Method for estimating underground water burial depth according to surface vegetation

Abstract

The invention relates to a method for estimating underground water burial depth according to earth surface vegetation. The method comprises the following steps: acquiring earth surface data; wherein the earth surface data comprises vegetation data, underground water data and environment data; extracting feature data of the earth surface data; training a long-short-term memory network model based on the feature data, and obtaining an underground water burial depth estimation model; and inputting surface data of a to-be-measured area into the underground water burial depth estimation model to obtain an underground water burial depth estimation result. The method provided by the invention not only improves the accuracy of groundwater burial depth estimation, but also realizes a rapid and automatic estimation process, and is of great significance to water resource management and environment monitoring.

Description

Technical Field

[0001] The invention relates to the technical field of hydrogeology, and in particular to a method for estimating the depth of groundwater based on surface vegetation. Background Art

[0002] As an important freshwater resource, groundwater is of great significance to maintaining ecological balance and human production and life. However, traditional groundwater depth monitoring methods, such as groundwater well measurement, borehole water level measurement, ground depression monitoring, etc., require a lot of manpower and material resources, especially in large areas or complex terrain areas, the monitoring cost increases significantly. Although some automated monitoring equipment such as float-type water level gauges can reduce the manpower burden, the purchase, installation and maintenance costs of the equipment are also relatively high. Manual observations need to be carried out on site regularly, and the data update speed is slow, which is difficult to meet the needs of real-time or high-frequency monitoring. Although some automated monitoring equipment can transmit data in real time, it is limited by factors such as equipment performance and communication conditions, and the data acquisition efficiency may be affected. The monitoring accuracy of groundwater levels is affected by a variety of environmental factors, such as geological conditions, hydrological conditions, climatic conditions, etc., which may lead to deviations in monitoring results.

[0003] In summary, traditional groundwater depth monitoring methods have problems such as high cost, low efficiency, and limited coverage, which makes it difficult to meet the needs of large-scale, real-time monitoring. With the rapid development of remote sensing technology, surface vegetation remote sensing images have become one of the important means to monitor the dynamic changes of groundwater. However, how to extract effective information from complex remote sensing images and accurately estimate the groundwater depth is still a problem that needs to be solved. Summary of the invention

[0004] The purpose of the present invention is to provide a method for estimating groundwater depth based on surface vegetation, so as to solve the problems existing in the above-mentioned prior art, utilize deep learning algorithm to perform automatic feature extraction and pattern recognition on surface vegetation remote sensing images, establish a nonlinear mapping relationship between surface vegetation and groundwater depth, and realize rapid and accurate estimation of groundwater depth.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A method for estimating groundwater depth based on surface vegetation, comprising:

[0007] Collecting surface data; wherein the surface data includes: vegetation data, groundwater data and environmental data;

[0008] Extracting characteristic data of the surface data;

[0009] Based on the characteristic data, a long short-term memory network model is trained to obtain a groundwater depth estimation model;

[0010] The surface data of the area to be measured is input into the groundwater depth estimation model to obtain the groundwater depth estimation result.

[0011] Optionally, the vegetation data includes: vegetation type, distribution and growth status data;

[0012] The groundwater data include: historical water level records and dynamic change trend data;

[0013] The environmental data include: environmental factor data corresponding to precipitation, evaporation, soil type, and topography.

[0014] Optionally, before extracting the characteristic data of the surface data, the method further includes: preprocessing the surface data;

[0015] Preprocessing the surface data includes:

[0016] Cleaning and validating the surface data;

[0017] The verified surface data is normalized.

[0018] Optionally, extracting characteristic data of the surface data includes:

[0019] The preprocessed surface data is input into a pre-trained convolutional neural network model to obtain vegetation coverage, vegetation type, vegetation growth status, vegetation moisture content and spectral characteristics.

[0020] Optionally, training a long short-term memory network model based on the feature data includes:

[0021] Performing time series processing on the characteristic data to obtain time series data;

[0022] Based on the time series data, a data set is constructed; wherein the data set includes: a training set, a validation set and a test set;

[0023] Using the training set to train the long short-term memory network model, and adjusting the model parameters through a back propagation algorithm;

[0024] Using the validation set to tune the hyperparameters of the trained model;

[0025] The verified model is evaluated using the test set to obtain the groundwater depth estimation model.

