Saline-alkali soil water and salt migration monitoring system and method
Through the hyperspectral image data decomposition and inversion model, the problem of extracting soil moisture and salt information in saline-alkali land in remote sensing technology is solved, and efficient, accurate and dynamic monitoring of water and salt migration monitoring in saline-alkali land is achieved, and saline-alkali land governance and agricultural management are supported.
Patent Information
- Application Number
- CN202411290057.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-14
- Publication Date
- 2024-11-15
AI Technical Summary
The prior art is difficult to accurately extract moisture and salt information from saline-alkali soil from remote sensing images. It is affected by the vegetation spectral characteristics and soil complexity, resulting in low monitoring accuracy and efficiency.
The soil spectral data is obtained by using hyperspectral image data decomposition technology, and an inversion model is constructed based on historical soil data, combined with independent component analysis, non-negative matrix decomposition and neural network methods to achieve accurate monitoring of soil moisture and salt.
It improves the accuracy and efficiency of water and salt migration monitoring, reduces cost and environmental interference, realizes dynamic monitoring of the water and salt conditions of the soil of saline-alkali land, and supports scientific governance decision-making and agricultural management.
Smart Images

Figure CN118961513A_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of water-salt monitoring, and in particular relates to a system and method for monitoring water-salt migration in saline-alkali land. Background Art
[0002] Water and salt migration in saline-alkali land refers to the phenomenon that water and salt migrate and redistribute in saline-alkali soil through various physical processes (such as diffusion, convection, etc.). This process is crucial for the management and improvement of saline-alkali land because it is directly related to the accumulation and distribution of soil salt, which in turn affects crop growth and land productivity. The monitoring method of water and salt migration in saline-alkali land is mainly field observation, which obtains data by directly measuring the changes in soil moisture and salt, but this method is difficult, time-consuming, labor-intensive and costly.
[0003] In the process of applying remote sensing technology to the monitoring of water and salt migration in saline-alkali land, a core technical challenge is how to accurately extract soil moisture and salt information from remote sensing images. The surface of saline-alkali land is often covered by various types of vegetation, and the growth conditions and types of these vegetation vary significantly, resulting in mixed spectral information of vegetation and soil in remote sensing images. The existence of this mixed spectrum makes the extraction of soil moisture and salt information complicated, because the spectral characteristics of vegetation may interfere with and affect the accurate identification of soil information, thereby reducing the accuracy of monitoring. In addition, different vegetation types and growth stages contribute differently to the spectrum, which further increases the complexity of spectral mixing, making it more difficult to separate soil moisture and salt information from these mixed spectra. In addition to the influence of vegetation, the complexity of saline-alkali soil itself is also a factor that cannot be ignored. Soil type, salt composition and content, soil moisture status, and soil surface crusting will all affect the characteristics of remote sensing reflectance spectra. Therefore, it is also necessary to study the spectral response characteristics of water and salt information under different saline-alkali soil backgrounds, establish targeted remote sensing interpretation models, and improve the accuracy and reliability of water and salt migration monitoring. Summary of the invention
[0004] The present invention aims to solve the deficiencies of the prior art and provides the following solutions:
[0005] A saline-alkali land water and salt migration monitoring system, comprising: a remote sensing data collection module, a model building module and a water and salt migration monitoring module;
[0006] The remote sensing data collection module is used to obtain the hyperspectral image data of the saline-alkali land to be monitored, and decompose the hyperspectral image data to obtain soil spectral data;
[0007] The model building module is used to collect historical soil data and historical soil spectral data, and build an inversion model based on the historical soil data and the historical soil spectral data;
[0008] The water-salt migration monitoring module completes water-salt migration monitoring based on the soil spectral data and the inversion model.
[0009] Preferably, the remote sensing data collection module comprises: a data collection unit, a first data processing unit and a data decomposition unit;
[0010] The data collection unit is used to collect the hyperspectral image data of the saline-alkali land to be monitored;
[0011] The first data processing unit is used to process the hyperspectral image data to obtain processed spectral data;
[0012] The data decomposition unit is used to decompose the processed spectral data to obtain soil spectral data.
[0013] Preferably, the model building module comprises: a historical data collection unit, a second data processing unit and a model building unit;
[0014] The historical data collection unit is used to collect the historical soil data and the historical soil spectrum data;
[0015] The second data processing unit is used to process the collected historical data to obtain processed data;
[0016] The model building unit builds and trains the inversion model based on the processed historical soil data and the historical soil spectrum.
