CNN-based intelligent water pollution identification system

Abstract

The invention discloses a CNN-based intelligent water pollution identification system in the field of environmental protection and intelligent monitoring, and the system comprises a data collection module, a data processing module, an interference elimination module, and a pollution identification module which is used for carrying out the accurate identification of pollutants through a convolutional neural network. The judgment logic of the interference elimination module is as follows: if the comparison data of the real-time data and the standard data are consistent, the surface of the water body is pollution-free; when the motion trail presents a feature from bottom to top, animal activity or bubble generation is indicated, and the motion trail is identified as an interference item; and when the motion trail is synchronous with the water surface fluctuation and changes along with the change of the sun position, the motion trail caused by the light and shadow change is identified as the interference item. According to the invention, through a plurality of modules, accurate identification of water pollutants is realized, interference of non-pollution factors is effectively eliminated, and monitoring accuracy and real-time performance are improved.

Description

Technical Field

[0001] The present invention belongs to the field of environmental protection and intelligent monitoring, and specifically is a CNN-based intelligent water pollution identification system. Background Art

[0002] With the rapid development of industrialization and urbanization, water pollution has become increasingly serious, posing a serious threat to the ecological environment and human health. Traditional water pollution monitoring methods mainly rely on manual sampling and laboratory analysis. This method has the disadvantages of long monitoring cycle, high cost, poor real-time performance, etc., which makes it difficult to meet the needs of modern environmental protection. At the same time, the surface of the water body is easily disturbed by a variety of non-pollution factors, such as animal activities, bubble generation, light and shadow changes, etc. These factors will interfere with the accurate identification of pollutants and lead to inaccurate monitoring results.

[0003] In recent years, with the rapid development of artificial intelligence technology, especially the widespread application of deep learning technology, new solutions have been provided for water pollution monitoring. As an important model of deep learning, convolutional neural network (CNN) has achieved remarkable results in image recognition, video analysis and other fields with its powerful feature extraction and nonlinear mapping capabilities. Applying CNN to water pollution identification can realize automatic analysis and processing of water surface images, extract characteristic information of pollutants, and thus realize rapid and accurate identification of pollutants.

[0004] However, most of the existing CNN-based water pollution identification systems have some problems, such as the inability to effectively eliminate the interference of non-pollution factors, insufficient real-time performance, and the need to improve identification accuracy. In order to solve the above problems, the present invention proposes a CNN-based smart water pollution identification system. Summary of the invention

[0005] The present invention proposes a CNN-based intelligent water pollution identification system, which realizes the accurate identification of water pollutants through multiple modules, while effectively eliminating the interference of non-pollution factors, and improving the accuracy and real-time performance of monitoring.

[0006] In order to achieve the above object, the technical solution of the present invention is as follows: A CNN-based smart water pollution identification system, comprising:

[0007] Data acquisition module, used to collect data on the surface of water bodies, including standard data and real-time data;

[0008] A data processing module is used to pre-process the data collected by the data acquisition module and identify the movement trajectory of the material in the real-time data;

[0009] Interference elimination module, used to compare real-time data with standard data, and distinguish and eliminate interference items based on motion trajectory characteristics;

[0010] The pollution identification module is used to conduct in-depth analysis of the data after eliminating interference and accurately identify pollutants using convolutional neural networks;

[0011] The judgment logic of the interference elimination module is: if the comparison data of real-time data and standard data are consistent, it means that there is no pollution on the water surface;

[0012] If the comparison data between real-time data and standard data are inconsistent, it is necessary to further analyze the movement trajectory of the real-time data material;

[0013] When the motion trajectory presents a bottom-up feature, it indicates animal activity or bubble generation and is identified as an interference item;

[0014] When the motion trajectory is synchronized with the fluctuations on the water surface and changes with the position of the sun, it indicates that it is caused by changes in light and shadow and is identified as an interference item.

[0015] The following beneficial effects are achieved by adopting the above scheme:

[0016] 1. High-precision pollution identification: By using convolutional neural networks (CNN) to deeply analyze pre-processed and interference-free data, the system of the present invention can accurately identify water pollutants. The powerful feature extraction capability of CNN enables it to capture the subtle features of pollutants, thereby improving the accuracy of identification.

