A remote monitoring and data analysis method and system based on ATP fluorescence detection and internet of things

Abstract

The application provides a remote monitoring and data analysis method and system based on ATP fluorescence detection and the Internet of Things. The original detection data generated by the ATP fluorescence detection sensor is collected to obtain stable and high signal-to-noise ratio processed detection data. The processed detection data is efficiently transmitted to a cloud data center through the Internet of Things technology to generate comprehensive detection data after safe transmission. The detection data is comprehensively analyzed to identify the change mode of the biological pollution level and generate an alarm notification record. A biological pollution prediction model is constructed to generate a biological pollution prediction report. The ATP fluorescence detection frequency settings of each monitoring point are adjusted, the detection frequency of the monitoring point predicted to be a high-risk area is increased, the detection frequency of a low-risk area is reduced, and an optimized monitoring strategy is generated to improve the response speed and resource utilization efficiency of the monitoring system. The technical scheme provided by the application improves the real-time performance, accuracy and response efficiency of biological pollution monitoring.

Description

TECHNICAL FIELD

[0001] The embodiments of the present application relate to the technical field of environmental monitoring and data analysis, and particularly relate to a remote monitoring and data analysis method and system based on ATP fluorescence detection and Internet of Things. BACKGROUND

[0002] In the fields of environmental monitoring, food safety, and public health, rapid and accurate detection and early warning of biological contamination are of great importance. With the development of Internet of Things technology, Adenosine Triphosphate (ATP) fluorescence detection sensors can be used to monitor microbial activity in the environment in real time. This technology can provide immediate data feedback to help managers respond quickly and reduce the occurrence and impact of pollution incidents.

[0003] There are various technical means for monitoring biological contamination in the current market, such as traditional laboratory testing methods and chemical indicators. In recent years, solutions combining ATP fluorescence detection technology and Internet of Things platforms have gradually emerged. These systems usually include a sensor network deployed in the field to collect raw data, cloud computing services to store and process large amounts of data, and advanced data analysis algorithms to identify potential risk patterns and generate alert notifications. In addition, some advanced applications integrate machine learning models to improve prediction accuracy and automation levels.

[0004] Although existing ATP fluorescence detection combined with Internet of Things methods have solved the problem of real-time monitoring to some extent, there are still some challenges. First, raw detection data is often disturbed by noise, and direct use may lead to inaccurate analysis results. Second, many current systems lack effective data preprocessing mechanisms, which cannot guarantee the security and integrity of information transmission. Third, although some systems introduce time series analysis or simple statistical methods for data analysis, their prediction ability is limited when facing complex and variable actual scenarios, making it difficult to effectively respond to sudden situations. Finally, in terms of alarm response, existing systems often use fixed threshold settings, lack flexibility, and cannot dynamically adjust alarm levels and response strategies according to actual conditions, leading to unreasonable resource allocation. SUMMARY

[0005] The embodiments of the present application provide a remote monitoring and data analysis method and system based on ATP fluorescence detection and Internet of Things to solve the problems of real-time, accuracy, and low response efficiency of biological contamination monitoring in the prior art.

[0006] In a first aspect, the embodiments of the present application provide a remote monitoring and data analysis method based on ATP fluorescence detection and Internet of Things, comprising:

[0007] Collect raw detection data generated by the ATP fluorescence detection sensor, the raw detection data including fluorescence intensity values and their accompanying environmental parameters, and perform real-time denoising processing on the raw detection data to obtain processed detection data that is stable and has a high signal-to-noise ratio;

[0008] According to the processed detection data, the data is efficiently transmitted to a cloud data center via Internet of Things technology, and the comprehensive detection data after secure transmission is generated together with the stored historical detection data;

[0009] Using time series analysis algorithm combined with environmental parameters in the comprehensive detection data after secure transmission, the detection data is comprehensively analyzed to identify the change pattern of biological pollution level, and when the detection data exceeds the preset threshold dynamically adjusted based on the historical detection data, a multi-level alarm response system is activated, which can select appropriate communication channels to send customized alarm information to designated management personnel according to different levels of alarm types, and generate alarm notification records;

[0010] Based on the alarm notification records, historical detection data and environmental parameters in the comprehensive detection data after secure transmission, combined with real-time detection data, a biological pollution prediction model is constructed using machine learning algorithm to generate a biological pollution prediction report;

[0011] Using the biological pollution prediction report, combined with the problem area information reflected in the alarm notification records, adjust the ATP fluorescence detection frequency settings of each monitoring point, increase the detection frequency of monitoring points in high-risk areas predicted by the report, and reduce the detection frequency of low-risk areas, to generate an optimized monitoring strategy to improve the response speed and resource utilization efficiency of the monitoring system.

[0012] Optionally, the time series analysis algorithm is used to combine the environmental parameters in the comprehensive detection data after secure transmission to comprehensively analyze the detection data and identify the change pattern of biological pollution level, and when the detection data exceeds the preset threshold dynamically adjusted based on the historical detection data, a multi-level alarm response system is activated, which can select appropriate communication channels to send customized alarm information to designated management personnel according to different levels of alarm types, and generate alarm notification records, including:

[0013] Using time series analysis algorithm to analyze the fluorescence intensity values and their accompanying environmental parameters in the comprehensive detection data after secure transmission, extract key indicators reflecting biological pollution level and their change trend over time, and obtain time series analysis results of biological pollution level;

[0014] Based on the time series analysis result of the biological pollution level, compared with the stored historical detection data, the preset warning threshold is dynamically adjusted by using a statistical method, to obtain a dynamic warning threshold that adapts to the change of the biological pollution level under different environmental conditions;

[0015] According to the dynamic warning threshold, the time series analysis result of the biological pollution level is monitored, and when the detected fluorescence intensity value exceeds the dynamic warning threshold, a multi-level alarm response mechanism is immediately started, to obtain an alarm trigger signal;

[0016] Based on the alarm trigger signal, the alarm level is automatically determined according to the severity of the alarm, the most suitable communication channel is selected to send customized alarm information containing the alarm level, specific pollution location, cause analysis and preliminary suggested countermeasures to the preset management personnel or emergency team, and an alarm notification record is generated.

