Soil total salt content inversion system and method based on improved spectral index

Abstract

The invention belongs to the technical field of soil inversion, and particularly relates to a soil total salt content inversion system and method based on an improved spectral index, and the method comprises the steps: obtaining a remote sensing image obtained through preprocessing of Sentinel-2; acquiring a sample of the soil sampling point and measuring the total salt content of the soil; analyzing spectral reflectivity characteristics of different salinized soil; analyzing the correlation between the spectral reflectivity and the total salt content of the soil; a wave band with the strongest correlation with the total salt content of the soil is introduced to improve a traditional spectral index; analyzing the correlation between the improved spectral index and the total salt content of the soil; constructing a nonlinear index inversion model of the total salt content of the soil based on the improved spectral index; and comparing different model precisions to obtain an optimal soil total salt content inversion model and drawing. According to the method, the problems that the precision of the soil total salt content inversion method based on the spectral index is low and the actual application requirement is difficult to meet are solved, and decision support is provided for dynamic monitoring of land salinization and sustainable utilization of land resources.

Description

Technical Field

[0001] The present invention belongs to the technical field of soil inversion, and particularly relates to a soil total salt content inversion system and method based on an improved spectral index. Background Art

[0002] Soil salinization threatens world food security, causes serious environmental degradation and hinders crop growth and yield. More than 420 million hectares of farmland worldwide are affected by salinization. 2 Soil salinization severely restricts the sustainable development of agriculture and the green development of the economy.

[0003] Timely and accurate monitoring of the dynamic changes in soil salinization and understanding its influencing factors are the prerequisites for preventing soil degradation. Field sampling analysis is a traditional and intuitive method for monitoring soil salinity, which can understand the dynamic changes of soil salinity in more detail. However, it is relatively time-consuming and laborious, and the number of sampling points is generally limited, so the spatial continuous changes of soil salinity in the area cannot be obtained. Remote sensing has high operational efficiency in agricultural monitoring, and has characteristics such as wide data sources, fast update speed and large extraction range. Compared with traditional soil salinity monitoring methods, remote sensing can quickly and accurately monitor the dynamic changes of soil salinity.

[0004] Exploring the optimal system and method for soil total salt content inversion is of great significance for the prevention and control of soil salinization. At present, soil total salt content inversion mainly focuses on three methods: spectral index, machine learning and deep learning. The spectral index method has become an important method for soil total salt content inversion due to advantages such as easy acquisition and processing of data, high calculation efficiency, low resource requirements, and strong model interpretability.

[0005] In the existing methods, the accuracy of the soil total salt content inversion method based on the spectral index is relatively low, which is difficult to meet the actual application requirements. To solve this problem, this paper proposes a soil total salt content inversion system and method based on an improved spectral index. Summary of the Invention

[0006] Aiming at the problems existing in the prior art, the purpose of the present invention is to provide a soil total salt content inversion system and method based on an improved spectral index.

[0007] To achieve the above purpose, the present invention provides the following technical solutions:

[0008] A soil total salt content inversion system and method based on an improved spectral index, comprising the following steps:

[0009] S1. Obtain remotely sensed images preprocessed by Sentinel-2;

[0010] S2. Obtain samples of soil sampling points and measure the soil total salt content;

[0011] S3. Analyze the spectral reflectance characteristics of different saline soils;

[0012] S4. Analyze the correlation between spectral reflectance and soil total salt content;

[0013] S5. Introduce the band with the strongest correlation with soil total salt content to improve the traditional spectral index;

[0014] S6. Analyze the correlation between the improved spectral index and soil total salt content;

[0015] S7. Construct a non - linear index inversion model of soil total salt content based on the improved spectral index;

[0016] S8. Compare the accuracies of different models, obtain the optimal soil total salt content inversion model and draw a graph.

