A water affairs detection data analysis system, method and medium based on DAS
Through the DAS-based water exploration data analysis system, using technologies such as data cleaning, synchronous noise removal and feature fusion, combined with deep learning, the problem of low efficiency of existing water exploration systems is solved, and efficient and accurate real-time monitoring of water areas is achieved.
Patent Information
- Application Number
- CN202410414381.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-08
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2044-04-08
AI Technical Summary
The existing water detection data analysis systems are challenging in real-time and accuracy, and the complex maintenance and management of large-scale water monitoring systems lead to low efficiency.
Adopt the DAS-based water detection data analysis system, including data cleaning module, synchronous noise removal module, feature fusion module, model training module and water detection module, through filtering noise removal, self-attention noise removal, multi-dimensional feature extraction and cross-modal Feature fusion, combined with deep learning methods, establish a water analysis model, and realize real-time distributed acoustic sensing detection.
It improves the accuracy and efficiency of water exploration and analysis, reduces labor costs, realizes real-time monitoring and analysis of water areas, and simplifies workflow.
Smart Images

Figure CN118312739B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of water affairs detection technology, and in particular to a DAS-based water affairs detection data analysis system, method and medium. Background Art
[0002] With the increasing shortage of water resources and water pollution, in order to ensure the safety of drinking water and maintain the water ecosystem, it is necessary to use technical means to monitor and analyze water resources, that is, water detection is needed.
[0003] Most of the existing water affairs detection data analysis systems are based on the Internet of Things sensor technology, that is, they use the Internet of Things technology to control multiple sensors to monitor water temperature, pH value, dissolved oxygen and other data in real time, and then realize water affairs data analysis. In actual applications, production plans are greatly affected by the economic environment. Water affairs detection data analysis systems based on Internet of Things sensor technology have certain challenges in real-time and accuracy, and cannot capture the dynamic changes of water resources in time. In addition, large-scale water affairs monitoring systems require complex maintenance and management, including regular maintenance of equipment and data processing, which may lead to low efficiency in water affairs detection and analysis. Summary of the invention
[0004] The present invention provides a DAS-based water affairs detection data analysis system, method and medium, the main purpose of which is to solve the problem of low efficiency in water affairs detection analysis.
[0005] To achieve the above object, the present invention provides a water affairs detection data analysis system based on DAS, the system includes a data cleaning module, a synchronous noise removal module, a feature fusion module, a model training module and a water affairs detection module, wherein:
[0006] The data cleaning module is used to respectively obtain historical water area acoustic data and historical water affairs data of the target water area, and sequentially filter and de-noise the historical water area acoustic data and perform sensor synchronization to obtain primary water area acoustic data;
[0007] The synchronous denoising module is used to synchronize the primary acoustic data of the water area into the synchronous acoustic data of the water area according to the historical water affairs data of the water area, and perform self-attention denoising on the synchronous acoustic data of the water area to obtain the standard acoustic data of the water area. When the synchronous denoising module performs self-attention denoising on the synchronous acoustic data of the water area to obtain the standard acoustic data of the water area, it is specifically used to:
[0008] The following acoustic feature algorithm is used to extract features from the water area synchronous acoustic data to obtain the water area acoustic features:
[0009]
[0010] Wherein, w(a,b) refers to the characteristic value of the acoustic characteristics of the water area at a scale of a and a displacement of b, a is the scale parameter, b is the displacement parameter, ∞ is the infinity symbol, v(t) refers to the acoustic data value of the synchronous acoustic data of the water area at time sequence t, t is the time sequence number, z() is the conjugate function symbol, e is the Euler number, i is the imaginary number symbol, ω 0 is the preset frequency parameter, d is the integral symbol;
[0011] Performing position coding on the acoustic features of the water area to obtain acoustic feature coding;
[0012] Performing multi-level residual connection and full connection operations on the acoustic feature code to obtain a standard acoustic feature code;
[0013] Performing a linear activation operation on the standard acoustic feature code to obtain water area standard acoustic data;
[0014] The feature fusion module is used to perform multi-dimensional feature extraction on the water area standard acoustic data to obtain a water area acoustic feature group, and perform cross-modal feature fusion on the water area acoustic feature group to form a water area acoustic standard feature;
[0015] The model training module is used to train the preset acoustic water affairs model into a water affairs analysis model by using the acoustic standard characteristics of the water area and the historical water affairs data of the water area;
[0016] The water affairs detection module is used to perform real-time distributed acoustic sensing detection on the target water area to obtain real-time water area acoustic data, and use the water affairs analysis model to analyze the real-time water affairs data corresponding to the real-time water area acoustic data.
[0017] Optionally, when the data cleaning module sequentially performs filtering, denoising and sensor synchronization on the historical water area acoustic data to obtain the water area primary acoustic data, it is specifically used to:
[0018] Extracting acoustic data numbers from the historical water area acoustic data, and splitting the historical water area acoustic data into water area acoustic data sets according to the acoustic data numbers;
[0019] Deduplication and outlier denoising are performed on the water area acoustic data set in turn to obtain a primary acoustic data set;
[0020] performing bandpass filtering on the primary acoustic data set to obtain a filtered acoustic data set;
[0021] performing temperature correction and sensitivity correction on the filtered acoustic data set in sequence to obtain a corrected acoustic data set;
[0022] extracting a time delay set from the corrected acoustic data set;
[0023] The corrected acoustic data set is time synchronized according to the time delay set to obtain a synchronized acoustic data set, and the synchronized acoustic data set is used as primary acoustic data of the water area.
