Water quality monitoring system, device and method for marine ecological protection

Abstract

The invention discloses a water quality monitoring system, device and method for marine ecological protection, and belongs to the technical field of marine ecological environment monitoring, the system comprises an edge calculation and preprocessing module, the edge calculation and preprocessing module receives first data output by a sensor, carries out signal processing on the first data and then outputs second data; the block chain module is used for receiving the second data, encrypting the second data, generating a hash identifier and outputting third data; and the data analysis module receives the third data, completes water quality analysis, classification and abnormality diagnosis through clustering and decision tree algorithms, and outputs a monitoring result. According to the invention, multi-parameter, multi-depth and wide-area collaborative ocean water quality real-time monitoring is realized, and the space-time density and processing efficiency of data acquisition are remarkably improved through combination of the modular sensor array and the edge computing nodes.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of marine ecological environment monitoring, and particularly relates to a water quality monitoring system, device and method for marine ecological protection. BACKGROUND

[0002] Marine water quality monitoring technology provides important support for marine ecological protection through the deployment of sensor networks and data analysis platforms. In the past few decades, the field has continued to develop, and based on sensing and information fusion methods, researchers have developed various types of monitoring systems, such as buoy monitoring stations, seafloor observation networks, remote sensing monitoring platforms, and unmanned shipborne mobile monitoring systems. Generally speaking, an ideal intelligent water quality monitoring system needs to have functions such as multi-parameter synchronous acquisition, high-precision sensing, anti-interference transmission, and intelligent analysis, which puts high requirements on its hardware stability, communication reliability, and algorithm generalization ability. Although basic monitoring functions can be achieved through modular integration and traditional data transmission technology, such systems often face problems such as data drift, transmission delay, sensor corrosion, and high energy consumption in complex marine environments, resulting in insufficient monitoring continuity and reduced data reliability.

[0003] However, existing systems and algorithms still have limitations such as response lag, single evaluation dimension, weak model interpretability, and insufficient cross-platform compatibility, especially in the face of complex hydrological conditions and sudden pollution events, lacking comprehensive perception and decision-making linkage capabilities. So far, there is no full-link intelligent marine water quality monitoring system that can simultaneously achieve wide-area coverage, high-precision perception, intelligent early warning, and autonomous decision-making. SUMMARY

[0004] The present application aims to provide a water quality monitoring system, device and method for marine ecological protection.

[0005] To achieve the above-mentioned purpose, the technical scheme of the present application is: A water quality monitoring system for marine ecological protection, comprising, an edge computing and preprocessing module that accepts first data output by a sensor, performs signal processing on the first data, and outputs second data; a blockchain module that accepts the second data, encrypts and generates a hash identifier for the second data, and outputs third data; a data analysis module that accepts the third data, completes water quality analysis, classification, and anomaly diagnosis through clustering and decision tree algorithms, and outputs monitoring results.

[0006] Further, the edge computing and preprocessing module comprises, A microprocessor, which adopts an embedded processor, a real-time data cleaning algorithm based on a sliding window and a wavelet transform compression algorithm; the microprocessor completes data denoising filtering and feature extraction through embedded algorithms; A data acquisition terminal, which synchronously analyzes sensor signals; A distributed coordinator, which constructs a local area network to realize cross-layer instruction interaction.

[0007] Further, the blockchain module includes a communication unit, which encrypts each batch of data through a lightweight consensus mechanism, generates a unique hash identifier and a timestamp for the data, and the expression is: , wherein, D is the data content, T is the timestamp, TD is the node identifier.

[0008] Further, the data analysis module is built-in with a Hadoop distributed file system storage architecture, a water quality clustering analysis model based on a Mahout framework, and a CART decision tree classifier, which stores and reads water quality monitoring data through the Hadoop distributed file system storage architecture, and relies on the water quality clustering analysis model based on the Mahout framework and the CART decision tree classifier to complete multi-source data fusion, intelligent judgment of water quality grades, and pollution risk early warning.

[0009] Further, K-Means algorithm is used for clustering, and the expression is: , wherein, is the preset number of clusters, is the i-th cluster, is its centroid.

