acquiring the spatial variation pattern is specifically:

[0012] S111: Construct an environmental factor dataset by collecting soil samples at different locations, testing soil organic matter content, recording environmental factors such as temperature, humidity, and land use type;

[0013] S112: Based on the environmental factor data set, with soil organic matter content as the dependent variable and environmental factors as the independent variables, perform multivariate linear regression analysis, determine key environmental factors through stepwise regression, calculate the contribution of key environmental factors, and obtain preliminary regression analysis results;

[0014] S113: Based on the preliminary regression analysis results, perform spatial variation analysis by identifying the distribution of soil organic matter using the formula:

[0015] ,

[0016] Get the spatial variation pattern;

[0017] in, represents the predicted soil organic matter content of the i-th sample point, represents the environmental factor temperature, Indicates the environmental factor humidity, represents the measured value of the nth environmental factor, represents the model intercept, represents the regression coefficient of temperature, represents the regression coefficient of humidity, Represents the regression coefficient of land use type.

[0018] As a further solution of the present invention, the steps for obtaining the spatial weighted results of the environmental factors are specifically as follows:

[0019] S121: Based on the spatial variation pattern, according to a preset spatial resolution, perform regional division, analyze the weight of each region and perform weighted processing on the data, output the weighted data according to a spatial grid, and obtain a regional weighted relationship;

[0020] S122: Based on the regional weighted relationship, extract the spatial changes of environmental factors in the region, analyze the spatial influence of environmental factors in the key area, determine the weighted contribution of environmental factors to the remote sensing index, rank them according to the contribution size, and generate the weighted contribution distribution result of environmental factors;

[0021] S123: Based on the weighted contribution distribution result of the environmental factors, the weighted coefficient of each environmental factor is adjusted, and the spatial weighted coefficient is optimized with reference to the difference in the impact of environmental factors on differentiated regions and combined with real-time geographic features. By adjusting the weight ratio, the spatial weighted result of the environmental factor is generated.

[0022] As a further solution of the present invention, the steps of obtaining the environmental factor weighted regression parameter matrix are specifically as follows:

[0023] S211: Based on the spatial weighted results of the environmental factors, the spatial weighted value of each environmental factor is adjusted by using bandwidth optimization through the spatial association between the environmental factors and the soil organic matter concentration, and weighted processing is performed to obtain the weighted value of the environmental factor through weight and spatial distance calculation;

[0024] S212: Based on the weighted values ​​of the environmental factors, combined with the soil organic matter concentration data, spatial regression analysis is performed, a regression model is constructed using the spatial data, the weight of the environmental factors in the model is adjusted, the model is tested cyclically, the optimal factor weights are determined, and the accuracy of the regression model is optimized, through the formula:

[0025] ,

[0026] Get the adjusted regression parameters;

[0027] Among them, R represents the adjusted regression parameter, is the weighted value of environmental factors, is the environmental factor data, is the spatial distance between the jth factor and the soil organic matter concentration, is the spatial distance adjustment coefficient, is the bandwidth optimization factor, is a constant adjustment term, and m represents the total number of factors;

[0028] S213: Optimizing the adjusted regression parameters, combining spatial factors, analyzing the spatial distribution characteristics of soil organic matter concentration, and generating an environmental factor weighted regression parameter matrix.

[0029] As a further solution of the present invention, the step of obtaining the predicted soil organic matter concentration value is specifically as follows:

[0030] S221: capturing environmental factor data of differentiated regions according to the environmental factor weighted regression parameter matrix, including temperature, humidity and soil pH value, and performing real-time monitoring on each environmental factor to obtain a standardized environmental factor data set;

[0031] S222: weighting the standardized environmental factor data set, and determining the weight of each factor through the original impact data of the environmental factor on the soil organic matter concentration, using the formula:

[0032] ,

[0033] Calculate the weighted sum of environmental factors;

[0034] Where A represents the weighted sum of environmental factors, , , represent the weights of temperature, humidity and pH value respectively, and a, b and c represent the corresponding standardized values ​​of environmental factors;

[0035] S223: Performing regression analysis on the standardized environmental factors according to the weighted sum of the environmental factors to generate a predicted soil organic matter concentration value.

[0036] As a further solution of the present invention, the steps for obtaining the soil organic matter weight adjustment list are specifically as follows:

[0037] S311: Based on the predicted soil organic matter concentration value and the original data of environmental factors, analyzing the influence of each environmental factor on the soil organic matter concentration, estimating the coefficient of each environmental factor using a multivariate linear regression model, and obtaining an initial estimated weight;

[0038] S312: Adjust the initial estimation weight, identify and analyze the environmental conditions of the differentiated areas, and use the formula:

[0039] ,

[0040] Generate a list of adjusted weights;

[0041] in, represents the adjusted weight, is the initial weight of the kth environmental factor, is the adjustment coefficient corresponding to the environmental factor, K is the total number of environmental factors;

[0042] S313: Utilizing the adjusted weight list and combining the soil organic matter concentration distribution characteristics of each region, the weight distribution of environmental factors in each region is optimized to obtain a soil organic matter weight adjustment list.

