Data processing method for long-term dynamic monitoring of water body

Abstract

The invention provides a data processing method for long-term dynamic monitoring of a water body, and relates to the field of computer data processing and analysis. Comprising the following steps: acquiring and standardizing sensor data to obtain standardized sensor data; performing feature extraction on the standardized sensor data to obtain water body feature data; modeling the dynamic evolution process of the water body, and performing high-dimensional orthogonal projection transformation on the water body characteristic data to obtain an optimized water body characteristic matrix; designing and solving a Wasserstein variational loss function based on the optimized water body characteristic matrix to obtain an optimal water body characteristic matrix; and performing water body change prediction based on the optimal water body characteristic matrix, and minimizing a prediction loss function to obtain an optimized water body change prediction state matrix. The problems that the long-term change trend of the water body is difficult to capture, the complex space-time dependence of the water body change cannot be effectively modeled, and the accuracy requirements of long-term water resource management and ecological monitoring cannot be met are solved.

Description

Technical Field

[0001] The present invention relates to the field of computer data processing and analysis, and in particular to a data processing method for long-term water body dynamic monitoring. Background Art

[0002] With global climate change and intensified human activities, the distribution and dynamics of surface water bodies are becoming increasingly complex, making long-term monitoring a crucial foundation for sustainable water resource utilization. However, as dynamic natural entities, water bodies' area, boundaries, and morphology are influenced by multiple factors, including precipitation, evaporation, groundwater recharge, snowmelt, and human regulation. This results in complex temporal and spatial patterns of water variation.

[0003] Furthermore, sensor data acquisition is subject to limitations in sensor performance and environmental interference. Sensor data at different time steps may exhibit deviations, impacting data consistency and, in turn, the accuracy of dynamic water monitoring. Therefore, a computational method is needed that can integrate time series information, improve data processing accuracy, reduce noise interference, and stably track long-term water trends to ensure scientific and reliable water monitoring. Summary of the Invention

[0004] The present invention provides a data processing method for long-term dynamic monitoring of water bodies to solve the problem that existing technologies are difficult to accurately capture the long-term change trends of water bodies. It also solves the problem that traditional water body monitoring methods are mostly based on simple statistical models or traditional regression methods, which cannot effectively model the complex spatiotemporal dependencies of water body changes, resulting in low prediction accuracy and difficulty in achieving long-term stable monitoring; at the same time, it also solves the defect that traditional water body monitoring methods are prone to computational instability when the data dimension is high, which affects the robustness and reliability of the monitoring results and cannot meet the accuracy requirements of long-term water resources management and ecological monitoring.

[0005] The present invention provides a data processing method for long-term water body dynamic monitoring, which specifically includes the following technical solutions: A data processing method for long-term water body dynamic monitoring includes the following steps: S1. Acquire and standardize sensor data to obtain standardized sensor data; perform feature extraction on the standardized sensor data to obtain water body feature data; model the dynamic evolution process of the water body, perform high-dimensional orthogonal projection transformation on the water body feature data, and obtain an optimized water body feature matrix; S2. Based on the optimized water body characteristic matrix, design and solve the Wasserstein variational loss function to obtain the optimal water body characteristic matrix; based on the optimal water body characteristic matrix, predict water body changes and minimize the prediction loss function to obtain the optimized water body change prediction state matrix.

[0006] Preferably, the S1 specifically includes: A time recursive model is established using a spatiotemporal multi-layer recursive variational optimization network to extract features from standardized sensor data and obtain water body characteristic data.

[0007] Preferably, the S1 specifically includes: In the process of establishing the time recursive model, the water body characteristic data is adaptively adjusted through the variational optimization mechanism, and the spatiotemporal recursive calculation framework is introduced. The water body state at the current moment and the water body characteristic data of the past time steps are combined to construct the dynamic evolution process of the water body and obtain the water body characteristic state matrix.

[0008] Preferably, the S1 specifically includes: Based on the water body characteristic state matrix, an orthogonal projection matrix is ​​constructed to project the water body characteristic data so that the water body characteristic data is retained after projection. The initial water body characteristic matrix is ​​set, and the regularization parameter is introduced to perform high-dimensional orthogonal projection transformation on the water body characteristic data.

