Intelligent prevention and control method and system integrating photoelectric sensor and big data analysis

Abstract

The invention discloses an intelligent prevention and control method and system integrating a photoelectric sensor and big data analysis. The method comprises the steps that data collected by a plurality of photoelectric sensor nodes deployed in a target area are acquired to obtain initial data, and the layout of photoelectric sensors is automatically adjusted according to environment changes under the condition of no human intervention to form a self-adaptive monitoring network to capture the environment changes; performing preprocessing and feature extraction on the initial data to obtain an extraction result; inputting the extraction result into a pre-trained multi-reference object relation segmentation model to extract an environment feature vector which may affect health so as to obtain a first feature vector; performing historical analysis and comparison to obtain an analysis result; generating a second feature vector; determining the final risk level and the position of the affected area; and outputting the obtained content. By implementing the method provided by the invention, the environment change can be monitored in real time, the monitoring strategy can be dynamically adjusted, and the disease transmission can be comprehensively monitored.

Description

Technical Field

[0001] The present invention relates to an intelligent prevention and control method, and more specifically to an intelligent prevention and control method and system integrating photoelectric sensors and big data analysis. Background Art

[0002] In recent years, with the acceleration of urbanization, urban population density has continued to increase, and this change has brought huge challenges to public health. In particular, the risk of outbreaks of sudden infectious diseases has increased. High population density allows viruses to spread more rapidly, while environmental changes and increased mobility of people caused by urbanization also provide a breeding ground for the spread of infectious diseases.

[0003] Traditional infectious disease monitoring methods focus on post-event statistical analysis and clinical reports. These methods can only provide relevant data after an epidemic breaks out, which is somewhat delayed and cannot timely reflect the trend and potential risks of infectious disease transmission. Clinical reports rely on active reports from medical institutions and patients, which makes it difficult to ensure the integrity and timeliness of the data. At the same time, traditional monitoring methods usually focus on the health status and condition data of patients, and lack the ability to monitor and warn of environmental factors.

[0004] In modern urban environments, in addition to the health status of the population, environmental factors such as air quality, climate change, and the status of public health facilities play an important role in the spread of diseases. However, traditional monitoring systems often fail to effectively integrate these environmental data, resulting in the inability to provide real-time feedback on potential health risks.

[0005] Therefore, it is necessary to design a new method to achieve real-time monitoring of environmental changes and dynamic adjustment of monitoring strategies, and to achieve all-round monitoring of disease transmission, including real-time monitoring and early warning of environmental factors, which will help improve the timeliness and accuracy of public health emergency responses and effectively reduce the risk of the spread of sudden diseases. Summary of the invention

[0006] The purpose of the present invention is to overcome the defects of the prior art and provide an intelligent prevention and control method and system integrating photoelectric sensors and big data analysis.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: an intelligent prevention and control method integrating photoelectric sensors and big data analysis, comprising:

[0008] Acquiring data collected by a plurality of photoelectric sensor nodes deployed in a target area to obtain initial data, wherein the photoelectric sensors automatically adjust their layout according to environmental changes without human intervention to form an adaptive monitoring network to capture environmental changes;

[0009] Preprocessing and feature extraction are performed on the initial data to obtain extraction results;

[0010] Inputting the extraction result into a pre-trained multi-reference object relationship segmentation model to extract environmental feature vectors that may affect health, so as to obtain a first feature vector;

[0011] Collecting data at different time points at the same location, generating data pairs of healthy state and unhealthy state, and performing comparative analysis on the data pairs to obtain analysis results;

[0012] generating a second feature vector according to the analysis result;

[0013] Determine a final risk level and a location of an affected area according to the first feature vector and the second feature vector;

[0014] The final risk level and the location of the affected area are output to the terminal for display on the terminal.

[0015] The present invention also provides an intelligent prevention and control system integrating photoelectric sensors and big data analysis, including:

[0016] A data acquisition unit, used to acquire data collected by a plurality of photoelectric sensor nodes deployed in a target area to obtain initial data, wherein the photoelectric sensors automatically adjust their layout according to environmental changes without human intervention to form an adaptive monitoring network to capture environmental changes;

[0017] An extraction unit, used for preprocessing and feature extraction of the initial data to obtain an extraction result;

[0018] A first vector determination unit, configured to input the extraction result into a pre-trained multi-reference object relationship segmentation model to extract an environmental feature vector that may affect health, so as to obtain a first feature vector;

[0019] A historical comparison unit is used to collect data at different time points in the same place, generate data pairs of healthy state and unhealthy state, and compare and analyze the data pairs to obtain analysis results;

[0020] A second vector generating unit, used for generating a second feature vector according to the analysis result;

[0021] a level and position determination unit, configured to determine a final risk level and a position of an affected area according to the first feature vector and the second feature vector;

[0022] The output unit is used to output the final risk level and the location of the affected area to the terminal for display on the terminal.

[0023] The present invention further provides a computer device, which includes a memory and a processor. The memory stores a computer program, and the processor implements the above method when executing the computer program.

[0024] The present invention also provides a storage medium, wherein the storage medium stores a computer program, and the computer program implements the above method when executed by a processor.

[0025] The beneficial effects of the present invention compared with the prior art are as follows: the present invention forms an adaptive monitoring network to capture dynamic changes in the environment by deploying photoelectric sensors that can automatically adjust their layout according to environmental changes in the target area; the collected initial data is preprocessed and feature extracted, and then input into a pre-trained model to identify environmental features that may affect health and form a first feature vector; the system also collects data from different time periods for comparative analysis, evaluates changes in health conditions, and then generates a second feature vector representing the degree of environmental deterioration; using the fusion results of the first and second feature vectors, the system determines the final risk level and the location of the affected area, supporting accurate public health emergency response; the entire process integrates GIS technology for spatial analysis and public participation mechanisms, ensuring the dynamic and adaptable nature of the monitoring strategy and helping to reduce the risk of disease transmission.

[0026] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying any creative work.

[0028] Figure 1 A schematic diagram of a flow chart of an intelligent prevention and control method integrating photoelectric sensors and big data analysis provided by an embodiment of the present invention;

[0029] Figure 2 A schematic block diagram of an intelligent prevention and control system integrating photoelectric sensors and big data analysis provided by an embodiment of the present invention;

[0030] Figure 3 A schematic block diagram of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0031] 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 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.

[0032] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0033] It should also be understood that the terms used in this specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. As used in the specification of the present invention and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0034] It should be further understood that the term "and / or" used in the present description and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0035] See also Figure 1 , Figure 1 A schematic flow chart of an intelligent prevention and control method integrating photoelectric sensors and big data analysis provided by an embodiment of the present invention. The intelligent prevention and control method integrating photoelectric sensors and big data analysis is applied to a server. The server interacts with the terminal and the photoelectric sensor, and uses multiple photoelectric sensor nodes deployed in the target area to automatically adapt to environmental changes, form a self-organizing monitoring network, collect and pre-process environmental data in real time, and analyze environmental factors that may affect health through feature extraction and multi-reference object relationship segmentation model, thereby generating a first feature vector. The system further collects data at different time points in the same place for comparative analysis, evaluates the changing trend of health status, and generates a second feature vector reflecting the degree of environmental deterioration. Combining the two feature vectors, the system determines the risk level and the specific location affected, and realizes spatial distribution analysis and high-risk area delineation through GIS technology to ensure that the monitoring strategy is dynamically adjusted to respond to the latest situation. Finally, the analysis results are sent to the terminal device for display, supporting timely public health decision-making. This method realizes all-round monitoring of disease transmission, improves the speed and accuracy of emergency response, and effectively reduces the risk of the spread of sudden diseases.

[0036] Figure 1FIG. 1 is a flow chart of an intelligent prevention and control method integrating photoelectric sensors and big data analysis provided by an embodiment of the present invention. Figure 1 As shown, the method includes the following steps S110 to S170.

[0037] S110, acquiring data collected by multiple photoelectric sensor nodes deployed in the target area to obtain initial data, wherein the photoelectric sensors automatically adjust their layout according to environmental changes without human intervention to form an adaptive monitoring network to capture environmental changes.

[0038] In this embodiment, the initial data includes various environmental parameters collected from multiple photoelectric sensor nodes in the target area. These data may specifically include but are not limited to the following types of information:

[0039] Light intensity and spectral information: used to monitor changes in lighting conditions, which may have an impact on the generation, distribution or degradation of certain types of pollutants.