[0017] Optionally, the fluorescence intensity value and its accompanying environmental parameters in the comprehensive detection data after the secure transmission are analyzed by using a time series analysis algorithm, key indicators reflecting the biological pollution level and their change trends over time are extracted, and a time series analysis result of the biological pollution level is obtained, including:

[0018] The fluorescence intensity value and its accompanying environmental parameters in the comprehensive detection data after the secure transmission are decomposed and processed by using a time series decomposition technique, to separate three components of long-term trend, seasonal fluctuation and random fluctuation, and obtain a preliminary decomposition result;

[0019] Based on the preliminary decomposition result, a statistical regression analysis method is applied to identify the main environmental factors affecting the change of the fluorescence intensity value, a mathematical model between the fluorescence intensity value and the environmental parameters is established, and an environmental factor influence model is obtained;

[0020] The environmental factor influence model is used to combine the long-term trend and seasonal fluctuation in the preliminary decomposition result to predict the biological pollution level, extract key indicators reflecting the biological pollution level, including average fluorescence intensity and fluorescence intensity change rate, and generate a key indicator set;

[0021] According to the key indicator set, the change trend of the biological pollution level over time is analyzed, and the rising, falling or stable state of the biological pollution level is identified, to generate a time series analysis result of the biological pollution level.

[0022] Optionally, the fluorescence intensity value and its accompanying environmental parameters in the comprehensive detection data after the secure transmission are analyzed by using a time series analysis algorithm, key indicators reflecting the biological pollution level and their change trends over time are extracted, and a time series analysis result of the biological pollution level is obtained, including:

[0023] Using time series decomposition technology, the fluorescence intensity value F(t) and its accompanying environmental parameters in the comprehensive detection data after secure transmission are decomposed to separate three components: long-term trend T(t), seasonal fluctuation S(t), and random fluctuation R(t). The preliminary decomposition result is obtained as F(t) = T(t) + S(t) + R(t), where t represents time, F(t) represents the fluorescence intensity value at time t, T(t) represents the long-term trend, indicating the long-term variation of the fluorescence intensity value, S(t) represents the seasonal fluctuation, indicating the periodic variation of the fluorescence intensity value, and R(t) represents the random fluctuation, indicating the random variation of the fluorescence intensity value.

[0024] Based on the preliminary decomposition results, statistical regression analysis was applied to identify the main environmental factors E affecting the change in fluorescence intensity. i (t), establish the fluorescence intensity value F(t) and environmental parameter E i The following is a nonlinear mathematical model between (t), and the model for the influence of environmental factors:

[0025]

[0026] Among them; E i (t) represents the environmental parameter, indicating the i-th environmental parameter at time t; β0 is the intercept term, β i γ i and δ i These are the linear, quadratic, and cubic regression coefficients of the i-th environmental parameter, respectively. α1 and α2 are the regression coefficients of the long-term trend and seasonal fluctuation, respectively. n is the number of environmental parameters, and ∈ is the error term, thus obtaining the environmental factor impact model.

[0027] Using the aforementioned environmental factor impact model, and combining the long-term trend T(t) and seasonal fluctuation S(t) from the preliminary decomposition results, the level of biological pollution is predicted, and key indicators reflecting the level of biological pollution are extracted:

[0028] The average fluorescence intensity is calculated using the following formula.

[0029]

[0030] The fluorescence intensity change rate ΔF(t) can be calculated using the following formula:

[0031]

[0032] Where Δt represents the time interval, and represents the time interval for calculating the rate of change of fluorescence intensity;

[0033] Weighted moving average fluorescence intensity WMA(F(t)):

[0034]

[0035] where w i is the weight, k is the length of the time window; i represents the index within the time window;

[0036] Exponential smoothing fluorescence intensity ES(F(t)) is calculated by the following formula:

[0037] ES(F(t)) = aF(t) + (1-a)ES(F(t-1))

[0038] where a is the smoothing factor, usually taking a value between 0 and 1; a set of key indicators is generated;

[0039] According to the set of key indicators, the trend of the biological pollution level over time is analyzed, and the rising, falling or stable state of the biological pollution level is identified, and a time series analysis result of the biological pollution level is generated.