[0043] As a further solution of the present invention, the steps for obtaining the optimized soil organic matter concentration data set are specifically as follows:

[0044] S321: Based on the soil organic matter weight adjustment list, extract the organic matter concentration data and spatial coordinates of the soil sampling points, correct the concentration values ​​point by point according to the weight adjustment principle, compare and remove abnormal data points, and obtain an adjusted soil sampling concentration data set;

[0045] S322: Based on the adjusted soil sampling concentration data set, the concentration distribution of continuous grids is analyzed through the concentration value and coordinate point association table, the blank areas are filled in by grid-by-grid interpolation, and the data is corrected in combination with terrain characteristic parameters, and the influence range of land use type is analyzed by weight superposition, and the gradient change of grid concentration is corrected to obtain gridded soil organic matter spatial distribution data;

[0046] S323: Based on the gridded soil organic matter spatial distribution data, extract the areas with drastic local concentration changes, perform multi-point resampling and correction on the concentration values ​​in combination with the soil characteristics in the area, merge the local concentration data with the grid data, and perform difference verification and data consistency check to generate an optimized soil organic matter concentration data set.

[0047] As a further solution of the present invention, the steps for obtaining the dynamic soil organic matter estimation result are specifically as follows:

[0048] S411: According to the optimized soil organic matter concentration data set, a spatial statistical model is used to adjust the spatial heterogeneity of the soil organic matter concentration, and the current soil organic matter concentration distribution result is obtained by referring to the spatial relationship of adjacent data points and the initial concentration value of the soil organic matter;

[0049] S412: Using the current soil organic matter concentration distribution result, dynamic optimization and error correction are performed, and the formula is applied:

[0050] ,

[0051] Get an estimate of soil organic matter concentration;

[0052] in, represents the estimated soil organic matter concentration, is the calculated initial concentration, is the reference soil organic matter concentration value, G is the error correction coefficient;

[0053] S413: Based on the estimated value of soil organic matter concentration and in combination with current environmental factors, feature extraction and updating of soil organic matter concentration are performed, and the features of environmental factors are integrated using data fusion technology to form a feature subset. The feature subset is optimized through a feature selection algorithm to generate a dynamic soil organic matter estimation result.

[0054] Compared with the prior art, the advantages and positive effects of the present invention are:

[0055] In the present invention, by extracting environmental factors and analyzing the spatial variation relationship between environmental factors and soil organic matter, the spatial distribution characteristics of soil organic matter are effectively revealed. Based on the spatial weighted results of environmental factors, their contribution to the estimation of soil organic matter concentration is optimized, and then the accuracy of the weighted values ​​of environmental factors is improved through bandwidth optimization adjustment, so that the estimation of soil organic matter concentration is more accurate, especially reflecting more detailed local changes on the spatial scale. When the estimation results of soil organic matter concentration are dynamically adjusted, the generation of soil organic matter weight adjustment list can identify and correct outliers in local areas, thereby enhancing the robustness of the overall estimation. The spatial heterogeneity adjustment and error correction methods improve the reliability of soil organic matter estimation results, can capture the complex changes in soil organic matter distribution on different spatial scales, and improve the monitoring accuracy, especially in areas where the soil organic matter concentration is relatively uneven. Through dynamic optimization and real-time updating, the limitations of static models are avoided, ensuring that the estimation results are more timely and accurate, thereby providing more accurate decision support for agricultural management and environmental protection. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 It is a schematic diagram of the main steps of the present invention;

[0057] Figure 2 is a flow chart of the spatial variation mode in the present invention;

[0058] Figure 3 It is a flow chart of the spatial weighted results of environmental factors in the present invention;

[0059] Figure 4 It is a flow chart of the weighted regression parameter matrix of environmental factors in the present invention;

[0060] Figure 5 A flow chart of the soil organic matter concentration value predicted in the present invention;

[0061] Figure 6 A flow chart of the soil organic matter weight adjustment list in the present invention;

[0062] Figure 7 A flow chart of the optimized soil organic matter concentration data set in the present invention;

[0063] Figure 8 It is a flow chart of the dynamic soil organic matter estimation results in the present invention. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0065] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, in the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0066] Embodiment 1:

[0067] See also Figure 1 The present invention provides a technical solution: a spatially variable multi-scale soil organic matter remote sensing estimation method, comprising the following steps:

[0068] S1: By collecting soil samples at different locations, testing the soil organic matter content and extracting environmental factors, conducting spatial difference analysis, analyzing the spatial variation relationship between environmental factors and soil organic matter, generating spatial variation patterns, optimizing the spatial weighted relationship between remote sensing index and environmental factors, and obtaining the spatial weighted results of environmental factors;

[0069] S2: Based on the spatial weighted results of environmental factors, the spatial weighted values ​​of environmental factors are adjusted through bandwidth optimization, the spatial regression parameters of environmental factors and soil organic matter concentration are analyzed, the weighted regression parameter matrix of environmental factors is obtained, the soil organic matter concentration of differentiated regions is estimated, and the predicted soil organic matter concentration value is obtained;

[0070] S3: Based on the predicted soil organic matter concentration values, adjust the spatial scale, optimize the weights of environmental factors in differentiated areas, identify the local impact of environmental factors on soil organic matter concentration, generate a list of soil organic matter weight adjustments, and update the soil organic matter concentration of the region in combination with the spatial change pattern to obtain the optimized soil organic matter concentration dataset;

[0071] S4: Based on the optimized soil organic matter concentration dataset, spatial heterogeneity adjustment is performed, the soil organic matter concentration of the region is dynamically optimized and error correction is performed, and the feature subset is extracted and adjusted and updated according to the characteristics of the current environmental factors to generate dynamic soil organic matter estimation results.