[0010] Further, it also includes an energy supply module, which includes a single-crystal silicon solar cell panel, a lithium ion battery pack, and an intelligent power management unit, and provides power energy for all hardware of the sensor module, the edge computing and preprocessing module, the blockchain module, and the data analysis module.

[0011] Further, it also includes a sensor module, which includes a water quality sensor, a pH sensor, and a dissolved oxygen sensor, which collects water quality parameters at different water depths and transmits the collected physical signals to the edge computing and preprocessing module.

[0012] ​Further, each sensor in the sensor module adopts a modular corrosion-resistant packaging structure, is coated with a bio-attachment-resistant material on the surface, and is subjected to signal conditioning and analog-to-digital conversion by a multi-channel data acquisition terminal.

[0013] A water quality monitoring device for marine ecological protection comprises, The energy supply module provides power energy for all hardware of the intelligent acquisition terminal; The intelligent acquisition terminal collects data through the sensor, performs edge computing and preprocessing, and then transmits the data after encryption; The buoy provides buoyancy, so that the entire device stably floats on the water surface.

[0014] A water quality monitoring method for marine ecological protection comprises, Step S1, collecting data of the sensor, and transmitting the collected signal to an edge computing and preprocessing module; Step S2, performing edge node preprocessing on the data; Step S3, after preprocessing, transmitting the data to a data analysis module through a blockchain module; Step S4, performing multi-modal fusion analysis on the data through the data analysis module to generate a water quality level evaluation result and a pollution risk map; Step S5, outputting early warning information to the user through a visual interface or an API interface.

[0015] Advantages, the water quality detection system for marine ecological protection realizes real-time monitoring of marine water quality with multiple parameters, multiple depths and wide-area cooperation, significantly improves the spatiotemporal density and processing efficiency of data acquisition by combining a modular sensor array with an edge computing node; the system realizes encrypted transmission and storage of monitoring data by means of blockchain technology, ensures that the data is real, reliable and traceable, and effectively overcomes the defects of easy tampering and poor transmission security of data in traditional systems; by introducing machine learning algorithms and intelligent decision-making models, the system can accurately evaluate and early warn of complex water quality conditions, providing scientific basis and decision support for marine ecological protection; the system has the characteristics of low power consumption, self-power supply, corrosion resistance and bio-attachment resistance, and can work stably in harsh marine environments for a long time, and is suitable for various application scenarios such as offshore, port and breeding area; the system uses an edge computing and cloud data analysis collaborative computing architecture to effectively reduce the data transmission bandwidth demand and cloud computing load, while supporting independent operation of edge nodes and temporary storage of data in a disconnected environment, significantly enhancing the adaptability and reliability of the system in the open sea and weak signal environment; the system not only realizes intelligent upgrading of water quality monitoring, but also provides efficient and reliable data support for marine ecological environment protection, disaster warning, emergency response and other business applications, and has important social and ecological values.

[0016] For the above features and advantages of the invention can be more obvious and easy to understand, below the embodiment, and with the accompanying drawings are described in detail as follows. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 For the structure of the water quality monitoring device for marine ecological protection of the present application.

[0018] Figure 2 For the structure of the intelligent acquisition terminal 2.

[0019] Figure 3 For the schematic diagram of the Hadoop distributed file system storage architecture in the data analysis module.

[0020] Figure 4 For the 6-mean water quality data clustering convergence process based on the K-Means algorithm.

[0021] Figure 5 For the water quality parameter association rule mining and FP tree construction based on the FP-Growth algorithm.

[0022] Figure 6 For the flow chart of the water quality detection method for marine ecological protection of the present application. DETAILED DESCRIPTION

[0023] In order to make the purpose and technical scheme of the embodiments of the present application more clear, the technical scheme of the embodiments of the present application will be described clearly and completely below in combination with the drawings of the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all the embodiments. Based on the described embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the present application.

[0024] Figure 1 For the structure of the water quality monitoring device for marine ecological protection of the present application. As Figure 1 shown, a water quality monitoring device for marine ecological protection, comprising: energy supply module 1, intelligent acquisition terminal 2, buoy 3, the device is laid in the target sea area to collect water quality parameters of different water depth.