[0040] Temperature and humidity data: Temperature and humidity are important factors that affect biological activity and chemical reaction rates in the environment and are critical for assessing the risk of disease transmission.

[0041] Air quality indicators: such as PM2.5 and PM10 particle concentrations, pollutant concentrations such as sulfur dioxide (SO2) and nitrogen oxides (NOx), and volatile organic compounds (VOCs) levels, which are key indicators of air quality and potential health risks.

[0042] Gas composition analysis: In addition to conventional pollutants, it may also include the presence and concentration changes of specific gases (such as carbon monoxide, ozone, etc.), which helps to discover abnormal emission sources or natural phenomena.

[0043] Meteorological parameters: wind speed, wind direction, rainfall, etc. This information is very important for understanding the diffusion pattern of pollutants.

[0044] Other physical characteristics: such as sound levels, electromagnetic radiation intensity, etc., although not directly related to health, may also become important considerations in specific situations.

[0045] Each photoelectric sensor node is equipped with multiple sensors that can simultaneously measure the above-mentioned environmental parameters, and all of these data are collected in real time and transmitted back to the central processing system as the basis for subsequent preprocessing, feature extraction and analysis. In addition, since photoelectric sensors have the ability to adjust themselves adaptively, they can automatically optimize their position and parameter settings according to environmental changes, ensuring that the collected data is always the most representative and timely.

[0046] In one embodiment, the above-mentioned photoelectric sensors automatically adjust their layout according to environmental changes without human intervention to form an adaptive monitoring network to capture environmental changes, including:

[0047] Select key locations within the target area for initial deployment of photoelectric sensors based on preset rules or expert knowledge. These rules take into account geographic features, historical environmental data, and the location of potential pollution sources. Each photoelectric sensor node is equipped with multiple sensors.

[0048] The data stream from each sensor node is analyzed in real time to identify significant changes in the environment, such as sudden increases in pollutant concentrations and sudden changes in temperature. When important environmental changes are detected, an adaptive adjustment program is initiated, which combines the current environmental conditions with historical data trends to calculate the optimal new sensor location and parameter configuration plan. In this process, multiple factors are considered, including but not limited to wind direction, wind speed, topography, building distribution, etc., to determine how to most effectively adjust the sensor layout.

[0049] According to the calculation results, a movement command or parameter update command is sent to a specific sensor node. After receiving the command, the sensor autonomously performs the corresponding action, including changing its own azimuth, tilt angle or adjusting the sensing sensitivity. The specific sensor can automatically complete repositioning or parameter optimization based on the built-in logic and external guidance;

[0050] After each adjustment, the effect of the new layout is evaluated, and if the expected performance indicators are not achieved, the iterative optimization process continues until the established criteria are met, and then the data collected by multiple optoelectronic sensor nodes deployed in the target area are obtained.

[0051] The entire process, from environmental change detection to final sensor adjustment, is completed automatically without direct human operation, which improves response speed and flexibility. The system has self-learning capabilities and can continuously optimize its monitoring strategy as time and environmental conditions change to maintain the best monitoring status. Through intelligent adjustment, it ensures that all important environmental changes can be captured in a timely manner, and maintains efficient monitoring effects even in complex and changing urban or natural environments.

[0052] It not only achieves efficient monitoring of environmental changes, but also greatly improves the intelligence level of the monitoring system and its ability to respond to emergencies.

[0053] Specifically, first of all, the data received from each sensor node may contain noise or outliers. Therefore, before any analysis, these raw data need to be cleaned and standardized. This includes removing invalid data, filling missing values, and smoothing to reduce the impact of random fluctuations.

[0054] Next, for the data collected by each sensor node, extract the feature vector that can reflect the environmental status. For example, for air quality monitoring, PM2.5 concentration, temperature, humidity, etc. are all important features. Then, through correlation analysis or other statistical methods, select the most representative and predictive features for subsequent analysis.

[0055] Establish a real-time monitoring system to analyze feature vectors online using machine learning models (such as support vector machines, random forests, neural networks, etc.) or rule-based algorithms. Once certain features are found to be out of the normal range or meet the pre-set abnormal patterns, an alarm is triggered and marked as "significant change".

[0056] When significant changes are identified, the adaptive adjustment process is immediately activated. The main task at this stage is to make a quick assessment of the current environmental conditions to determine which areas are affected and to what extent. Geographic Information System (GIS) technology can be used to assist in locating the affected areas and combine weather forecast information to predict future trends.

[0057] At the same time, the stored historical data is called for comparative analysis. These historical data not only include the occurrence of similar events in the past, but also include the sensor layout and effect evaluation at that time. By studying historical data, it can help to better understand the characteristics of the current event and guide the next step.

[0058] Clarify the optimization goal, such as minimizing the risk of pollution spread, maximizing monitoring coverage, reducing energy consumption, etc. Depending on the specific application scenario, there may be multiple mutually restrictive goals that need to be considered comprehensively.

[0059] Based on the above goals, one or more mathematical models are established to describe the relationship between sensor layout and expected performance. Commonly used modeling methods include linear programming, integer programming, genetic algorithms, particle swarm optimization, etc. These models should be able to take into account factors such as geographical restrictions (such as building obstructions) and physical constraints (such as the maximum detection distance of the sensor).

[0060] Use computer simulation tools to test various proposed layout solutions and observe their performance in different scenarios. If a solution fails to achieve the expected effect, return to the previous step and continue to optimize until the most satisfactory solution is found.

[0061] Finally, the best sensor location and parameter setting scheme is selected based on the simulation results. Subsequently, the system automatically sends instructions to the selected sensors to start working according to the new arrangement. In this process, a feedback mechanism can also be added so that the system can continuously self-correct and improve in practice.

[0062] For example: sensor nodes collect this data through built-in microcontrollers or processing units, and send the data to the central processing system or edge computing device through wireless communication networks (such as LoRa, ZigBee, 5G, etc.).

[0063] The raw data needs to be preprocessed to remove noise and fill in missing values. Common methods include filtering, normalization, standardization, etc. For example, a low-pass filter can be used to remove short-term noise, or the mean substitution method can be used to fill in missing data.

[0064] For the data of each sensor node, feature extraction is performed using statistical or machine learning methods.

[0065] For example:

[0066] By calculating the mean, standard deviation, and rate of change of sensor data, key features of the environmental state can be extracted. Using the moving window method, trend changes in time series data can be calculated. For example, if the temperature change exceeds a certain threshold, or the air quality index exceeds the standard, it can be considered that a significant change has occurred.

[0067] Use anomaly detection algorithms (such as Z-score, model-based detection methods, etc.) to identify abnormal fluctuations in environmental data and determine whether significant changes have occurred. These changes may include sudden increases in pollution sources, climate change, changes in traffic density, etc. The purpose of the algorithm is to reduce false positives and false negatives, ensuring that the response mechanism is only activated when real environmental changes occur.

[0068] When the system detects significant environmental changes, it starts the adaptive adjustment process. This process can be completed in the following steps:

[0069] There are two aspects to assessing the current environmental conditions: Current data analysis: By analyzing the data from each sensor in real time, we can understand the specific conditions of the current environment (such as the location of the pollution source, temperature, humidity, etc.). Historical data trend comparison: By comparing with historical data, we can identify the trend of change. For example, if the air quality in a certain area suddenly deteriorates, but historical data shows that the area is usually not polluted, the system will identify the change as abnormal and start the adjustment process.

[0070] Calculate the optimal new sensor layout and parameter configuration: Based on environmental changes and historical data, the system calculates the optimal new sensor layout and configuration. This process can be achieved through the following methods:

[0071] Optimization algorithm: Use optimization algorithms (such as genetic algorithms, particle swarm optimization algorithms, simulated annealing algorithms, etc.) to optimize the location and parameters of sensors based on current environmental data and historical data trends. For example, increase the density of sensors near a pollution source, or adjust the sensitivity of sensors to adapt to new environmental changes. Environmental modeling: Through environmental modeling technology, a virtual environmental model is established to simulate the impact of different sensor layouts on environmental monitoring effects. The model can include climate models, pollutant diffusion models, etc., to simulate which locations and configurations can achieve the best monitoring effects under different environmental conditions. Objective function design: Design an objective function (such as sensor data coverage, detection sensitivity, energy consumption, etc.), and use optimization algorithms to minimize or maximize the objective function to find the optimal sensor layout and configuration.