[0017] Preferably, the water-salt migration monitoring module comprises: a soil data monitoring unit and a migration calculation unit;
[0018] The soil data monitoring unit is used to obtain soil moisture data and soil salinity data of the saline-alkali land to be monitored based on the soil spectral data and the inversion model;
[0019] The migration calculation unit calculates water-salt migration data based on the soil moisture data and the soil salinity data to complete water-salt migration monitoring.
[0020] The present invention also provides a method for monitoring water and salt migration in saline-alkali land, the monitoring method being applied to any of the monitoring systems described above, comprising the following steps:
[0021] Acquire hyperspectral image data of the saline-alkali land to be monitored, and decompose the hyperspectral image data to obtain soil spectral data;
[0022] Collecting historical soil data and historical soil spectral data, and constructing an inversion model based on the historical soil data and the historical soil spectral data;
[0023] Water and salt migration monitoring is completed based on the soil spectral data and the inversion model.
[0024] Preferably, the method for obtaining the soil spectral data comprises:
[0025] Collecting the hyperspectral image data of the saline-alkali land to be monitored;
[0026] Processing the hyperspectral image data to obtain processed spectral data;
[0027] The processed spectral data are decomposed to obtain soil spectral data.
[0028] Preferably, the method for constructing the inversion model comprises:
[0029] Collecting the historical soil data and the historical soil spectrum data;
[0030] Process the collected historical data to obtain processed data;
[0031] The inversion model is constructed and trained based on the processed historical soil data and the historical soil spectra.
[0032] Preferably, the method for monitoring water and salt migration includes:
[0033] Acquiring soil moisture data and soil salinity data of the saline-alkali land to be monitored based on the soil spectral data and the inversion model;
[0034] The water-salt migration data is calculated based on the soil moisture data and the soil salinity data to complete the water-salt migration monitoring.
[0035] Compared with the prior art, the present invention has the following beneficial effects:
[0036] The present invention uses remote sensing technology to obtain hyperspectral image data, and obtains soil spectral data through data decomposition. Compared with the traditional sampling and analysis method, this process greatly improves the efficiency and coverage of data acquisition, while reducing costs and interference with the environment; in terms of model construction, the system collects historical soil data and spectral data to build an inversion model, which can more accurately predict and simulate the water-salt migration process. The establishment of this model not only improves the accuracy of prediction, but also can better understand the long-term trend and dynamic changes of soil water-salt migration through the accumulation and analysis of historical data, and provides more scientific decision-making support for the management and improvement of saline-alkali land; in addition, the water-salt migration monitoring module of the system can monitor soil moisture and salt data in real time, calculate water-salt migration data through the migration calculation unit, and realize dynamic monitoring of water-salt conditions in saline-alkali land. This real-time monitoring capability enables management departments to adjust agricultural measures such as irrigation and drainage in a timely manner, effectively control the process of soil salinization, and improve land utilization and agricultural production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0037] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0038] Figure 1 Schematic diagram of the system structure of an embodiment of the present invention. DETAILED DESCRIPTION
[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0040] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.
[0041] Embodiment 1
[0042] In this embodiment, if Figure 1 As shown, a saline-alkali land water and salt migration monitoring system includes: a remote sensing data collection module, a model building module and a water and salt migration monitoring module.
[0043] The remote sensing data collection module is used to obtain the hyperspectral image data of the saline-alkali land to be monitored, and decompose the hyperspectral image data to obtain soil spectral data.
[0044] The remote sensing data collection module includes: a data collection unit, a first data processing unit and a data decomposition unit. The data collection unit is used to collect the hyperspectral image data of the saline-alkali land to be monitored; the first data processing unit is used to process the hyperspectral image data to obtain processed spectral data; the data decomposition unit is used to decompose the processed spectral data to obtain soil spectral data.