[0017] Furthermore, in S1, the specific steps to obtain the remotely sensed image for pre - processing are as follows:

[0018] S1.1. Image download: Select the L2A - level data of Sentinel - 2 that has been radiometrically and atmospherically corrected, and obtain it from the Copernicus Open Access Hub of the European Space Agency;

[0019] S1.2. Image selection rule: Select the images during the period when the ground is bare, as close as possible to the actual soil sampling time and with a cloud cover of less than 10%;

[0020] S1.3. Image resampling: Import the downloaded L2A - level data into the SNAP software for resampling, and the sampling resolution is mainly for the optical band (10m).

[0021] Furthermore, in S2, the specific steps to measure the soil total salt content are as follows:

[0022] S2.1. Use a soil drill to take soil samples at the soil sampling points. Air - dry the taken soil samples, pass them through a 1 - mm sieve, then prepare a soil extract with a soil - water ratio of 1:5, and measure the electrical conductivity using a conductivity meter;

[0023] S2.2. The calculation formula for converting soil electrical conductivity into soil total salt content is:

[0024] C s = 3.7657EC 1：5 - 0.2405

[0025] Where C s is the soil total salt content, g / kg; EC 1：5 is the electrical conductivity of the soil extract with a soil - water ratio of 1:5, dS / m.

[0026] Furthermore, in S5, the specific calculation steps of the traditional spectral index are as follows:

[0027] S5.1. The 12 selected traditional spectral indices are respectively the Brightness Index (BI); multiple salinity indices (Salinity Indices, SI, SI1, SI2, SI3, SI4, SI5); Perpendicular Dryness Index (PDI); Normalized Difference Vegetation Index (NDVI); Enhanced Vegetation Index (EVI); Difference Vegetation Index (DVI); Normalized Difference Salinity Index (NDSI).

[0028] S5.2. The calculation formulas for traditional spectral indices are as follows:

[0029]

[0030]

[0031]

[0032]

[0033]

[0034]

[0035]

[0036]

[0037]

[0038]

[0039] DVI = NIR - R

[0040]

[0041] In the formula, B is the reflectance of the B2 blue band; G is the reflectance of the B3 green band; R is the reflectance of the B4 red band; NIR is the reflectance of the B8 near-infrared band.

[0042] Furthermore, in S5, the improvement principle formula is as follows:

[0043] GIndex = Index + C

[0044] GIndex = Index × C

[0045] Where Index is the traditional spectral index, GIndex is the improved spectral index, and C is the band with the strongest correlation with the total soil salt content.

[0046] Furthermore, in S7, the specific steps for constructing a non - linear index inversion model of the total soil salt content based on the improved spectral index are as follows:

[0047] S7.1: Select the improved spectral index with the strongest correlation with the total soil salt content and its corresponding traditional spectral index as independent variables and introduce them into the non - linear index model to construct an inversion model of the total soil salt content;

[0048] S7.2: The calculation formula of the non - linear index model is as follows:

[0049] C s = Ae Bx

[0050] Where C s is the total soil salt content, g·kg -1 ; A and B are undetermined coefficients; x is the traditional spectral index or the improved spectral index.

[0051] Furthermore, in S8, to compare the accuracy evaluation parameters of different models, the coefficient of determination (R 2 ), mean absolute error (MAE), root mean square error (RMSE), and ratio of performance to inter - quartile range (RPIQ) are used. The specific calculation formulas are:

[0052]

[0053]

[0054]

[0055]

[0056] Where N is the number of observation points; X i is the measured value of the total soil salt content, g·kg -1 ; Y i is the predicted value of the total soil salt content, g·kg -1 ; is the average value of the total soil salt content, g·kg -1 ; Q 3 is the 75% quantile of the total soil salt content data, g·kg -1 ; Q 1 is the 25% quantile of the total soil salt content data, g·kg -1 , Q3 -Q 1 is the interquartile range.