[0072] Based on the calculation results, the system sends instructions to specific sensor nodes for adaptive adjustments. These instructions include:

[0073] Sensor position adjustment: Movement command: If the sensor has the ability to move (such as mounted on a drone, robot or other platform), a movement command is sent to guide it to move to a new location. Deployment and expansion: If the sensor cannot be moved, a new sensor needs to be deployed to the best location to expand the network coverage area.

[0074] Sensor parameter adjustment: Sensitivity adjustment: Adjust the sensitivity of the sensor according to the new environmental conditions. For example, when the intensity of the pollution source increases, the sensitivity of the sensor can be increased to detect the change earlier. Azimuth and tilt adjustment: If the sensor is directional, its azimuth or tilt angle can be adjusted to make it more accurate in monitoring the target area. Sampling frequency adjustment: Adjust the data sampling frequency of the sensor according to the nature of the environmental changes. For example, when a major environmental event occurs, increase the data collection frequency to monitor sudden changes with higher accuracy.

[0075] After the sensors have been adjusted, the system needs to evaluate the effect of the new layout. This can be done in the following ways:

[0076] Check whether the new layout can effectively cover the target area and improve the accuracy of the data. Ensure that the sensor can achieve the predetermined performance indicators (such as data accuracy, response time, monitoring accuracy, etc.) after adjustment. If the new layout fails to achieve the expected results, continue to optimize the adjustment plan. This process can be achieved through technologies such as machine learning, allowing the system to gradually improve the accuracy and efficiency of the adjustment after multiple adjustments.

[0077] In addition, each sensor node is equipped with a GPS module to obtain accurate geographic location information. For indoor environments or environments with poor GPS signals, UWB (ultra-wideband) or other indoor positioning technologies can be used as a supplement. Low-power wide area network (LPWAN) technologies such as LoRaWAN and NB-IoT are used to ensure long-distance stable communication. At the same time, relay nodes are deployed in key areas to expand network coverage and reduce communication blind spots.

[0078] Set up multiple sensor nodes at important monitoring points to form a redundant group. Even if a node fails, other nodes can continue to work to ensure data continuity. Develop a self-healing algorithm to automatically find alternative paths or reconfigure the network topology when a communication link is detected to maintain the connectivity and stability of the entire system. Design a built-in diagnostic program to regularly check the sensor status and immediately trigger an alarm and report to the central control system once an abnormality is found. Remotely repair software problems through OTA (over-the-air) updates. Implement a local caching mechanism, where each sensor node stores the collected data locally and uploads it to the cloud server after communication is restored to normal to avoid data loss caused by temporary network disconnection. The central control system dynamically adjusts the task allocation of each node according to the real-time monitoring situation, such as increasing the sampling frequency or changing the monitoring direction, to ensure efficient resource utilization while improving response speed.

[0079] S120: Preprocess and extract features of the initial data to obtain extraction results.

[0080] In this embodiment, the extraction result refers to a data set obtained after a series of preprocessing and feature extraction operations, which can more accurately reflect the characteristics of environmental changes. This data set not only contains key information extracted from the original initial data, but also has been optimized to enhance its effectiveness and usability for specific applications.

[0081] In one embodiment, the above-mentioned step S120 may include steps S121 - S122 .

[0082] S121, performing data cleaning, standardization, time series synchronization and data segmentation on the initial data to obtain a preprocessing result;

[0083] S122, extracting static features and dynamic features and fusing multi-source data from the preprocessing results to obtain extraction results.

[0084] Specifically, photoelectric sensors may be subject to external interference (such as electromagnetic interference, physical obstacles, etc.) during the acquisition process, resulting in outliers or noise in the data. Use filtering algorithms (such as low-pass filters, Gaussian filters) to smooth the data and reduce the impact of random fluctuations. Data loss may occur due to equipment failure or other reasons. Interpolation methods (linear interpolation, spline interpolation, etc.), model-based methods (such as K nearest neighbor algorithm, regression analysis), or infer missing values ​​from data at adjacent time points can be used.

[0085] In order to ensure that the data between different sensor nodes are comparable and that the data input to the machine learning model are in the same scale range, the raw data needs to be standardized (z-score standardization) or normalized (Min-Max scaling).

[0086] Ensure that the data from all sensor nodes are aligned on the same time base, eliminating the time deviation problem caused by sampling frequency differences or transmission delays.

[0087] For data that is recorded continuously for a long time, the data can be divided into several segments according to a specific time window (such as every hour, every day) to facilitate subsequent analysis.

[0088] Calculate the statistical characteristics of the photoelectric intensity in each time window, such as the mean, variance, peak value, and range. These indicators can help describe the changing characteristics of the ambient light. Through Fourier transform (FFT) or other spectrum analysis methods, the time domain signal is converted into a frequency domain representation to identify the periodic components and their intensities, which is particularly useful for detecting environmental changes.

[0089] Use methods such as moving average and exponentially weighted moving average to analyze the trend of data changes over time and capture the long-term evolution law. Set a threshold or use a mutation detection algorithm (such as the CUSUM cumulative sum control chart). When an event that significantly deviates from the normal range is monitored, mark the moment as a potential important change point. Apply unsupervised learning methods such as cluster analysis (K-means, DBSCAN) and principal component analysis (PCA) to discover hidden patterns or structures in the data, which helps to understand the essential characteristics of environmental changes.

[0090] If the system contains not only photoelectric sensors, but also other types of sensors (such as temperature, humidity, air quality, etc.), then multi-source data fusion can be considered. This can be achieved through joint probability distribution modeling, Bayesian networks, etc., comprehensively considering the possibility of multiple factors affecting health.

[0091] In summary, preprocessing and feature extraction aim to improve data quality and dig out key information that can effectively characterize environmental changes and possible health risks. The results of this process, the "extraction results", provide a solid foundation for subsequent steps, enabling the system to more accurately assess health risks and provide timely warnings.

[0092] S130: Input the extraction result into a pre-trained multi-reference object relationship segmentation model to extract environmental feature vectors that may affect health, so as to obtain a first feature vector.

[0093] In this embodiment, the first feature vectors include environmental feature vectors that affect health. These feature vectors not only describe specific environmental conditions, but also imply the relationship between these conditions and health risks.

[0094] In one embodiment, the above-mentioned step S130 may include steps S131 - S132 .

[0095] S131. For each input sample in the extraction result, the multi-reference relationship segmentation model segments the input sample and compares it with each pre-defined reference one by one. Through the self-attention mechanism and multi-layer perceptron technology, several dimensions are selected from all environmental features that may affect health to form a multi-dimensional feature space.

[0096] In this embodiment, for each input sample in the extraction result, the multi-reference object relationship segmentation model first performs segmentation processing on it. This process involves breaking down complex environmental data into smaller, more manageable parts and comparing them one by one with multiple pre-defined reference objects. The reference objects referred to here include but are not limited to ideal health environment parameters, historical averages, environmental parameters in extreme cases, and health thresholds set for different populations.

[0097] By introducing the self-attention mechanism, the model can focus on those environmental features that have the most significant impact on health. At the same time, using the multi-layer perceptron technology (MLP, Multi-Layer Perceptron), the model can automatically learn and select several key dimensions to form a multi-dimensional feature space. The core of this step is to capture the complex interactions between different environmental variables, rather than just evaluating the impact of each variable separately.

[0098] S132. Vectorize the feature space to obtain a first feature vector.

[0099] In this embodiment, once the multi-dimensional feature space is constructed, the next step is to convert it into a form that is easy to handle - the "first eigenvector". During this conversion process, the model performs a feature mapping operation to compress the high-dimensional original feature space into a low-dimensional but information-rich representation. This nonlinear transformation helps to simplify subsequent calculations while retaining important information in the original data.

[0100] Based on the multi-reference setting, the model then evaluates the degree of association between current environmental conditions and health risks. Specifically, it identifies which environmental characteristics are most likely to cause health problems and assigns corresponding weights based on the importance of these characteristics. This step not only considers the role of individual environmental factors, but also explores the interaction patterns between them.

[0101] The final output is one or a group of eigenvectors representing the possible health impacts under the current environmental conditions - the "first eigenvector". These vectors not only contain the importance scores of various environmental factors, but also reflect the complex interaction patterns between them. The first eigenvector is an important bridge connecting environmental monitoring and health early warning, providing timely and targeted guidance and suggestions for decision makers.

