Multi-sensor water quality automatic detection dynamic feedback system and method

Abstract

The invention discloses a multi-sensor water quality automatic detection dynamic feedback system and method, relates to the technical field of automatic monitoring, and is used for solving the problems of accuracy and timeliness of water quality monitoring in a complex water area environment. By reasonably arranging the sensors and carrying out data fusion and evaluation, efficient monitoring and accurate judgment of water quality are realized. Firstly, the system selects and arranges multiple types of sensors according to factors such as water area characteristics and climate conditions, and collects water quality parameter data; and secondly, through a data processing and fusion module, a dynamic time warping algorithm is used to align time sequence data of water quality parameters, and a weighted summation method is combined to complete fusion processing of different water quality data. And finally, the water quality evaluation and decision-making module performs comprehensive analysis on the fused data by adopting an evidence theory, judges the water quality grade according to the maximum probability, and provides dynamic feedback according to a monitoring result.

Description

Technical Field

[0001] The present invention relates to the field of automatic monitoring technology, and more specifically, to a multi-sensor water quality automatic detection dynamic feedback system and method. Background Art

[0002] With the acceleration of industrialization and the increasing severity of environmental pollution, water pollution has become an environmental problem that needs to be solved urgently worldwide. Water quality monitoring, as an important means of assessing water pollution, plays a vital role in environmental protection, resource management and public health. However, traditional water quality monitoring methods have problems such as low monitoring frequency, narrow coverage, and poor adaptability to complex water environments. In particular, conventional sensors are difficult to achieve accurate monitoring under the influence of multiple factors such as water flow velocity, temperature gradient, and water geology.

[0003] Existing technologies often rely on fixed-position sensor layouts and cannot dynamically adjust monitoring strategies, resulting in blind spots in data collection and affecting the accuracy and timeliness of monitoring results. Especially in some special waters, such as deep sea or complex waters, factors such as water flow intensity and geological characteristics have a greater impact on water quality, and traditional monitoring technologies are difficult to provide real-time and comprehensive data feedback.

[0004] In view of the above problems, the present invention proposes a solution. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a multi-sensor water quality automatic detection dynamic feedback system and method to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions: In a preferred embodiment, it comprises: Step 1: Divide the detected water area into multiple separate water areas and assign different sensor layout schemes according to different separate water areas; Step 2: Screen the fusionable waters and use the data fusion technology SVM to integrate the outputs of sensors in the fusionable waters; Step 3: Use DS evidence theory to synthesize the evidence from multiple sensors and finally determine the water quality level.

[0007] In a preferred embodiment, in step 1, the detected water area is physically divided evenly or unevenly; According to the water flow speed and direction in each individual water area, the CFD model outputs the quantitative value of the hydrodynamic stratification intensity; the pH value in the individual water area is monitored in real time, the upper and lower limits of the critical pH value are found, and the critical pH value range in each individual water area is determined; the changes in the methane gas concentration in each individual water area are monitored, and the decomposition rate of natural gas hydrate in each individual water area is calculated; the temperature and pressure data around the hydrothermal vents in each individual water area are monitored and recorded as the activity intensity data of the seafloor hydrothermal vents; The comprehensive regional score Sf of each individual water area is determined by weighted average of the quantitative values ​​of the hydrodynamic stratification intensity, the critical pH value range of the colloidal phase change, the decomposition rate of the natural gas hydrate, and the activity intensity of the seafloor hydrothermal vents. The comprehensive regional score Sf of each individual water area is compared with the water area score threshold Yz. When the comprehensive regional score Sf of the individual water area is greater than or equal to the water area score threshold Yz, it indicates that the individual water area is a high-dynamic water area; when the comprehensive regional score Sf of the individual water area is less than the water area score threshold Yz, it indicates that the individual water area is a low-dynamic water area.

[0008] In a preferred embodiment, in step 1, the overlapping waters of high-dynamic waters and low-dynamic waters are determined, and when the flow velocity change rate of the overlapping waters is greater than the flow velocity change threshold value YL, it is recorded as a high-dynamic overlapping waters, and the sensor configuration scheme of the high-dynamic waters is preferentially adopted for the high-dynamic overlapping waters; when the flow velocity change rate of the overlapping waters is less than the flow velocity change threshold value YL, it is recorded as a low-dynamic overlapping waters, and the sensor configuration scheme of the low-dynamic waters is preferentially adopted for the low-dynamic overlapping waters; when the flow velocity change rate of the overlapping waters is equal to YL, this type of overlapping waters is recorded as an S waters, and a hybrid sensor scheme is adopted in the S waters.