[0024] Optionally, when the data cleaning module sequentially performs data deduplication and outlier denoising on the water area acoustic data set to obtain a primary acoustic data set, it is specifically used to:
[0025] Performing hash coding on the water area acoustic data set to obtain an acoustic hash coding set;
[0026] Deduplication of the water area acoustic data set is performed according to the acoustic hash code set to obtain a deduplication acoustic data set;
[0027] Performing frequency domain transformation on the deduplication acoustic data set to obtain a deduplication acoustic frequency domain set;
[0028] Performing spectrum analysis on the deduplicated acoustic frequency domain set to obtain an acoustic frequency domain feature set;
[0029] Performing outlier detection on the acoustic frequency domain feature set in turn to obtain an outlier frequency domain feature group;
[0030] The deduplicated acoustic data set is subjected to outlier denoising according to the outlier frequency domain feature group to obtain a primary acoustic data set.
[0031] Optionally, when synchronizing the primary acoustic data of the water area into synchronized acoustic data of the water area according to the historical water affairs data of the water area, the synchronous noise removal module is specifically used to:
[0032] Extracting water affairs data timestamp from the historical water affairs data of the water area;
[0033] Extracting acoustic data timestamps from the primary acoustic data of the water area;
[0034] The water area synchronous acoustic data is filtered out from the water area primary acoustic data according to the water affairs data timestamp and the acoustic data timestamp.
[0035] Optionally, when the feature fusion module extracts multi-dimensional features from the water area standard acoustic data to obtain the water area acoustic feature group, it is specifically used to:
[0036] Performing time-domain statistical analysis on the standard acoustic data of the water area to obtain acoustic time-domain characteristics;
[0037] Performing frequency domain transformation and frequency domain feature extraction on the standard acoustic data of the water area in sequence to obtain acoustic frequency domain features;
[0038] Extracting sensor position data from the water area standard acoustic data, and converting the sensor position data into acoustic space features;
[0039] The acoustic time domain features, the acoustic frequency domain features and the acoustic space features are aggregated into a water area acoustic feature group.
[0040] Optionally, when the feature fusion module performs cross-modal feature fusion on the water area acoustic feature group into a water area acoustic standard feature, it is specifically used to:
[0041] Performing global pooling on the water area acoustic feature group to obtain a pooled acoustic feature group;
[0042] The pooled acoustic feature group is fused into the water area acoustic standard feature using the following acoustic feature fusion algorithm:
[0043]
[0044] Where S refers to the standard acoustic feature of the water area, j refers to the feature dimension number, E refers to the total dimension of the features in the pooled acoustic feature group, tanh() is the hyperbolic tangent function, Q j It refers to the j-th dimension feature of the first pooled acoustic feature in the pooled acoustic feature group, α, β, γ are the weight matrices of the acoustic feature fusion algorithm, T is the transposition symbol, d() is the feature dimension function, R j refers to the j-th dimension feature of the second pooled acoustic feature in the pooled acoustic feature group, P j It refers to the j-th dimension feature of the third pooled acoustic feature in the pooled acoustic feature group.
[0045] Optionally, when the model training module trains the preset acoustic water affairs model into a water affairs analysis model using the water area acoustic standard features and the historical water affairs data, it is specifically used to:
[0046] Using a preset acoustic water affairs model to perform regression analysis on the acoustic standard characteristics of the water area, to obtain predicted water affairs data of the water area;
[0047] The marginal loss value between the predicted water affairs data of the water area and the historical water affairs data of the water area is calculated using the following marginal loss function:
[0048]
[0049] Wherein, L refers to the edge loss value, k refers to the data sequence number, K refers to the total number of data, and the total number of data of the predicted water affairs data in the water area is equal to the total number of data of the historical water affairs data in the water area, D k It refers to the kth data value in the water affairs data of the predicted water area, Y krefers to the kth data value in the historical water affairs data of the water area, ε is a preset constant, λ is a preset loss weight, and Δ is the Laplace operator symbol;
[0050] The model parameters of the acoustic water service model are iteratively updated according to the edge loss value to obtain a water service analysis model.
[0051] Optionally, when the model training module uses a preset acoustic water affairs model to perform regression analysis on the acoustic standard characteristics of the water area to obtain predicted water affairs data of the water area, it is specifically used to:
[0052] Performing residual convolution on the acoustic standard characteristics of the water area using a preset acoustic water affairs model to obtain residual acoustic characteristics;
[0053] Using the acoustic water service model to regress and map the residual acoustic features to obtain mapped acoustic features;
[0054] Using the acoustic water service model, the mapped acoustic features are multi-level upsampled to obtain primary water service data;
[0055] The acoustic water affairs model is used to perform convolution and zero filling on the primary water affairs data to obtain predicted water affairs data of the water area.
[0056] In order to solve the above problems, the present invention also provides a water affairs detection data analysis method based on DAS, the method comprising:
[0057] Acquire historical water area acoustic data and historical water affairs data of the target water area respectively, and perform filtering, noise removal and sensor synchronization on the historical water area acoustic data in turn to obtain primary water area acoustic data;
[0058] Synchronizing the primary acoustic data of the water area into synchronized acoustic data of the water area according to the historical water affairs data of the water area, and performing self-attention denoising on the synchronized acoustic data of the water area to obtain standard acoustic data of the water area;
[0059] Performing multi-dimensional feature extraction on the water area standard acoustic data to obtain a water area acoustic feature group, and performing cross-modal feature fusion on the water area acoustic feature group to form a water area acoustic standard feature;
[0060] Using the acoustic standard features of the water area and the water affairs data of the historical water area to train a preset acoustic water affairs model into a water affairs analysis model;
[0061] The target water area is detected by real-time distributed acoustic sensing to obtain real-time water area acoustic data, and the water affairs analysis model is used to analyze the real-time water affairs data corresponding to the real-time water area acoustic data.
[0062] In order to solve the above problems, the present invention also provides a computer-readable storage medium, in which at least one computer program is stored. The at least one computer program is executed by a processor in an electronic device to implement the above-mentioned DAS-based water detection data analysis system.