[0009] Preferably, the S1 specifically includes: The standardized sensor data matrix is ​​optimized through the orthogonal projection matrix to obtain the optimized water body characteristic matrix.

[0010] Preferably, the S2 specifically includes: Water body changes are predicted based on the optimal water body characteristic matrix to predict the future water body state; the water body characteristic data of the current time step are combined with the water body characteristic data of the past time step and the historical optimal water body characteristic matrix, and the orthogonal projection matrix is ​​added to calculate the water body change prediction state matrix.

[0011] Preferably, the S2 specifically includes: Based on the water body change prediction state matrix, a prediction loss function is introduced to compare the water body change prediction state matrix with the actual water body state matrix, and the prediction error is calculated and minimized.

[0012] Preferably, the S2 specifically includes: The prediction loss function is minimized to obtain the optimized water body change prediction state matrix, which is used as the prediction result of long-term water body dynamic monitoring and provides the water body state distribution at future moments.

[0013] The beneficial effects of the technical solution of the present invention are: 1. The time recursion model is established by using the space-time multi-layer recursive variational optimization network, high-precision extraction of water body feature data is realized, the accuracy of water body classification is improved by introducing time dependence, the misclassification caused by short-term environmental changes is reduced, the continuity of water body features is enhanced, and the stability of water body dynamic monitoring is ensured.

[0014] 2. In the water body classification stage, the water body feature optimization method based on high-dimensional orthogonal projection is proposed, the orthogonal projection matrix is constructed, the water body feature data is projected into a new feature space, the influence of background noise on water body classification is reduced, the robustness of water body classification is improved, and the error caused by water body boundary blur is reduced.

[0015] 3. After water body classification is completed, the Wasserstein variational loss function constraint method is introduced to optimize the distribution of water body classification, so that the consistency of water body classification results at different time steps in the mathematical space is ensured, the water body classification deviation caused by sensor performance, environmental interference and other factors is reduced, and the stability and interpretability of water body classification are improved.

[0016] 4. In the aspect of long-term water body dynamic change prediction, the orthogonal projection matrix is introduced to constrain the change range of water body features, so that the water body change prediction state matrix is more consistent with the actual water body change trend, and the accuracy of water body change prediction is effectively improved. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 A flow chart of the data processing method for long-term water body dynamic monitoring is described. DETAILED DESCRIPTION

[0018] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined object, the technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present application belongs.

[0020] The specific scheme of the data processing method for long-term water body dynamic monitoring provided by the present application will be described in detail below with reference to the drawings.

[0021] Referring to the drawings Figure 1, which shows a flow chart of a data processing method for long-term water body dynamic monitoring provided by an embodiment of the present invention, the method comprising the following steps: S1. Acquire and standardize sensor data to obtain standardized sensor data; perform feature extraction on the standardized sensor data to obtain water body feature data; model the dynamic evolution process of the water body, perform high-dimensional orthogonal projection transformation on the water body feature data, and obtain an optimized water body feature matrix; Through long-term monitoring of water bodies in the study area using hydrological sensors, current meters, and GNSS receivers, sensor data for the study area's water bodies is acquired. This sensor data includes water depth, water velocity, water quality parameters (such as dissolved oxygen, turbidity, and pH), precipitation, and evaporation. During the sensor data acquisition process, existing data cleaning algorithms are applied to remove abnormal data to ensure input data quality. All sensor data undergoes calibration and regularization to ensure temporal alignment and consistency of sensor data at different time steps, ensuring data comparability during subsequent water body dynamic monitoring.

[0022] The acquisition of sensor data is affected by multiple factors, including sensor noise, environmental interference, differences in sampling time, and changes in environmental conditions. This can lead to significant deviations in sensor data acquired at different times for the same water body. Therefore, it is necessary to standardize the sensor data to ensure that sensor data at different time steps have consistent numerical ranges and statistical characteristics, thereby preventing subsequent water body classification and prediction accuracy from being affected by differences in the numerical distribution of sensor data.