[0102] The training process of the multi-reference relationship segmentation model:

[0103] Collect time series data containing a variety of environmental parameters and health monitoring data at corresponding time points (such as disease incidence, number of hospital visits, etc.). Ensure that the data has sufficient time span and geographical coverage to support the effectiveness and universality of the model.

[0104] Features that are helpful in predicting health risks are extracted from the raw data, including static and dynamic features, and appropriate preprocessing (such as normalization, dimensionality reduction, etc.) is performed on them to optimize model performance.

[0105] Select appropriate algorithms based on the characteristics of the problem, such as support vector machine (SVM), random forest, neural network (such as CNN, LSTM), etc. Use cross-validation and other techniques to tune model parameters and ultimately determine the optimal model structure.

[0106] Introduce multiple references or benchmark conditions during the model design phase so that the model can more accurately assess the potential impact of the current environment on health. These references can be historical averages, parameter configurations under ideal health environments, etc. Specifically, choosing the right reference in practical applications is crucial to the accuracy and effectiveness of the multi-reference relationship segmentation model. Here are some guidelines and considerations to help you choose the most appropriate reference:

[0107] Ideal health environment reference: The ideal health environment reference should be based on long-term scientific research results and reflect the environmental conditions that are most conducive to human health. For example, the air quality index (AQI) can refer to the standards issued by the World Health Organization (WHO) or other authoritative organizations.

[0108] Considering the geographical and climatic differences in different regions, the ideal health environment parameters may be different. Therefore, the actual local conditions should be considered when making the selection.

[0109] Historical average references usually use data from the same season or time period in the past few years to ensure that the data is representative. For example, when analyzing the impact of summer air quality, you can choose the average PM2.5 concentration in the summer of the past five years as a reference. Make sure that the selected historical data comes from the same or similar geographical location to facilitate comparison of whether the current environmental status deviates from the normal state.

[0110] Extreme event records include environmental parameters corresponding to extreme weather events or other special circumstances recorded in history, such as high temperatures during heat waves, high humidity after floods, etc. These data help understand the health response of the population under extreme conditions and predict the risks under similar circumstances in the future. Extreme events that occur less frequently but have significant impacts on health are selected as references to better assess rare but potentially dangerous environmental changes.

[0111] Health threshold reference for different groups of people: Taking into account the differences in environmental sensitivity of different age groups, genders, underlying disease states, etc., personalized health thresholds are set for different groups of people as reference standards. For example, the elderly, children, and people with respiratory diseases may need lower safety thresholds.

[0112] The setting of health thresholds should be based on existing medical research and epidemiological survey results to ensure their scientificity and rationality.

[0113] Collect the latest research results in related fields, especially research reports on the impact of specific environmental factors on human health. Summarize historical environmental monitoring data and public health statistics in the target area to ensure that the time span of the data is long enough and covers a wide range. Conduct a preliminary analysis of the collected data to determine which parameters have the greatest impact on health, and preliminarily select reference objects based on this. Use statistical methods or machine learning algorithms to verify the validity of the selected reference objects, such as by building a predictive model to test the performance of the model under different reference objects. With the continuous accumulation of new data and technological advances, regularly review and update the selection of reference objects to ensure that they always reflect the latest scientific understanding and technical levels. According to changes in actual application scenarios (such as the emergence of new health threats), introduce new reference objects or adjust the standards of existing reference objects in a timely manner.

[0114] The selection of appropriate reference objects requires not only the support of scientific theories, but also the combination of practical application background and regional characteristics. By comprehensively applying the above principles and methods, the accuracy of the multi-reference object relationship segmentation model in the analysis of the association between environmental characteristics and health risks can be effectively improved, thus providing more reliable support for public health management and individual health management.

[0115] For example, suppose you are analyzing the air quality of a city during the summer and its impact on human health. The multi-reference relational segmentation model works like this:

[0116] Reference for ideal healthy environment: Use the PM2.5 safety limit recommended by the organization.

[0117] Reference to historical average value: The average PM2.5 concentration in summer of the city over the past five years is used.

[0118] Reference for special groups: Taking into account that the elderly are more vulnerable to the effects of air pollution, a lower health threshold is set for residents over 60 years old.

[0119] Dynamically adjust weights: Analyze the interaction between the current PM2.5 concentration and other environmental factors (such as temperature and humidity), and adjust their weights in the first eigenvector accordingly.

[0120] The multi-reference object relationship segmentation model can maintain high accuracy and adaptability in complex and changing real-world environments, providing strong support for public health management and individual health management. This method not only improves the accuracy of health risk prediction, but also provides a solid theoretical basis for formulating scientific and reasonable preventive measures.

[0121] Through the above steps, the first eigenvector is obtained, where the specific environmental conditions refer to physical, chemical or biological factors that can be directly measured and observed. For example, the concentration of pollutants in the air (such as PM2.5, PM10, sulfur dioxide, nitrogen oxides, etc.), the content of harmful substances in water bodies, the degree of soil pollution, noise level, radiation intensity, and meteorological parameters such as temperature and humidity.

[0122] Health risk associations: These feature vectors are not just independent data points, they also imply potential links with human health conditions. For example, long-term exposure to high concentrations of fine particulate matter (PM2.5) may increase the risk of respiratory diseases; excessive lead in drinking water may affect children's cognitive development; the incidence of heat stroke or heat stroke may increase under hot weather conditions, etc.

[0123] Comprehensive impact assessment: The first eigenvector generated by analyzing a large amount of data through machine learning algorithms is actually a quantitative representation of the combined impact of multiple environmental factors. This means that each eigenvalue not only reflects the status of a single environmental factor, but also takes into account the interaction between different factors and their cumulative effects on human health.

[0124] Using such feature vectors, a model can be built to predict the health risks faced by residents in a specific environment. For example, the feature vectors can be updated in combination with real-time monitoring data, and the public health early warning system can be adjusted accordingly to take measures in advance to reduce adverse effects. For individuals, customized health management plans can also be provided based on the environmental feature vectors of the individual's area. For example, protective equipment, dietary advice, or exercise schedules suitable for local conditions can be recommended to reduce health risks caused by environmental factors.

[0125] By introducing multiple reference settings, the model compares current environmental conditions with ideal health standards, historical data, and thresholds for specific populations, directly linking environmental characteristics with potential health risks. The self-attention mechanism enables the model to focus on the features that have the most significant impact on health, while the multi-layer perceptron (MLP) technology captures the complex interactions between different environmental variables, ensuring that the first eigenvector contains filtered key information. The feature mapping and vectorization steps compress high-dimensional data into a low-dimensional but information-rich first eigenvector, which comprehensively reflects the environmental status and its potential impact on health. Finally, the health risk association analysis evaluates the quantitative relationship between environmental conditions and health risks, further strengthening the importance scores of environmental features in the first eigenvector and their interaction patterns.

[0126] S140, collecting data at different time points at the same location, and generating data pairs of healthy status and unhealthy status, and performing comparative analysis on the data pairs to obtain analysis results.

[0127] In this embodiment, it specifically refers to the comprehensive evaluation of environmental condition changes and their impact on health obtained by comparing data pairs at different time points (healthy state period and unhealthy state period) at the same location. In more detail:

[0128] For each monitored environmental indicator (such as air quality, noise level, humidity, etc.), calculate the change in value from the selected healthy state period to the unhealthy state period. This includes but is not limited to the absolute change, relative change percentage, etc.

[0129] Apply statistical methods or machine learning algorithms to analyze these data over time to identify possible long-term trends. This process aims to find out which factors show persistent changes rather than just short-term fluctuations, thereby helping to identify those factors that truly have a long-term impact on health.

[0130] Combine the two aspects above to form a more comprehensive understanding. For example, if an environmental indicator not only shows a significant rate of change, but also has a clear long-term upward or downward trend, it may be a key area that needs special attention.

[0131] In one embodiment, the above-mentioned step S140 may include steps S141 - S143 .

[0132] S141. Determine a healthy state period as a reference to obtain healthy state data, and use this to select unhealthy state data.

[0133] In this embodiment, one or more time periods need to be selected as references for "health status". These time periods should be identified as periods when environmental conditions are within an ideal or safe range, i.e., no significant health risk reports are made during this period or environmental monitoring indicators are within the preset safety threshold range. The selected health status period is used to obtain a data set representing normal or safe environmental conditions, which will serve as the basis for subsequent comparative analysis.