[0063] The present invention obtains primary acoustic data of the water area by filtering, denoising and sensing synchronization on the historical water area acoustic data in sequence, which can realize primary data cleaning of the historical water area acoustic data, improve the accuracy of subsequent model training and the accuracy of the final water affairs detection and analysis, and synchronize the primary acoustic data of the water area into synchronized acoustic data of the water area according to the historical water area water affairs data, so as to realize data mapping between the historical water area water affairs data and the primary acoustic data of the water area, thereby facilitating the establishment of subsequent model data sets, and obtains standard acoustic data of the water area by performing self-attention denoising on the synchronized acoustic data of the water area, and can realize intelligent denoising in combination with the self-attention mechanism to improve the accuracy of the acoustic data, and obtains a water area acoustic feature group by performing multi-dimensional feature extraction on the standard acoustic data of the water area, which can extract the features of the standard acoustic data of the water area in the three feature dimensions of time domain, frequency domain and space, and can realize feature fusion of acoustic features in the three dimensions of frequency domain, time domain and space by performing cross-modal feature fusion on the water area acoustic feature group into standard acoustic features of the water area, thereby improving the accuracy of subsequent water affairs detection.
[0064] By using the acoustic standard features of the water area and the water affairs data of the historical water area to train the preset acoustic water affairs model into a water affairs analysis model, it is possible to realize the correlation analysis between the water area acoustic data and the water affairs data, and obtain the mapping relationship between the water area acoustic data and the water affairs data. By relying on the deep learning method, water affairs detection is realized, and the efficiency of water affairs detection is improved. By performing real-time distributed acoustic sensing detection on the target water area, real-time water area acoustic data is obtained, and the actual acoustic data of the target water area can be obtained. By using the water affairs analysis model to analyze the real-time water affairs data corresponding to the real-time water area acoustic data, real-time monitoring and analysis of the target water area can be realized, reducing the workflow of water affairs analysis, reducing the cost of manpower detection, and improving the efficiency of water affairs detection. Therefore, the water affairs detection data analysis system, method and medium based on DAS proposed in the present invention can solve the problem of low efficiency in water affairs detection and analysis. BRIEF DESCRIPTION OF THE DRAWINGS
[0065] Figure 1 A functional module diagram of a water affairs detection data analysis system based on DAS provided in one embodiment of the present invention;
[0066] Figure 2 A schematic diagram of a flow chart of a DAS-based water affairs detection data analysis method provided in one embodiment of the present invention;
[0067] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments. DETAILED DESCRIPTION
[0068] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.
[0069] The embodiment of the present application provides a water detection data analysis system based on DAS. The execution subject of the water detection data analysis system based on DAS includes but is not limited to at least one of the electronic devices such as a server, a terminal, etc. that can be configured to execute the system provided by the embodiment of the present application. In other words, the water detection data analysis system based on DAS can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms.
[0070] like Figure 1 , which is a functional module diagram of a DAS-based water affairs detection data analysis system provided by an embodiment of the present invention.
[0071] The DAS-based water affairs detection data analysis system 100 of the present invention can be installed in an electronic device. According to the functions to be implemented, the DAS-based water affairs detection data analysis system 100 can include a data cleaning module 101, a synchronous noise removal module 102, a feature fusion module 103, a model training module 104 and a water affairs detection module 105. The module of the present invention can also be called a unit, which refers to a series of computer program segments that can be executed by an electronic device processor and can complete fixed functions, which are stored in the memory of the electronic device.
[0072] In this embodiment, the functions of each module / unit are as follows:
[0073] The data cleaning module 101 is used to respectively obtain historical water area acoustic data and historical water affairs data of the target water area, and perform filtering, noise removal and sensor synchronization on the historical water area acoustic data in turn to obtain primary water area acoustic data.
[0074] In the embodiment of the present invention, the target water area refers to the water area where water data detection is required, which may be an underground river, an underground pipeline, a water supply system or a drainage system, etc. Due to factors such as cost and detection time, water data detection for the target water area has certain limitations, so it is necessary to use distributed acoustic sensing (DAS) for water detection.
[0075] In detail, the historical water area acoustic data refers to the data obtained by detecting and monitoring the target water area using distributed acoustic sensors in the past time period, and the historical water area water affairs data refers to the data obtained by water affairs detection and analysis of the target water area using traditional water quality monitoring systems in the past time period. The traditional water quality monitoring system refers to the detection of water affairs data such as dissolved oxygen, pH value, turbidity, ammonia nitrogen and flow rate of the target water area; the historical water area acoustic data and historical water affairs data of the target water area can be obtained respectively by using a database retrieval system or a crawler data crawling system.
[0076] In the embodiment of the present invention, when the data cleaning module 101 sequentially performs filtering, noise removal and sensor synchronization on the historical water area acoustic data to obtain the primary water area acoustic data, it is specifically used to:
[0077] Extracting acoustic data numbers from the historical water area acoustic data, and splitting the historical water area acoustic data into water area acoustic data sets according to the acoustic data numbers;
[0078] Deduplication and outlier denoising are performed on the water area acoustic data set in turn to obtain a primary acoustic data set;
[0079] performing bandpass filtering on the primary acoustic data set to obtain a filtered acoustic data set;
[0080] performing temperature correction and sensitivity correction on the filtered acoustic data set in sequence to obtain a corrected acoustic data set;
[0081] extracting a time delay set from the corrected acoustic data set;
[0082] The corrected acoustic data set is time synchronized according to the time delay set to obtain a synchronized acoustic data set, and the synchronized acoustic data set is used as primary acoustic data of the water area.
[0083] In detail, the acoustic data number refers to the numbering data of the historical water area acoustic data when it is stored, and each water area acoustic data in the water area acoustic data set represents data obtained by performing a distributed acoustic sensing detection.