[0023] A time-recursive model is established using a spatiotemporal multi-layer recursive variational optimization network to extract features from standardized sensor data to obtain water body characteristic data. Adaptive adjustment of water body characteristic data is achieved through a variational optimization mechanism to capture the long-term trends and short-term dynamics of water body changes and construct a complete spatiotemporal evolution trajectory of the water body. By introducing a spatiotemporal recursive computational framework, the current state of the water body not only depends on the current sensor data, but also combines water body characteristic data from multiple time steps in the past, effectively modeling the dynamic evolution process of the water body and improving the accuracy and stability of classification. Construct a recursive equation for the characteristic state of the water body:

[0024] in, yes Water body characteristic state matrix at the moment; is the hyperbolic tangent activation function; is the state recursive weight matrix used to calculate the time dependence of water body characteristics, which is learned by the existing gradient descent method and initialized with Xavier; yes The water body feature state matrix at time t is denoted as is the input data weight matrix, which is used to map the contribution of the input normalized sensor data to the current water body feature state, and is learned by the existing gradient descent method. is the normalized sensor data matrix at time t is denoted as is the bias term, which is used to adjust the calculation result, and is learned by the existing gradient descent method, with a value range of ; is the recursive decay factor, which is used to control the influence weight of historical sensor data, and is obtained by the existing gradient descent method, with a value range of ; is the historical time step, representing the number of past time steps. is the historical weight matrix at the previous time steps. is the normalized sensor data matrix at time t is denoted as is the two-norm. is the numerical stability term to prevent numerical overflow in calculation, which can be taken as ; is the activation function. is the weight parameter of the activation function calculation, which is learned by the existing gradient descent method.

[0025] The above water body feature state recursive equation is an improved scheme based on the existing technology, which organically integrates the following three key technologies: state update , historical regularization term (VAE / VRNN-like) , gating mechanism (GRU / LSTM-like) . This integration significantly improves the accuracy and robustness of water body feature state modeling.

[0026] By calculating the water body feature state matrix at the current time step, while considering the normalized sensor data at the previous time steps, a time recursive model is established to extract features from the normalized sensor data, obtain the water body feature state matrix, and thus obtain the water body feature data, enhancing the time sequence information of water body classification.

[0027] In order to further optimize the water body classification boundary, the water body feature data needs to be subjected to high-dimensional orthogonal projection transformation, so as to reduce the interference of non-water body features and improve the accuracy of water body classification; an orthogonal projection matrix is constructed, the water body feature data is projected into a new feature space, and the water body feature data is maximally preserved after projection, while the non-water body feature data is effectively suppressed, so as to improve the separability of the water body class. The non-water body feature data is noise data in the water body feature data. An initial water body feature matrix is set , which contains the sampling point information of all sensor data processed by water body classification. In order to ensure the stability of the transformed water body feature data, orthogonal projection calculation is required for the water body feature data, so that the water body feature data remains independent in the projected space while reducing the similarity with the non-water body feature data sampling points. The calculation of the orthogonal projection matrix is based on the classical orthogonal complementary projection formula , with a regularization term , and the implementation form is as follows:

[0028] wherein, is the orthogonal projection matrix; is the unit matrix; is a regularization parameter for preventing the singularity problem of the orthogonal projection matrix, and is set to ; is the initial water body feature matrix, which is a static water body feature matrix obtained by standardizing and extracting features from the original sensor data, and is used to calculate the distribution characteristics of the sampling points of the sensor data processed by water body classification. The feature extraction can use the existing principal component analysis method; is the adjustment coefficient of the water body feature state matrix, which is obtained by fitting calculation, and the value range is ; represents the transpose of a matrix; is the square of the two-norm.

[0029] By projecting the water body feature data into an orthogonal subspace, the background information in the sampling points of the sensor data processed by water body classification is removed, and the robustness of water body classification is improved. Subsequently, the standardized sensor data matrix is optimized by the orthogonal projection matrix to obtain the optimized water body feature matrix :

[0030] By optimizing water body classification through orthogonal projection transformation, the water body feature matrix after orthogonal projection transformation is more separable, the spectral confusion between water body and non-water body feature data sampling points is reduced, and the classification accuracy is improved.

[0031] S2, design and solve the Wasserstein variation loss function based on the optimized water body feature matrix to obtain the optimal water body feature matrix; based on the optimal water body feature matrix, water body change prediction is carried out, and the prediction loss function is minimized to obtain the optimized water body change prediction state matrix.