[0025] More specifically, the energy supply module 1 provides power energy for the hardware of the intelligent acquisition terminal 2, which is the guarantee for the operation of the whole monitoring device; the intelligent acquisition terminal 2 collects data through the sensor, performs edge computing and preprocessing, and then transmits the data after encryption; the buoy 3 provides buoyancy, so that the whole device is stably floating on the water surface.

[0026] Figure 2 For the structure of the intelligent acquisition terminal 2, as Figure 2As shown, the intelligent acquisition terminal 2 is divided into three layers, the A layer includes a communication unit, a distributed coordinator, a microprocessor, and a data acquisition terminal; the B layer includes a water quality sensor, and the C layer includes a PH sensor and a dissolved oxygen sensor.

[0027] Further, a water quality monitoring system for marine ecological protection includes an energy supply module, a sensor module, an edge computing and preprocessing module, a blockchain module, and a data analysis module. The sensor module, the edge computing and preprocessing module, and the blockchain module are integrated in the intelligent acquisition terminal 2; the data analysis module is deployed in the cloud.

[0028] More specifically, the energy supply module provides power energy for all hardware of the sensor module, the edge computing and preprocessing module, the blockchain module, and the data analysis module, which guarantees the operation of the entire monitoring system.

[0029] Further, the sensor module collects water quality parameters at different water depths and transmits the collected signals to the edge computing and preprocessing module.

[0030] Further, the edge computing and preprocessing module accepts the first data output by the sensor, processes the first data, and outputs the second data.

[0031] Further, the blockchain module accepts the second data, encrypts and generates a hash identifier for the second data, and outputs the third data.

[0032] Further, the data analysis module accepts the third data, completes water quality analysis, classification, and anomaly diagnosis through clustering and decision tree algorithms, and outputs the monitoring results.

[0033] Further, the energy supply module includes a monocrystalline silicon solar cell panel, a lithium ion battery pack, and an intelligent power management unit (PMU).

[0034] Further, the output power of the solar panel can be represented as: , wherein, is the conversion efficiency, is the effective area, is the solar irradiance.

[0035] Further, the intelligent power management unit dynamically switches the power supply mode according to the system load and supports remote monitoring and scheduling.

[0036] Further, the energy supply module needs to meet the energy balance of the system, and the relationship can be represented by the following formula: , wherein, is the energy balance of the system in a time period Energy generated by the inner solar panels, and are the period The remaining energy of the battery at the beginning and end, is the total energy consumption of the system in the period .

[0037] Further, the intelligent power management unit dynamically manages the collection, storage and consumption of energy according to the energy balance relationship, ensuring that the system can still work continuously in rainy weather. The total energy consumption of the system can be estimated as: , where , , are the average power consumption of the sensor module, edge computing and preprocessing module and communication unit, respectively.

[0038] Further, the sensor module includes water quality sensors, pH sensors, and dissolved oxygen sensors.

[0039] Further, each sensor adopts a modular corrosion-resistant packaging structure, is coated with a bio-attachment-resistant material on the surface, and is subjected to signal conditioning and analog-to-digital conversion through a multi-channel data acquisition terminal.

[0040] Further, the edge computing and preprocessing module includes a microprocessor, a data acquisition terminal, and a distributed coordinator.

[0041] Further, the microprocessor is an embedded processor, which can be an ARM Cortex-A53, and has a real-time data cleaning algorithm and a wavelet transform compression algorithm based on a sliding window built-in. The microprocessor completes data denoising filtering and feature extraction through embedded algorithms.

[0042] More specifically, the real-time data cleaning algorithm uses an adaptive threshold filtering method based on a sliding window, and the specific principle is as follows: First, define the sliding window, let the current time be , the sliding window contains the last sampling points, denoted as: , where is the sampling point at the time, and is the original water quality parameter reading collected by the sensor at the time.

[0043] Further, the dynamic threshold is calculated, for the sliding window , the mean and the standard deviation are calculated, and the expression is: , , in, For the first Raw sensor readings at each moment, such as pH value, dissolved oxygen concentration, etc.

[0044] Furthermore, the dynamic threshold interval is defined as: , in, This is the sensitivity coefficient, set according to the sensor type and environmental stability.