[0084] In the embodiment of the present invention, when the data cleaning module 101 sequentially performs data deduplication and outlier denoising on the water area acoustic data set to obtain a primary acoustic data set, it is specifically used to:
[0085] Performing hash coding on the water area acoustic data set to obtain an acoustic hash coding set;
[0086] Deduplication of the water area acoustic data set is performed according to the acoustic hash code set to obtain a deduplication acoustic data set;
[0087] Performing frequency domain transformation on the deduplication acoustic data set to obtain a deduplication acoustic frequency domain set;
[0088] Performing spectrum analysis on the deduplicated acoustic frequency domain set to obtain an acoustic frequency domain feature set;
[0089] Performing outlier detection on the acoustic frequency domain feature set in turn to obtain an outlier frequency domain feature group;
[0090] The deduplicated acoustic data set is subjected to outlier denoising according to the outlier frequency domain feature group to obtain a primary acoustic data set.
[0091] Specifically, the water area acoustic data set can be hash-encoded using the MD5 algorithm, the SHA-256 algorithm or the CityHash algorithm to obtain an acoustic hash code set. The data cleaning module 101 deduplicates the water area acoustic data set according to the acoustic hash code set to obtain a deduplicated acoustic data set. Specifically, the repeated acoustic hash codes in the acoustic hash code set are used as repeated acoustic hash codes, and the deduplicated acoustic data corresponding to the repeated acoustic hash codes are deleted from the water area acoustic data set to obtain a deduplicated acoustic data set.
[0092] In detail, the deduplicated acoustic data set can be transformed in the frequency domain using a fast Fourier transform algorithm to obtain a deduplicated acoustic frequency domain set. The deduplicated acoustic frequency domain set can be spectrally analyzed using a dominant frequency analysis method or a harmonic analysis method to obtain an acoustic frequency domain feature set. The acoustic frequency domain feature set can be detected for outliers using an outlier detection method based on a Z score, a box plot method, or feature clustering to obtain an outlier frequency domain feature group.
[0093] In detail, the primary acoustic data set can be bandpass filtered using a bandpass filter to obtain a filtered acoustic data set, the filtered acoustic data set can be temperature corrected using pre-acquired temperature sensor data, the sensitivity correction coefficient corresponding to the filtered acoustic data set can be calculated using pre-acquired response data of each sensor, and the filtered acoustic data set can be sensitivity corrected according to the sensitivity correction coefficient to obtain a corrected acoustic data set.
[0094] Specifically, a time delay set can be extracted from the correction acoustic data set by calculating a cross-correlation function, wherein each time delay in the time delay set is a time delay between each light sensor in each correction acoustic data in the corresponding correction acoustic data set; and time synchronization of the correction acoustic data set according to the time delay set refers to adjusting the timing between each sensor data of the corresponding correction acoustic data in the correction acoustic data set according to the corresponding time delay in the time delay set.
[0095] In an embodiment of the present invention, the data cleaning module 101 obtains primary acoustic data of the water area by filtering, denoising and sensing the historical water area acoustic data in sequence, thereby achieving primary data cleaning of the historical water area acoustic data, improving the accuracy of subsequent model training and the accuracy of the final water affairs detection and analysis.
[0096] The synchronous denoising module 102 is used to synchronize the primary acoustic data of the water area into synchronized acoustic data of the water area according to the historical water affairs data of the water area, and perform self-attention denoising on the synchronized acoustic data of the water area to obtain standard acoustic data of the water area.
[0097] In the embodiment of the present invention, the water area synchronous acoustic data refers to a portion of the water area primary acoustic data corresponding to the historical water area water affairs data in the water area primary acoustic data.
[0098] In the embodiment of the present invention, when synchronizing the primary acoustic data of the water area into the synchronized acoustic data of the water area according to the historical water affairs data of the water area, the synchronous noise removal module 102 is specifically used to:
[0099] Extracting water affairs data timestamp from the historical water affairs data of the water area;
[0100] Extracting acoustic data timestamps from the primary acoustic data of the water area;
[0101] The water area synchronous acoustic data is filtered out from the water area primary acoustic data according to the water affairs data timestamp and the acoustic data timestamp.
[0102] In detail, the water affairs data timestamp refers to the timestamp corresponding to when the historical water affairs data of the water area is obtained. When the synchronous noise removal module 102 filters out the water area synchronous acoustic data from the water area primary acoustic data according to the water affairs data timestamp and the acoustic data timestamp, it specifically refers to selecting the acoustic data timestamp corresponding to the water affairs data timestamp as the target acoustic data timestamp, and taking part of the acoustic data in the water area primary acoustic data corresponding to the target acoustic data timestamp as the water area synchronous acoustic data.
[0103] In the embodiment of the present invention, when the synchronous denoising module 102 performs self-attention denoising on the water area synchronous acoustic data to obtain the water area standard acoustic data, it is specifically used to:
[0104] The following acoustic feature algorithm is used to extract features from the water area synchronous acoustic data to obtain the water area acoustic features:
[0105]
[0106] Wherein, w(a,b) refers to the characteristic value of the acoustic characteristics of the water area at a scale of a and a displacement of b, a is the scale parameter, b is the displacement parameter, ∞ is the infinity symbol, v(t) refers to the acoustic data value of the synchronous acoustic data of the water area at time sequence t, t is the time sequence number, z() is the conjugate function symbol, e is the Euler number, i is the imaginary number symbol, ω 0 is the preset frequency parameter, d is the integral symbol;
[0107] Performing position coding on the acoustic features of the water area to obtain acoustic feature coding;
[0108] Performing multi-level residual connection and full connection operations on the acoustic feature code to obtain a standard acoustic feature code;
[0109] A linear activation operation is performed on the standard acoustic feature code to obtain water area standard acoustic data.
[0110] In detail, by utilizing an acoustic feature algorithm to extract features from the water area synchronous acoustic data, the water area acoustic features are obtained, and feature data of the water area synchronous acoustic data at different scales and displacements can be extracted, thereby characterizing the acoustic features of the water area.