[0017] 2. Real-time data monitoring and analysis: The data acquisition module can obtain real-time data on the water surface, including standard data and real-time data, to ensure the timeliness of monitoring. At the same time, the data processing module quickly pre-processes and identifies trajectories of real-time data, providing a solid foundation for subsequent pollution identification.

[0018] 3. Intelligent interference elimination: The interference elimination module effectively eliminates interference caused by non-pollution factors such as animal activities, bubble generation, light and shadow changes, etc. by comparing real-time data with standard data and combining characteristic analysis of material movement trajectories. This mechanism greatly improves the system's ability to identify real pollution events and reduces false alarms and missed alarms.

[0019] 4. Automation and intelligence: The entire system realizes a fully automated process from data collection, processing, interference elimination to pollution identification, and can complete complex monitoring tasks without human intervention. This not only reduces the burden of manpower, but also improves the continuity and stability of monitoring.

[0020] 5. Broad application prospects: The system of the present invention can be widely used in pollution monitoring of various water bodies such as rivers, lakes, and reservoirs, providing timely and accurate pollution information for environmental protection departments, helping to formulate scientific governance measures and protect the safety and sustainable development of water resources.

[0021] In summary, the CNN-based smart water pollution identification system has brought revolutionary changes to the field of water pollution monitoring with its high precision, real-time, intelligence and broad application prospects, and is of great significance for protecting water resources and maintaining ecological balance.

[0022] Furthermore, the identification logic of the pollution identification module is:

[0023] When the trajectory of the object is from top to bottom, CNN and image features are used to judge that it is floating objects such as fallen leaves or plastics, which are pollutants.

[0024] When the movement trajectory of a substance shows a gradual expansion from small to large, accompanied by obvious color changes, it is judged to be oil pollution, which indicates a pollutant.

[0025] Beneficial effects: By combining the movement trajectory of the substance (such as from top to bottom, from small to gradually expanding) and image features (such as color changes, shape, etc.), the pollution identification module can more accurately determine the type of pollutant. This multi-dimensional analysis method reduces the possibility of misjudgment and improves the accuracy of identification. When the system detects suspected pollutants, it can quickly perform image analysis and processing through CNN to determine the nature of the pollutant in real time. This rapid response capability helps to take timely countermeasures to prevent the spread of pollution. The recognition logic is not limited to a single type of pollutant, but can identify multiple types of pollutants, such as fallen leaves, floating plastics, and oil stains. This diverse recognition capability makes the system more comprehensive and practical.

[0026] Furthermore, the real-time data collected in the data collection module includes, but is not limited to: image data of the water surface, water temperature data and water flow rate data.

[0027] Beneficial effects: By simultaneously collecting multiple types of data such as images, temperature and flow rate, the system can gain a more comprehensive understanding of the state of the water body. This multi-dimensional monitoring capability helps to more accurately identify pollution incidents, because different types of pollutants tend to produce specific changes in the surface image, temperature or flow rate of the water body. Image data provides intuitive information on the surface of the water body, such as color, shape, floating objects, etc., which helps to directly observe signs of pollution. Water temperature and flow rate data provide indirect information on the physical state of the water body, which is equally important in pollution identification. Combining multiple data for analysis can significantly improve the accuracy of identification. Multiple types of data provide rich materials for in-depth analysis of water pollution. For example, by comparing image data at different time points, the diffusion path of pollutants can be traced; by analyzing the changing patterns of water temperature and flow rate, the nature and source of pollutants can be inferred. This in-depth analysis helps to formulate more effective governance measures.

[0028] Furthermore, the data processing module identifies the movement trajectory of the substance in the real-time data by analyzing the position change of the substance in the image sequence of the image data in combination with the morphology and size characteristics of the substance, and calculating its movement trajectory.

[0029] Beneficial effects: Combining the morphology and size characteristics of the substance for analysis can more effectively distinguish the target substance from the complex background and accurately track its position changes in the time series image. This multi-dimensional recognition method reduces the possibility of misjudgment and missed judgment and improves the accuracy of motion trajectory recognition. Faced with complex and changeable water environments, such as changes in lighting, waves, etc., the system can more robustly identify the motion trajectory of the substance by comprehensively considering the morphology, size and position changes of the substance. This robustness ensures the stable operation and accurate monitoring of the system under various environmental conditions. By accurately identifying the motion trajectory of the substance, the system can analyze key information such as the source, diffusion path and speed of the substance. This is of great significance for understanding complex pollution scenarios, assessing the scope of pollution impact, and formulating effective response measures.