[0072] The spatial change pattern includes change trend, affected area, and pattern type. The spatial weighted results of environmental factors include weighted index, spatial relationship, and adjustment parameters. The weighted regression parameter matrix of environmental factors includes regression coefficient, variable weight, and statistical significance. The predicted soil organic matter concentration values ​​include estimated values, analysis areas, and prediction parameters. The soil organic matter weight adjustment list includes weight distribution, adjustment range, and influencing factors. The optimized soil organic matter concentration data set includes local weights, optimized areas, and adjustment results. The dynamic soil organic matter estimation results include dynamic valuation, corrected errors, and feature adjustments.

[0073] See also Figure 2 , the steps to obtain the spatial variation pattern are as follows:

[0074] S111: Construct an environmental factor dataset by collecting soil samples at different locations, testing soil organic matter content, recording environmental factors such as temperature, humidity, and land use type;

[0075] Before conducting the multivariate linear regression analysis, soil samples were first systematically collected from different geographical locations. This involved selecting representative locations to ensure the diversity and breadth of the samples, thereby obtaining a statistically significant data set. The soil organic matter content of each sample was measured, and the associated environmental factors, such as temperature, humidity, and land use type, were recorded to ensure the accuracy of subsequent analysis. This process required detailed data entry and preliminary data review to eliminate any errors or deviations. In order to further analyze the data, after entering the data, preliminary data processing was required, including data cleaning and preliminary statistical analysis, such as calculating statistical parameters such as the mean, median, and standard deviation of each environmental factor. The ultimate goal was to construct an environmental factor data set for subsequent regression analysis.

[0076] S112: Based on the environmental factor data set, with soil organic matter content as the dependent variable and environmental factors as the independent variables, perform multivariate linear regression analysis, determine key environmental factors through stepwise regression, calculate the contribution of key environmental factors, and obtain preliminary regression analysis results;

[0077] After completing the construction of the environmental factor data set, perform multiple linear regression analysis. First, you need to select appropriate statistical software or tools for analysis, then import the sorted data set, select soil organic matter content as the dependent variable, and each environmental factor as the independent variable, and use the stepwise regression method to identify significant environmental factors. This requires performing a series of statistical tests, such as ANOVA analysis and T test, to determine which factors statistically significantly affect soil organic matter content, and calculate the regression coefficient of the significant factor, that is, the influence of each factor on the dependent variable. This process not only involves complex calculations, but also requires strict verification and adjustment of the calculation results to ensure the accuracy and reliability of the model, and obtain preliminary regression analysis results, which provide a scientific basis for further spatial analysis.

[0078] S113: Based on the preliminary regression analysis results, perform spatial variation analysis by identifying the distribution of soil organic matter using the formula:

[0079] ,

[0080] Get the spatial variation pattern;

[0081] in, represents the predicted soil organic matter content of the i-th sample point, represents the environmental factor temperature, Indicates the environmental factor humidity, represents the measured value of the nth environmental factor, represents the model intercept, represents the regression coefficient of temperature, represents the regression coefficient of humidity, Represents the regression coefficient of land use type.

[0082] The benefit of the formula is that it can describe and predict in detail how soil organic matter content is affected by the combined influence of various environmental factors through the multiple linear regression model, and can precisely regulate and manage factors to optimize soil quality;

[0083] Set in a specific experimental area, measure the soil organic matter content at a certain point 10%, environmental factor temperature 20℃, humidity 30%, land use type For cultivated land, the values ​​are obtained through actual monitoring data. According to the established regression model, the regression coefficient of the known temperature is 0.05, the regression coefficient of humidity is 0.03, and the regression coefficient of land use type is 0.2, the intercept is 1, substitute the numerical value into the calculation :

[0084] ,

[0085] The results showed that the predicted soil organic matter content was 3.1% under the environmental factors, which indicated that further investigation and control strategies were needed to improve the soil organic matter content in the area.

[0086] See also Figure 3 , the specific steps for obtaining the spatial weighted results of environmental factors are:

[0087] S121: Based on the spatial variation pattern, according to the preset spatial resolution, the region is divided, the weight of each region is analyzed and the data is weighted, and the weighted data is output according to the spatial grid to obtain the regional weighted relationship;

[0088] Define the size of the spatial grid and divide the region based on it, ensure that the boundaries of each grid are clear and the grid size is consistent, so as to facilitate the uniformity and accuracy of subsequent data processing. Collect geographical and environmental data in each grid area, including terrain, vegetation coverage and soil type, for subsequent weight analysis. The weight of each region is calculated based on factors such as its environmental sensitivity and the intensity of human activities to ensure the objectivity and scientificity of data processing. Multiply the weight of each region with the original data for weighted processing to ensure that the actual influence of the data in spatial analysis is accurately reflected. Output the weighted data according to the spatial grid to obtain the regional weighted relationship to ensure the spatial continuity of the data and the integrity of the analysis.

[0089] S122: Based on the regional weighted relationship, extract the spatial changes of environmental factors in the region, analyze the spatial influence of environmental factors in key areas, determine the weighted contribution of environmental factors to remote sensing index, rank them according to the contribution size, and generate the weighted contribution distribution results of environmental factors;

[0090] Using geographic information system (GIS) technology, based on the existing regional weighted relationship, the spatial distribution and changing trends of various environmental factors are extracted. The distribution density of environmental factors in each region is calculated through spatial statistical methods to determine their actual impact on the remote sensing index. According to the calculation results, the influence of environmental factors will be assigned different weight values ​​and ranked accordingly to generate the weighted contribution distribution results of environmental factors. The ranking of environmental factors not only assists in identifying key factors, but also facilitates monitoring which factors have a greater spatial influence in a specific area.