[0111] Specifically, the sine function and cosine function can be used to perform position encoding on the acoustic features of the water area to obtain acoustic feature encoding, the encoder module of the Transformer model can be used to perform multi-level residual connection and full connection operations on the acoustic feature encoding to obtain standard acoustic feature encoding, and the decode module can be used to perform linear activation operations on the standard acoustic feature encoding to obtain standard acoustic data of the water area.
[0112] In an embodiment of the present invention, the synchronous denoising module 102 can realize data mapping of historical water area water affairs data and water area primary acoustic data by synchronizing the water area primary acoustic data into water area synchronous acoustic data according to the historical water area water affairs data, thereby facilitating the establishment of subsequent model data sets, and by performing self-attention denoising on the water area synchronous acoustic data to obtain water area standard acoustic data, intelligent denoising can be realized in combination with the self-attention mechanism to improve the accuracy of acoustic data.
[0113] The feature fusion module 103 is used to perform multi-dimensional feature extraction on the water area standard acoustic data to obtain a water area acoustic feature group, and perform cross-modal feature fusion on the water area acoustic feature group to form a water area acoustic standard feature.
[0114] In the embodiment of the present invention, the water area acoustic feature group includes multi-dimensional features of the water area standard acoustic data, such as propagation time features and frequency features of sound waves.
[0115] In the embodiment of the present invention, when the feature fusion module 103 extracts multi-dimensional features from the water area standard acoustic data to obtain the water area acoustic feature group, it is specifically used to:
[0116] Performing time-domain statistical analysis on the standard acoustic data of the water area to obtain acoustic time-domain characteristics;
[0117] Performing frequency domain transformation and frequency domain feature extraction on the standard acoustic data of the water area in sequence to obtain acoustic frequency domain features;
[0118] Extracting sensor position data from the water area standard acoustic data, and converting the sensor position data into acoustic space features;
[0119] The acoustic time domain features, the acoustic frequency domain features and the acoustic space features are aggregated into a water area acoustic feature group.
[0120] In detail, the time domain statistical analysis refers to the extraction of amplitude data, energy data, peak data and root mean square data of the standard acoustic data of the water area. The fast Fourier transform method can be used to perform frequency domain transformation on the standard acoustic data of the water area. The frequency domain feature extraction refers to the extraction of features such as main frequency, frequency band energy, and frequency domain statistics.
[0121] Specifically, when the feature fusion module 103 performs cross-modal feature fusion on the water area acoustic feature group into a water area acoustic standard feature, it is specifically used to:
[0122] Performing global pooling on the water area acoustic feature group to obtain a pooled acoustic feature group;
[0123] The pooled acoustic feature group is fused into the water area acoustic standard feature using the following acoustic feature fusion algorithm:
[0124]
[0125] Where S refers to the standard acoustic feature of the water area, j refers to the feature dimension number, E refers to the total dimension of the features in the pooled acoustic feature group, tanh() is the hyperbolic tangent function, Q jIt refers to the j-th dimension feature of the first pooled acoustic feature in the pooled acoustic feature group, α, β, γ are the weight matrices of the acoustic feature fusion algorithm, T is the transposition symbol, d() is the feature dimension function, R j refers to the j-th dimension feature of the second pooled acoustic feature in the pooled acoustic feature group, P j It refers to the j-th dimension feature of the third pooled acoustic feature in the pooled acoustic feature group.
[0126] Specifically, the pooled acoustic feature group is fused into the water area acoustic standard feature through the acoustic feature fusion algorithm, which can be combined with the attention algorithm to realize the feature fusion of acoustic features in three dimensions: frequency domain, time domain and space, retain the key elements in the features, and thus improve the accuracy of subsequent water affairs detection.
[0127] In an embodiment of the present invention, the feature fusion module 103 obtains a water area acoustic feature group by performing multi-dimensional feature extraction on the water area standard acoustic data, and can extract features of the water area standard acoustic data in three feature dimensions: time domain, frequency domain, and space. By performing cross-modal feature fusion of the water area acoustic feature group into water area acoustic standard features, feature fusion of acoustic features in three dimensions: frequency domain, time domain, and space can be achieved, thereby improving the accuracy of subsequent water affairs detection.
[0128] The model training module 104 is used to train the preset acoustic water affairs model into a water affairs analysis model by using the water area acoustic standard characteristics and the historical water affairs data of the water area.
[0129] In an embodiment of the present invention, the acoustic water management model may be a deep learning model such as a support vector machine model or a random forest model.
[0130] In the embodiment of the present invention, when the model training module 104 trains the preset acoustic water affairs model into a water affairs analysis model using the water area acoustic standard features and the historical water affairs data, it is specifically used to:
[0131] Using a preset acoustic water affairs model to perform regression analysis on the acoustic standard characteristics of the water area, to obtain predicted water affairs data of the water area;
[0132] The marginal loss value between the predicted water affairs data of the water area and the historical water affairs data of the water area is calculated using the following marginal loss function:
[0133]
[0134] Wherein, L refers to the edge loss value, k refers to the data sequence number, K refers to the total number of data, and the total number of data of the predicted water affairs data in the water area is equal to the total number of data of the historical water affairs data in the water area, D kIt refers to the kth data value in the water affairs data of the predicted water area, Y k refers to the kth data value in the historical water affairs data of the water area, ε is a preset constant, λ is a preset loss weight, and Δ is the Laplace operator symbol;
[0135] The model parameters of the acoustic water service model are iteratively updated according to the edge loss value to obtain a water service analysis model.
[0136] Specifically, by calculating the edge loss value between the predicted water area water affairs data and the historical water area water affairs data through the edge loss function, the output space can be fitted, the gradient smoothing term is increased, the gradient explosion problem is better handled, and the robustness of the loss function is improved.