[0032] Based on the optimized water body feature matrix, the Wasserstein variation loss function is designed to ensure the stability of water body classification and reduce the classification deviation between different time steps; due to the influence of factors such as sensor performance and environmental changes in the collection process of sensor data at different time steps, even the water body in the same area may have a large fluctuation in its characteristics; therefore, the Wasserstein variation loss function is used to constrain the distribution of water body classification, so that the water body classification results at different time steps remain consistent in mathematical space, thereby improving the stability of classification. The Wasserstein variation loss function is defined as:

[0033] wherein, is the Wasserstein variation loss function; is the Frobenius norm; is the Wasserstein distance regularization factor, which is obtained by expert experience method, and the value range is ; is the target water body label data at the moment , which is obtained by manual labeling of historical sensor data.

[0034] The above Wasserstein variation loss function is the combination of the square of the Frobenius norm and the Wasserstein regularization term .

[0035] The Wasserstein variation loss function is solved by the existing gradient descent method to obtain the optimal water body feature matrix . The Wasserstein variation loss function is constrained to make the water body classification more stable and reduce the uncertainty of data classification.

[0036] Based on the optimal water body feature matrix, water body change prediction is carried out to predict the future water body state; due to the strong correlation of the spatio-temporal change of water body, the prediction process not only needs to consider the water body feature data at the current time step, but also needs to combine the past The water body feature data of a time step and the optimal water body feature matrix at a historical time are modeled to enhance the time series dependency of the prediction. In addition, in order to ensure the smoothness of the prediction and reduce the influence of error accumulation, a orthogonal projection matrix is added to constrain the change range of the water body feature, so that it can more accurately reflect the water body change trend. Finally, the water body change prediction formula establishes a recursive model by integrating the information of multiple time steps to predict the future water body state. The water body change prediction formula is as follows:

[0037] wherein, is the water body change prediction state matrix at the time t, each element of the water body change prediction state matrix represents the water body change state value of a sampling point predicted at the current time step; is the weight parameter of the water body change prediction, which is obtained by expert experience method, and is initialized according to Xavier or He; is the number of historical time steps for predicting the water body change, i.e. the size of the time window for backtracking; , are the dynamic adjustment coefficients of the water body feature state matrix and the optimal water body feature matrix at the first and the th historical time step, respectively, which are obtained by fitting calculation, and the value range is ; is the water body feature state matrix at the time t; is the optimal water body feature matrix at the time t; is the gradient adjustment factor, which is obtained by the existing gradient descent method, and the value range is ; is the optimal water body feature matrix at the time t; is the optimal water body feature matrix at the time t; is the time weighted state input in the recurrent neural network (RNN), which is used for memory and modeling of time series; is the multi-source feature fusion in the Transformer, which combines feature transformation such as PCA or orthogonal projection, and introduces the historical optimal water body feature from the historical optimal water body feature matrix into the prediction; is the gradient normalization (such as ResNet / LayerNorm) and smooth trend detection (such as time difference normalization), which is used to capture short-term fluctuations and prevent gradient explosion / disappearance.

[0038] ​​​​​​Since the dynamic change of water body has time-dependent and is affected by various external factors, there may be dramatic fluctuations or cumulative drift in the data during prediction, which affects the stability of long-term prediction. Therefore, a prediction loss function is introduced to compare the predicted state matrix of water body change with the actual state matrix of water body, calculate the prediction error and minimize the processing to ensure the stability and rationality of long-term water body change trend. The mathematical expression of the prediction loss function is as follows:

[0039] wherein, is the prediction loss function, which is used to measure the prediction error of the predicted state matrix of water body change at time , and to regularize the constraint of the predicted state matrix of water body change to enhance the stability of long-term water body dynamic monitoring; is the target water body label data at time , i.e. the actual state matrix of water body at time ; is the total time step of prediction, i.e. the time series length considered in long-term water body monitoring; is the time decay parameter, which is used to determine the weighting effect of prediction error in time dimension, and is obtained by expert experience method, with the value range of ; is the recursive regularization parameter, which is used to control the smoothness of the predicted state matrix of water body change, and ensure the stability of water body change trend in time dimension, and is obtained by expert experience method, with the value range of ; is the predicted state matrix of water body change at time ; is the time weight decay term of prediction error, which is used to control the focus of prediction accuracy at different times, represents the deviation degree of future water body state (such as water area, water quality index) at the current prediction time, is a time focus weight, with the maximum weight at the middle time of prediction sequence; is the change amplitude regularization term, which is used to control the smoothness of prediction sequence and prevent drastic change, is used to reflect whether the continuity of prediction result in time is good.