[0045] Furthermore, outliers are detected and corrected for sampling points. : like If the value is not found, it is considered normal data and the original value is retained. like If the value exceeds the threshold, it is identified as an outlier and corrected using the median within the window. , Simultaneously, anomaly flags are recorded for subsequent sensor health status assessment.

[0046] Furthermore, to reduce computational complexity, a recursive update mechanism is used to update the mean and standard deviation: , , This recursive formula can update window statistics in a constant time, meeting the real-time requirements of edge devices.

[0047] Furthermore, to address the issue of slow sensor drift, the algorithm introduces an exponentially weighted moving average (EWMA) as a baseline: , in, for The exponentially weighted moving average at time t, As a smoothing factor, for The exponentially weighted moving average at time 1 is the dynamic baseline of the previous time step.

[0048] Furthermore, the baseline drift is simultaneously deducted during the cleaning process: , in, The drift calibration cycle is typically calculated using 24 hours of data points.

[0049] More specifically, the specific calculations of the wavelet transform compression algorithm are as follows: Let the original data sequence collected by the sensor be... After wavelet transform compression, the following is obtained: , in, For wavelet basis functions, For scale parameters, This is a translation parameter. The compressed data packet size is reduced by approximately 60%, significantly reducing the transmission load.

[0050] Furthermore, the data acquisition terminal simultaneously analyzes the sensor signals.

[0051] Furthermore, the distributed coordinator constructs a local area network to enable cross-layer command interaction.

[0052] Furthermore, the blockchain module includes a communication unit. After receiving data from the edge computing and preprocessing module, the blockchain module encrypts each batch of data using a lightweight consensus mechanism, generating a unique hash identifier and timestamp for the data. The expression is: , in, D For data content, T For timestamps, TD This is the node identifier.

[0053] More specifically, it ensures the immutability of data. If data is modified during transmission, its hash identifier will change synchronously. The data analysis module can quickly identify data anomalies by verifying the hash value, ensuring the authenticity and integrity of transmitted data and avoiding data distortion caused by human or environmental interference.

[0054] The data analysis module incorporates a Hadoop Distributed File System (HDFS) storage architecture, a water quality clustering analysis model based on the Mahout framework, and a CART decision tree classifier. It stores and retrieves water quality monitoring data through the Hadoop Distributed File System storage architecture and completes multi-source data fusion, intelligent water quality level determination, and pollution risk early warning based on the water quality clustering analysis model based on the Mahout framework and the CART decision tree classifier.

[0055] Figure 3This diagram illustrates the Hadoop Distributed File System (HDFS) storage architecture within the data analysis module. Within this architecture, clients initiate file access requests to the NameNode, which manages metadata. The NameNode queries the metadata and returns the file storage location to the client. Simultaneously, the NameNode continuously interacts with multiple DataNodes (DNodes) that store actual data blocks, issuing control commands and reporting node status. DataNodes also synchronize and replicate data blocks. Finally, based on the obtained storage location information, the client directly reads the file data blocks from the corresponding DataNode, enabling efficient and reliable access to massive amounts of data and ensuring high system reliability and throughput.

[0056] Furthermore, the water quality clustering analysis model based on the Mahout framework is specifically as follows: let the water quality parameter vector be... Clustering is performed using the K-Means algorithm, and its expression is: , in, The preset number of clusters, For the first One cluster, Its center of mass.

[0057] Furthermore, the objective function of the K-Means clustering algorithm is to minimize the sum of squared distances from all sample points to their respective cluster centers, and its expression is: , in, The preset number of clusters, The total number of water quality parameter samples. Representing the The feature vector of each sample point Representing the The center point of each cluster, It is a binary indicator variable, when the sample points Belongs to the Its value is 1 when a cluster is selected, and 0 otherwise. It is continuously updated through iterative optimization. and until the objective function The data converges, thus completing the clustering and classification of water quality data, and enabling automatic identification of different water quality levels.

[0058] Figure 4 This diagram illustrates the convergence process of clustering 6-means water quality data based on the K-Means algorithm. Figure 4As shown, by iteratively calculating the distance between each data point and the cluster center, reassigning cluster types, and updating the cluster centers, stable convergent cluster centers are finally obtained. This process enables automatic grouping of water quality data of different states, providing a cluster analysis basis for the system to complete water quality level determination and abnormal pollution area identification.