[0009] In a preferred embodiment, in step 2, a multi-parameter dynamic time warping algorithm is used to calculate the DTW distance of water quality parameter data, and a weighted sum is performed to calculate the final weighted distance value; based on the historical data K-means clustering, a fusion threshold Yr is set, and when the final weighted distance value is less than or equal to the fusion threshold Yr, it indicates that the separate water area is a fusionable water area, and when the final weighted distance value is greater than the fusion threshold Yr, it indicates that the separate water area is an independently processed water area; In the fusionable water area, the vector machine SVM is used to perform preliminary classification processing on the water quality parameter data and output the posterior probability after classification; The correlation between the water quality parameter data of each sensor was calculated by Pearson correlation coefficient, and the weighted average calculation method was used to deal with conflicting evidence; The water quality parameter data obtained in the water area is independently processed and double-sliced ​​by time and space; the national secret SM4 algorithm is used to encrypt the data block, and the key is dynamically generated by the local security chip and stored in isolation.

[0010] In a preferred embodiment, in step 3, the DS evidence theory is used to convert the posterior probability output by the SVM into the basic distribution probability BPA required by the DS evidence theory. Based on the synthesized BPA, a set of water quality levels is set, and the weighted BPA value of each water quality level is calculated by the weighted average decision rule, and the water quality level with the highest BPA value is selected as the final judgment.

[0011] In a preferred embodiment, the method for obtaining the water area score threshold Yz is as follows: The mean, standard deviation, median and quartiles of the hydrodynamic stratification intensity, critical pH range of colloidal phase change, natural gas hydrate decomposition rate and seafloor hydrothermal vent activity intensity data were calculated. K-means clustering was used to analyze the mean, standard deviation, median and quartiles of the influencing parameters to determine the water area score threshold Yz.

[0012] In a preferred embodiment, the flow rate change threshold value YL is obtained as follows: The multi-frequency acoustic current meter AquaRAP was used to measure the water velocity in the overlapping waters, the velocity change rate L1 per unit time was calculated, and the velocity change threshold YL was set.

[0013] In a preferred embodiment, it includes: a data acquisition module, a data fusion module, a decision module, and a signal connection between the modules of the GIS module; The GIS module is mainly used to generate detailed water area geographic information maps based on the detected water area in combination with the water flow simulation model FVCOM, and to clarify the geographical boundaries of the entire detected water area; The data acquisition module is mainly used to divide the detected water area into multiple separate water areas and allocate different sensor layout schemes according to different separate water areas; The data fusion module is mainly used to screen the fusionable waters and integrate the outputs of sensors in the fusionable waters using the data fusion technology SVM; The decision-making module mainly evaluates water quality based on the fused data and uses evidence theory to output the final water quality grade.

[0014] The present invention discloses a multi-sensor water quality automatic detection dynamic feedback system and method, which relates to the field of automatic monitoring technology, and is used to solve the accuracy and timeliness problems of water quality monitoring in complex water environments; by reasonably arranging sensors and performing data fusion and evaluation, efficient monitoring and accurate judgment of water quality are achieved. First, the system selects and arranges multiple types of sensors according to factors such as water characteristics and climatic conditions to collect water quality parameter data. Secondly, through the data processing and fusion module, the dynamic time warping algorithm is used to align the time series data of water quality parameters, and the weighted summation method is combined to complete the fusion processing of different water quality data. Finally, the water quality assessment and decision-making module uses evidence theory to conduct a comprehensive analysis of the fused data, judges the water quality level according to the maximum probability, and provides dynamic feedback based on the monitoring results. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a schematic diagram of the structure of the multi-sensor water quality automatic detection dynamic feedback system of the present invention.

[0016] Figure 2 This is an operation flow chart of the multi-sensor water quality automatic detection dynamic feedback method of the present invention. DETAILED DESCRIPTION