[0091] S123: Based on the weighted contribution distribution results of environmental factors, the weighted coefficient of each environmental factor is adjusted, and the spatial weighted coefficient is optimized by referring to the difference in the impact of environmental factors on differentiated regions and combining real-time geographical features. By adjusting the weight ratio, the spatial weighted result of environmental factors is generated;

[0092] According to the weighted contribution distribution results of environmental factors, the influence of each environmental factor in different regions is evaluated, especially in areas with significant differences. According to the evaluation results, the original weighting coefficient is appropriately adjusted to ensure that the weight of each environmental factor is more in line with the influence of actual geographical and ecological characteristics. During the adjustment process, real-time geographical feature data are considered, such as real-time meteorological changes, vegetation growth conditions, etc. The spatial weighting coefficient of environmental factors is optimized through data, and the spatial weighted results of environmental factors are generated by adjusting the weight ratio to ensure that the responsiveness and adaptability of the model more accurately reflect the changes in spatial data under different environmental conditions.

[0093] See also Figure 4 , the specific steps for obtaining the environmental factor weighted regression parameter matrix are:

[0094] S211: Based on the spatial weighted results of environmental factors, through the spatial association between environmental factors and soil organic matter concentration, the spatial weighted value of each environmental factor is adjusted by bandwidth optimization, and weighted processing is performed. The weighted value of the environmental factor is obtained by weight and spatial distance calculation;

[0095] According to the spatial correlation between environmental factors and soil organic matter concentration, the weighted values ​​of environmental factors are first calculated. The weights are determined based on the distribution characteristics of each environmental factor in the region and the correlation with the soil organic matter concentration. The relationship between environmental factors and soil organic matter concentration is calculated through spatial distance, and an appropriate weighting function is selected. This function can be represented by a spatial regression model. The weight coefficient of each factor is obtained based on the regression model. According to the regression model, the weight of the environmental factor is estimated by the stepwise regression method or the least squares method, including collecting data from multiple monitoring points in the region, normalizing it, and adjusting the values ​​of the environmental factors to the same order of magnitude. According to the spatial position of each factor, the environmental factors are weighted using the weighted distance function to calculate the weighted values ​​of the environmental factors. After all environmental factors are weighted, they are used in the next step of regression analysis to characterize the contribution of each environmental factor to the soil organic matter concentration.

[0096] S212: Based on the weighted values ​​of environmental factors and combined with soil organic matter concentration data, spatial regression analysis is performed. The regression model is constructed using spatial data, the weight of environmental factors in the model is adjusted, the model is tested cyclically, the optimal factor weights are determined, and the accuracy of the regression model is optimized. The formula is:

[0097] ,

[0098] Get the adjusted regression parameters;

[0099] Among them, R represents the adjusted regression parameter, is the weighted value of environmental factors, is the environmental factor data, is the spatial distance between the jth factor and the soil organic matter concentration, is the spatial distance adjustment coefficient, is the bandwidth optimization factor, is a constant adjustment term, and m represents the total number of factors;

[0100] The benefit of the formula is that it comprehensively considers the weighted value of environmental factors, spatial distance and bandwidth optimization coefficients, and can more accurately reflect the impact of environmental factors on soil organic matter concentration. The core innovation of the formula is to introduce the spatial distance adjustment coefficient and bandwidth optimization factor , which quantifies the impact of environmental factors at different distances on regression results and can effectively optimize the weighting coefficients in regression analysis;

[0101] represents the weighted value of the jth environmental factor, which is calculated by spatial regression in the previous step and reflects the contribution of environmental factors to soil organic matter concentration;

[0102] represents the measured value of the jth environmental factor, which is derived from the field monitoring data;

[0103] is the spatial distance between the jth environmental factor and the soil organic matter concentration. The specific location of each monitoring point is obtained through GPS positioning during calculation;

[0104] and is the adjustment coefficient in the spatial regression model, Control the influence of spatial distance on weighted value, This determines the effect of bandwidth optimization;

[0105] It is a constant term used to compensate for other influencing factors not included in the model;

[0106] Assume that there are three environmental factors in a certain area, and calculate their weighted values ​​respectively: , , The measured values ​​of environmental factors are , , , and the corresponding spatial distances are = 5 meters, = 3 meters, =8 meters, set the space distance adjustment coefficient =0.2, bandwidth optimization factor =1.5, constant term =0.1, the calculation process is as follows:

[0107] ,

[0108] First calculate the weight of each item:

[0109] ,

[0110] Similarly, calculate the second and third terms:

[0111] ,

[0112] ,

[0113] Finally add all the terms together and add the constant term :

[0114] ,

[0115] The results showed that the regression parameter between environmental factors and soil organic matter concentration was 0.2816, which reflected the influence of environmental factors on soil organic matter concentration in the selected area and provided a quantitative basis for subsequent analysis.

[0116] S213: Optimize the adjusted regression parameters, combine the spatial factors, analyze the spatial distribution characteristics of soil organic matter concentration, and generate the environmental factor weighted regression parameter matrix;

[0117] The regression parameters are optimized through the adjusted regression parameters. During the optimization process, the contribution of the weighted regression parameters is evaluated. The regression parameters are iteratively adjusted using stepwise regression or weighted regression analysis methods. For each regression parameter, it is corrected according to its correlation with the soil organic matter concentration so that the optimized coefficient is more in line with the actual situation. A suitable regression optimization algorithm is selected to gradually adjust each regression parameter. By comparing the results before and after the correction, the optimal regression parameter value is determined and matched with the spatial distribution characteristics of the soil organic matter concentration. According to the optimized regression parameter matrix, the weighted regression analysis of the spatial factors is further performed to obtain a more accurate regression model. The analysis process is completed and the weighted regression parameter matrix of the environmental factors is generated.