[0059] Meanwhile, the FP-Growth algorithm is used to mine the correlation rules between parameters and generate a rule library for the simultaneous occurrence of low dissolved oxygen and high turbidity, so as to diagnose water quality anomalies. Figure 5 This is a schematic diagram of water quality parameter association rule mining and FP-tree construction based on the FP-Growth algorithm, as shown below. Figure 5 As shown, the core implementation process of mining water quality parameter association rules based on the FP-Growth algorithm mainly includes raw transaction data, an item header table, an FP tree, and a node linked list. The raw data area uses 10 transaction records to simulate multi-dimensional water quality parameter combinations at different monitoring times (e.g., A, B, C, D, and E respectively map to key water quality indicators such as pH, dissolved oxygen, and conductivity). The item header table counts the frequency of occurrence of each water quality parameter and arranges them in descending order, establishing associations with corresponding nodes in the FP tree through pointers. The FP tree uses null as the root node and constructs a tree structure according to frequency priority, transforming the discrete water quality parameter transaction data into a compact tree storage structure. The node linked list realizes the horizontal connection of nodes with the same water quality parameter in the FP tree, supporting the algorithm to quickly mine frequent co-occurrence patterns between water quality parameters by traversing the linked list, thereby identifying potential associations between pollutants.

[0060] Furthermore, the CART decision tree classifier uses Gini impurity as the criterion for selecting feature split points during its construction. For the dataset... D The formula for calculating the impurity of the Gini coefficient is: , in, The number of sample classes in the dataset. For the first Class samples in the dataset D The proportion it accounts for.

[0061] Furthermore, if the dataset is divided according to a certain value of feature A... D Split into two subsets D 1 and D 2. The overall Gini impurity after the split is: , Furthermore, the algorithm selection makes The smallest feature and its split point are used as the optimal splitting scheme to recursively construct a decision tree, ultimately achieving the classification and pollution risk warning of newly collected water quality data.

[0062] Figure 6 This is a flowchart of a water quality testing method for marine ecological protection according to the present invention, as shown below. Figure 6 As shown, the process of a water quality testing method for marine ecological protection according to the present invention includes the following steps.

[0063] Step S1: Collect data from the sensor and transmit the collected signals to the edge computing and preprocessing module; Step S2: Perform edge node preprocessing on the data; Step S3: After preprocessing, the data is transmitted to the data analysis module via the blockchain module; Step S4: Perform multimodal fusion analysis on the data through the data analysis module to generate water quality level assessment results and pollution risk map; Step S5: Output warning information to the user through a visual interface or API (Application Programming Interface).

[0064] This invention discloses a water quality monitoring system for marine ecological protection, which has been field-tested in a sea area of ​​the Yellow Sea in China. The system was deployed at depths of 5m, 10m, and 15m, with a sampling interval of 10 minutes. The test results show that the system can operate continuously and stably for 72 hours, with a data transmission success rate of 99.2% and a water quality anomaly identification accuracy rate exceeding 90%, verifying its reliability and practicality in complex marine environments.

[0065] This invention presents a water quality monitoring system for marine ecological protection, enabling real-time monitoring of marine water quality across multiple parameters, depths, and wide areas. By combining a modular sensor array with edge computing nodes, the system significantly improves the spatiotemporal density and processing efficiency of data acquisition. Utilizing blockchain technology, the system encrypts and transmits monitoring data, ensuring its authenticity, reliability, and traceability, effectively overcoming the shortcomings of traditional systems such as susceptibility to data tampering and poor transmission security. By introducing machine learning algorithms and intelligent decision-making models, the system can accurately assess complex water quality conditions and provide early risk warnings, offering scientific evidence and decision support for marine ecological protection. The system is characterized by low power consumption and self-powered operation. With its anti-corrosion and anti-biofouling properties, the system can operate stably in harsh marine environments for extended periods, making it suitable for various applications such as nearshore areas, ports, and aquaculture zones. Through a collaborative computing architecture combining edge computing and cloud data analysis, the system effectively reduces data transmission bandwidth requirements and cloud computing load. It also supports independent operation and data storage of edge nodes in offline environments, significantly enhancing the system's adaptability and reliability in offshore and weak signal environments. The system constructed in this invention not only achieves an intelligent upgrade for water quality monitoring but also provides efficient and reliable data support for marine ecological environment protection, disaster early warning, and emergency response, demonstrating significant social and ecological benefits.