[0134] S142: Calculate the change rate of each monitoring indicator according to the unhealthy state data and the healthy state data.

[0135] In this embodiment, once the data set of the healthy state is determined, the next step is to select the data of the unhealthy state from different time points in the same geographical location. Unhealthy state refers to the time period when the environmental conditions deviate from the normal or safe range. For each monitoring indicator (such as air quality, humidity, temperature, etc.), calculate the rate of change from the healthy state to the unhealthy state. The rate of change can be an absolute change, a percentage change, or other suitable way to measure the degree of change. This step is intended to quantify the changes in environmental conditions, thereby providing a specific numerical basis for further analysis.

[0136] S143. Using statistical methods or machine learning algorithms to identify long-term trends based on the data of the unhealthy state and the data of the healthy state, and combining them with the rate of change to form an analysis result.

[0137] In this embodiment, the last step is to use statistical methods or machine learning algorithms to analyze the change patterns of the data over time to identify any possible long-term trends. This method can help distinguish between accidental short-term fluctuations and persistent long-term changes. The identification of long-term trends combined with the rate of change of each monitoring indicator can more accurately reveal which environmental characteristics are most likely to affect health and assess the extent of these impacts. The final analysis results include not only the changes in each monitoring indicator, but also their trends over time and a comprehensive evaluation of their potential impact on health.

[0138] Through the above steps, it is possible to systematically compare environmental conditions at different points in time in the same place, identify key factors that may cause health problems, and assess their importance and impact. Such analysis is of great significance for understanding and predicting possible health risks in a specific environment.

[0139] In addition, Apache Spark is used as the main large-scale data processing engine because its support for in-memory computing makes it perform well in processing large-scale data sets. Hadoop HDFS can be used to store raw data for a long time, while Spark Streaming can realize real-time data stream processing. Build a multi-node distributed computing cluster, including a master node (Master Node) responsible for coordinating task allocation, and several worker nodes (Worker Nodes) to process data in parallel. Considering the cost-effectiveness ratio, you can choose the elastic computing resources provided by cloud service providers (such as AWS, Azure, Google Cloud).

[0140] The Do-Calculus theoretical framework is introduced to analyze the relationship between variables under intervention conditions, thereby revealing causal effects. Specifically, the potential causal paths between environmental factors and health outcomes can be represented by constructing directed acyclic graphs (DAGs). Causal models based on observational and experimental data are developed using machine learning libraries (such as Scikit-Learn, TensorFlow) combined with causal reasoning toolkits (such as DoWhy). These models can help understand which environmental characteristics are most likely to cause the occurrence of specific health problems and assess their extent of impact. In order to better understand and communicate causal relationships, interactive charts are created using visualization tools (such as Plotly, Matplotlib) to display health risk prediction results under different environmental conditions. In addition, natural language descriptions can be generated to provide clear and easy-to-understand insights for non-technical personnel.

[0141] Perform necessary cleanup, normalization, and dimensionality reduction operations on input data to reduce unnecessary computational burden. At the same time, consider using data compression technology to reduce storage requirements and transmission overhead. Regularly evaluate the performance of existing models, adjust hyperparameter settings based on feedback, and introduce new features or improve algorithm logic to continuously improve model accuracy and generalization capabilities.

[0142] S150: Generate a second feature vector according to the analysis result.

[0143] In one embodiment, the above-mentioned step S150 may include steps S151 - S152 .

[0144] S151. Allocate a value representing the degree of deterioration of each monitoring indicator according to the analysis result; for multiple monitoring indicators, assign different weights and then perform weighted summation to form a comprehensive index.

[0145] In this step, based on the previous analysis of changes in environmental conditions and their impact on health, the system assigns a value to each monitored indicator (e.g., air quality, noise level, humidity, etc.) that directly reflects the degree of deterioration or improvement of the indicator from a healthy state to an unhealthy state. This value can be determined in a variety of ways, such as using a rate of change calculated using a statistical model, an expert assessment score, or other suitable methods.

[0146] Next, considering that different monitoring indicators may have different potential impacts on health, these indicators will be assigned corresponding weights. The weight setting can be based on historical data analysis, professional field knowledge, or trained through machine learning algorithms. Once the weight of each indicator is determined, a comprehensive index can be calculated by weighted summation. The purpose of this step is to integrate the information of multiple monitoring indicators into a single quantitative value, thereby simplifying the subsequent processing process.

[0147] In this step, based on the previous analysis of environmental condition changes and their impact on health, the system not only assigns a value to each monitoring indicator (e.g., air quality, noise level, humidity, etc.) to directly reflect the degree of deterioration or improvement of the indicator from the healthy state period to the unhealthy state period, but also further considers the following information to construct a richer and more explanatory second feature vector:

[0148] Timestamp and geographic location information: Timestamp: records the measurement time and the time window for calculating the change rate of each monitoring indicator, which helps track the time trend of air quality changes. Geographic location information: clearly indicates the specific location where data was collected, which is crucial for cross-regional comparisons.

[0149] Rate of change and volatility: Rate of change: In addition to the absolute change value, it also includes the speed of change of pollutant concentration per unit time. Volatility: describes the degree of fluctuation of pollutant concentration over a period of time. High volatility may indicate unstable atmospheric conditions.

[0150] Evaluation of the effectiveness of air quality improvement measures: Policy response: If new environmental regulations are implemented or emission reduction measures are taken, the impact of these measures on air quality can be evaluated. Natural resilience: Measures the role of natural factors in improving air quality and the ability of the ecosystem to purify pollutants on its own.

[0151] Related health risk indicators: Disease incidence: Linked to public health data in the same period, such as the incidence of respiratory diseases or other diseases related to air pollution. Exposure of susceptible populations: Considering that children, the elderly and people with chronic diseases are more susceptible to air pollution, the exposure of specific groups is added as an additional dimension.

[0152] Economic and social costs: Economic losses: Estimate the direct economic losses caused by deteriorating air quality. Social response: Pay attention to the public's attention to air quality issues and social reactions.

[0153] Model Confidence and Uncertainty: Confidence Intervals: Provide confidence intervals for each predicted value. Uncertainties: Identify and quantify factors that may bias the predictions.

[0154] In this way, the value assigned to each monitoring indicator not only reflects the degree of deterioration, but also contains multi-dimensional information, thus forming a richer comprehensive index. The weight setting of different monitoring indicators also takes into account the influence of the above aspects, ensuring the final

[0155] S151. Normalize the generated comprehensive index so that the data of all regions fall within the same range to obtain a second eigenvector.

[0156] In order to ensure the comparability of data between different regions and make the comprehensive index of all regions fall within the same range, it is necessary to normalize the comprehensive index obtained in the above steps. Normalization is a common data preprocessing method that can convert the original data to a specific range (usually the [0,1] interval) to eliminate the impact of dimension and improve algorithm performance.

[0157] Specifically, the normalization method used in this embodiment may be Min-Max Normalization, that is: Among them, OriginalValue is the unprocessed comprehensive index, Min Value and Max Value are the minimum and maximum values ​​of the comprehensive index in all regions respectively. After such a conversion, the comprehensive index of all regions will be mapped to a fixed range between 0 and 1, and the final vector is the "second eigenvector" referred to.

[0158] Through the above two steps, not only the effective integration of multi-dimensional environmental monitoring data is achieved, but also the data between different regions can be directly compared, which provides a solid foundation for further analyzing the pattern of environmental quality changes and formulating targeted health protection measures.

[0159] In addition, in order to make the second eigenvector more comprehensive in reflecting the changes in air quality and its potential impact, the aforementioned additional information (such as timestamp, geographic location information, rate of change, volatility, policy response effects, etc.) is also retained during the normalization process. This information together constitutes a multidimensional eigenvector, which can not only serve as a comprehensive index to characterize the historical changes in air quality, but also become a bridge connecting multiple fields such as environmental science, public health, economics and sociology, supporting more complex and in-depth analysis. This multidimensional eigenvector can provide a more solid data foundation for the formulation of effective environmental protection strategies.

[0160] In summary, by including the above information, the second eigenvector not only realizes the effective integration of multidimensional environmental monitoring data, but also ensures that data between different regions can be directly compared, providing a solid foundation for further analyzing environmental quality change patterns and formulating targeted health protection measures.

[0161] S160: Determine a final risk level and a location of an affected area according to the first feature vector and the second feature vector.