[0030] Furthermore, it also includes an alarm module, which is used to immediately send out an alarm signal when the pollution identification module identifies pollutants, to remind relevant personnel to pay attention and take corresponding treatment measures.

[0031] Beneficial effects: The instant response capability of the alarm module ensures that relevant personnel can be quickly notified when the system detects pollutants. This timely alarm mechanism helps to shorten the response time of pollution incidents and reduce the risk of pollution spread, thereby protecting the safety of water resources. The alarm signal can quickly trigger the emergency response process, allowing relevant personnel to quickly take necessary treatment measures. This includes but is not limited to dispatching cleanup teams, starting emergency treatment equipment, shutting down contaminated water sources, etc., thereby effectively controlling the spread of pollution sources and reducing the impact of pollution on the environment and human health. In some cases, pollution incidents may be accompanied by the release of toxic and hazardous substances, posing a threat to personnel and the environment. The timely alarm of the alarm module helps relevant personnel take protective measures in a timely manner to ensure personal safety and reduce the risk of environmental pollution.

[0032] Furthermore, it also includes a user interaction module for generating and displaying pollution reports, which include the quantity, distribution range, real-time data and alarm information of pollutants; it can also be used to adjust the system's monitoring parameters, select specific monitoring areas or set alarm thresholds.

[0033] Beneficial effects: The user interaction module can generate detailed pollution reports, including key information such as the number of pollutants, distribution range, real-time data, and alarm information. The transparent display of this information helps relevant personnel to fully understand the pollution situation and provide a more accurate and comprehensive basis for decision-making. The user interaction module allows users to adjust the system's monitoring parameters, select specific monitoring areas, or set alarm thresholds. This high degree of flexibility enables the system to be customized according to different scenarios and needs, improving the system's adaptability and practicality.

[0034] Furthermore, it also includes a historical data query module, which is used to query water pollution conditions, monitoring data and processing records in any time period in the past through the user interaction module.

[0035] Beneficial effects: Historical data is an important resource for analyzing pollution trends and evaluating governance effects. Through the historical data query module, users can easily obtain data over a long period of time, conduct in-depth analysis and mining, discover potential pollution patterns and trends, and provide a scientific basis for formulating long-term governance strategies. By querying historical data, users can understand the pollution situation and treatment needs in different time periods and regions, so as to more reasonably allocate monitoring and treatment resources. For example, increase the monitoring frequency and intensity during high-incidence pollution periods or key monitoring areas to ensure timely detection and treatment of pollution incidents.

[0036] Furthermore, it also includes a remote monitoring module, which is used to remotely access a remote terminal via the Internet, view the water pollution situation in the monitoring area in real time, monitor the video, receive alarm information, and remotely control the start and stop of the data acquisition module and the alarm module.

[0037] Beneficial effects: The remote monitoring module breaks the geographical restrictions, allowing users to access the system through the Internet anytime and anywhere, and view the water pollution in the monitoring area in real time. This greatly improves the efficiency and flexibility of monitoring, ensuring the timely discovery and handling of pollution incidents. When the system detects pollutants or abnormal conditions, the remote monitoring module can immediately push alarm information to relevant personnel and allow them to remotely control the start and stop of the data acquisition module and the alarm module. Through remote monitoring, users can reduce the number of on-site inspections and personnel requirements, thereby reducing operating costs. At the same time, remote monitoring can reduce monitoring errors and alarm omissions caused by personnel misoperation or negligence, and improve the reliability and stability of the system.

[0038] Furthermore, it also includes a multi-source data fusion processing module for integrating water quality data and meteorological data, and through comprehensive analysis and processing, comprehensively assessing the water condition and its potential pollution risks.

[0039] Beneficial effects: The multi-source data fusion processing module can integrate water quality data and meteorological data, such as temperature, humidity, wind speed, and precipitation, which may affect the diffusion, transformation, and degradation of water pollutants. By comprehensively analyzing these data, the status and trend of water pollution can be more accurately assessed. A single data source may have errors or limitations, and the fusion of multi-source data can reduce the errors caused by a single data source and improve the accuracy of the assessment results through the complementarity and verification between different data sources.