[0118] See also Figure 5 , the specific steps for obtaining the predicted soil organic matter concentration value are:

[0119] S221: According to the environmental factor weighted regression parameter matrix, the environmental factor data of the differentiated regions are captured, including temperature, humidity and soil pH value, and each environmental factor is monitored in real time to obtain a standardized environmental factor data set;

[0120] When conducting environmental monitoring, we first collect environmental factor data for differentiated areas, including temperature, humidity and soil pH. Each monitoring point is equipped with a sensor to capture data in real time. The data generated by the sensor is first stored in the regional data center. Inside the data center, each factor data is recorded according to the timestamp, and data cleaning and preliminary analysis are performed. The cleaning process includes removing outliers, correcting deviations and smoothing data. All data are converted into a format suitable for further analysis, that is, standardization. Standardization is to convert data into a unit-free and dimensionless form through the Z-score method, thereby eliminating the influence of different units and magnitudes in the original data and obtaining a standardized environmental factor data set.

[0121] S222: Weight allocation is performed on the standardized environmental factor data set. The weight of each factor is determined based on the original impact data of environmental factors on soil organic matter concentration. The formula is:

[0122] ,

[0123] Calculate the weighted sum of environmental factors;

[0124] Where A represents the weighted sum of environmental factors, , , represent the weights of temperature, humidity and pH value respectively, and a, b and c represent the corresponding standardized values ​​of environmental factors;

[0125] By adjusting the weights, the influence of different environmental factors on soil organic matter concentration can be accurately reflected, increasing the adaptability and prediction accuracy of the model;

[0126] , , It is obtained through historical data analysis, which involves collecting environmental factor data and corresponding data on soil organic matter concentration in the same area over a period of time in the past, using statistical or machine learning methods to analyze the relationship between each factor and soil organic matter concentration, and based on this, estimating the impact weight of each factor;

[0127] Assume that the standardized environmental factor data are obtained through actual measurement, for example, temperature a=0.5, humidity b=0.3, soil pH c=0.2, and the corresponding weights are obtained through historical data analysis, for example =0.6, =0.3, =0.1, substitute the value into the formula and calculate:

[0128] ,

[0129] The calculation result A=0.41 represents the weighted sum of environmental factors. By taking the weighted average of different environmental factors according to their actual impact, this value can be directly used for decision support to guide agricultural production and environmental management.

[0130] S223: Performing regression analysis on the standardized environmental factors based on the weighted sum of the environmental factors to generate predicted soil organic matter concentration values;

[0131] Using the weighted sum of environmental factors, a linear regression model is first introduced. The model relies on the weighted average of environmental factors to predict the concentration of soil organic matter. Through regression analysis, the coefficients of each environmental factor are evaluated to determine the contribution of the final result, including the estimation of the coefficients and the calculation of standard errors. Through statistical indicators, the model's fit and predictive ability can be evaluated. The obtained regression model not only provides the specific contribution value of each factor, but also can further adjust the model parameters and optimize the model performance by comparing the predicted values ​​with the measured values, thus obtaining the predicted soil organic matter concentration value, which will be directly applied to the decision-making process of soil management and agricultural production.

[0132] See also Figure 6 , the specific steps for obtaining the soil organic matter weight adjustment list are:

[0133] S311: Based on the predicted soil organic matter concentration value and the original data of environmental factors, the influence of each environmental factor on the soil organic matter concentration is analyzed, and the coefficient of each environmental factor is estimated using a multivariate linear regression model to obtain the initial estimated weight;

[0134] Weight calculation is performed based on the predicted value of soil organic matter and historical data of environmental factors. The impact of various environmental factors on soil organic matter concentration is analyzed through a multivariate linear regression model. The model is trained using existing data sets to estimate the coefficient of each factor in the model. The coefficient reflects the contribution of each environmental factor to the change in soil organic matter concentration. The coefficient is used to form the initial estimated weight, which provides basic data for subsequent steps. The coefficient of each factor is obtained through statistical analysis of environmental data, including data collection and processing of factors such as temperature, humidity, and pH value. The data is collected by sensors in different areas and at different times to ensure the representativeness and accuracy of the data. By analyzing the regression coefficients of various environmental factors, the specific impact of each factor on soil organic matter concentration can be clarified, providing a scientific basis for the next step of weight adjustment.

[0135] S312: Adjust the initial estimation weights, identify and analyze the environmental conditions of the differentiated areas, using the formula:

[0136] ,

[0137] Generate a list of adjusted weights;

[0138] in, represents the adjusted weight, is the initial weight of the kth environmental factor, is the adjustment coefficient corresponding to the environmental factor, K is the total number of environmental factors;

[0139] The benefit of the formula is that by considering the region-specific environmental factor adjustment coefficient, it can more accurately reflect the impact of environmental characteristics of different regions on soil organic matter concentration, thereby achieving a more customized environmental management strategy;

[0140] First, set the specific parameters, set K = 3, the regional environmental factors include temperature, humidity and pH value, and the initial weights are , , The adjustment coefficients are , , , calculated according to the formula:

[0141] ,

[0142] The results show that the adjusted average weight is 0.34, reflecting the average impact weight of each environmental factor on soil organic matter concentration after considering regional specific adjustments. This calculation process not only shows how to adjust the weights of each factor through specific parameters and actual data, but also highlights how the adjusted weights can better reflect regional environmental characteristics.