[0017] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] Example The present invention discloses a multi-sensor water quality automatic detection dynamic feedback method, such as Figure 2 As shown, including: Step 1: Collect the water quality parameter data of pH, conductivity and temperature in the tested water area; The detected waters are physically divided into multiple separate waters. Different sensor layout schemes are selected in different separate waters according to the internal factors of each separate water, such as the hydrodynamic stratification intensity, the critical pH of the colloidal phase change, and the external factors of the waters, such as the decomposition rate of natural gas hydrates and the intensity of the activity of the seafloor hydrothermal vents. In addition, each sensor is connected through a wireless sensor network in each separate water. First, the GIS module ArcGIS is used in combination with the water flow simulation model FVCOM to perform spatial analysis on the inspected water area, generate a detailed water area geographic information map, and clarify the geographical boundaries of the entire inspected water area; secondly, the entire inspected water area is physically divided evenly or unevenly in the GIS module. For example, a water area with a volume of 100 cubic meters is evenly physically divided into four separate water areas with a volume of 25 cubic meters, and each separate water area is numbered, for example, S1, S2, S3; its geographical location in the entire inspected water area is recorded to facilitate the deployment of sensors in separate water areas.

[0019] Furthermore, for each individual water area, the velocity and direction of the water flow in each individual water area are measured by a flow meter and a flow direction sensor, and the collected velocity and direction data are input into a computational fluid dynamics (CFD) model to simulate the stratification of the water body. Then, the CFD model outputs the quantitative values ​​of the hydrodynamic stratification intensity, the number of stratification layers and the thickness, and generates a hydrodynamic stratification intensity report for each individual water area. At the same time, a pH sensor is used to monitor the pH value in the individual water area in real time, and the real-time pH value is compared with the experimental data in combination with the general colloidal phase change experimental results to find the upper and lower limits of the critical pH value and determine the critical pH value range in each individual water area. Then, a gas sensor is used to monitor the methane gas in each individual water area. The gas concentration changes and the gas concentration change rate are calculated. Combined with geological data, including formation pressure and temperature, the pressure coefficient and temperature coefficient in the geological data are calculated according to the formula: Xs=d1×(dz1-dz0), where Xs represents the pressure or temperature coefficient, d1 is a fitting constant, dz1 represents the current pressure or temperature, and dz0 represents the reference pressure or temperature. Then the gas hydrate decomposition rate in each separate water area is determined by the formula: gas hydrate decomposition rate=gas concentration change rate×(pressure coefficient+temperature coefficient). At the same time, the temperature and pressure data around the hydrothermal vents in each separate water area are monitored using temperature sensors and pressure sensors, and the changes in temperature and pressure are recorded. Furthermore, mathematical formulas or empirical formulas are used to normalize the hydrodynamic stratification intensity, the critical pH range of colloidal phase transition, the decomposition rate of natural gas hydrates, and the intensity of seafloor hydrothermal vent activity data and convert them into quantitative values. The specific formula is: In the formula, is the normalized value; then, the comprehensive regional score Sf of each individual water area is determined by weighted average calculation according to the quantified values ​​of hydrodynamic stratification intensity, critical pH value range of colloidal phase change, natural gas hydrate decomposition rate, and seafloor hydrothermal vent activity intensity data, specifically according to the formula: Sf=Q1×Sc+Q2×Jt+Q3×Tf+Q4×Ry, where Sc represents the quantified value of hydrodynamic stratification intensity, Q1 represents the weight of the quantified value of hydrodynamic stratification intensity, Jt represents the quantified value of the critical pH value range of colloidal phase change, Q2 represents the weight of the quantified value of the critical pH value range of colloidal phase change, Tf represents the quantified value of natural gas hydrate decomposition rate, Q3 represents the weight of the quantified value of natural gas hydrate decomposition rate, Ry represents the quantified value of seafloor hydrothermal vent activity intensity, and Q4 represents the weight of the quantified value of seafloor hydrothermal vent activity intensity; Next, the data cleaning tool Pandas library was used to remove missing values ​​and duplicate data from the collected data on hydrodynamic stratification intensity, critical pH range of colloidal phase change, decomposition rate of natural gas hydrate, and activity intensity of seafloor hydrothermal vents, and a clean and complete data set of influencing parameters was generated; the mean, standard deviation, median and quartile of the data on hydrodynamic stratification intensity, critical pH range of colloidal phase change, decomposition rate of natural gas hydrate, and activity intensity of seafloor hydrothermal vents were calculated, and K-means clustering was used to analyze the mean, standard deviation, median and quartile of the influencing parameters to determine the water area score threshold Yz, and the comprehensive regional score Sf of each individual water area was compared with the water area score Yz. When the comprehensive regional score Sf of an individual water area is greater than or equal to the water area score threshold Yz, it indicates that the individual water area belongs to a high-dynamic water area; when the comprehensive regional score Sf of an individual water area is less than the water area score threshold Yz, it indicates that the individual water area belongs to a low-dynamic water area; Sensor layout plan for highly dynamic waters: Use solid-state pH electrode sensors with temperature compensation, arrange them in a hexagonal close-packed pattern, and deploy liftable probes; when the temperature gradient of the monitoring channel is greater than 0.5℃ / m or the conductivity change rate is greater than 10μS / cm / min, the sampling time needs to be switched to full parameter sampling 5 times per second for 5 minutes.