[0137] In detail, when the model training module 104 uses the preset acoustic water affairs model to perform regression analysis on the water area acoustic standard features to obtain the predicted water affairs data of the water area, it is specifically used to:
[0138] Performing residual convolution on the acoustic standard characteristics of the water area using a preset acoustic water affairs model to obtain residual acoustic characteristics;
[0139] Using the acoustic water service model to regress and map the residual acoustic features to obtain mapped acoustic features;
[0140] Using the acoustic water service model, the mapped acoustic features are multi-level upsampled to obtain primary water service data;
[0141] The acoustic water affairs model is used to perform convolution and zero filling on the primary water affairs data to obtain predicted water affairs data of the water area.
[0142] In detail, the residual convolution layer (Residual ConvolutionalLayer) of the acoustic water conservancy model can be used to perform residual convolution on the standard acoustic features of the water area to obtain residual acoustic features. The kernel function of the acoustic water conservancy model can be used to perform regression mapping on the residual acoustic features to obtain mapped acoustic features. The zero padding (Zero Padding) algorithm of the acoustic water conservancy model can be used to perform convolution and zero padding on the primary water conservancy data to obtain predicted water area water conservancy data.
[0143] Specifically, when the model training module 104 iteratively updates the model parameters of the acoustic water service model according to the edge loss value to obtain the water service analysis model, it is specifically used to:
[0144] Determine whether the edge loss value is greater than a preset loss value threshold;
[0145] If not, then using an adaptive learning rate algorithm to iteratively update the model parameters of the acoustic water affairs model according to the edge loss value, and returning to the step of performing regression analysis on the acoustic standard characteristics of the water area using the preset acoustic water affairs model;
[0146] If so, the updated acoustic water services model is used as the water services analysis model.
[0147] In detail, the adaptive learning rate algorithm may be an Adam (Adaptive Moment Estimation) algorithm, an RMSprop (Root Mean Square Propagation) algorithm or an Adagrad (Adaptive Gradient Algorithm) algorithm.
[0148] In an embodiment of the present invention, the model training module 104 trains the preset acoustic water affairs model into a water affairs analysis model by utilizing the water area acoustic standard features and the historical water area water affairs data, thereby realizing the correlation analysis between the water area acoustic data and the water affairs data, and obtaining the mapping relationship between the water area acoustic data and the water affairs data, thereby realizing water affairs detection by means of deep learning methods and improving the efficiency of water affairs detection.
[0149] The water affairs detection module 105 is used to perform real-time distributed acoustic sensing detection on the target water area to obtain real-time water area acoustic data, and use the water affairs analysis model to analyze the real-time water affairs data corresponding to the real-time water area acoustic data.
[0150] In an embodiment of the present invention, when the water affairs detection module 105 performs real-time distributed acoustic sensing detection on the target water area, it specifically refers to using distributed acoustic sensing (DAS) to perform real-time acoustic sensing detection on the target water area. The method of using the water affairs analysis model to analyze the real-time water affairs data corresponding to the real-time water area acoustic data is consistent with the method of using a preset acoustic water affairs model in the model training module 104 to perform regression analysis on the acoustic standard characteristics of the water area to obtain predicted water affairs data of the water area.
[0151] In an embodiment of the present invention, the water affairs detection module 105 obtains real-time water area acoustic data by performing real-time distributed acoustic sensor detection on the target water area, and can obtain the actual acoustic data of the target water area. By using the water affairs analysis model to analyze the real-time water affairs data corresponding to the real-time water area acoustic data, real-time monitoring and analysis of the target water area can be achieved, thereby reducing the workflow of water affairs analysis, reducing the cost of manpower detection, and improving the efficiency of water affairs detection.
[0152] Reference Figure 2 FIG. 1 is a flow chart of a DAS-based water affairs detection data analysis method according to an embodiment of the present invention. In this embodiment, the DAS-based water affairs detection data analysis method includes:
[0153] S1. Acquire historical water area acoustic data and historical water affairs data of the target water area respectively, and perform filtering, noise removal and sensor synchronization on the historical water area acoustic data in turn to obtain primary water area acoustic data.
[0154] S2. Synchronize the primary acoustic data of the water area into synchronized acoustic data of the water area according to the historical water affairs data of the water area, perform self-attention denoising on the synchronized acoustic data of the water area, and obtain standard acoustic data of the water area.
[0155] S3. Perform multi-dimensional feature extraction on the water area standard acoustic data to obtain a water area acoustic feature group, and perform cross-modal feature fusion on the water area acoustic feature group to form a water area acoustic standard feature.
[0156] S4. Using the acoustic standard features of the water area and the historical water affairs data of the water area, the preset acoustic water affairs model is trained into a water affairs analysis model.
[0157] S5. Perform real-time distributed acoustic sensing detection on the target water area to obtain real-time water area acoustic data, and use the water affairs analysis model to analyze the real-time water affairs data corresponding to the real-time water area acoustic data.
[0158] The present invention obtains primary acoustic data of the water area by filtering, denoising and sensing synchronization on the historical water area acoustic data in sequence, which can realize primary data cleaning of the historical water area acoustic data, improve the accuracy of subsequent model training and the accuracy of the final water affairs detection and analysis, and synchronize the primary acoustic data of the water area into synchronized acoustic data of the water area according to the historical water area water affairs data, so as to realize data mapping between the historical water area water affairs data and the primary acoustic data of the water area, thereby facilitating the establishment of subsequent model data sets, and obtains standard acoustic data of the water area by performing self-attention denoising on the synchronized acoustic data of the water area, and can realize intelligent denoising in combination with the self-attention mechanism to improve the accuracy of the acoustic data, and obtains a water area acoustic feature group by performing multi-dimensional feature extraction on the standard acoustic data of the water area, which can extract the features of the standard acoustic data of the water area in the three feature dimensions of time domain, frequency domain and space, and can realize feature fusion of acoustic features in the three dimensions of frequency domain, time domain and space by performing cross-modal feature fusion on the water area acoustic feature group into standard acoustic features of the water area, thereby improving the accuracy of subsequent water affairs detection.