[0040] The optimized predicted state matrix of water body change obtained by minimizing the prediction loss function can be used as the prediction result of long-term water body dynamic monitoring, and provides the water body state distribution at future time. Finally, the optimized predicted state matrix of water body change can be used to draw the water body change trend chart, calculate the water area change rate, analyze the long-term trend, and provide data support for water resource management and ecological protection decision-making, to realize the data processing of long-term water body dynamic monitoring.

[0041] In summary, a data processing method for long-term water body dynamic monitoring is completed.

[0042] The order of the embodiments is merely for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0043] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment mainly describes the difference from other embodiments.

[0044] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalent; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A data processing method for long-term water body dynamic monitoring, characterized in that: The following steps are involved: S1. Acquire and standardize sensor data to obtain standardized sensor data; perform feature extraction on the standardized sensor data to obtain water body feature data; model the dynamic evolution process of the water body, perform high-dimensional orthogonal projection transformation on the water body feature data, and obtain an optimized water body feature matrix; S2. Based on the optimized water body characteristic matrix, design and solve the Wasserstein variational loss function to obtain the optimal water body characteristic matrix; Based on the optimal water body characteristic matrix, water body change prediction is performed, and the prediction loss function is minimized to obtain the optimized water body change prediction state matrix.

2. The data processing method for long-term water body dynamic monitoring according to claim 1 is characterized in that: Said S1 specifically includes: A time recursive model is established using a spatiotemporal multi-layer recursive variational optimization network to extract features from standardized sensor data and obtain water body characteristic data.

3. The data processing method for long-term water body dynamic monitoring according to claim 2, characterized in that: Said S1 specifically includes: In the process of establishing the time recursive model, the water body characteristic data is adaptively adjusted through the variational optimization mechanism, and the spatiotemporal recursive calculation framework is introduced. The water body state at the current moment and the water body characteristic data of the past time steps are combined to construct the dynamic evolution process of the water body and obtain the water body characteristic state matrix.

4. The data processing method for long-term water body dynamic monitoring according to claim 3 is characterized in that: Said S1 specifically includes: Based on the water body characteristic state matrix, an orthogonal projection matrix is ​​constructed to project the water body characteristic data so that the water body characteristic data is retained after projection. The initial water body characteristic matrix is ​​set, and the regularization parameter is introduced to perform high-dimensional orthogonal projection transformation on the water body characteristic data.

5. The data processing method for long-term water body dynamic monitoring according to claim 4 is characterized in that: Said S1 specifically includes: The standardized sensor data matrix is ​​optimized through the orthogonal projection matrix to obtain the optimized water body characteristic matrix.

6. The data processing method for long-term water body dynamic monitoring according to claim 1, characterized in that: Said S2 specifically includes: Water body changes are predicted based on the optimal water body characteristic matrix to predict the future water body state; the water body characteristic data of the current time step are combined with the water body characteristic data of the past time step and the historical optimal water body characteristic matrix, and the orthogonal projection matrix is ​​added to calculate the water body change prediction state matrix.

7. The data processing method for long-term water body dynamic monitoring according to claim 6, characterized in that: Said S2 specifically includes: Based on the water body change prediction state matrix, a prediction loss function is introduced to compare the water body change prediction state matrix with the actual water body state matrix, and the prediction error is calculated and minimized.

8. The data processing method for long-term water body dynamic monitoring according to claim 7, characterized in that: Said S2 specifically includes: The prediction loss function is minimized to obtain the optimized water body change prediction state matrix, which is used as the prediction result of long-term water body dynamic monitoring and provides the water body state distribution at future moments.

Images