[0057] The beneficial effects of the present invention are as follows:

[0058] In the present invention, based on the improved spectral index combined with the non-linear index model, the model with the highest accuracy is selected to invert the total soil salt content, ensuring the accuracy and reliability of the inversion result;

[0059] In the present invention, the inversion of the total soil salt content based on the improved spectral index is carried out, so as to realize the quantitative analysis of soil salinization under complex environmental conditions, which is of great significance to precision agriculture and ecological environment governance. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to make the technical solutions of the embodiments of the present invention clearer, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0061] Figure 1 is a flowchart of a system and method for inverting the total soil salt content based on an improved spectral index provided by the present invention;

[0062] Figure 2 is a characteristic diagram of the spectral reflectance of different saline soils provided by the present invention;

[0063] Figure 3 is a correlation coefficient diagram of spectral reflectance and total soil salt content provided by the present invention;

[0064] Figure 4 is a correlation diagram of the improved spectral index and the total soil salt content provided by the present invention;

[0065] Figure 5 is a precision diagram of the verification set of the total soil salt content inversion model provided by the present invention;

[0066] Figure 6 is a spatial distribution diagram of the total soil salt content provided by the present invention; DETAILED DESCRIPTION OF THE EMBODIMENTS

[0067] In order to make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0068] Example 1 Refer toFigure 1 As shown in Figure 1 , a soil total salt content inversion system and method based on an improved spectral index includes the following steps:

[0069] Step 1: Obtain remotely sensed images preprocessed by Sentinel-2;

[0070] Select L2A data of Sentinel-2 that has been radiometrically corrected and atmospherically corrected;

[0071] Select images in October during the period of bare ground surface, with the image time close to the actual soil sampling time and the cloud cover of the image less than 10%;

[0072] Import the downloaded L2A data into the SNAP software for resampling, with the sampling resolution mainly in the optical band (10m).

[0073] Step 2: Obtain samples at soil sampling points and measure the total salt content of the soil;

[0074] At the soil sampling points, soil samples are taken using a soil drill, with the sampling depth of 0 - 20 cm. The collected soil samples are air-dried and passed through a 1mm sieve, then prepared into a soil extract with a soil-water ratio of 1:5, and the conductivity is measured using a conductivity meter and converted into the total salt content of the soil.

[0075] Step 3: Analyze the spectral reflectance characteristics of different salinized soils;

[0076] Soil sampling points are divided into five categories according to the total salt content of the soil: non-saline soil (<1g·kg -1 ), slightly saline soil (1 - 2g·kg -1 ), moderately saline soil (2 - 4g·kg -1 ), severely saline soil (4 - 6g·kg -1 ), and saline soil (>6g·kg -1 );

[0077] Import the obtained remotely sensed images into ENVI to extract the spectral reflectance of soil sampling points;

[0078] Classify the extracted spectral reflectance according to the degree of soil salinization and take the average value;

[0079] After Savitzky-Golay smoothing processing, draw the spectral curve of remote sensing reflectance, and the results are as shown in Figure 2 .

[0080] Step 4: Analyze the correlation between spectral reflectance and soil total salt content;

[0081] Conduct a correlation analysis on the total salt content and spectral reflectance of the 0 - 20 cm soil layer at soil sampling points;

[0082] The B1 band is mainly used to monitor the nearshore water body and aerosols in the atmosphere. Therefore, the analysis of this band is ignored. The B2 (blue), B3 (green), B4 (red), B5 (red edge 1), B6 (red edge 2), B7 (red edge 3), B8 (near-infrared), B8A (narrow near-infrared), B9 (water vapor), B11 (short-wave infrared 1), and B12 (short-wave infrared 2) bands are selected for correlation analysis. The results are as Figure 3 .

[0083] Step 5: Introduce the band with the strongest correlation with the total soil salt content to improve the traditional spectral index;

[0084] The 12 selected traditional spectral indices are the Brightness Index (BI); Salinity Indices (SI, SI1, SI2, SI3, SI4, SI5); Perpendicular Dryness Index (PDI); Normalized Difference Vegetation Index (NDVI); Enhanced Vegetation Index (EVI); Difference Vegetation Index (DVI); Normalized Difference Salinity Index (NDSI);

[0085] Taking the correlation coefficient between the spectral reflectance and the total soil salt content reaching 0.6 as the evaluation criterion, the bands with a correlation coefficient greater than or equal to 0.6 are introduced into the traditional spectral index to obtain the improved spectral index, as shown in Table 1.