[0159] By using the acoustic standard features of the water area and the water affairs data of the historical water area to train the preset acoustic water affairs model into a water affairs analysis model, it is possible to realize the correlation analysis between the water area acoustic data and the water affairs data, and obtain the mapping relationship between the water area acoustic data and the water affairs data. By relying on the deep learning method, water affairs detection is realized, and the efficiency of water affairs detection is improved. By performing real-time distributed acoustic sensing detection on the target water area, real-time water area acoustic data is obtained, and the actual acoustic data of the target water area can be obtained. By using the water affairs analysis model to analyze the real-time water affairs data corresponding to the real-time water area acoustic data, real-time monitoring and analysis of the target water area can be realized, reducing the workflow of water affairs analysis, reducing the cost of manpower detection, and improving the efficiency of water affairs detection. Therefore, the DAS-based water affairs detection data method proposed in the present invention can solve the problem of low efficiency when performing water affairs detection and analysis.
[0160] In detail, the DAS-based water affairs detection data analysis method in the embodiment of the present invention is used in the same manner as above. Figure 1 The water detection data analysis system based on DAS has the same technical means as described in the text and can produce the same technical effects, so I will not go into details here.
[0161] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement:
[0162] Acquire historical water area acoustic data and historical water affairs data of the target water area respectively, and perform filtering, noise removal and sensor synchronization on the historical water area acoustic data in turn to obtain primary water area acoustic data;
[0163] Synchronizing the primary acoustic data of the water area into synchronized acoustic data of the water area according to the historical water affairs data of the water area, and performing self-attention denoising on the synchronized acoustic data of the water area to obtain standard acoustic data of the water area;
[0164] Performing multi-dimensional feature extraction on the water area standard acoustic data to obtain a water area acoustic feature group, and performing cross-modal feature fusion on the water area acoustic feature group to form a water area acoustic standard feature;
[0165] Using the acoustic standard features of the water area and the water affairs data of the historical water area to train a preset acoustic water affairs model into a water affairs analysis model;
[0166] The target water area is detected by real-time distributed acoustic sensing to obtain real-time water area acoustic data, and the water affairs analysis model is used to analyze the real-time water affairs data corresponding to the real-time water area acoustic data.
[0167] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, systems and systems can be implemented in other ways. For example, the system embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.
[0168] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0169] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.
[0170] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.
[0171] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.
[0172] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, system, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.
[0173] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units or systems stated in the system embodiment can also be implemented by one unit or system through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any specific order.
[0174] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.
Claims
1. A water affairs detection data analysis system based on DAS, characterized in that: The system includes a data cleaning module, a synchronous noise removal module, a feature fusion module, a model training module and a water affairs detection module, wherein: The data cleaning module is used to respectively obtain historical water area acoustic data and historical water affairs data of the target water area, and sequentially filter and de-noise the historical water area acoustic data and perform sensor synchronization to obtain primary water area acoustic data; The synchronous denoising module is used to synchronize the primary acoustic data of the water area into the synchronous acoustic data of the water area according to the historical water affairs data of the water area, and perform self-attention denoising on the synchronous acoustic data of the water area to obtain the standard acoustic data of the water area. When the synchronous denoising module performs self-attention denoising on the synchronous acoustic data of the water area to obtain the standard acoustic data of the water area, it is specifically used to: The following acoustic feature algorithm is used to extract features from the water area synchronous acoustic data to obtain the water area acoustic features: Wherein, w(a,b) refers to the characteristic value of the acoustic characteristics of the water area at a scale of a and a displacement of b, a is the scale parameter, b is the displacement parameter, ∞ is the infinity symbol, v(t) refers to the acoustic data value of the synchronous acoustic data of the water area at time sequence number t, t is the time sequence number, z() is the conjugate function symbol, e is the Euler number, i is the imaginary number symbol, ω0 is the preset frequency parameter, and d is the integral symbol; Performing position coding on the acoustic features of the water area to obtain acoustic feature coding; Performing multi-level residual connection and full connection operations on the acoustic feature code to obtain a standard acoustic feature code; Performing a linear activation operation on the standard acoustic feature code to obtain water area standard acoustic data; The feature fusion module is used to perform multi-dimensional feature extraction on the water area standard acoustic data to obtain a water area acoustic feature group, and perform cross-modal feature fusion on the water area acoustic feature group to form a water area acoustic standard feature; The model training module is used to train the preset acoustic water affairs model into a water affairs analysis model by using the acoustic standard characteristics of the water area and the historical water affairs data of the water area; The water affairs detection module is used to perform real-time distributed acoustic sensing detection on the target water area to obtain real-time water area acoustic data, and use the water affairs analysis model to analyze the real-time water affairs data corresponding to the real-time water area acoustic data.
2. The DAS-based water affairs detection data analysis system according to claim 1, characterized in that: The data cleaning module is specifically used to filter, remove noise and synchronize sensors on the historical water area acoustic data in sequence to obtain the primary water area acoustic data: Extracting acoustic data numbers from the historical water area acoustic data, and splitting the historical water area acoustic data into water area acoustic data sets according to the acoustic data numbers; Deduplication and outlier denoising are performed on the water area acoustic data set in turn to obtain a primary acoustic data set; performing bandpass filtering on the primary acoustic data set to obtain a filtered acoustic data set; performing temperature correction and sensitivity correction on the filtered acoustic data set in sequence to obtain a corrected acoustic data set; extracting a time delay set from the corrected acoustic data set; The corrected acoustic data set is time synchronized according to the time delay set to obtain a synchronized acoustic data set, and the synchronized acoustic data set is used as primary acoustic data of the water area.