[0143] S313: using the adjusted weight list and combining the soil organic matter concentration distribution characteristics of each region, optimizing the weight distribution of environmental factors in each region, and obtaining a soil organic matter weight adjustment list;

[0144] First, the concentration distribution data of soil organic matter in each region was analyzed, and soil sample data was collected through the geographic information system (GIS) and remote sensing technology. The data was spatially analyzed to identify high-value and low-value areas of soil organic matter. Then, the weights of various environmental factors were adjusted according to the data to ensure that the weight distribution was more in line with the actual soil environmental conditions. The specific data of each region, such as soil moisture, temperature range and pH value changes, were taken into consideration. By calculating the weight product of the environmental factors in each region, a soil organic matter weight adjustment list was formed, which was applied to the formulation of soil management and improvement measures, so as to achieve the purpose of precise policy implementation. This refinement process highlights how to adjust the weights through geographic information data and ensure that the weights can actually reflect the environmental characteristics of different regions, so as to manage soil more accurately.

[0145] See also Figure 7 , the steps for obtaining the optimized soil organic matter concentration dataset are as follows:

[0146] S321: Based on the soil organic matter weight adjustment list, extract the organic matter concentration data and spatial coordinates of the soil sampling points, correct the concentration values ​​point by point according to the weight adjustment principle, compare and remove abnormal data points, and obtain the adjusted soil sampling concentration data set;

[0147] In the process of correcting the concentration value point by point according to the weight adjustment principle, it is necessary to read the original concentration data of the sampling points one by one, and make adjustments item by item in combination with the weight factors for different regions or soil types in the weight adjustment list, including weighted calculation of the weight coefficient for the data of each sampling point, multiplying the weight factor by the initial concentration value of the sampling point and reassigning it according to the adjustment rules. For repeated data at the same sampling point under different sampling times or methods, it is necessary to calculate its weighted average to eliminate data bias. For abnormal data points, the upper and lower quartile method is used to calculate the concentration distribution range of all sampling points. Concentration values ​​exceeding three times the interquartile range are identified as abnormal data and replaced by the concentration mean of their spatially adjacent points to ensure the continuity and rationality of the data after elimination, forming an adjusted soil sampling concentration data set as the basis for subsequent analysis.

[0148] S322: Based on the adjusted soil sampling concentration data set, the concentration value and coordinate point association table are used to analyze the continuous grid concentration distribution, and the blank areas are filled in by grid-by-grid interpolation. The data is corrected in combination with terrain characteristic parameters, and the influence range of land use type is analyzed by weight superposition. The gradient change of grid concentration is corrected to obtain the gridded soil organic matter spatial distribution data;

[0149] The sampling concentration data are accurately matched with the geographic coordinates to form a corresponding relationship table between the geographic location and concentration value of each sampling point. In the process of grid-by-grid interpolation, the boundary of each grid is determined according to the grid division rules, and the concentration value of the unsampled grid is calculated in combination with the sampling point data of the adjacent grids. In the interpolation calculation, the spatial distance weight of the sampling points is selected as the basis for interpolation, and the points with closer distances are given higher influence values. For blank areas, the terrain characteristic parameters in the area, such as slope and altitude differences, are considered to correct the interpolation results and avoid unreasonable concentration value distribution caused by terrain influence. In the weighted overlay analysis of land use types, the influence range of each type is layered and overlaid on the grid concentration value, and weight coefficients are assigned to different land types such as cultivated land and forest land. The grid concentration is adjusted according to the regional overlay rules, and its concentration gradient change is corrected in units of grids to obtain gridded soil organic matter spatial distribution data.

[0150] S323: Based on the gridded soil organic matter spatial distribution data, extract the area with drastic local concentration changes, perform multi-point resampling and correction on the concentration value in combination with the soil characteristics in the area, merge the local concentration data with the grid data, perform difference verification and data consistency check, and generate an optimized soil organic matter concentration data set;

[0151] First, the identification criteria for drastic concentration changes are defined, and the areas within the grid where the concentration gradient exceeds a certain threshold are marked as drastic change areas. When performing multi-point resampling of the area, additional sampling points need to be accurately selected based on the geographical location to supplement the spatial resolution of the data. The mean and variance are calculated through multiple sampling to adjust the stability of the concentration data in the area. When correcting the concentration value in combination with the soil characteristics in the area, the physical and chemical properties data of the soil, such as soil texture, pH value, etc., can be referred to, and the attributes and concentration values ​​are linearly mapped. The data values ​​with abnormal fluctuations are corrected through regression analysis. After the local concentration data is corrected, it is re-matched and merged with the original gridded data, and the difference analysis method is used to compare the new and old data grid by grid to confirm the consistency and effectiveness of the data update, generate an optimized soil organic matter concentration data set, and provide accurate data support for further ecological or agricultural research.