[0086] Table 1 Formula Table of Improved Spectral Index

[0087] Step 6: Analyze the correlation between the improved spectral index and the total soil salt content;

[0088] Use the band operation tool in ENVI to calculate the improved spectral index;

[0089] Extract the improved spectral index of the soil sampling points;

[0090] Calculate the correlation coefficient between different improved spectral indices and the total soil salt content. The results are as Figure 4 .

[0091] Step 7: Construct a non-linear index inversion model for the total soil salt content based on the improved spectral index;

[0092] The rule for constructing the inversion of soil total salt content is that the ratio of the number of training set samples to the number of validation set samples is 2:1;

[0093] Select the improved spectral index with the strongest correlation with soil total salt content and its corresponding traditional spectral index as independent variables and introduce them into the non - linear index model to construct the soil total salt content inversion model. The results are shown in Table 2.

[0094] Table 2 Expression of soil total salt content inversion model and training set accuracy

[0095] Step 8: Compare the accuracies of different models, obtain the optimal soil total salt content inversion model and draw a graph;

[0096] To evaluate the accuracy of the soil salt content inversion model constructed by the improved spectral index, the accuracy verification is carried out on the selected validation set samples, and it is concluded that the accuracy of the soil total salt content inversion model constructed based on the improved spectral index has been improved well. The results are as Figure 5 ;

[0097] Draw the soil total salt content distribution map based on the soil total salt content inversion model constructed with the optimal improved spectral index. The results are as Figure 6 .

[0098] The present invention proposes a soil total salt content inversion system and method based on an improved spectral index. First, the acquisition and pre - processing of remote sensing images are carried out, and the soil total salt content of soil samples obtained from field experiments is measured. Then, the important relationship between spectral reflectance and soil total salt content is analyzed, and the band with the strongest correlation with soil total salt content is selected to improve the traditional spectral index. After that, the band operation tool of ENVI software is used to calculate the traditional spectral index and the improved spectral index. The calculated traditional spectral index and improved spectral index are used as independent variables to introduce into the non - linear index model to construct the soil total salt content inversion model. Finally, various parameters are used to evaluate the accuracy of various models, confirm the optimal soil total salt content inversion model and draw the spatial distribution map of soil total salt content in the area. It solves the problem that the accuracy of the soil total salt content inversion method based on spectral index is too low to meet the actual application requirements, and provides decision - making support for the dynamic monitoring of land salinization and the sustainable utilization of land resources.

[0099] In this article, specific examples are used to elaborate on the principle and implementation method of the present invention. Each embodiment is described in a progressive manner, and the key point of each embodiment is to illustrate the differences from other embodiments. The descriptions of the above embodiments are only used to help understand the method of the present invention and its core idea. For any person skilled in the art within the technical scope disclosed by the present invention, any equivalent replacement or change made according to the technical solution and inventive concept of the present invention should be covered within the protection scope of the present invention.

Claims

1. A soil total salt inversion system and method based on improved spectral index, characterized in that: The following steps are involved: S1, obtain remote sensing images preprocessed by Sentinel-2; S2, obtaining samples from soil sampling points and determining the total salt content of the soil; S3. Analyze the spectral reflectance characteristics of different salinized soils; S4. Analyze the correlation between spectral reflectance and soil total salt content; S5, introduce the band with the strongest correlation with soil total salt content to improve the traditional spectral index; S6. Analyze the correlation between the improved spectral index and the total salt content of the soil; S7. Construct a nonlinear index inversion model of soil total salt content based on the improved spectral index; S8. Compare the accuracy of different models, obtain the optimal soil total salt inversion model and draw a graph.