[0040] Optionally, based on the environmental parameters in the alarm notification record, historical detection data and comprehensive detection data after safe transmission, combined with real-time detection data, a biological pollution prediction model is constructed by using a machine learning algorithm, and a biological pollution prediction report is generated, including:

[0041] The environmental parameters in the alarm notification record, historical detection data, comprehensive detection data after safe transmission and real-time detection data are collected and preprocessed to obtain a high-quality data set;

[0042] According to the high-quality data set, feature selection and conversion are performed, the key features affecting biological pollution prediction are determined, and part of the features are converted or encoded as necessary to obtain feature data after feature selection and conversion;

[0043] Using the feature data after feature selection and conversion, support vector machine is selected as the machine learning algorithm to construct the biological pollution prediction model, and the appropriate kernel function is selected and the related parameters are adjusted to optimize the model performance;

[0044] Based on the constructed biological pollution prediction model, the model is trained and verified by using the cross-validation method to ensure that the model has good generalization ability and prediction accuracy, and an optimized biological pollution prediction model is obtained;

[0045] Using the optimized biological pollution prediction model, new or future data is predicted to generate a detailed biological pollution prediction report, and the content of the biological pollution prediction report includes the predicted time range, the prediction result, the potential high-risk area and the corresponding recommended measures.

[0046] Optionally, the feature data after feature selection and conversion is used to select a support vector machine as a machine learning algorithm to construct a biological contamination prediction model, and a suitable kernel function is selected to optimize the performance of the model, including:

[0047] The feature data after feature selection and conversion is used to select a support vector machine as a machine learning algorithm to model the biological contamination prediction task, and a preliminary vector machine model is obtained.

[0048] According to the requirements of the biological contamination prediction task, a suitable kernel function is selected for the vector machine model, and the selection of the kernel function is based on the nonlinear mapping ability and generalization ability of the model, so as to obtain a vector machine model with strong nonlinear mapping ability, thereby optimizing the performance of the model.

[0049] The preliminary vector machine model is used to adjust the penalty parameter and the kernel function parameter, and the adjustment process adopts a parameter optimization method to find the best parameter combination in a predefined parameter range, thereby obtaining an optimized parameter combination.

[0050] Optionally, the feature data after feature selection and conversion is used to select a support vector machine as a machine learning algorithm to construct a biological contamination prediction model, and a suitable kernel function is selected to optimize the performance of the model, including:

[0051] The feature data after feature selection and conversion is used to select a support vector machine as a machine learning algorithm to model the biological contamination prediction task, and a preliminary vector machine model is obtained.

[0052] The basic form of the vector machine model f(x) is represented as:

[0053]

[0054] Where N represents the number of support vectors, α i is the Lagrange multiplier, y i is the label of each support vector (+1 or -1), K(x, x i ) is the kernel function, and b is the bias term.

[0055] According to the requirements of the biological contamination prediction task, a suitable kernel function is selected for the vector machine model, and the selection of the kernel function is based on the nonlinear mapping ability and generalization ability of the model, so as to obtain a vector machine model with strong nonlinear mapping ability; common kernel functions include linear kernel, polynomial kernel and radial basis function kernel; the mathematical expression of the radial basis function kernel K(x, x i ) is:

[0056] K(x, x i ) = exp(-γ‖x-x i∥ 2 )

[0057] wherein, γ is a kernel function parameter, ∥x-x i ∥ represents the Euclidean distance between two vectors.

[0058] Optionally, the biological contamination prediction report is used in combination with the problem area information reflected in the alarm notification record to adjust the ATP fluorescence detection frequency settings of each monitoring point, increase the detection frequency for monitoring points predicted to be high-risk areas, and reduce the detection frequency for low-risk areas, to generate an optimized monitoring strategy to improve the response speed and resource utilization efficiency of the monitoring system, including:

[0059] The future biological contamination risk of each monitoring point is evaluated using the prediction results, prediction confidence, and potential impact analysis in the biological contamination prediction report, to obtain a risk area evaluation result;

[0060] Based on the risk area evaluation result, the actual contamination risk level of each monitoring point is comprehensively analyzed in combination with the problem area information reflected in the alarm notification record, to generate a risk assessment report for each monitoring point;

[0061] According to the risk assessment report for each monitoring point, for monitoring points predicted to be high-risk areas, the detection frequency is increased by adjusting the ATP fluorescence detection frequency settings to ensure that biological contamination events can be discovered and responded to in a timely manner, and a monitoring strategy for high-risk areas is generated;

[0062] For monitoring points predicted to be low-risk areas, the detection frequency is appropriately reduced based on the risk assessment report for each monitoring point to reduce resource waste and improve the operating efficiency of the monitoring system, and a monitoring strategy for low-risk areas is generated;

[0063] The monitoring strategy for high-risk areas and the monitoring strategy for low-risk areas are integrated to form a comprehensive and optimized monitoring strategy to improve the response speed and resource utilization efficiency of the monitoring system and ensure effective monitoring and management of biological contamination.

[0064] Optionally, the processed detection data is efficiently transmitted to a cloud data center via Internet of Things technology together with stored historical detection data to generate comprehensive detection data after secure transmission, including:

[0065] The processed detection data is used to form a structured detection data packet to obtain a structured detection data packet;

[0066] Based on the structured detection data packet, a secure encryption protocol in the Internet of Things technology is used to encrypt the detection data packet, ensuring the security of the data in the transmission process, preventing data from being stolen or tampered with, and generating an encrypted detection data packet;

[0067] The encrypted detection data packet is efficiently transmitted to a cloud data center through an Internet of Things network, data compression technology is used in the transmission process to reduce data transmission volume, and error control technology is used to ensure the integrity of data transmission, obtaining the detection data after safe transmission;

[0068] In the cloud data center, the detection data after safe transmission is integrated with the stored historical detection data to ensure the consistency and continuity of new and old data, and the comprehensive detection data after safe transmission is generated.