[0040] Furthermore, it also includes an intelligent prediction module, which is used to predict the probability of pollution occurring in the future through a machine learning algorithm based on historical pollution data and current environmental parameters, and display the prediction results through a user interaction module.

[0041] Beneficial effects: The intelligent prediction module can identify potential pollution risks in advance, so that relevant departments have enough time to take preventive measures, such as strengthening monitoring, adjusting pollution discharge plans, and preparing emergency supplies, thereby effectively reducing the occurrence of pollution incidents. Through early warning, the damage caused by pollution incidents to the environment and human health can be minimized, and economic losses and social impacts can be reduced. The prediction results are based on a large amount of historical data and real-time environmental parameters, calculated by machine learning algorithms, and have high accuracy and reliability. This provides decision makers with a scientific and objective reference basis and improves decision-making efficiency and accuracy. According to the prediction results, monitoring, governance and emergency resources can be allocated more reasonably to ensure rapid response at critical moments and improve resource utilization efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1This is a basic framework diagram of an embodiment of the CNN-based smart water pollution identification system of the present invention.

[0043] Figure 2 This is a flow chart of an embodiment of the CNN-based smart water pollution identification system of the present invention.

[0044] Figure 3 This is an overall framework diagram of an embodiment of the CNN-based smart water pollution identification system of the present invention.

[0045] Figure 4 It is an axonometric diagram of the data acquisition structure in the embodiment of the CNN-based smart water pollution identification system of the present invention.

[0046] The reference numerals in the drawings of the specification include: 1, housing; 2, connecting rod; 3, insertion rod; 4, vertical rod; 5, fan blade; 6, mounting groove; 101, through groove. DETAILED DESCRIPTION

[0047] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0048] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0049] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0050] The following is further described in detail through specific implementation methods:

[0051] Embodiment 1, basically as attached Figure 1 , Figure 2 and Figure 3 As shown: A CNN-based smart water pollution identification system, including:

[0052] The data acquisition module is used to collect data on the surface of the water body, including standard data and real-time data; the real-time data includes but is not limited to: image data of the water body surface, water body temperature data and water body flow rate data.

[0053] The data processing module is used to pre-process the data collected by the data acquisition module and identify the movement trajectory of the material in the real-time data; that is, by combining the morphology and size characteristics of the material, analyzing the position change of the material in the image sequence of the image data, and calculating its movement trajectory.

[0054] The interference elimination module is used to compare the real-time data with the standard data, and distinguish and eliminate interference items based on the motion trajectory characteristics; the judgment logic of the interference elimination module is:

[0055] If the comparison data of real-time data and standard data are consistent, it means that there is no pollution on the water surface; if the comparison data of real-time data and standard data are inconsistent, it is necessary to further analyze the movement trajectory of the real-time data material; when the movement trajectory presents a bottom-up feature, it indicates animal activity or bubble generation, which is identified as an interference item; when the movement trajectory is synchronized with the fluctuation of the water surface and changes with the position of the sun, it means that it is caused by changes in light and shadow and is identified as an interference item.

[0056] The pollution identification module is used to conduct in-depth analysis of the data after eliminating interference and accurately identify pollutants using a convolutional neural network. The identification logic of the pollution identification module is:

[0057] When the movement trajectory of the substance is from top to bottom, CNN and combined with image features are used to judge that it is floating objects such as fallen leaves or plastic, which are represented as pollutants; when the movement trajectory of the substance is gradually expanding from small to large, accompanied by obvious color changes, it is judged to be oil pollution, which is represented as a pollutant.

[0058] It also includes an alarm module, which is used to send out an alarm signal immediately when the pollution identification module identifies pollutants, so as to remind relevant personnel to pay attention and take corresponding treatment measures.

[0059] It also includes a user interaction module for generating and displaying pollution reports, which include the quantity, distribution range, real-time data and alarm information of pollutants; it can also be used to adjust the system's monitoring parameters, select specific monitoring areas or set alarm thresholds.

[0060] The specific implementation process is as follows: First, install a CNN-based smart water pollution identification system in the waters that need to be monitored, including necessary hardware equipment (such as high-definition cameras and various sensors, etc.) and software systems. Through the user interaction module, operators can set the system's monitoring parameters, such as image acquisition frequency, temperature and flow rate monitoring thresholds, alarm thresholds, etc. At the same time, specific monitoring areas can be selected, such as a section of a river, a specific area of ​​a lake, etc.