3. The DAS-based water affairs detection data analysis system according to claim 2, characterized in that: When the data cleaning module sequentially removes data duplication and outlier noise from the water area acoustic data set to obtain a primary acoustic data set, it is specifically used to: Performing hash coding on the water area acoustic data set to obtain an acoustic hash coding set; Deduplication of the water area acoustic data set is performed according to the acoustic hash code set to obtain a deduplication acoustic data set; Performing frequency domain transformation on the deduplication acoustic data set to obtain a deduplication acoustic frequency domain set; Performing spectrum analysis on the deduplicated acoustic frequency domain set to obtain an acoustic frequency domain feature set; Performing outlier detection on the acoustic frequency domain feature set in turn to obtain an outlier frequency domain feature group; The deduplicated acoustic data set is subjected to outlier denoising according to the outlier frequency domain feature group to obtain a primary acoustic data set.
4. The DAS-based water affairs detection data analysis system according to claim 1, characterized in that: When the synchronous noise removal module synchronizes the primary acoustic data of the water area into the synchronized acoustic data of the water area according to the historical water affairs data of the water area, it is specifically used to: Extracting water affairs data timestamp from the historical water affairs data of the water area; Extracting acoustic data timestamps from the primary acoustic data of the water area; The water area synchronous acoustic data is filtered out from the water area primary acoustic data according to the water affairs data timestamp and the acoustic data timestamp.
5. The DAS-based water affairs detection data analysis system according to claim 1, characterized in that: When the feature fusion module extracts multi-dimensional features from the water area standard acoustic data to obtain the water area acoustic feature group, it is specifically used to: Performing time-domain statistical analysis on the standard acoustic data of the water area to obtain acoustic time-domain characteristics; Performing frequency domain transformation and frequency domain feature extraction on the standard acoustic data of the water area in sequence to obtain acoustic frequency domain features; Extracting sensor position data from the water area standard acoustic data, and converting the sensor position data into acoustic space features; The acoustic time domain features, the acoustic frequency domain features and the acoustic space features are aggregated into a water area acoustic feature group.
6. The DAS-based water affairs detection data analysis system according to claim 1, characterized in that: When the feature fusion module fuses the water area acoustic feature group into a water area acoustic standard feature through cross-modal feature fusion, it is specifically used to: Performing global pooling on the water area acoustic feature group to obtain a pooled acoustic feature group; The pooled acoustic feature group is fused into the water area acoustic standard feature using the following acoustic feature fusion algorithm: Where S refers to the standard acoustic feature of the water area, j refers to the feature dimension number, E refers to the total dimension of the features in the pooled acoustic feature group, tanh() is the hyperbolic tangent function, Q j It refers to the j-th dimension feature of the first pooled acoustic feature in the pooled acoustic feature group, α, β, γ are the weight matrices of the acoustic feature fusion algorithm, T is the transposition symbol, d() is the feature dimension function, R j refers to the j-th dimension feature of the second pooled acoustic feature in the pooled acoustic feature group, P j It refers to the j-th dimension feature of the third pooled acoustic feature in the pooled acoustic feature group.
7. The DAS-based water affairs detection data analysis system according to claim 1, characterized in that: When the model training module trains the preset acoustic water affairs model into a water affairs analysis model using the water area acoustic standard features and the historical water affairs data of the water area, it is specifically used to: Using a preset acoustic water affairs model to perform regression analysis on the acoustic standard characteristics of the water area, to obtain predicted water affairs data of the water area; The marginal loss value between the predicted water affairs data of the water area and the historical water affairs data of the water area is calculated using the following marginal loss function: Wherein, L refers to the edge loss value, k refers to the data sequence number, K refers to the total number of data, and the total number of data of the predicted water affairs data in the water area is equal to the total number of data of the historical water affairs data in the water area, D k It refers to the kth data value in the water affairs data of the predicted water area, Y k refers to the kth data value in the historical water affairs data of the water area, ε is a preset constant, λ is a preset loss weight, and Δ is the Laplace operator symbol; The model parameters of the acoustic water service model are iteratively updated according to the edge loss value to obtain a water service analysis model.
8. The DAS-based water affairs detection data analysis system according to claim 7, characterized in that: The model training module is specifically used to perform regression analysis on the acoustic standard characteristics of the water area using the preset acoustic water affairs model to obtain predicted water affairs data of the water area: Performing residual convolution on the acoustic standard characteristics of the water area using a preset acoustic water affairs model to obtain residual acoustic characteristics; Using the acoustic water service model to regress and map the residual acoustic features to obtain mapped acoustic features; Using the acoustic water service model, the mapped acoustic features are multi-level upsampled to obtain primary water service data; The acoustic water affairs model is used to perform convolution and zero filling on the primary water affairs data to obtain predicted water affairs data of the water area.
9. A water affairs detection data analysis method based on DAS, characterized in that: The method comprises: Acquire historical water area acoustic data and historical water affairs data of the target water area respectively, and perform filtering, noise removal and sensor synchronization on the historical water area acoustic data in turn to obtain primary water area acoustic data; Synchronizing the primary acoustic data of the water area into synchronized acoustic data of the water area according to the historical water affairs data of the water area, and performing self-attention denoising on the synchronized acoustic data of the water area to obtain standard acoustic data of the water area; Performing multi-dimensional feature extraction on the water area standard acoustic data to obtain a water area acoustic feature group, and performing cross-modal feature fusion on the water area acoustic feature group to form a water area acoustic standard feature; Using the acoustic standard features of the water area and the water affairs data of the historical water area to train a preset acoustic water affairs model into a water affairs analysis model; The target water area is detected by real-time distributed acoustic sensing to obtain real-time water area acoustic data, and the water affairs analysis model is used to analyze the real-time water affairs data corresponding to the real-time water area acoustic data.
10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the DAS-based water detection data analysis system according to any one of claims 1 to 8 is implemented.
Citation Information
Cited By
Water affair detection data analysis system and method based on DAS and medium
CN120653925A