[0069] In a second aspect, the embodiments of the present application provide a remote monitoring and data analysis system based on ATP fluorescence detection and the Internet of Things, comprising:

[0070] A collection and processing module is configured to collect raw detection data generated by an ATP fluorescence detection sensor, the raw detection data including fluorescence intensity values and accompanying environmental parameters, and to perform real-time denoising processing on the raw detection data to obtain processed detection data that is stable and has a high signal-to-noise ratio;

[0071] A transmission module is configured to efficiently transmit the processed detection data to a cloud data center via Internet of Things technology to generate comprehensive detection data after safe transmission together with stored historical detection data;

[0072] An analysis and activation module is configured to use a time series analysis algorithm to analyze the detection data in combination with the environmental parameters in the comprehensive detection data after safe transmission, identify the change pattern of the biological contamination level, activate a multi-level alarm response system when the detection data exceeds a preset threshold dynamically adjusted based on the historical detection data, and generate an alarm notification record.

[0073] A construction module is configured to use a machine learning algorithm to construct a biological contamination prediction model based on the alarm notification record, the historical detection data, the environmental parameters in the comprehensive detection data after safe transmission, and real-time detection data, and generate a biological contamination prediction report.

[0074] The adjusting generation module is configured to utilize the biological contamination prediction report, combine the problem area information reflected in the alarm notification record, adjust the ATP fluorescence detection frequency setting of each monitoring point, increase the detection frequency of the monitoring point predicted to be a high-risk area, and decrease the detection frequency of a low-risk area, to generate an optimized monitoring strategy, so as to improve the response speed and resource utilization efficiency of the monitoring system.

[0075] In the embodiment of the present application, raw detection data generated by the ATP fluorescence detection sensor is collected, the raw detection data including fluorescence intensity values and accompanying environmental parameters, and the raw detection data is subjected to real-time denoising processing to obtain processed detection data that is stable and has a high signal-to-noise ratio; the processed detection data is efficiently transmitted to a cloud data center via Internet of Things technology, and comprehensive detection data after safe transmission is generated together with stored historical detection data; the detection data is comprehensively analyzed by using a time series analysis algorithm in combination with the environmental parameters in the comprehensive detection data after safe transmission, a change pattern of the biological contamination level is identified, and when the detection data exceeds a preset threshold dynamically adjusted based on the historical detection data, a multi-level alarm response system is activated, the multi-level alarm response system can select a suitable communication channel to send customized alarm information to designated management personnel according to different levels of alarm types, and an alarm notification record is generated; a biological contamination prediction model is constructed by using a machine learning algorithm based on the alarm notification record, the historical detection data, and the environmental parameters in the comprehensive detection data after safe transmission in combination with real-time detection data, and a biological contamination prediction report is generated; the ATP fluorescence detection frequency setting of each monitoring point is adjusted by utilizing the biological contamination prediction report in combination with the problem area information reflected in the alarm notification record, the detection frequency of the monitoring point predicted to be a high-risk area is increased, and the detection frequency of a low-risk area is decreased, to generate an optimized monitoring strategy, so as to improve the response speed and resource utilization efficiency of the monitoring system.

[0076] The technical solution of the present application has the following beneficial effects:

[0077] The embodiment of the present application improves the real-time performance, accuracy, and response efficiency of biological contamination monitoring by combining ATP fluorescence detection and Internet of Things technology, optimizes the monitoring strategy, and improves the resource utilization efficiency.

[0078] Further, the embodiment of the present application enhances the sensitivity and adaptability of the biological contamination monitoring system by time series analysis and dynamic adjustment of the alarm threshold, ensures the timeliness and accuracy of the alarm information, and effectively improves the early warning capability for biological contamination events.

[0079] Further, the biological pollution prediction model is constructed by adopting the machine learning algorithm, so that the prediction accuracy of the future biological pollution trend is improved, a detailed prediction report is generated, a scientific basis is provided for decision makers, and effective prevention and control measures can be taken in advance.

[0080] These and other aspects of the present application will become more fully understood from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0081] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0082] Figure 1 A flowchart of a remote monitoring and data analysis method based on ATP fluorescence detection and Internet of Things provided by the embodiments of the present application;

[0083] Figure 2 A structural schematic diagram of a remote monitoring and data analysis system based on ATP fluorescence detection and Internet of Things provided by the embodiments of the present application;

[0084] Figure 3 A structural schematic diagram of a computing device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0085] In order to make the person skilled in the art better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the drawings in the embodiments of the present application.

[0086] In some processes described in the specification and claims of the present application and the above-mentioned drawings, a plurality of operations appearing in a specific order are included, but it should be clearly understood that these operations can be executed or in parallel without the order appearing in the text. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes can include more or fewer operations, and the operations can be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second" and the like in this paper are used to distinguish different messages, devices, modules, etc., and do not represent the order of sequence, nor do "first" and "second" represent different types.

[0087] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of the present application.

[0088] Figure 1 A flowchart of a remote monitoring and data analysis method based on ATP fluorescence detection and the Internet of Things is provided for the embodiments of the present application, as shown in Figure 1 The method comprises the following steps.

[0089] 101. Collecting raw detection data generated by an ATP fluorescence detection sensor, the raw detection data including fluorescence intensity values and accompanying environmental parameters, and performing real-time denoising processing on the raw detection data to obtain processed detection data that is stable and has a high signal-to-noise ratio.

[0090] In this step, the fluorescence detection sensor can measure the content of adenosine triphosphate (ATP) in the water sample. ATP is the energy source of all living cells, and its presence indicates biological activity. The raw detection data includes fluorescence intensity values and accompanying environmental parameters, which are used to evaluate the activity of microorganisms in the water body.