[0061] The data acquisition module starts automatically to capture the image data of the water surface, water temperature and flow rate data. These data are transmitted to the data processing module in real time, where the initially collected non-polluted water data is used as standard data. The data processing module first performs pre-processing operations such as denoising and contrast enhancement on the received image data to improve the accuracy of subsequent processing. Through image processing technology, combined with the morphology and size characteristics of the substance, the position change of the substance in the continuous image sequence is analyzed to calculate its motion trajectory.

[0062] The interference elimination module compares the real-time data with the standard data. For example, if the system detects that the water temperature has risen slightly but is still within the normal range, and the movement trajectory of the object in the image is caused by changes in light and shadow (such as changes in the reflection of the water surface caused by changes in the position of the sun), it is judged as an interference item and eliminated.

[0063] If the system detects that the motion trajectory presents a bottom-up feature (that is, the motion trajectory obviously moves from the bottom of the water body or deeper to the water surface), and the motion trajectory should maintain a certain continuity, that is, the substance shows an upward movement trend in continuous image frames, rather than being random or disordered, the morphology and speed characteristics of the substance will be further analyzed. If the morphology is streamlined or irregular, and the speed is relatively fast or accompanied by dynamic changes, then it is likely to be a sign of animal activity. If the morphology is a circular or elliptical bubble group, and the speed is relatively slow and uniform, then it is more likely to be the result of bubble generation, and it is judged as an interference item and excluded.

[0064] The pollution identification module uses a convolutional neural network (CNN) to perform in-depth analysis on the data after eliminating interference. CNN extracts feature information from the image through multi-layer convolution, pooling and other operations, and compares it with the pre-trained pollutant model.

[0065] Floating object identification: If the system detects that the movement trajectory of a substance is from top to bottom, and the image features show texture and color similar to fallen leaves or plastic, the CNN model determines that it is floating object pollution and marks it in the pollution report.

[0066] Oil pollution identification: If the movement trajectory of a substance in a certain area shows a gradual expansion from a small size, accompanied by an obvious color change (such as from a transparent to a dark oil film), the CNN model will identify it as oil pollution and immediately issue an alarm.

[0067] Once the pollution identification module confirms the presence of pollutants, the alarm module immediately sends out an alarm signal and notifies relevant personnel through voice, text messages, emails, etc. After receiving the alarm, relevant personnel quickly check the pollution report to understand the quantity, distribution range and other detailed information of the pollutants, and initiate emergency plans, such as sending a cleaning team to deal with the problem and shutting down the relevant water source.

[0068] The user interaction module automatically generates a detailed pollution report based on the pollution identification results, including the type, quantity, distribution range, real-time data screenshots, alarm time, etc. of the pollutants. The administrator can adjust the system's monitoring parameters, select new monitoring areas, or set stricter alarm thresholds at any time through the user interaction module to meet different monitoring needs.

[0069] For example, suppose that in the monitoring area of ​​a city lake, the system captures a picture of a black oil film on the surface of the water through a camera, and detects that the water temperature in the area has risen slightly. The data processing module analyzes the movement trajectory of the oil film from small to gradually expanding, accompanied by a deepening of color. The pollution identification module uses the CNN model to identify that this is oil pollution and immediately triggers an alarm. After receiving the alarm, the administrator checks the pollution report and finds that the oil has spread to a certain range. The cleaning team is immediately organized to deal with it, and the surrounding water intakes are closed to prevent oil from entering the water supply system.

[0070] Embodiment 2 is different from the above embodiment in that, as shown in the attached Figure 3 As shown: it also includes a historical data query module, which is used to query water pollution conditions, monitoring data and processing records in any time period in the past through the user interaction module.

[0071] The specific implementation process is as follows: the system will store all collected monitoring data (including real-time data and standard data), pollution identification results, alarm records, treatment measures and other information in a database or data warehouse. These data should be stored in chronological order and contain sufficient metadata for subsequent query and analysis. The user interaction module will provide an intuitive query interface, allowing users to enter query conditions, such as start time, end time, monitoring area, pollutant type, etc. The interface should be designed to be concise and easy for users to operate. The query results will be presented to users in the form of charts, tables or maps. For example, a line chart can be used to show the changing trend of numerical data such as water temperature and flow rate; a heat map or scatter plot can be used to show the distribution range of pollutants; a table can be used to list detailed monitoring data and treatment records, etc.