[0162] In one embodiment, the above-mentioned step S160 may include steps S161 - S163 .

[0163] S161. Fuse the first eigenvector with the second eigenvector to obtain a fusion result.

[0164] In this embodiment, the previously constructed first eigenvector (mainly reflecting the basic state of environmental monitoring indicators) is fused with the second eigenvector (comprehensively reflecting the multidimensional information of air quality changes and their potential impacts). The fusion process is not just a simple addition or merging of the two vectors, but a specific algorithm (such as principal component analysis PCA, factor analysis FA, etc.) is used to comprehensively consider all dimensions in the two eigenvectors, aiming to capture the complex interaction between the two and generate a more comprehensive and explanatory new eigenvector, that is, the fusion result. This fusion result not only retains the important characteristics of the original data, but also can more accurately reflect the overall trend of environmental conditions and the possible health and socioeconomic impacts.

[0165] Specifically, in this embodiment, in order to more accurately capture the overall trend of environmental conditions and the possible health and socioeconomic impacts, the previously constructed first eigenvector (mainly reflecting the basic state of environmental monitoring indicators) and the second eigenvector (comprehensively reflecting the multidimensional information of air quality changes and their potential impacts) are fused. This fusion is not just a simple addition or merging operation, but through a series of carefully designed data processing steps, it is ensured that the fused result can retain the key characteristics of the original data and reveal the complex interaction between the two.

[0166] According to the data properties and technical requirements of the two feature vectors, a variety of advanced data analysis techniques can be selected for fusion, such as principal component analysis (PCA), factor analysis (FA), linear discriminant analysis (LDA), etc. These algorithms can effectively reduce the dimension, remove redundant information, and find the feature combination that best represents the essence of the data.

[0167] Before the formal fusion, the two feature vectors need to be preprocessed, including but not limited to standardization / normalization, to ensure that the values ​​at different scales will not cause deviations in the fusion results; at the same time, check whether there are missing values ​​or outliers, and take appropriate measures to correct them.

[0168] Once ready, the selected algorithm can be used to perform the fusion. For example, in the case of PCA, the covariance matrix between the two eigenvectors is calculated first, and then the eigenvalues ​​and eigenvectors are solved, and a new coordinate system is reconstructed based on this. In this new coordinate system, the original features are rearranged and combined into a new set of unrelated variables, which are the fused eigenvectors. For FA, it attempts to extract the common factors hidden behind the observed variables, thereby achieving data dimensionality reduction while retaining the original information to the greatest extent.

[0169] After the fusion is completed, the quality of the fusion results needs to be further evaluated. For example, cross-validation methods can be used to test whether the model prediction performance has been improved, or visualization tools can be used to visually compare the changes in data distribution before and after fusion to ensure that the newly generated feature vectors have indeed enhanced the depth of understanding of environmental conditions and associated impacts.

[0170] The final fusion result not only inherits the basic environmental status information carried by the first eigenvector, but also incorporates the rich details about air quality changes and their potential impacts in the second eigenvector. This makes the fused eigenvector a powerful data representation, which can be used for subsequent risk assessment, affected area location and other advanced applications, and also provides valuable data support for interdisciplinary research, such as connecting environmental science, public health, economics and sociology, and promoting more innovative discoveries and decision-making that serves social development.

[0171] S162: Input the fusion result into the risk assessment model for processing to obtain the final risk level.

[0172] In this embodiment, the fusion result obtained in the previous step is used as input and sent to a pre-trained risk assessment model for processing. The risk assessment model is established based on machine learning or other advanced statistical methods. It can learn risk patterns under different environmental conditions based on historical data, and predict the risk level in the current or future period of time. Specifically, the model will calculate one or more risk scores based on the characteristic values ​​in the fusion results, and these scores will be mapped to a predefined risk level system (such as low, medium, and high risks). In addition, in order to improve the prediction accuracy, the model will also be dynamically adjusted in combination with the latest weather forecast information, policy changes and other factors. The final output is a clear risk level, which provides a scientific basis for the subsequent adoption of corresponding protective measures.

[0173] Specifically, the training process of the risk assessment model is as follows:

[0174] First, a large amount of historical data covering multiple dimensions needs to be collected. These data include but are not limited to the first eigenvector (the basic status of environmental monitoring indicators) and the second eigenvector (multidimensional information on changes in air quality and their potential impacts). In addition, consideration should be given to introducing data on other relevant factors, such as weather forecast information, policy changes, etc. The original data should be cleaned to remove outliers and noise, fill in missing data points, and ensure the quality and integrity of the data. The values ​​in the two eigenvectors should be standardized or normalized to ensure that data of different scales will not deviate during the fusion process.

[0175] Through domain expert knowledge and technical means (such as PCA, FA, etc.), the key features that best represent environmental conditions and health impacts are selected. Advanced data analysis techniques (such as PCA, FA, LDA, etc.) are used to comprehensively consider all dimensions of the two feature vectors, capture the complex interactions between the two, and generate a more comprehensive and explanatory new feature vector - the fusion result. This process not only retains the important characteristics of the original data, but also enhances the understanding of the overall trend of environmental conditions.

[0176] Choose an appropriate machine learning or statistical model architecture based on the task requirements. For example, you can use decision trees, random forests, SVMs, neural networks, etc. For time series prediction problems, RNNs, LSTMs, or GRUs may be more appropriate. Choose an appropriate loss function for the optimization objective, such as mean squared error (MSE) for regression tasks and cross entropy loss for classification tasks. Find the best hyperparameter combination to improve model performance through grid search, random search, or Bayesian optimization. Train the selected model using the preprocessed and feature engineered dataset. During this period, an early stopping strategy can be used to prevent overfitting.

[0177] In order to adapt to changing actual conditions, the model should have the ability to continuously learn, that is, to gradually update its own parameters as new data arrives without having to retrain the entire model. Regularly update the data set used to train the model, especially adding the latest weather forecast information, policy changes and other factors to ensure that the model always makes predictions based on the most cutting-edge information.

[0178] Use methods such as cross-validation to evaluate the generalization ability of the model to ensure that its performance on unseen data is stable and reliable. If possible, you can also apply the model to independent data sets or real-world scenarios to further test its effectiveness and practicality.

[0179] The risk scores output by the model are mapped to a predefined risk level system (such as low, medium, and high risk) to facilitate understanding and taking appropriate protective measures. GIS technology is used to conduct spatial distribution analysis, spatial correlation research between feature vectors, boundary delineation and buffer zone setting of high-risk areas, as well as public participation and feedback mechanisms to determine the location of affected areas.

[0180] In summary, the training of a "risk assessment model" is a systematic process that not only relies on high-quality data preparation and feature engineering, but also requires careful selection of algorithms that are suitable for the task requirements and continuous iterative optimization to improve the performance of the model. Ultimately, such a model can provide accurate risk predictions, support scientific decision-making, and promote the development of public health management and services.

[0181] S163. Determine the location of affected areas through spatial distribution analysis integrating GIS technology, spatial correlation study between feature vectors, boundary delineation and buffer zone setting of high-risk areas, and public participation and feedback mechanism.

[0182] Using GIS technology, we can draw a spatial distribution map represented by each feature vector, which can intuitively show the differences in environmental health impacts in different regions. By superimposing map layers at different time points, we can track the development trajectory of risk areas and help identify which places are most likely to become new hotspots.

[0183] Spatial association analysis: Study the spatial association between the first eigenvector and the second eigenvector to find out whether there is a specific pattern or regularity between the two. For example, are certain types of environmental degradation always accompanied by increased health problems among residents in a specific area? This analysis can help reveal hidden spatial relationships and better locate the communities most threatened.

[0184] Boundary delineation and buffer zone setting: Based on the above analysis results, the exact boundaries of the high-risk areas are clearly marked on the map. In addition, a safety buffer zone of a certain width can be set up around these core areas to take more proactive preventive measures to prevent the risk from spreading to surrounding areas.

[0185] Public participation and feedback mechanism: Encourage local residents to participate in the risk assessment process and collect their opinions and suggestions, especially for small-scale or temporary risk sources that may be overlooked. At the same time, it is also necessary to establish effective information feedback channels so that the public can be informed of the latest risk status and respond accordingly.

[0186] In summary, through the above process, the first eigenvector of environmental health impact and the second eigenvector of health status change can be effectively combined to accurately determine the final risk level and the location of the affected area. This method not only overcomes the lag and one-sidedness problems existing in traditional infectious disease monitoring methods, but also provides strong technical support for urban public health management.