[0091] First, install ATP fluorescence detection sensors at water quality monitoring points to collect fluorescence signals and environmental parameters of water samples in real time. Second, to improve data quality, perform real-time denoising processing on the collected data, such as using a filter to remove noise, to ensure that the data used for subsequent analysis is stable and has a high signal-to-noise ratio.

[0092] In the present example, assume that in a city water supply system, multiple key nodes are deployed with ATP fluorescence detection sensors. These sensors automatically collect water samples every hour and record fluorescence intensity values as well as environmental parameters such as water temperature and pH value. Through real-time denoising algorithms, noise caused by device vibration or external interference is removed to obtain purer data.

[0093] 102. According to the processed detection data, efficiently transmit it to a cloud data center via Internet of Things technology, together with stored historical detection data, to generate comprehensive detection data after secure transmission;

[0094] In this step, the processed detection data is transmitted to a cloud data center via Internet of Things technology, together with historical detection data, to form a comprehensive detection data set. This step ensures the secure transmission and centralized management of data.

[0095] First, using Internet of Things technologies such as LoRaWAN, NB-IoT, and other low-power wide-area network technologies, the processed detection data is transmitted from the field sensors to the cloud server. Second, encryption technology is used during data transmission to ensure security.

[0096] In the examples of the present application, in the above-mentioned urban water supply system, each ATP fluorescence detection sensor transmits the processed data to the cloud data center through the built-in LoRaWAN module. The cloud server receives and stores these data, and at the same time, combines with the historical data of the past few months to form a complete water quality monitoring database.

[0097] Optionally, in step 102, the processed detection data is transmitted to the cloud data center via Internet of Things technology, and the stored historical detection data is combined to generate comprehensive detection data after secure transmission, including: using the processed detection data to form a structured detection data packet, obtaining a structured detection data packet; based on the structured detection data packet, using a secure encryption protocol in the Internet of Things technology to encrypt the detection data packet, ensuring the security of the data during transmission, preventing data from being stolen or tampered with, generating an encrypted detection data packet; the encrypted detection data packet is transmitted to the cloud data center through the Internet of Things network, data compression technology is used to reduce data transmission volume during transmission, and error control technology is used to ensure data transmission integrity, obtaining detection data after secure transmission; in the cloud data center, the detection data after secure transmission is integrated with the stored historical detection data to ensure consistency and continuity of new and old data, and to generate comprehensive detection data after secure transmission.

[0098] In this step, the structured detection data packet refers to the organization of the processed detection data according to a certain format and structure, which is convenient for subsequent data processing and analysis. The secure encryption protocol is a technical means to protect data from being stolen or tampered with during transmission, ensuring the security of the data. Data compression technology can reduce the amount of data transmission and improve transmission efficiency. Error control technology ensures the integrity of data transmission through checksum and other methods to prevent data loss or damage. These technologies work together to ensure that the transmission process from the sensor to the cloud data center is both efficient and secure.

[0099] Firstly, the processed detection data is structured into packets for efficient transmission and processing. Then, the detection data packets are encrypted using security encryption protocols in the Internet of Things technology to ensure data security during transmission. Next, the encrypted detection data packets are efficiently transmitted to the cloud data center through the Internet of Things network. Data compression techniques are used to reduce data transmission volume, and error control techniques are used to ensure data transmission integrity. Finally, the securely transmitted detection data is integrated with the stored historical detection data in the cloud data center to ensure consistency and continuity of the data, generating comprehensive detection data after secure transmission.

[0100] In the water quality monitoring project of a city's water supply system, multiple key nodes are equipped with ATP fluorescence detection sensors. These sensors automatically collect fluorescence intensity values and environmental parameters of water samples every hour and perform real-time denoising processing. The processed data is organized into structured data packets, each containing a timestamp, sensor ID, fluorescence intensity value, and environmental parameters. Next, the data packets are encrypted using the Advanced Encryption Standard encryption algorithm to ensure data security during transmission. The encrypted data packets are efficiently transmitted to the cloud data center through the LoRaWAN network. During transmission, GZIP compression technology is used to reduce data transmission volume, and cyclic redundancy check technology is used to ensure data transmission integrity. Upon arrival at the cloud data center, the system integrates the newly received data with historical data, updating the records in the database to ensure data consistency and continuity. This not only improves the security and efficiency of data transmission, but also provides a reliable foundation for subsequent data analysis and prediction.

[0101] 103、Utilizing time series analysis algorithms in combination with environmental parameters in the comprehensive detection data after secure transmission, the detection data is analyzed comprehensively to identify patterns of biological contamination level changes. When the detection data exceeds the preset threshold dynamically adjusted based on historical detection data, a multi-level alarm response system is activated. The multi-level alarm response system can select appropriate communication channels to send customized alarm information to designated management personnel according to different levels of alarm types, generating alarm notification records.

[0102] In this step, time series analysis algorithms are used to identify trends and patterns in data over time, and environmental parameters can be combined to more accurately assess changes in biological contamination levels. When the detection data exceeds the dynamically adjusted preset threshold, a multi-level alarm response system is triggered.

[0103] First, time series analysis method is used to analyze the trend of fluorescence intensity value over time. Combined with environmental parameters such as temperature, pH value, etc., the early warning threshold is dynamically adjusted. Once the detection data exceeds this threshold, the multi-level alarm response system is activated immediately, sending alarm information to the management personnel.