[0072] First, based on the start time and end time entered by the user, all data within the time period is filtered out from the database. If the user also enters other query conditions (such as monitoring area, pollutant type, etc.), the system will further filter the data to narrow the scope of the query results. The system will integrate the filtered data and display it in the format specified by the user. At the same time, the system can also provide a data export function, allowing users to save query results to local files.

[0073] Embodiment 3 is different from the above embodiment in that, as shown in the attached Figure 3 As shown: It also includes a remote monitoring module, which is used to remotely access the remote terminal through the Internet, view the water pollution situation in the monitoring area in real time, monitor the video and receive alarm information, and remotely control the start and stop of the data acquisition module and the alarm module.

[0074] The specific implementation process is as follows: The remote terminal (such as PC, mobile phone or tablet, etc.) is connected to the server or cloud platform of the smart water pollution identification system through the Internet. Use authentication mechanisms (such as operator name and password, token, etc.) to ensure the security of remote access. The remote monitoring module obtains water pollution data in the monitoring area from the server or cloud platform in real time, including water quality parameters, identification results of pollutants, etc. These data are displayed on the remote terminal in the form of charts, images, etc., and the operator can intuitively see the real-time status of water pollution. When the pollution identification module identifies pollutants and triggers an alarm, the remote monitoring module is responsible for pushing the alarm information to the remote terminal in real time. The alarm information includes the type, location, severity, etc. of the pollutant, as well as recommended countermeasures.

[0075] The operator can remotely control the start and stop of the data acquisition module through the remote terminal to adjust the frequency or range of data acquisition when necessary. Similarly, the operator can also remotely control the start and stop of the alarm module, for example, temporarily shut down the alarm in certain circumstances to avoid false alarms or excessive alarms.

[0076] Embodiment 4 is different from the above embodiment in that, as shown in the attached Figure 3 As shown: It also includes a multi-source data fusion processing module, which is used to integrate water quality data and meteorological data, and through comprehensive analysis and processing, comprehensively evaluate the water condition and its potential pollution risks.

[0077] The specific implementation process is as follows: The multi-source data fusion processing module is an important part of the smart water pollution identification system. It is responsible for real-time or regular acquisition of water quality data (mainly from water quality monitoring stations or online monitoring equipment, including but not limited to key indicators such as pH value, dissolved oxygen, turbidity, ammonia nitrogen and heavy metal content) and meteorological data (mainly from meteorological stations or satellite remote sensing, including meteorological elements such as temperature, precipitation, wind speed, wind direction, humidity and light intensity) from various data sources. Convert data from different sources and formats into a unified format and standard to facilitate subsequent data analysis and processing.

[0078] Based on water quality data, statistical methods, machine learning algorithms and other technologies are used to quantitatively evaluate water quality and identify possible pollutants and their concentrations in the water. Combined with meteorological data, the impact of climate conditions on water quality is analyzed. For example, rainfall may increase surface runoff, which in turn brings in pollutants; high temperatures may accelerate the rate of certain chemical reactions in the water. Taking into account water quality data and meteorological data, risk assessment models are used to predict and evaluate potential pollution risks that water bodies may face. The assessment results include pollution risk level, impact range, and duration.

[0079] Embodiment 5 is different from the above embodiment in that, as shown in the attached Figure 3 As shown: It also includes an intelligent prediction module, which is used to predict the probability of pollution in the future through a machine learning algorithm based on historical pollution data and current environmental parameters, and display the prediction results through a user interaction module.

[0080] The specific implementation process is as follows: extract historical pollution data from the database, including the time, location, type of pollutants, concentration, and related environmental parameters of the pollution. Obtain current environmental parameters in real time through the data acquisition module, such as temperature, precipitation, wind speed, wind direction, and water quality indicators. According to the characteristics of the prediction task and the characteristics of the data, select a suitable machine learning algorithm as the prediction model. Commonly used algorithms include time series analysis, regression models, classification algorithms (such as logistic regression, random forest, support vector machine, etc.) and deep learning models (such as LSTM, GRU and other recurrent neural networks).