[0152] See also Figure 8 , the specific steps for obtaining dynamic soil organic matter estimation results are as follows:

[0153] S411: Based on the optimized soil organic matter concentration dataset, the spatial statistical model is used to adjust the spatial heterogeneity of the soil organic matter concentration, and the current soil organic matter concentration distribution result is obtained by referring to the spatial relationship of adjacent data points and the initial concentration value of the soil organic matter;

[0154] From the locally optimized data set, soil organic matter concentration data in a small area is selected. The data has been preliminarily measured and recorded in the data system. Each data point has a geographic tag so that the concentration value of each point can be accurately adjusted. The spatial statistical model is used to analyze the spatial relationship between data points. The model automatically adjusts the data between adjacent points based on geographic proximity to reduce spatial errors and improve the spatial consistency of the overall data. The model input includes the soil organic matter concentration of each point and its geographic coordinates. The model output is an adjusted concentration distribution map to ensure that the spatial heterogeneity of the data is fully considered and generate the current soil organic matter concentration distribution results, which will be used in subsequent optimization and analysis steps.

[0155] S412: Use the current soil organic matter concentration distribution results to perform dynamic optimization and error correction, and apply the formula:

[0156] ,

[0157] Get an estimate of soil organic matter concentration;

[0158] in, represents the estimated soil organic matter concentration, is the calculated initial concentration, is the reference soil organic matter concentration value, G is the error correction coefficient;

[0159] The benefit of the formula is that by dynamically adjusting the error correction factor G, the concentration estimate of soil organic matter can be adjusted according to the actual measurement deviation of the soil sample and previous data to adapt to geographical and environmental changes and enhance the application value of the data;

[0160] Set the adjusted soil organic matter concentration obtained from the site 15%, referring to soil organic matter concentration The error correction factor G is 12% and is determined to be 0.8 based on the data changes in the past year. The calculation process is:

[0161] ,

[0162] The results showed that the estimated soil organic matter concentration was 12.6%, and this value will be used for further analysis and decision support, indicating that through error correction, it is possible to get closer to the actual soil organic matter reference value, which is crucial to improving the accuracy and efficiency of soil management.

[0163] S413: Based on the estimated value of soil organic matter concentration and combined with the current environmental factors, the characteristics of soil organic matter concentration are extracted and updated, and the characteristics of environmental factors are integrated using data fusion technology to form a feature subset. The feature subset is optimized through the feature selection algorithm to generate dynamic soil organic matter estimation results;

[0164] Using the data after dynamic optimization and error correction, combined with current environmental factors such as temperature, humidity and soil type, the optimized soil organic matter concentration is feature extracted. The environmental factor data is obtained through field sensors and recorded in the environmental monitoring database. The feature extraction uses multivariate statistical analysis methods to ensure that the most relevant features are extracted from complex environmental data. Data fusion technology is used to integrate features, and the most effective feature subset is selected through algorithm optimization to reduce the complexity of the model and improve the accuracy of the prediction. The feature selection algorithm analyzes the correlation between each feature and soil organic matter concentration, selects the environmental factors most relevant to the changes in soil organic matter, and generates dynamic soil organic matter estimation results. The current environmental data and adjusted soil data are used to provide scientific decision support for soil management.

[0165] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. A spatially variable multi-scale soil organic matter remote sensing estimation method, characterized in that: The following steps are involved: By collecting soil samples at different locations, testing soil organic matter content and extracting environmental factors, conducting spatial difference analysis, analyzing the spatial variation relationship between environmental factors and soil organic matter, generating spatial variation patterns, optimizing the spatial weighted relationship between remote sensing index and environmental factors, and obtaining spatial weighted results of environmental factors; Based on the spatial weighted results of the environmental factors, the spatial weighted values ​​of the environmental factors are adjusted through bandwidth optimization, the spatial regression parameters of the environmental factors and the soil organic matter concentration are analyzed, the weighted regression parameter matrix of the environmental factors is obtained, the soil organic matter concentration of the differentiated regions is estimated, and the predicted soil organic matter concentration value is obtained; The steps for obtaining the environmental factor weighted regression parameter matrix are specifically as follows: Based on the spatial weighted results of the environmental factors, through the spatial association between the environmental factors and the soil organic matter concentration, the spatial weighted value of each environmental factor is adjusted by using bandwidth optimization, and weighted processing is performed, and the weighted value of the environmental factor is obtained by weight and spatial distance calculation; Based on the weighted values ​​of the environmental factors, combined with the soil organic matter concentration data, spatial regression analysis is performed, the regression model is constructed using spatial data, the proportion of environmental factors in the model is adjusted, the model is tested cyclically, the optimal factor weights are determined, and the accuracy of the regression model is optimized. The formula is: ， Get the adjusted regression parameters; Among them, R represents the adjusted regression parameter, is the weighted value of environmental factors, is the environmental factor data, is the spatial distance between the jth factor and the soil organic matter concentration, is the spatial distance adjustment coefficient, is the bandwidth optimization factor, is a constant adjustment term, and m represents the total number of factors; Optimizing the adjusted regression parameters, combining spatial factors, analyzing the spatial distribution characteristics of soil organic matter concentration, and generating an environmental factor weighted regression parameter matrix; The steps for obtaining the predicted soil organic matter concentration value are specifically as follows: According to the environmental factor weighted regression parameter matrix, the environmental factor data of the differentiated regions are captured, including temperature, humidity and soil pH value, and each environmental factor is monitored in real time to obtain a standardized environmental factor data set; The standardized environmental factor data set is weighted, and the weight of each factor is determined by the original impact data of environmental factors on soil organic matter concentration, using the formula: ， Calculate the weighted sum of environmental factors; Among them, A represents the weighted sum of environmental factors, , , represent the weights of temperature, humidity and pH value respectively, and a, b and c represent the corresponding standardized values ​​of environmental factors; According to the weighted sum of the environmental factors, a regression analysis is performed on the standardized environmental factors to generate a predicted soil organic matter concentration value; Based on the predicted soil organic matter concentration value, the spatial scale is adjusted, the weights of environmental factors in differentiated areas are optimized, the local effects of environmental factors on soil organic matter concentration are identified, a soil organic matter weight adjustment list is generated, and the soil organic matter concentration of the region is updated in combination with the spatial change pattern to obtain an optimized soil organic matter concentration dataset; The steps for obtaining the soil organic matter weight adjustment list are specifically as follows: Based on the predicted soil organic matter concentration value and the original data of environmental factors, the influence of each environmental factor on the soil organic matter concentration is analyzed, and the coefficient of each environmental factor is estimated using a multivariate linear regression model to obtain an initial estimated weight; Adjust the initial estimated weights, identify and analyze the environmental conditions of the differentiated areas, using the formula: ， Generate a list of adjusted weights; in, represents the adjusted weight, is the initial weight of the kth environmental factor, is the adjustment coefficient corresponding to the environmental factor, K is the total number of environmental factors; Using the adjusted weight list, combined with the soil organic matter concentration distribution characteristics of each region, the weight distribution of environmental factors in each region is optimized to obtain a soil organic matter weight adjustment list; The steps for obtaining the optimized soil organic matter concentration data set are specifically as follows: Based on the soil organic matter weight adjustment list, the organic matter concentration data and spatial coordinates of the soil sampling points are extracted, the concentration values ​​are corrected point by point according to the weight adjustment principle, and the abnormal data points are compared and removed to obtain an adjusted soil sampling concentration data set; Based on the adjusted soil sampling concentration data set, the concentration distribution of continuous grids is analyzed through the concentration value and coordinate point association table, the blank areas are filled in by grid-by-grid interpolation, and the data is corrected in combination with terrain characteristic parameters. The influence range of land use type is analyzed by weight superposition, the gradient change of grid concentration is corrected, and the gridded soil organic matter spatial distribution data is obtained; Based on the gridded soil organic matter spatial distribution data, extract the area where the local concentration changes dramatically, perform multi-point resampling and correction on the concentration value in combination with the soil characteristics in the area, merge the local concentration data with the grid data, perform difference verification and data consistency check, and generate an optimized soil organic matter concentration data set; Based on the optimized soil organic matter concentration dataset, spatial heterogeneity adjustment is performed, the soil organic matter concentration of the region is dynamically optimized and error correction is performed, and the feature subset is extracted and adjusted and updated according to the characteristics of the current environmental factors to generate dynamic soil organic matter estimation results.