[0066] Although the present invention has been disclosed above by way of embodiments, it is not intended to limit the present invention. Anyone skilled in the art can make some modifications and refinements without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A water quality monitoring system for marine ecological protection, characterized in that, include, An edge computing and preprocessing module receives first data output from a sensor, performs signal processing on the first data, and outputs second data. The blockchain module accepts the second data, encrypts the second data, generates a hash identifier, and outputs the third data; The data analysis module receives the third data, performs water quality analysis, classification, and anomaly diagnosis using clustering and decision tree algorithms, and outputs monitoring results.

2. The water quality monitoring system for marine ecological protection as described in claim 1, characterized in that, The edge computing and preprocessing module includes, The microprocessor is an embedded processor that incorporates a real-time data cleaning algorithm based on a sliding window and a wavelet transform compression algorithm. The microprocessor performs data noise reduction filtering and feature extraction through embedded algorithms; A data acquisition terminal, which synchronously analyzes sensor signals; A distributed coordinator, which constructs a local area network to enable cross-layer command interaction.

3. The water quality monitoring system for marine ecological protection as described in claim 1, characterized in that, The blockchain module includes a communication unit. The blockchain module encrypts each batch of data using a lightweight consensus mechanism, generating a unique hash identifier and timestamp for the data. The expression is: ， in, D For data content, T For timestamps, TD This is the node identifier.

4. The water quality monitoring system for marine ecological protection as described in claim 1, characterized in that, The data analysis module incorporates a Hadoop distributed file system storage architecture, a water quality clustering analysis model based on the Mahout framework, and a CART decision tree classifier. The data analysis module stores and retrieves water quality monitoring data through the Hadoop distributed file system storage architecture, and completes multi-source data fusion, intelligent water quality level determination, and pollution risk early warning based on the water quality clustering analysis model based on the Mahout framework and the CART decision tree classifier.

5. The water quality monitoring system for marine ecological protection as described in claim 4, characterized in that, Clustering is performed using the K-Means algorithm, and its expression is: ， in, The preset number of clusters, For the first One cluster, Its center of mass.

6. The water quality monitoring system for marine ecological protection as described in claim 1, characterized in that, It also includes an energy supply module, which comprises a monocrystalline silicon solar panel, a lithium-ion battery pack, and an intelligent power management unit. The energy supply module provides power to all hardware components of the sensor module, the edge computing and preprocessing module, the blockchain module, and the data analysis module.

7. The water quality monitoring system for marine ecological protection as described in claim 1, characterized in that, It also includes a sensor module, which includes a water quality sensor, a pH sensor, and a dissolved oxygen sensor. The sensor module collects water quality parameters at different water depths and transmits the collected physical signals to the edge computing and preprocessing module.

8. The water quality monitoring system for marine ecological protection as described in claim 7, characterized in that, Each sensor in the sensor module adopts a modular anti-corrosion packaging structure, with its surface coated with anti-bioadhesion material, and performs signal conditioning and analog-to-digital conversion through a multi-channel data acquisition terminal.

9. A water quality monitoring device for marine ecological protection, characterized in that, include, The energy supply module provides electrical energy to the hardware of the intelligent data acquisition terminal; The intelligent data acquisition terminal collects data through sensors, performs edge computing and preprocessing, and then encrypts the data before transmission. The buoy provides buoyancy, allowing the entire device to float stably on the water surface.

10. A water quality monitoring method for marine ecological protection, characterized in that, include, Step S1: Collect data from the sensor and transmit the collected signals to the edge computing and preprocessing module; Step S2: Perform edge node preprocessing on the data; Step S3: After preprocessing, the data is transmitted to the data analysis module via the blockchain module; Step S4: Perform multimodal fusion analysis on the data through the data analysis module to generate water quality level assessment results and pollution risk map; Step S5: Output the warning information to the user through a visual interface or API interface.

Images