[0020] Sensor layout plan for low-dynamic waters: MEMS sensors are deployed in a fixed grid of 100m×100m. To avoid surface turbulence, sensor nodes are installed at a depth of 2m±0.3m. When the temperature difference between adjacent sensor nodes is detected to be >0.3℃ or the conductivity difference is >5%, all parameters are sampled once a minute for 1 hour.

[0021] Considering the diffusion of water body fluidity and water quality changes, it is impossible to achieve absolute isolation between individual water areas, resulting in overlapping water areas between individual water areas. Therefore, the overlapping water areas of high-dynamic water areas and low-dynamic water areas are first determined through the GIS module, and the water velocity of the overlapping water areas is measured using the multi-frequency acoustic current meter AquaRAP. The velocity change rate L1 per unit time is calculated, specifically based on the formula: L1=Δv / Δt, where Δv represents the change in velocity per unit time, and Δt represents the time interval; a velocity change threshold YL is set. When the velocity change rate of the overlapping water area is greater than YL, it indicates that the overlapping water area is close to a high-dynamic water area and is recorded as a high-dynamic overlapping water area. For the high-dynamic overlapping water area, the sensor configuration scheme for the high-dynamic water area is preferentially adopted; when When the velocity change rate of the overlapping water area is less than YL, it indicates that the overlapping water area is close to the low-dynamic water area, which is recorded as the low-dynamic overlapping water area. In this case, the sensor configuration scheme of the low-dynamic overlapping water area is preferentially adopted for the low-dynamic overlapping water area. When the velocity change rate of the overlapping water area is equal to YL, it indicates that there are both high-dynamic changes and low-dynamic changes in the overlapping water area. This type of overlapping water area is recorded as the S water area. A hybrid sensor scheme is adopted in the S water area: including a velocity meter sensor, a flow direction sensor, a dynamic sensor, and a pH sensor, a dissolved oxygen sensor and other static sensors; and an adaptive acquisition frequency based on sensor data analysis is set: according to the change of water flow velocity in the S water area, the water flow prediction model based on machine learning integrated in the sensor network is used to automatically adjust the acquisition frequency of each sensor. Furthermore, the Pandas library is used to clean the water quality parameter data collected by each sensor, and the precise time protocol PTP is used to synchronize the time of each sensor node to ensure the consistency of the water quality parameter data of each sensor.

[0022] Step 2: Use data fusion technology to integrate the outputs of multiple sensors; The multi-parameter dynamic time warping algorithm is used to calculate the DTW distance of water quality parameter data and weighted sum it: the specific formula is: Dmdtw=Qp×Dph+Qd×Dddl+Qc×Dc+Ql×lj, where Dph represents the DTW distance of pH data, Dddl represents the DTW distance of conductivity, Dc represents the DTW distance of temperature data, Qp represents the DTW distance weight of pH data, Qd represents the DTW distance weight of conductivity, Qc represents the DTW distance weight of temperature data, lj represents the path penalty term, constraining the slope of the timing alignment path to be ≤2 to prevent excessive distortion, Ql represents the weight of the path penalty term, and Dmdtw represents the final weighted distance value; Based on the K-means clustering of historical data, 75% of the minimum spacing is set as the fusion threshold Yr. When Dmdtw is less than or equal to the fusion threshold Yr, it means that the single water area is a fusion water area. When Dmdtw is greater than the fusion threshold Yr, it means that the single water area is an independent processing water area. In the fusionable water area, the vector machine (SVM) is first used to perform preliminary classification processing on the water quality parameter data.

[0023] Specifically, the particle swarm optimization algorithm PSO is used to optimize the penalty factor and kernel parameter of SVM for the water quality parameter data collected by each sensor. By finding the optimal penalty factor C and kernel parameter in the parameter space, the classification error is minimized and the classification accuracy is maximized. The water quality parameter data is classified by the optimized SVM, and the posterior probability after classification is output. The SVM classification formula is: Among them, f(x) represents the posterior probability after classification of water quality parameter data, ω represents the weight vector of the hyperplane, ϕ(x) represents the high-dimensional feature space after mapping, and b represents the bias.