[0026] Optionally, in the long short-term memory network model, an attention mechanism is introduced between the LSTM layer and the output layer to enhance the model's attention to preset key information by calculating preset importance weights of input features, and residual connections are added between LSTM layers to directly add the output of the previous layer to the input of the next layer.

[0027] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements a method for estimating groundwater depth based on surface vegetation.

[0028] An electronic device includes a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, a method for estimating groundwater depth based on surface vegetation is implemented.

[0029] The beneficial effects of the present invention are:

[0030] The present invention trains a long short-term memory network model based on feature data to obtain a groundwater depth estimation model; inputs surface data of the area to be measured into the groundwater depth estimation model to obtain a groundwater depth estimation result; uses a deep learning algorithm to automatically extract features and recognize patterns in surface vegetation remote sensing images, establishes a nonlinear mapping relationship between surface vegetation and groundwater depth, and realizes rapid and accurate estimation of groundwater depth. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. 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.

[0032] Figure 1 The present invention is a flowchart of a method for estimating groundwater depth based on surface vegetation according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] 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.

[0034] 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.

[0035] like Figure 1 As shown, this embodiment proposes a method for estimating groundwater depth based on surface vegetation, including:

[0036] Collecting surface data; wherein the surface data includes: vegetation data, groundwater data and environmental data;

[0037] Extracting characteristic data of the surface data;

[0038] Based on the characteristic data, a long short-term memory network model is trained to obtain a groundwater depth estimation model;

[0039] The surface data of the area to be measured is input into the groundwater depth estimation model to obtain the groundwater depth estimation result.

[0040] This embodiment trains a long short-term memory network model based on feature data to obtain a groundwater depth estimation model; inputs surface data of the area to be measured into the groundwater depth estimation model to obtain a groundwater depth estimation result; uses a deep learning algorithm to automatically extract features and recognize patterns in surface vegetation remote sensing images, establishes a nonlinear mapping relationship between surface vegetation and groundwater depth, and realizes rapid and accurate estimation of groundwater depth.

[0041] Furthermore, the vegetation data includes: vegetation type, distribution and growth status data;

[0042] The groundwater data include: historical water level records and dynamic change trend data;

[0043] The environmental data include: environmental factor data corresponding to precipitation, evaporation, soil type, and topography.

[0044] Specifically, in this embodiment, vegetation data includes: obtaining vegetation type, distribution and growth status data in the study area through remote sensing technology (such as satellite images, drone aerial photography), extracting key indicators such as vegetation index (such as NDVI) to reflect the growth status and coverage of vegetation.

[0045] Groundwater data include: monitoring data collected from groundwater observation wells, including historical water level records and dynamic change trends. These data will serve as a benchmark for model training and verification.

[0046] Environmental data include: collecting environmental factor data such as precipitation, evaporation, soil type, topography, etc. in the study area. These data are crucial to understanding the relationship between vegetation and groundwater. Obtaining surface vegetation information through non-contact means such as remote sensing images avoids the cost of a large amount of manual observation and data collection required in traditional methods.

[0047] Furthermore, before extracting the characteristic data of the surface data, the method further includes: preprocessing the surface data;

[0048] Preprocessing the surface data includes:

[0049] Cleaning and validating the surface data;

[0050] The verified surface data is normalized.

[0051] Specifically, in this embodiment, the collected data is cleaned and verified to remove errors and outliers, and the data is normalized or standardized to ensure that data from different sources can be compared and analyzed on the same scale.

[0052] Furthermore, extracting characteristic data of the surface data includes:

[0053] The preprocessed surface data is input into a pre-trained convolutional neural network model to obtain vegetation coverage, vegetation type, vegetation growth status, vegetation moisture content and spectral characteristics.