2. According to claim 1, a soil total salt inversion system and method based on improved spectral index is characterized in that: In S1, the specific steps of obtaining the remote sensing image for preprocessing are as follows: S1.1 Image download: Select Sentinel-2 L2A data that has been corrected for radiation and atmosphere, and obtain it from the European Space Agency Copernicus Open Access Center; S1.2, Image selection rules: Select images from the period of bare surface, as close as possible to the actual soil sampling time and with cloud cover less than 10%; S1.

3. Image resampling: Import the downloaded L2A-level data into the SNAP software for resampling. The sampling resolution is mainly based on the optical band (10m).

3. The soil total salt inversion system and method based on improved spectral index according to claim 1 is characterized in that: In S2, the specific steps for determining the total salt content of the soil are as follows: S2.

1. Soil sampling points: Soil samples were collected using a soil drill. The soil samples were air-dried and passed through a 1 mm sieve to prepare a soil extract with a soil-water ratio of 1:

5. The conductivity was then measured using a conductivity meter. S2.

2. The calculation formula for converting soil conductivity into total soil salt content is: <h2 style=";text-align:left;direction:ltr">C<h2 style=";text-align:left;direction:ltr"> s <h2 style=";text-align:left;direction:ltr"> =3.7657EC<h2 style=";text-align:left;direction:ltr"> 1：5 <h2 style=";text-align:left;direction:ltr"> -0.2405 In the formula, C s is the total salt content of soil, g / kg; EC 1：5 It is the electrical conductivity of soil extract with a soil-water ratio of 1:5, dS / m.

4. The soil total salt inversion system and method based on improved spectral index according to claim 1 is characterized in that: In S5, the specific calculation steps of the traditional spectral index are as follows: S5.1, the 12 traditional spectral indices selected are Brightness Index (BI); Salinity Indices (SI, SI1, SI2, SI3, SI4, SI5); Perpendicular Dryness Index (PDI); Normalized Difference Vegetation Index (NDVI); Enhanced Vegetation Index (EVI); Difference Vegetation Index (DVI); Normalized Difference Salinity Index (NDSI); S5.

2. The traditional spectral index calculation formula is as follows: Wherein, B is the reflectivity of the blue band of B2; G is the reflectivity of the green band of B3; R is the reflectivity of the red band of B4; NIR is the reflectivity of the near-infrared band of B8.

5. The soil total salt inversion system and method based on improved spectral index according to claim 1 is characterized in that: In S5, the improved principle formula is as follows: GIndex=Index+C GIndex=Index×C Where Index is the traditional spectral index; GIndex is the improved spectral index; C is the band with the strongest correlation with the total salt content of the soil.

6. The soil total salt inversion system and method based on improved spectral index according to claim 1 is characterized in that: In S7, the specific steps of constructing the soil total salt content nonlinear index inversion model based on the improved spectral index are as follows: S7.

1. Select the improved spectral index with the strongest correlation with the total salt content of soil and its corresponding traditional spectral index as independent variables to introduce the nonlinear index model to construct the inversion model of the total salt content of soil; S7.

2. The calculation formula of the nonlinear exponential model is as follows: C s =Ae Bx In the formula, C s is the total salt content of the soil, g·kg -1 ; A and B are unknown coefficients; x is the traditional spectral index or improved spectral index.

7. The soil total salt inversion system and method based on improved spectral index according to claim 1 is characterized in that: In S8, the determination coefficient (R 2 ), mean absolute error (MAE), root mean square error (RMSE) and interquartile performance ratio (RPIQ). The specific calculation formula is: Where N is the number of observation points; X i is the measured value of total soil salt content, g·kg -1 ; Y i is the predicted value of total soil salt content, g·kg -1 ; is the average value of total soil salt content, g·kg -1 ; Q3 is the quantile of 75% of the total soil salt content data, g·kg -1 ; Q1 is the 25% quantile of the total soil salt content data, g·kg -1 , Q3-Q1 is the interquartile range.

Images