[0024] It should be noted that when there is conflicting evidence in the water quality parameter data from different sensors, the correlation between the water quality parameter data of each sensor is calculated by the Pearson correlation coefficient, and the correlation matrix of evidence is constructed according to the formula:

[0025] Among them, r represents the support of sensor water quality parameter data, Xi and Yi represent the sampling values ​​of water quality parameter data set, and Indicates the average value of the sampled values; According to the support r of the sensor water quality parameter data in the correlation matrix, the weighted average calculation method is used to process the conflicting evidence, according to the formula: Among them, w represents the weighted value of the support of water quality parameter data, ri represents the weight of each piece of evidence, and xi represents the evidence value corresponding to each piece of evidence; The water quality parameter data obtained in the independent processing water area is double-sharded by time and space; the national secret SM4 algorithm is used to encrypt the data block, and the key is dynamically generated by the local security chip and stored in isolation; Step 3: Use DS evidence theory to synthesize the evidence from multiple sensors and finally determine the water quality level.

[0026] First, it is necessary to transform the posterior probability of the water quality parameter data collected by each sensor to construct the basic probability BPA required by the DS evidence theory.

[0027] Specifically, a basic event set G={A1, A2, A3...} is set for each water quality parameter data to represent different water quality levels; the posterior probability value output by each sensor data is mapped to a probability distribution m(G) of a basic event, that is: m(Ai)=f(xi), where i=(1, 2, 3...), m(Ai) represents the probability BPA of occurrence of water quality level Ai, and xi represents the posterior probability of the i-th water quality parameter data.

[0028] Dempster's combination rule is used to synthesize two BPAs. The formula is as follows: Among them, m(Ai) and m(Bi) represent the basic probability distribution from different sensors, and m(Ai)⊕m(Bi) represents the combined BPA; based on the synthesized BPA, the water quality level set H={Grade A, Grade B, Grade C...} is set, and the weighted BPA value of each water quality level is calculated by the weighted average decision rule, according to the formula:

[0029] in, Wn represents the basic probability distribution value of each sensor for the water quality level, and Wn represents the weight of each sensor; the weighted BPA value corresponding to each water quality level (such as A, B, C, etc.) is obtained, and the water quality level with the highest BPA value is selected as the final judgment; for example, the final calculated weighted BPA value is: BPAf={A: 0.7, B: 0.2, C: 0.1}, and the result will be output as A-level water quality.

[0030] The present invention also discloses a multi-sensor water quality automatic detection dynamic feedback system to implement the method described in the above embodiment, such as Figure 1 As shown, it includes: data acquisition module, data fusion module, decision module and signal connection between GIS modules; The GIS module is mainly used to generate detailed water area geographic information maps based on the detected water area in combination with the water flow simulation model FVCOM, and to clarify the geographical boundaries of the entire detected water area; The data acquisition module is mainly used to divide the detected water area into multiple separate water areas and allocate different sensor layout schemes according to different separate water areas; The data fusion module is mainly used to screen the fusionable waters and integrate the outputs of sensors in the fusionable waters using the data fusion technology SVM; The decision-making module mainly evaluates water quality based on the fused data and uses evidence theory to output the final water quality grade.

[0031] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0032] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0033] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0034] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0035] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0036] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. Multi-sensor water quality automatic detection dynamic feedback method, It is characterized by; including: Step 1: Divide the detected water area into multiple separate water areas and assign different sensor layout schemes according to different separate water areas; Step 2: Screen the fusionable waters and use the data fusion technology SVM to integrate the outputs of sensors in the fusionable waters; Step 3: Use DS evidence theory to synthesize the evidence from multiple sensors and finally determine the water quality level.

2. The multi-sensor water quality automatic detection dynamic feedback method according to claim 1 is characterized by: In step 1, the inspected water area is physically divided uniformly or non-uniformly; According to the water flow speed and direction in each individual water area, the CFD model outputs the quantitative value of the hydrodynamic stratification intensity; the pH value in the individual water area is monitored in real time, the upper and lower limits of the critical pH value are found, and the critical pH value range in each individual water area is determined; the changes in the methane gas concentration in each individual water area are monitored, and the decomposition rate of natural gas hydrate in each individual water area is calculated; the temperature and pressure data around the hydrothermal vents in each individual water area are monitored and recorded as the activity intensity data of the seafloor hydrothermal vents; The comprehensive regional score Sf of each individual water area is determined by weighted average of the quantitative values ​​of the hydrodynamic stratification intensity, the critical pH value range of the colloidal phase change, the decomposition rate of the natural gas hydrate, and the activity intensity of the seafloor hydrothermal vents. The comprehensive regional score Sf of each individual water area is compared with the water area score threshold Yz. When the comprehensive regional score Sf of the individual water area is greater than or equal to the water area score threshold Yz, it indicates that the individual water area is a high-dynamic water area; when the comprehensive regional score Sf of the individual water area is less than the water area score threshold Yz, it indicates that the individual water area is a low-dynamic water area.