[0045] In the present embodiment, the data collection unit obtains the hyperspectral image data of the saline-alkali area to be monitored through a satellite remote sensing platform, and sets the spatial resolution to 30 meters, the spectral resolution to 10 nanometers, and the band range to 400-2500 nanometers. The first data processing unit first uses an atmospheric correction model to perform atmospheric correction on the obtained hyperspectral image data, and then uses wavelet transform to denoise the corrected data to obtain processed spectral data. The data decomposition unit first uses an ICA (independent component analysis) algorithm to perform a preliminary decomposition of the processed spectral data, specifically: find a separation matrix W so that the spectral data after the matrix transformation is as statistically independent as possible. The objective function of ICA can be expressed as maximizing non-Gaussianity, and mutual information is usually used as a metric. In the present embodiment, the objective function of ICA is:
[0046]
[0047] Where W represents the separation matrix, n represents the number of independent components in the processed spectral data, E represents the expected value operation, and W i is the i-th row of the separation matrix, T represents the transpose operation of the matrix, represents the direction of the i-th independent component, x represents the processed spectral data, G represents the nonlinear function, and v represents the random variable of Gaussian distribution; then the FastICA (fast independent component analysis) algorithm is used for iterative optimization to obtain the initial end member spectrum and the corresponding abundance map. Specifically: the FastICA algorithm finds the independent components of the data by iteratively updating the weight vector w, and the update rule is:
[0048]
[0049] Among them, w new Represents the updated weight vector, w oldRepresents the weight vector before updating, α represents the learning rate, g represents the derivative of the nonlinear function G. By iterating this process, w is gradually optimized until it converges to independent components, and the initial end-member spectrum and the corresponding abundance map are obtained; then the NMF (non-negative matrix factorization) algorithm is used to optimize the initial end-member spectrum and abundance map, and non-negative constraints are set. The objective function is to minimize the spectral reconstruction error. The multiplication update rule is used to iteratively solve the problem. When the reconstruction error is less than 01, it is judged to have converged, and the optimized spectral end-member matrix and abundance matrix, i.e., soil spectral data, are obtained.
[0050] The model building module is used to collect historical soil data and historical soil spectral data, and to build an inversion model based on the historical soil data and historical soil spectral data.
[0051] The model building module includes: a historical data collection unit, a second data processing unit and a model building unit. The historical data collection unit is used to collect historical soil data and historical soil spectrum data; the second data processing unit is used to process the collected historical data to obtain processed data; the model building unit builds and trains the inversion model based on the processed historical soil data and historical soil spectrum.
[0052] In this embodiment, the historical soil data collected by the historical data collection unit are the historical soil moisture data and historical soil salinity data of the saline-alkali land to be monitored. The second data processing unit is used to identify abnormal values in the historical data that do not meet the expected range and delete the abnormal values. The abnormal values may be caused by sensor errors, transmission errors or environmental factors, and then check the missing values in the historical data and fill them with the average value of the data near the missing values to complete the data cleaning; then use the principal component analysis method to extract features from the cleaned data set, and by calculating the eigenvalues and eigenvectors, select the first 5 principal components as key spectral features to obtain processed data, that is, processed historical soil data. The model building module is used to build and train the inversion model. Specifically, in the model training stage, three models, support vector machine, random forest and BP neural network, are constructed respectively. The three models are trained using the historical soil data and historical soil spectral data of the processed data, wherein: the support vector machine adopts the radial basis kernel function, the penalty factor C=10, the random forest adopts 500 decision trees, the neural network adopts a 3-layer hidden layer structure, the learning rate is 01, and the number of iterations is 1000, to generate three preliminary estimation models; then the accuracy of the three preliminary models is evaluated through 5-fold cross validation, and the average accuracy of the support vector machine, random forest and neural network is 85, 88 and 90 respectively; then, according to the verification results, the neural network model with the highest average accuracy is selected for optimization, and the parameters such as the number of hidden layer nodes and the learning rate are adjusted through the grid search algorithm, and finally the number of hidden layer nodes is determined to be 64, the learning rate is 005, and the accuracy of the optimized neural network model on the validation set is improved to 92, and the root mean square error is reduced to 08, and the final inversion model is obtained.
[0053] The water and salt migration monitoring module completes water and salt migration monitoring based on soil spectral data and inversion models.
[0054] The water-salt migration monitoring module includes: a soil data monitoring unit and a migration calculation unit. The soil data monitoring unit is used to obtain soil moisture data and soil salinity data of the saline-alkali land to be monitored based on soil spectral data and an inversion model. In this embodiment, the collected soil spectral data is input into the inversion model to obtain soil moisture data and soil salinity data of the saline-alkali land to be monitored; the migration calculation unit calculates water-salt migration data based on soil moisture data and soil salinity data to complete water-salt migration monitoring.
[0055] Embodiment 2
[0056] In this embodiment, a method for monitoring water and salt migration in saline-alkali land comprises the following steps:
[0057] S1. Obtain hyperspectral image data of the saline-alkali land to be monitored, and decompose the hyperspectral image data to obtain soil spectral data.