[0187] S170: Output the final risk level and the location of the affected area to the terminal for display on the terminal.

[0188] Outputting the final risk level and the location of the affected area and displaying it on the terminal ensures the immediacy of information transmission, allowing relevant departments and individuals to respond quickly to potential threats. This process enhances information transparency, builds public trust in management agencies, and promotes community participation in risk prevention and control measures. At the same time, using geolocation technology to accurately identify risk areas helps users intuitively understand their own safety status and take appropriate actions. In addition, presenting complex information through data visualization lowers the threshold for understanding and improves the readability and practicality of information.

[0189] The above-mentioned intelligent prevention and control method integrating photoelectric sensors and big data analysis forms an adaptive monitoring network to capture dynamic changes in the environment by deploying photoelectric sensors in the target area that can automatically adjust their layout according to environmental changes. The collected initial data is preprocessed and feature extracted, and then input into a pre-trained model to identify environmental features that may affect health and form a first eigenvector. The system also collects data from different time periods for comparative analysis to evaluate changes in health status and generate a second eigenvector representing the degree of environmental deterioration. Using the fusion results of the first and second eigenvectors, the system determines the final risk level and the location of the affected area to support precise public health emergency response. The entire process integrates GIS technology for spatial analysis and public participation mechanisms, ensuring the dynamic and adaptable nature of the monitoring strategy and helping to reduce the risk of disease transmission.

[0190] Figure 2 is a schematic block diagram of an intelligent prevention and control system 300 integrating photoelectric sensors and big data analysis provided by an embodiment of the present invention. Figure 2 As shown, corresponding to the above intelligent prevention and control method integrating photoelectric sensors and big data analysis, the present invention also provides an intelligent prevention and control system 300 integrating photoelectric sensors and big data analysis. The intelligent prevention and control system 300 integrating photoelectric sensors and big data analysis includes a unit for executing the above intelligent prevention and control method integrating photoelectric sensors and big data analysis, and the system can be configured in the system. Specifically, please refer to Figure 2 The intelligent prevention and control system 300 integrating photoelectric sensors and big data analysis includes a data acquisition unit 301, an extraction unit 302, a first vector determination unit 303, a history comparison unit 304, a second vector generation unit 305, a level position determination unit 306 and an output unit 307.

[0191] The data acquisition unit 301 is used to acquire data collected by multiple photoelectric sensor nodes deployed in the target area to obtain initial data, wherein the photoelectric sensors automatically adjust their layout according to environmental changes without human intervention to form an adaptive monitoring network to capture environmental changes; the extraction unit 302 is used to preprocess and extract features of the initial data to obtain extraction results; the first vector determination unit 303 is used to input the extraction results into a pre-trained multi-reference object relationship segmentation model to extract environmental feature vectors that may affect health to obtain a first feature vector; the history comparison unit 304 is used to collect data at different time points in the same place and generate data pairs of healthy status and unhealthy status, and compare and analyze the data pairs to obtain analysis results; the second vector generation unit 305 is used to generate a second feature vector according to the analysis results; the level position determination unit 306 is used to determine the final risk level and the location of the affected area according to the first feature vector and the second feature vector; the output unit 307 is used to output the final risk level and the location of the affected area to the terminal for display on the terminal.

[0192] In one embodiment, the data acquisition unit 301 is used to:

[0193] In the target area, key locations are selected for the initial deployment of photoelectric sensors based on preset rules or expert knowledge. These rules take into account geographical features, historical environmental data, and the location of potential pollution sources. Each photoelectric sensor node is equipped with multiple sensors. The data streams from each sensor node are analyzed in real time to identify significant changes in the environment. When important environmental changes are detected, an adaptive adjustment program is started. The program combines the current environmental conditions with historical data trends to calculate the optimal new sensor location and parameter configuration plan. Based on the calculation results, a movement instruction or parameter update instruction is sent to a specific sensor node. After receiving the instruction, the sensor autonomously performs the corresponding action, including changing its own azimuth, tilt angle, or adjusting the sensing sensitivity. The specific sensor can automatically complete repositioning or parameter optimization based on built-in logic and external guidance. After each adjustment, the effect of the new layout is evaluated. If the expected performance indicators are not achieved, the iterative optimization process continues until the established standards are met, and then the data collected by multiple photoelectric sensor nodes deployed in the target area are obtained.

[0194] In one embodiment, the extraction unit 302 is used to:

[0195] The initial data is cleaned, standardized, time-series synchronized, and segmented to obtain a preprocessing result; static features and dynamic features are extracted and multi-source data is fused to obtain an extraction result.

[0196] In one embodiment, the first vector determining unit 303 is used to:

[0197] For each input sample in the extraction result, the multi-reference relationship segmentation model segments the input sample and compares it with each pre-defined reference one by one. Through the self-attention mechanism and multi-layer perceptron technology, several dimensions are selected from all environmental features that may affect health to form a multi-dimensional feature space; the feature space is vectorized to obtain a first feature vector, wherein the first feature vector includes environmental feature vectors that affect health. These feature vectors not only describe specific environmental conditions, but also imply the relationship between these conditions and health risks.

[0198] In one embodiment, the history comparison unit 304 is used to:

[0199] Determine a healthy status period as a reference to obtain healthy status data, and use this to select unhealthy status data; calculate the change rate of each monitoring indicator based on the unhealthy status data and the healthy status data; use statistical methods or machine learning algorithms to identify long-term trends based on the unhealthy status data and the healthy status data, and combine them with the change rate to form an analysis result.

[0200] In one embodiment, the second vector generating unit 305 is used to:

[0201] According to the analysis results, a numerical value representing the degree of deterioration is assigned to each monitoring indicator; for multiple monitoring indicators, different weights are assigned and then weighted summed to form a comprehensive index; the generated comprehensive index is normalized so that the data of all regions fall within the same range to obtain a second eigenvector.

[0202] In one embodiment, the level position determination unit 306 is configured to:

[0203] The first eigenvector is fused with the second eigenvector to obtain a fusion result; the fusion result is input into the risk assessment model for processing to obtain the final risk level; the location of the affected area is determined by integrating spatial distribution analysis of GIS technology, spatial correlation research between eigenvectors, boundary delineation and buffer zone setting of high-risk areas, and public participation and feedback mechanism.

[0204] It should be noted that technical personnel in the relevant field can clearly understand that the specific implementation process of the above-mentioned intelligent prevention and control system 300 integrating photoelectric sensors and big data analysis and each unit can refer to the corresponding description in the aforementioned method embodiment. For the convenience and conciseness of the description, it will not be repeated here.

[0205] The above-mentioned intelligent prevention and control system 300 integrating photoelectric sensors and big data analysis can be implemented in the form of a computer program. The computer program can be used in Figure 3 Runs on the computer device shown.

[0206] See also Figure 3 , Figure 3 5 is a schematic block diagram of a computer device provided in an embodiment of the present application. The computer device 500 may be a server, wherein the server may be an independent server or a server cluster composed of multiple servers.

[0207] See also Figure 3 The computer device 500 includes a processor 502 , a memory and a network interface 505 connected via a system bus 501 , wherein the memory may include a non-volatile storage medium 503 and an internal memory 504 .

[0208] The non-volatile storage medium 503 can store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions, which, when executed, can enable the processor 502 to execute an intelligent prevention and control method integrating photoelectric sensors and big data analysis.

[0209] The processor 502 is used to provide computing and control capabilities to support the operation of the entire computer device 500 .

[0210] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute an intelligent prevention and control method that integrates photoelectric sensors and big data analysis.

[0211] The network interface 505 is used to communicate with other devices over the network. Figure 3 The structure shown in the figure is merely a block diagram of a portion of the structure related to the solution of the present application, and does not constitute a limitation on the computer device 500 to which the solution of the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0212] The processor 502 is used to run the computer program 5032 stored in the memory to implement the following steps:

[0213] Acquire data collected by multiple photoelectric sensor nodes deployed in the target area to obtain initial data, wherein the photoelectric sensors automatically adjust their layout according to environmental changes without human intervention to form an adaptive monitoring network to capture environmental changes; preprocess and extract features of the initial data to obtain extraction results; input the extraction results into a pre-trained multi-reference object relationship segmentation model to extract environmental feature vectors that may affect health to obtain a first feature vector; collect data at different time points in the same location, and generate data pairs of healthy and unhealthy states, and compare and analyze the data pairs to obtain analysis results; generate a second feature vector based on the analysis results; determine the final risk level and the location of the affected area based on the first feature vector and the second feature vector; output the final risk level and the location of the affected area to a terminal for display on the terminal.