3. The multi-sensor water quality automatic detection dynamic feedback method according to claim 2 is characterized by: In step 1, the overlapping waters of high dynamic waters and low dynamic waters are determined. When the velocity change rate of the overlapping waters is greater than the velocity change threshold value YL, it is recorded as a high dynamic overlapping waters, and the sensor configuration scheme of the high dynamic waters is preferentially adopted for the high dynamic overlapping waters; when the velocity change rate of the overlapping waters is less than the velocity change threshold value YL, it is recorded as a low dynamic overlapping waters, and the sensor configuration scheme of the low dynamic waters is preferentially adopted for the low dynamic overlapping waters; when the velocity change rate of the overlapping waters is equal to YL, this type of overlapping waters is recorded as an S waters, and a hybrid sensor scheme is adopted in the S waters.

4. The multi-sensor water quality automatic detection dynamic feedback method according to claim 3, characterized in that; In step 2, a multi-parameter dynamic time warping algorithm is used to calculate the DTW distance of water quality parameter data, and the weighted sum is used to calculate the final weighted distance value; based on the K-means clustering of historical data, a fusion threshold Yr is set. When the final weighted distance value is less than or equal to the fusion threshold Yr, it indicates that the separate water area is a fusionable water area. When the final weighted distance value is greater than the fusion threshold Yr, it indicates that the separate water area is an independently processed water area. In the fusionable water area, the vector machine SVM is used to perform preliminary classification processing on the water quality parameter data and output the posterior probability after classification; The correlation between the water quality parameter data of each sensor was calculated by Pearson correlation coefficient, and the weighted average calculation method was used to deal with conflicting evidence; The water quality parameter data obtained in the water area is independently processed and double-sliced ​​by time and space; the national secret SM4 algorithm is used to encrypt the data block, and the key is dynamically generated by the local security chip and stored in isolation.

5. The multi-sensor water quality automatic detection dynamic feedback method according to claim 4 is characterized in that: In step 3, the DS evidence theory is used to convert the posterior probability output by SVM into the basic assignment probability BPA required by the DS evidence theory. Based on the synthesized BPA, a set of water quality levels is set, and the weighted BPA value of each water quality level is calculated by the weighted average decision rule. The water quality level with the highest BPA value is selected as the final judgment.

6. The multi-sensor water quality automatic detection dynamic feedback method according to claim 2 is characterized by: The method for obtaining the water area score threshold Yz is as follows: The mean, standard deviation, median and quartiles of the hydrodynamic stratification intensity, critical pH range of colloidal phase change, natural gas hydrate decomposition rate and seafloor hydrothermal vent activity intensity data were calculated. K-means clustering was used to analyze the mean, standard deviation, median and quartiles of the influencing parameters to determine the water area score threshold Yz.

7. The multi-sensor water quality automatic detection dynamic feedback method according to claim 3 is characterized by: The method for obtaining the flow rate change threshold YL is as follows: The multi-frequency acoustic current meter AquaRAP was used to measure the water velocity in the overlapping waters, the velocity change rate L1 per unit time was calculated, and the velocity change threshold YL was set.

8. A multi-sensor water quality automatic detection dynamic feedback system, used to implement the multi-sensor water quality automatic detection dynamic feedback method according to any one of claims 1 to 7 above, It is characterized by: Including: data acquisition module, data fusion module, decision module, signal connection between GIS modules; The GIS module is mainly used to generate detailed water area geographic information maps based on the detected water area in combination with the water flow simulation model FVCOM, and to clarify the geographical boundaries of the entire detected water area; The data acquisition module is mainly used to divide the detected water area into multiple separate water areas and allocate different sensor layout schemes according to different separate water areas; The data fusion module is mainly used to screen the fusionable waters and integrate the outputs of sensors in the fusionable waters using the data fusion technology SVM; The decision-making module mainly evaluates water quality based on the fused data and uses evidence theory to output the final water quality grade.

Images