[0058] The method for obtaining soil spectral data includes: collecting hyperspectral image data of saline-alkali land to be monitored; processing the hyperspectral image data to obtain processed spectral data; and decomposing the processed spectral data to obtain soil spectral data.
[0059] S2. Collect historical soil data and historical soil spectral data, and construct an inversion model based on the historical soil data and historical soil spectral data.
[0060] The method for constructing an inversion model includes: collecting historical soil data and historical soil spectral data; processing the collected historical data to obtain processed data; and constructing and training an inversion model based on the processed historical soil data and historical soil spectra.
[0061] S3. Complete water and salt migration monitoring based on soil spectral data and inversion model.
[0062] The method for monitoring water and salt migration includes: obtaining soil moisture data and soil salinity data of the saline-alkali land to be monitored based on soil spectral data and an inversion model; calculating water and salt migration data based on the soil moisture data and soil salinity data to complete water and salt migration monitoring.
[0063] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.
Claims
1. A saline-alkali land water and salt migration monitoring system, characterized in that: include: Remote sensing data collection module, model building module and water and salt transport monitoring module; The remote sensing data collection module is used to obtain the hyperspectral image data of the saline-alkali land to be monitored, and decompose the hyperspectral image data to obtain soil spectral data; The model building module is used to collect historical soil data and historical soil spectral data, and build an inversion model based on the historical soil data and the historical soil spectral data; The water-salt migration monitoring module completes water-salt migration monitoring based on the soil spectral data and the inversion model.
2. A saline-alkali land water and salt migration monitoring system according to claim 1, characterized in that: The remote sensing data collection module includes: a data collection unit, a first data processing unit and a data decomposition unit; The data collection unit is used to collect the hyperspectral image data of the saline-alkali land to be monitored; The first data processing unit is used to process the hyperspectral image data to obtain processed spectral data; The data decomposition unit is used to decompose the processed spectral data to obtain soil spectral data.
3. A saline-alkali land water and salt migration monitoring system according to claim 1, characterized in that: The model building module includes: a historical data collection unit, a second data processing unit and a model building unit; The historical data collection unit is used to collect the historical soil data and the historical soil spectrum data; The second data processing unit is used to process the collected historical data to obtain processed data; The model building unit builds and trains the inversion model based on the processed historical soil data and the historical soil spectrum.
4. A saline-alkali land water and salt migration monitoring system according to claim 1, characterized in that: The water and salt migration monitoring module includes: a soil data monitoring unit and a migration calculation unit; The soil data monitoring unit is used to obtain soil moisture data and soil salinity data of the saline-alkali land to be monitored based on the soil spectral data and the inversion model; The migration calculation unit calculates water-salt migration data based on the soil moisture data and the soil salinity data to complete water-salt migration monitoring.
5. A method for monitoring water and salt migration in saline-alkali land, the monitoring method being applied to the monitoring system according to any one of claims 1 to 4, characterized in that: The following steps are involved: Acquire hyperspectral image data of the saline-alkali land to be monitored, and decompose the hyperspectral image data to obtain soil spectral data; Collecting historical soil data and historical soil spectral data, and constructing an inversion model based on the historical soil data and the historical soil spectral data; Water and salt migration monitoring is completed based on the soil spectral data and the inversion model.
6. A method for monitoring water and salt migration in saline-alkali land according to claim 5, characterized in that: The method for obtaining the soil spectral data comprises: Collecting the hyperspectral image data of the saline-alkali land to be monitored; Processing the hyperspectral image data to obtain processed spectral data; The processed spectral data are decomposed to obtain soil spectral data.
7. The method for monitoring water and salt migration in saline-alkali land according to claim 5, characterized in that: The method for constructing the inversion model comprises: Collecting the historical soil data and the historical soil spectrum data; Process the collected historical data to obtain processed data; The inversion model is constructed and trained based on the processed historical soil data and the historical soil spectra.
8. The method for monitoring water and salt migration in saline-alkali land according to claim 5, characterized in that: The method for monitoring water and salt migration comprises: Acquiring soil moisture data and soil salinity data of the saline-alkali land to be monitored based on the soil spectral data and the inversion model; The water-salt migration data is calculated based on the soil moisture data and the soil salinity data to complete the water-salt migration monitoring.
Citation Information
Cited By
Soil salinity and alkalinity evaluation method based on quaternary geological feature analysis
CN119918813A
Saline-alkali soil water-salt motion simulation method based on big data analysis
CN120995943A
Simulation method for soil water and salt movement in saline-alkali soil based on big data analysis
CN120995943B