[0214] It should be understood that in the embodiment of the present application, the processor 502 may be a central processing unit (CPU), and the processor 502 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0215] It can be understood by those skilled in the art that all or part of the processes in the method for implementing the above embodiment can be completed by instructing the relevant hardware through a computer program. The computer program includes program instructions, and the computer program can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiment of the above method.

[0216] Therefore, the present invention also provides a storage medium. The storage medium may be a computer-readable storage medium. The storage medium stores a computer program, wherein when the computer program is executed by a processor, the processor executes the following steps:

[0217] Acquire data collected by multiple photoelectric sensor nodes deployed in the target area to obtain initial data, wherein the photoelectric sensors automatically adjust their layout according to environmental changes without human intervention to form an adaptive monitoring network to capture environmental changes; preprocess and extract features of the initial data to obtain extraction results; input the extraction results into a pre-trained multi-reference object relationship segmentation model to extract environmental feature vectors that may affect health to obtain a first feature vector; collect data at different time points in the same location, and generate data pairs of healthy and unhealthy states, and compare and analyze the data pairs to obtain analysis results; generate a second feature vector based on the analysis results; determine the final risk level and the location of the affected area based on the first feature vector and the second feature vector; output the final risk level and the location of the affected area to a terminal for display on the terminal.

[0218] It should be noted that the functions or steps that can be implemented by the above storage medium or computer device can be correspondingly referred to the relevant descriptions on the server side and the client side in the aforementioned method embodiment. In order to avoid repetition, they will not be described one by one here.

[0219] The storage medium may be a USB flash drive, a mobile hard disk, a read-only memory (ROM), a magnetic disk, or an optical disk, etc., which are computer-readable storage media that can store program codes.

[0220] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. 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 the present invention.

[0221] In the several embodiments provided by the present invention, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are only schematic. For example, the division of each unit is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0222] The steps in the method of the embodiment of the present invention can be adjusted in order, combined and deleted according to actual needs. The units in the system of the embodiment of the present invention can be combined, divided and deleted according to actual needs. In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0223] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, terminal, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention.

[0224] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed by the present invention, and these modifications or replacements should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. An intelligent prevention and control method integrating photoelectric sensors and big data analysis, characterized in that: include: Acquiring data collected by a plurality of photoelectric sensor nodes deployed in a target area to obtain initial data, wherein the photoelectric sensors automatically adjust their layout according to environmental changes without human intervention to form an adaptive monitoring network to capture environmental changes; Preprocessing and feature extraction are performed on the initial data to obtain extraction results; Inputting the extraction result into a pre-trained multi-reference object relationship segmentation model to extract environmental feature vectors that may affect health, so as to obtain a first feature vector; Collecting data at different time points at the same location, generating data pairs of healthy state and unhealthy state, and performing comparative analysis on the data pairs to obtain analysis results; generating a second feature vector according to the analysis result; Determine a final risk level and a location of an affected area according to the first feature vector and the second feature vector; The final risk level and the location of the affected area are output to the terminal for display on the terminal.

2. The intelligent prevention and control method integrating photoelectric sensors and big data analysis according to claim 1 is characterized in that: The photoelectric sensors automatically adjust their layout according to environmental changes without human intervention, forming an adaptive monitoring network to capture environmental changes, including: Select key locations within the target area for initial deployment of photoelectric sensors based on preset rules or expert knowledge. These rules take into account geographic features, historical environmental data, and the location of potential pollution sources. Each photoelectric sensor node is equipped with multiple sensors. Real-time analysis of data streams from each sensor node to identify significant changes in the environment; when significant environmental changes are detected, an adaptive adjustment program is initiated, which combines current environmental conditions with historical data trends to calculate the optimal new sensor location and parameter configuration plan; According to the calculation results, a movement command or parameter update command is sent to a specific sensor node. After receiving the command, the sensor autonomously performs the corresponding action, including changing its own azimuth, tilt angle or adjusting the sensing sensitivity. The specific sensor can automatically complete repositioning or parameter optimization based on the built-in logic and external guidance; After each adjustment, the effect of the new layout is evaluated, and if the expected performance indicators are not achieved, the iterative optimization process continues until the established criteria are met, and then the data collected by multiple optoelectronic sensor nodes deployed in the target area are obtained.

3. The intelligent prevention and control method integrating photoelectric sensors and big data analysis according to claim 1 is characterized in that: The preprocessing and feature extraction of the initial data to obtain an extraction result includes: Performing data cleaning, standardization, time series synchronization, and data segmentation on the initial data to obtain a preprocessing result; Static features and dynamic features are extracted from the preprocessing results, and multi-source data is fused to obtain extraction results.

4. The intelligent prevention and control method integrating photoelectric sensors and big data analysis according to claim 1 is characterized in that: The step of inputting the extraction result into a pre-trained multi-reference object relationship segmentation model to extract an environmental feature vector that may affect health, so as to obtain a first feature vector, comprises: For each input sample in the extraction result, the multi-reference object relationship segmentation model segments the input sample and compares it with each pre-defined reference object one by one. Through the self-attention mechanism and multi-layer perceptron technology, several dimensions are selected from all environmental features that may affect health to form a multi-dimensional feature space. The feature space is vectorized to obtain a first feature vector, wherein the first feature vector includes environmental feature vectors that affect health. These feature vectors not only describe specific environmental conditions, but also imply the relationship between these conditions and health risks.

5. The intelligent prevention and control method integrating photoelectric sensors and big data analysis according to claim 1 is characterized in that: The data collected at different time points at the same location are generated into data pairs of healthy state and unhealthy state, and the data pairs are compared and analyzed to obtain analysis results, including: Determine a healthy state period as a reference to obtain healthy state data and use it to select unhealthy state data; Calculate the change rate of each monitoring indicator according to the unhealthy state data and the healthy state data; Based on the data of the unhealthy state and the data of the healthy state, a statistical method or a machine learning algorithm is used to identify the long-term trend, and the trend is combined with the rate of change to form an analysis result.

6. The intelligent prevention and control method integrating photoelectric sensors and big data analysis according to claim 5 is characterized in that: Generating a second feature vector according to the analysis result includes: According to the analysis results, a numerical value representing the degree of deterioration is assigned to each monitoring indicator; for multiple monitoring indicators, different weights are assigned and then weighted summed to form a comprehensive index; The generated comprehensive index is normalized so that the data of all regions fall within the same range to obtain the second eigenvector.

7. The intelligent prevention and control method integrating photoelectric sensors and big data analysis according to claim 1 is characterized in that: Determining a final risk level and a location of an affected area according to the first feature vector and the second feature vector includes: Fusing the first eigenvector with the second eigenvector to obtain a fusion result; Inputting the fusion result into the risk assessment model for processing to obtain a final risk level; The location of the affected area is determined through spatial distribution analysis integrating GIS technology, spatial association study between feature vectors, boundary delineation and buffer zone setting of high-risk areas, and public participation and feedback mechanism.

8. Intelligent prevention and control system integrating photoelectric sensors and big data analysis, characterized in that: include: A data acquisition unit, used to acquire data collected by a plurality of photoelectric sensor nodes deployed in a target area to obtain initial data, wherein the photoelectric sensors automatically adjust their layout according to environmental changes without human intervention to form an adaptive monitoring network to capture environmental changes; An extraction unit, used for preprocessing and feature extraction of the initial data to obtain an extraction result; A first vector determination unit, configured to input the extraction result into a pre-trained multi-reference object relationship segmentation model to extract an environmental feature vector that may affect health, so as to obtain a first feature vector; A historical comparison unit is used to collect data at different time points in the same place, generate data pairs of healthy state and unhealthy state, and compare and analyze the data pairs to obtain analysis results; A second vector generating unit, used for generating a second feature vector according to the analysis result; a level and position determination unit, configured to determine a final risk level and a position of an affected area according to the first feature vector and the second feature vector; The output unit is used to output the final risk level and the location of the affected area to the terminal for display on the terminal.

9. A computer device, characterized in that: The computer device comprises a memory and a processor, the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 7 when executing the computer program.

10. A storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images