[0081] Extract features that have a significant impact on the prediction results from the acquired data, including trend features and periodic features in historical pollution data and key indicators in current environmental parameters. Use historical pollution data and corresponding environmental parameters as training sets to train the selected machine learning model. Improve the prediction accuracy and generalization ability of the model by adjusting model parameters and optimizing algorithms. Input the current environmental parameters into the trained machine learning model for real-time prediction. The prediction results include the probability of water pollution in the future, possible types of pollutants, and the degree of pollution. The prediction results are presented to users in the form of charts, reports, etc. through the user interaction module. The display content should include the predicted time range, pollution probability, types of pollutants, degree of pollution, and corresponding recommended measures. The intelligent prediction module can be applied to multiple scenarios such as water pollution warning, water quality management, and pollution prevention and control, providing decision support for water management departments.

[0082] Embodiment 6 is different from the above embodiment in that, as shown in the attached Figure 4 As shown: the data acquisition module uses a data acquisition structure to collect data, the data acquisition structure includes a shell 1, the outer wall of the shell 1 is hinged with two connecting rods 2, the connecting rod 2 is a retractable structure, the connecting rod 2 is fixedly connected to an insertion rod 3 at one end away from the shell 1, a through slot 101 is provided at the bottom of the shell 1, a plurality of vertical rods 4 are fixedly connected in the through slot 101, the vertical rods 4 are rotatably matched with fan blades 5, the vertical rods 4 are fixedly connected with Hall sensors, the fan blades 5 are matched with magnets, the Hall sensor signal is connected to the controller and transmits real-time monitoring data to the controller, the controller is used to determine the rotation direction of the fan blades 5 according to the real-time monitoring data of the Hall sensor.

[0083] Both side walls of the shell 1 are provided with installation grooves 6, and sensor components are installed in the installation grooves 6. The sensor components include infrared sensors, temperature sensors, flow rate sensors, image sensors and strain sensors. The sensor components are connected to the controller signal, and the real-time monitoring data of the sensor components are transmitted to the data processing module through the controller.

[0084] The specific implementation process is as follows: when it is necessary to monitor the water area, the data acquisition structure is installed in the water area to be monitored. That is, according to the area of ​​the water area to be monitored, the length of the connecting rod 2 is adjusted, and then the plug rod 3 is fixed to the shore of the water area, and the shell 1 is placed on the water surface of the water area. The sensor assembly on the shell 1 monitors the water surface in real time. The infrared sensor is used to collect the movement data of objects on the water surface, the temperature sensor is used to monitor the water temperature, the flow rate sensor is used to monitor the flow speed of the water surface in real time, the image sensor collects the image data of the water surface in real time, and the strain sensor is used to collect the impact of the water surface fluctuation on the surface of the shell 1, so that the controller analyzes and determines the movement trend and law of the water surface fluctuation.

[0085] When the water area fluctuates, the water may directly act on the shell 1 with a large impact force, which poses a threat to the stability of the shell 1 and the accuracy of data collection. The fan blade 5 utilizes the principle of fluid mechanics to disperse and guide the water that impacts the shell 1, so that the impact force of the water is dispersed, thereby reducing the direct impact on the shell 1. The rotational coordination between the fan blade 5 and the vertical rod 4 allows the fan blade 5 to rotate freely under the action of the water body, which not only enhances the diversion effect, but also improves the adaptability of the shell 1 to the impact of the water body. No matter how the water body fluctuations change, the fan blade 5 can respond flexibly and keep the shell 1 relatively stable. Traditional data acquisition structures often lack sufficient adaptability when facing water body fluctuations. They may cause structural instability due to water flow, thereby affecting the accuracy of data collection. For example, if the structure does not have sufficient buffering or a mechanism to disperse the water body, the water body fluctuation may directly impact the sensor or cause the structure to tilt or even overturn. However, due to the diversion effect of the fan blade 5, the water body of this structure has been effectively dispersed before hitting the shell 1, thereby reducing the additional pressure caused by the obstruction of the water body. This not only protects the housing 1 and the internal sensors from damage, but also improves the continuity and accuracy of data collection.

[0086] At the same time, when the fan blade 5 rotates, the magnet on the fan blade 5 will periodically approach and move away from the Hall sensor, thereby generating a changing voltage signal. The controller can determine the rotation direction of the fan blade 5 by analyzing the voltage signal, thereby determining the movement direction of the water.