[0104] In the examples of the present application, in the water quality monitoring system, first, through time series analysis, it is found that the fluorescence intensity value in a certain area has increased significantly in the past week, combined with environmental parameter analysis, it is determined that this is due to the increase of microbial activity. Second, when the fluorescence intensity value exceeds the threshold dynamically adjusted according to historical data, the system automatically triggers an alarm, notifies the relevant management personnel through SMS and email.

[0105] Optionally, in step 103, the time series analysis algorithm is used to analyze the detection data in combination with the environmental parameters in the comprehensive detection data after security transmission, to identify the change pattern of biological pollution level. When the detection data exceeds the preset threshold dynamically adjusted based on historical detection data, the multi-level alarm response system is activated, which can select the appropriate communication channel to send customized alarm information to the designated management personnel according to different levels of alarm type, and generate alarm notification records, including: using time series analysis algorithm to analyze the fluorescence intensity value and its accompanying environmental parameters in the comprehensive detection data after security transmission, extracting key indicators reflecting biological pollution level and their change trend over time, obtaining time series analysis results of biological pollution level; based on the time series analysis results of biological pollution level, comparing with the stored historical detection data, dynamically adjusting the preset warning threshold using statistical method, obtaining dynamic warning threshold adapting to the change of biological pollution level under different environmental conditions; according to the dynamic warning threshold, monitoring the time series analysis results of the biological pollution level, when the detected fluorescence intensity value exceeds the dynamic warning threshold, starting the multi-level alarm response mechanism immediately, obtaining the alarm trigger signal; based on the alarm trigger signal, automatically determining the alarm level according to the severity of the alarm, selecting the most appropriate communication channel to send customized alarm information containing alarm level, specific pollution location, cause analysis and preliminary suggested countermeasures to the preset management personnel or emergency team, generating alarm notification records.

[0106] Optionally, in step 103, the time series analysis algorithm is used to analyze the fluorescence intensity values and their accompanying environmental parameters in the comprehensive detection data after secure transmission, extract key indicators reflecting the biological pollution level and their trends over time, and obtain the time series analysis results of the biological pollution level, including: using time series decomposition techniques to decompose and process the fluorescence intensity values and their accompanying environmental parameters in the comprehensive detection data after secure transmission, separate the long-term trend, seasonal fluctuations and random fluctuations into three components, and obtain the preliminary decomposition results; based on the preliminary decomposition results, applying statistical regression analysis methods to identify the main environmental factors affecting the change of fluorescence intensity values, establish a mathematical model between fluorescence intensity values and environmental parameters, and obtain the environmental factor influence model; using the environmental factor influence model, combining the long-term trend and seasonal fluctuations in the preliminary decomposition results, predicting the biological pollution level, extracting key indicators reflecting the biological pollution level, including average fluorescence intensity and fluorescence intensity change rate, generating a key indicator set; according to the key indicator set, analyzing the trend of biological pollution level over time, identifying the rising, falling or stable state of biological pollution level, and generating the time series analysis results of biological pollution level.

[0107] In this step, the time series analysis algorithm is a statistical method used to analyze and predict data that changes over time. In water quality detection, time series analysis can be used to identify the trend of fluorescence intensity values and their accompanying environmental parameters over time. Key indicators include average fluorescence intensity and fluorescence intensity change rate, which reflect the changes in biological pollution level. Time series decomposition techniques decompose data into long-term trends, seasonal fluctuations and random fluctuations, which help to better understand the changing patterns of data. Statistical regression analysis methods are used to establish a mathematical model between fluorescence intensity values and environmental parameters, further revealing the influence of environmental factors on fluorescence intensity.

[0108] First, the time series decomposition technique is used to decompose the fluorescence intensity values and their accompanying environmental parameters in the comprehensive detection data after secure transmission, separating the long-term trend, seasonal fluctuations, and random fluctuations into three components to obtain the preliminary decomposition results. Then, statistical regression analysis is applied to identify the main environmental factors affecting the fluorescence intensity value changes and establish a mathematical model between the fluorescence intensity value and the environmental parameters. Next, the model is used in combination with the long-term trend and seasonal fluctuations in the preliminary decomposition results to predict the biological contamination level and extract key indicators reflecting the biological contamination level to generate a key indicator set. According to the key indicator set, the trend of the biological contamination level over time is analyzed to identify the rising, falling, or stable state of the biological contamination level, and a time series analysis result of the biological contamination level is generated. Based on these analysis results, the historical detection data is compared, and the statistical method is used to dynamically adjust the preset warning threshold to obtain a dynamic warning threshold that adapts to the changes in the biological contamination level under different environmental conditions. When the detected fluorescence intensity value exceeds the dynamic warning threshold, a multi-level alarm response mechanism is immediately started, and an alarm trigger signal is generated. Finally, the alarm level is automatically determined according to the severity of the alarm, the most suitable communication channel is selected to send customized alarm information containing the alarm level, specific contamination location, cause analysis, and preliminary recommended response measures to the preset management personnel or emergency team, and an alarm notification record is generated.

[0109] In the water quality monitoring project of a city's water supply system in the embodiments of the present application, ATP fluorescence detection sensors are deployed at multiple key nodes, and fluorescence intensity values and environmental parameters of water samples are automatically collected every hour. These data are processed and transmitted to a cloud data center. In the cloud, time series decomposition techniques are used to decompose the fluorescence intensity values and their accompanying environmental parameters, separating the long-term trend, seasonal fluctuations, and random fluctuations into three components. For example, through decomposition, it is found that the fluorescence intensity value has obvious seasonal fluctuations in summer due to the increase in temperature.