[0054] Specifically, in this embodiment, feature extraction is to process the pre-processed surface data using a convolutional neural network model to obtain a feature vector; the specific process of feature extraction is as follows:

[0055] Input layer: The preprocessed surface data is used as the input of CNN. These surface image data include information of multiple bands, such as red, green, blue and other visible light bands, as well as near infrared, short-wave infrared and other bands, which are used to reflect the growth status and moisture content of vegetation.

[0056] Convolutional layer: The input image data is convolved through multiple convolutional layers to extract local features in the image. Each convolutional layer contains multiple convolution kernels, which slide on the image and perform dot product operations to generate new feature maps. As the convolutional layers deepen, the extracted features gradually change from low-level to high-level, and from concrete to abstract.

[0057] Activation layer: After the convolution layer, an activation layer (such as a ReLU layer) is added to increase the nonlinear ability of the model. The activation function can perform nonlinear transformations on the output of the convolution layer, allowing the model to learn more complex feature representations.

[0058] Pooling layer: In order to reduce the dimension of the feature map and reduce the amount of calculation, a pooling layer (such as a maximum pooling layer) is added between the convolutional layers. The pooling layer extracts important feature information and reduces redundancy by downsampling the feature map.

[0059] Fully connected layer: After multiple convolutional layers and pooling layers, a fully connected layer is added to integrate the extracted feature vectors. Each node in the fully connected layer is connected to all nodes in the previous layer to learn the nonlinear combination relationship between features.

[0060] The extracted feature vector is the output result of CNN after feature extraction of remote sensing images. It contains a variety of feature information of surface vegetation. These feature vectors may include but are not limited to the following: Vegetation coverage: reflects the density and distribution range of surface vegetation. Vegetation type: distinguishes different types of vegetation, such as forests, grasslands, farmlands, etc. Vegetation growth status: reflects the growth rate and health status of vegetation. Vegetation moisture content: indirectly reflects the replenishment of groundwater to vegetation. Spectral characteristics: including spectral reflectance and absorption rate of different bands, these features are of great significance for distinguishing vegetation types and growth conditions. The extracted feature vector is automatically learned based on the training data and model structure.

[0061] In this embodiment, deep learning technology is used to automatically extract the complex features of surface vegetation, combined with real data on groundwater depth, to achieve high-precision groundwater depth estimation.

[0062] Furthermore, training the long short-term memory network model based on the feature data includes:

[0063] Performing time series processing on the characteristic data to obtain time series data;

[0064] Based on the time series data, a data set is constructed; wherein the data set includes: a training set, a validation set and a test set;

[0065] Using the training set to train the long short-term memory network model, and adjusting the model parameters through a back propagation algorithm;

[0066] Using the validation set to tune the hyperparameters of the trained model;

[0067] The verified model is evaluated using the test set to obtain the groundwater depth estimation model.

[0068] Specifically, in this embodiment, the pre-processed data set is divided into a training set, a validation set, and a test set. The model is trained using the training set data, and the model parameters are adjusted by the back propagation algorithm to minimize the prediction error. The hyperparameters of the model (such as learning rate, number of hidden layer nodes, etc.) are tuned using the validation set data to improve the generalization ability of the model. The trained model is evaluated using the test set data to calculate evaluation indicators such as prediction error and accuracy. The prediction results of the model are compared with the actual observation data to verify the accuracy and reliability of the model. The input of the long short-term memory network model is an image feature sequence, so these features may represent the vegetation conditions in different areas or at different time points in the image. The long short-term memory network model will learn the relationship between vegetation conditions and groundwater depth through these feature sequences, and estimate the groundwater depth accordingly. Ultimately, the output of the model may be a summary of the entire sequence (such as the average groundwater depth of the entire area) or a separate prediction of each time step in the sequence (such as the groundwater depth at each time point or each area).

[0069] By utilizing the time series data of remote sensing images, dynamic monitoring and real-time updating of groundwater depth can be achieved, thus improving the timeliness of monitoring.

[0070] Furthermore, in the long short-term memory network model, an attention mechanism is introduced between the LSTM layer and the output layer, and the model's attention to preset key information is enhanced by calculating the preset importance weights of the input features. A residual connection is added between the LSTM layers, and the output of the previous layer is directly added to the input of the next layer.