[0087] The movement trajectory of the material obtained by the infrared sensor is used to analyze the movement direction and speed of the material. The movement trajectory of the material is compared with other data in the sensor component. If the movement trajectory of the material is consistent with or highly correlated with the direction of the water surface fluctuation or water flow, it may be passive movement; if the trajectory has no obvious correlation with the fluctuation or flow direction or is opposite, it may be autonomous movement. If the movement speed of the material is much greater than or much less than the flow velocity of the surrounding water body, it may be autonomous movement; if the speed is similar, it may be passive movement. At the same time, the Hall sensor is used to monitor the rotation direction and speed of the fan blade 5 to further verify the movement nature of the material. If the movement direction of the material is consistent with the rotation direction of the fan blade 5 and the speed is similar, the judgment of passive movement is also supported.

[0088] The above is only an embodiment of the present invention, and the common knowledge such as the known specific structure and / or characteristics in the scheme is not described in detail here. It should be pointed out that for those skilled in the art, several deformations and improvements can be made without departing from the structure of the present invention, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A CNN-based smart water pollution identification system, characterized in that: include: Data acquisition module, used to collect data on the surface of water bodies, including standard data and real-time data; A data processing module is used to pre-process the data collected by the data acquisition module and identify the movement trajectory of the material in the real-time data; Interference elimination module, used to compare real-time data with standard data, and distinguish and eliminate interference items based on motion trajectory characteristics; The pollution identification module is used to conduct in-depth analysis of the data after eliminating interference and accurately identify pollutants using convolutional neural networks; The judgment logic of the interference elimination module is: if the comparison data of real-time data and standard data are consistent, it means that there is no pollution on the water surface; If the comparison data between real-time data and standard data are inconsistent, it is necessary to further analyze the movement trajectory of the real-time data material; When the motion trajectory presents a bottom-up feature, it indicates animal activity or bubble generation and is identified as an interference item; When the motion trajectory is synchronized with the fluctuations on the water surface and changes with the position of the sun, it indicates that it is caused by changes in light and shadow and is identified as an interference item.

2. The CNN-based smart water pollution identification system according to claim 1 is characterized by: The identification logic of the pollution identification module is: When the trajectory of the object is from top to bottom, CNN and image features are used to judge that it is floating objects such as fallen leaves or plastics, which are pollutants. When the movement trajectory of a substance shows a gradual expansion from small to large, accompanied by obvious color changes, it is judged to be oil pollution, which indicates a pollutant.

3. The CNN-based smart water pollution identification system according to claim 1 is characterized by: The real-time data collected in the data collection module include but are not limited to: image data of the water surface, water temperature data and water flow rate data.

4. The CNN-based smart water pollution identification system according to claim 1 is characterized by: The data processing module identifies the movement trajectory of the substance in the real-time data by analyzing the position change of the substance in the image sequence of the image data by combining the morphology and size characteristics of the substance and calculating its movement trajectory.

5. The CNN-based smart water pollution identification system according to claim 1 is characterized by: It also includes an alarm module, which is used to send out an alarm signal immediately when the pollution identification module identifies pollutants, so as to remind relevant personnel to pay attention and take corresponding treatment measures.

6. The CNN-based smart water pollution identification system according to claim 1 is characterized by: It also includes a user interaction module for generating and displaying pollution reports, which include the quantity, distribution range, real-time data and alarm information of pollutants; it can also be used to adjust the system's monitoring parameters, select specific monitoring areas or set alarm thresholds.

7. The CNN-based smart water pollution identification system according to claim 1 is characterized by: It also includes a historical data query module, which is used to query water pollution conditions, monitoring data and processing records in any time period in the past through the user interaction module.

8. The CNN-based smart water pollution identification system according to claim 1 is characterized by: It also includes a remote monitoring module, which is used to remotely access a remote terminal via the Internet, view the water pollution situation in the monitoring area in real time, monitor videos, receive alarm information, and remotely control the start and stop of the data acquisition module and the alarm module.

9. The CNN-based smart water pollution identification system according to claim 1 is characterized by: It also includes a multi-source data fusion processing module, which is used to integrate water quality data and meteorological data, and through comprehensive analysis and processing, comprehensively assess the water condition and its potential pollution risks.

10. The CNN-based smart water pollution identification system according to claim 1 is characterized by: It also includes an intelligent prediction module, which is used to predict the probability of pollution in the future through a machine learning algorithm based on historical pollution data and current environmental parameters, and display the prediction results through a user interaction module.

Images