[0110] Next, statistical regression analysis is applied to identify temperature as the main environmental factor affecting the change in fluorescence intensity value and establish a mathematical model between the fluorescence intensity value and temperature. Using the model in combination with the long-term trend and seasonal fluctuations, future fluorescence intensity values are predicted, and key indicators such as average fluorescence intensity and fluorescence intensity change rate are extracted. According to these key indicators, the trend of the biological contamination level over time is analyzed, and it is identified that the biological contamination level in some areas has a rising trend.

[0111] Based on these analysis results, the preset alert threshold is dynamically adjusted compared with historical data. For example, during the summer high temperature period, the alert threshold is appropriately increased to adapt to the higher natural background fluorescence level. When the fluorescence intensity value of a certain area exceeds the dynamic alert threshold, the multi-level alarm response mechanism is immediately started. The system automatically determines the alarm level according to the severity of the alarm and selects the most appropriate communication channel to send alarm information to the preset management personnel or emergency team. The alarm information includes the alarm level, the specific pollution location, the cause analysis, and the preliminary recommended response measures. In this way, the management personnel can take prompt action to prevent further deterioration of the biological pollution event.

[0112] The present application considers that in water quality detection, the fluorescence intensity value F(t) is a key indicator for evaluating the biological pollution level in the water body. The fluorescence intensity value is affected by various environmental factors such as temperature, pH value, etc. Through time series analysis, the fluorescence intensity value can be decomposed into long-term trend T(t), seasonal fluctuation S(t) and random fluctuation R(t), so as to better understand the change pattern of the data. Further, by establishing a nonlinear mathematical model between the fluorescence intensity value and the environmental parameters, the future fluorescence intensity value can be predicted, and the key indicators reflecting the biological pollution level and their change trend over time can be extracted.

[0113] Optionally, in step 103, the fluorescence intensity value and its accompanying environmental parameters in the comprehensive detection data after secure transmission are analyzed using a time series analysis algorithm to extract key indicators reflecting the biological pollution level and their change trend over time, obtaining time series analysis results of the biological pollution level, including:

[0114] Using time series decomposition technology, the fluorescence intensity value F(t) and its accompanying environmental parameters in the comprehensive detection data after secure transmission are decomposed and processed to separate the long-term trend T(t), seasonal fluctuation S(t) and random fluctuation R(t) into three components, obtaining a preliminary decomposition result, i.e. F(t) = T(t) + S(t) + R(t), where t represents time, F(t) represents the fluorescence intensity value, F(t) represents the fluorescence intensity at time t; T(t) represents the long-term trend, which represents the long-term change trend of the fluorescence intensity value; S(t) represents the seasonal fluctuation, which represents the periodic change of the fluorescence intensity value; R(t) represents the random fluctuation, which represents the random change part of the fluorescence intensity value;

[0115] Based on the preliminary decomposition result, a statistical regression analysis method is applied to identify the main environmental factors E i (t) affecting the change of the fluorescence intensity value, and a nonlinear mathematical model between the fluorescence intensity value F(t) and the environmental parameters E i (t) is established. The following is the environmental factor influence model:

[0116]

[0117] wherein; E i (t) represents environmental parameters, represents the i-th environmental parameter at time t; β0 is an intercept term, β i , γ i and δ i are the regression coefficients of the i-th environmental parameter for the linear, quadratic and cubic terms, respectively, α1 and α2 are the regression coefficients of long-term trend and seasonal fluctuations, respectively, n is the number of environmental parameters, ∈ is an error term, and an environmental factor influence model is obtained;

[0118] Using the environmental factor influence model, combined with the long-term trend T(t) and seasonal fluctuations S(t) in the preliminary decomposition result, the biological contamination level is predicted, and the key indicators reflecting the biological contamination level are extracted:

[0119] The average fluorescence intensity is calculated by the following formula:

[0120]

[0121] The fluorescence intensity change rate ΔF(t) is calculated by the following formula:

[0122]

[0123] Where Δt represents the time interval, and represents the time interval for calculating the fluorescence intensity change rate;

[0124] The weighted moving average fluorescence intensity WMA(F(t)) is:

[0125]

[0126] Where w i is the weight, k is the length of the time window; i represents the index within the time window;

[0127] The exponential smoothing fluorescence intensity ES(F(t)) is calculated by the following formula:

[0128] ES(F(t)) = αF(t) + (1-α)ES(F(t-1))

[0129] Where α is the smoothing factor, usually taking a value between 0 and 1; a set of key indicators is generated;

[0130] According to the set of key indicators, the change trend of the biological contamination level over time is analyzed, and the rising, falling or stable state of the biological contamination level is identified, and a time series analysis result of the biological contamination level is generated.

[0131] Decomposing the fluorescence intensity value into long-term trends, seasonal fluctuations, and random fluctuations helps to separate the change patterns at different time scales. By identifying the main environmental factors that affect the fluorescence intensity value and establishing a nonlinear mathematical model, the variation law of the fluorescence intensity value can be described more accurately. By calculating the average fluorescence intensity, the fluorescence intensity change rate, the weighted moving average fluorescence intensity, and the exponential smoothing fluorescence intensity, key indicators reflecting the biological pollution level can be extracted for trend analysis.