2. The spatially variable multi-scale soil organic matter remote sensing estimation method according to claim 1 is characterized in that: The steps of acquiring the spatial variation pattern are specifically as follows: By collecting soil samples at different locations, testing soil organic matter content, recording environmental factors such as temperature, humidity and land use type, and constructing an environmental factor dataset; According to the environmental factor data set, multiple linear regression analysis is performed by taking soil organic matter content as the dependent variable and environmental factors as the independent variables, determining key environmental factors through stepwise regression, calculating the contribution of key environmental factors, and obtaining preliminary regression analysis results; Based on the results of the preliminary regression analysis, spatial variability analysis was performed by identifying the distribution of soil organic matter using the formula: ， Get the spatial variation pattern; in, represents the predicted soil organic matter content of the i-th sample point, represents the environmental factor temperature, Indicates the environmental factor humidity, represents the measured value of the nth environmental factor, represents the model intercept, represents the regression coefficient of temperature, represents the regression coefficient of humidity, Represents the regression coefficient of land use type.

3. The spatially variable multi-scale soil organic matter remote sensing estimation method according to claim 2 is characterized in that: The steps for obtaining the spatial weighted results of the environmental factors are specifically as follows: Based on the spatial variation pattern, the region is divided according to the preset spatial resolution, the weight of each region is analyzed and the data is weighted, and the weighted data is output according to the spatial grid to obtain the regional weighted relationship; Based on the regional weighted relationship, the spatial changes of environmental factors in the region are extracted, the spatial influence of environmental factors in the key areas is analyzed, the contribution of environmental factors to the weighted remote sensing index is determined, and the weighted contribution distribution results of environmental factors are generated according to the contribution size; Based on the weighted contribution distribution results of the environmental factors, the weighted coefficient of each environmental factor is adjusted. Referring to the differences in the impact of environmental factors on differentiated regions and combined with real-time geographic features, the spatial weighted coefficient is optimized, and the spatial weighted results of environmental factors are generated by adjusting the weight ratio.

4. The spatially variable multi-scale soil organic matter remote sensing estimation method according to claim 1 is characterized in that: The steps for obtaining the dynamic soil organic matter estimation result are specifically as follows: According to the optimized soil organic matter concentration data set, a spatial statistical model is used to adjust the spatial heterogeneity of the soil organic matter concentration, and the current soil organic matter concentration distribution result is obtained by referring to the spatial relationship of adjacent data points and the initial concentration value of the soil organic matter; The current soil organic matter concentration distribution result is used to perform dynamic optimization and error correction, and the formula is applied: ， Get an estimate of soil organic matter concentration; in, represents the estimated soil organic matter concentration, is the calculated initial concentration, is the reference soil organic matter concentration value, G is the error correction coefficient; Based on the estimated value of soil organic matter concentration and combined with current environmental factors, feature extraction and updating of soil organic matter concentration are performed, and data fusion technology is used to integrate environmental factor features to form feature subsets. The feature subsets are optimized through feature selection algorithms to generate dynamic soil organic matter estimation results.

Images