[0071] Specifically, in the long short-term memory network model of this embodiment, the input layer: receives the original time series data; LSTM layer: Attention mechanism: introduces the attention mechanism between the LSTM layer and the output layer, and enhances the model's attention to key information by calculating the importance weights of the input features. Residual connection: Add residual connection between LSTM layers, and add the output of the previous layer directly to the input of the next layer to solve the problem of gradient disappearance or gradient explosion in deep network training. Dynamic number of layers and units: According to the characteristics and performance of the data set, dynamically adjust the number of LSTM layers and the number of units in each layer to achieve the optimal model structure. Output layer: Receive the output of the LSTM layer and convert the output of the LSTM layer into the final prediction result.

[0072] This embodiment also proposes a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method for estimating the groundwater depth based on the surface vegetation is implemented.

[0073] This embodiment also proposes an electronic device, including a processor and a memory, wherein the memory stores a computer program, and when the processor executes the computer program, the method for estimating the groundwater depth based on the surface vegetation is implemented.

[0074] This embodiment collects remote sensing image data of surface vegetation, meteorological data, soil moisture and other multivariate information, and uses them as model input after preprocessing. Through the learning and training of the deep learning model, a nonlinear mapping relationship between surface vegetation characteristics and groundwater depth is established to achieve high-precision groundwater depth prediction. The method of this embodiment not only improves the accuracy of groundwater depth estimation, but also realizes a fast and automated estimation process, which is of great significance to water resources management and environmental monitoring.

[0075] This embodiment provides accurate groundwater depth information to water resource management departments, which helps them formulate scientific and reasonable water resource management and protection strategies and promote the sustainable use of water resources.

[0076] 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 method for estimating groundwater depth based on surface vegetation, characterized in that: include: Collect surface data; The surface data include: vegetation data, groundwater data and environmental data; Extracting characteristic data of the surface data; Based on the characteristic data, a long short-term memory network model is trained to obtain a groundwater depth estimation model; The surface data of the area to be measured is input into the groundwater depth estimation model to obtain the groundwater depth estimation result.

2. The method for estimating groundwater depth based on surface vegetation according to claim 1, characterized in that: The vegetation data includes: vegetation type, distribution and growth status data; The groundwater data include: historical water level records and dynamic change trend data; The environmental data include: environmental factor data corresponding to precipitation, evaporation, soil type, and topography.

3. The method for estimating groundwater depth based on surface vegetation according to claim 1, characterized in that: Before extracting the characteristic data of the surface data, the method further includes: preprocessing the surface data; Preprocessing the surface data includes: Cleaning and validating the surface data; The verified surface data is normalized.

4. The method for estimating groundwater depth based on surface vegetation according to claim 3, characterized in that: Extracting characteristic data of the surface data includes: The preprocessed surface data is input into a pre-trained convolutional neural network model to obtain vegetation coverage, vegetation type, vegetation growth status, vegetation moisture content and spectral characteristics.

5. The method for estimating groundwater depth based on surface vegetation according to claim 1, characterized in that: Training the long short-term memory network model based on the feature data includes: Performing time series processing on the characteristic data to obtain time series data; Based on the time series data, a data set is constructed; wherein the data set includes: a training set, a validation set and a test set; Using the training set to train the long short-term memory network model, and adjusting the model parameters through a back propagation algorithm; Using the validation set to tune the hyperparameters of the trained model; The verified model is evaluated using the test set to obtain the groundwater depth estimation model.

6. The method for estimating groundwater depth based on surface vegetation according to claim 5, characterized in that: In the long short-term memory network model, an attention mechanism is introduced between the LSTM layer and the output layer, and the model's attention to preset key information is enhanced by calculating the preset importance weights of the input features. A residual connection is added between the LSTM layers, and the output of the previous layer is directly added to the input of the next layer.

7. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

8. An electronic device, characterized in that: The method comprises a processor and a memory, wherein a computer program is stored in the memory, and when the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

Images