[0132] The design reasons for each term of the formula are briefly introduced as follows:

[0133] In the fluorescence intensity value formula F(t), β i E i (t) represents the linear effect of the environmental parameter E i (t) on the fluorescence intensity value F(t), capturing the basic effect of environmental factors; γ i E i (t) 2 This term represents the quadratic nonlinear effect of the environmental parameter E i (t) on the fluorescence intensity value F(t), capturing the nonlinear effect of environmental factors; δ i E i (t) 3 This term represents the cubic nonlinear effect of the environmental parameter E i (t) on the fluorescence intensity value F(t), further capturing more complex nonlinear effects; α1T(t) This term represents the effect of long-term trend T(t) on the fluorescence intensity value F(t), capturing the long-term change trend of the data; α2S(t) This term represents the effect of seasonal fluctuations S(t) on the fluorescence intensity value F(t), capturing the periodic changes of the data.

[0134] The design reasons for each term of the formula are briefly introduced as follows:

[0135] In the fluorescence intensity change rate formula ΔF(t), F(t-Δt) This term represents the fluorescence intensity value at the previous time point (t-Δt), used to calculate the fluorescence intensity change rate.

[0136] The design reasons for each term of the formula are briefly introduced as follows:

[0137] In the weighted moving average fluorescence intensity formula WMA(F(t)), F(t-Δt) This term represents the weighted fluorescence intensity value within the time window, used to calculate the weighted moving average fluorescence intensity, smoothing short-term fluctuations.

[0138] The design reasons for each term of the formula are briefly introduced as follows:

[0139] ​In the exponential smoothing fluorescence intensity formula ES(F(t), the term αF(t) represents the weight part of the fluorescence intensity value F(t) at the current time point t for exponential smoothing, reflecting the influence of the latest data; the term (1-α)ES(F(t-1)) represents the weight part of the exponential smoothing fluorescence intensity value at the previous time point t-1 for exponential smoothing, reflecting the influence of historical data.

[0140] The following briefly introduces the acquisition method of each parameter of the formula:

[0141] Among them, the fluorescence intensity value F(t) is collected in real time by an ATP fluorescence detection sensor; the environmental parameter E i (t) is collected in real time by a sensor or other monitoring equipment, such as temperature, pH value, etc.; the long-term trend T(t), the seasonal fluctuation S(t), and the random fluctuation R(t): are extracted from the fluorescence intensity value by time series decomposition technology; the regression coefficients β0, β i , γ i , δ i , α1, α2: are fitted by statistical regression analysis method (such as multiple regression); the weight w i is determined according to experience or optimization algorithm; the smoothing factor α usually takes a value between 0 and 1, and the specific value is adjusted according to the actual situation.

[0142] Suppose in a water quality monitoring project of a city water supply system, there are the following data:

[0143] The fluorescence intensity value F(t) is the data collected every hour;

[0144] The environmental parameter E1(t) is temperature, and E2(t) is pH value;

[0145] The time interval Δt = 1 hour;

[0146] The weight w i = 1 (simple moving average);

[0147] The smoothing factor α = 0.5;

[0148] The following is a data example:

[0149] Time t Fluorescence intensity F(t) temperature E1(t) [pH value E2(t)] 1 10 20 7.2 2 12 22 7.1 3 14 24 7.0 4 16 25 6.9 5 18 26 6.8

[0150] The following is the time series decomposition:

[0151] Suppose it is obtained by STL decomposition:

[0152] The long-term trend T(t) = [10, 10, 10, 10, 10]

[0153] The seasonal fluctuation S(t) = [0, 1, 2, 3, 4]

[0154] Random fluctuations R(t) = [0, 1, 2, 3, 4]

[0155] Here is the model of environmental factors:

[0156] Assume that through regression analysis we get:

[0157] β0 = 8;

[0158] β1 = 0.5, γ1 = 0.01, δ1 = 0.001;

[0159] β2 = 0.3, γ2 = 0.005, δ2 = 0.0005;

[0160] α1 = 0.1, α2 = 0.2;

[0161] Here are the key indicators calculated:

[0162] Average fluorescence intensity

[0163]

[0164] Fluorescence intensity change rate ΔF(t):

[0165]

[0166] Here is the weighted moving average fluorescence intensity WMA(F(t)):

[0167]

[0168] Here is the exponential smoothing fluorescence intensity ES(F(t)):

[0169] ES(F(1)) = 10

[0170] ES(F(2)) = 0.5 * 12 + (1 - 0.5) * 10 = 11

[0171] ES(F(3)) = 0.5 * 14 + (1 - 0.5) * 11 = 12.5

[0172] ES(F(4)) = 0.5 * 16 + (1 - 0.5) * 12.5 = 14.25

[0173] ES(F(5)) = 0.5 * 18 + (1 - 0.5) * 14.25 = 16.125

[0174] Average fluorescence intensity This indicates that the average level of fluorescence intensity during this period is 14.

[0175] The rate of change of fluorescence intensity ΔF(t) = 2 indicates that the fluorescence intensity increases by 2 units per hour, indicating that the level of biological contamination is gradually increasing.

[0176] The weighted moving average fluorescence intensity WMA(F(t)) and the exponential smoothing fluorescence intensity ES(F(t)) both show a gradual increase in fluorescence intensity over time, further confirming the rise in the level of biological contamination.

[0177] Through the above calculations, it can be concluded that the level of biological contamination in this area is showing an upward trend during this period, which needs to be closely monitored and appropriate measures taken.

[0178] 104、Based on the environmental parameters in the alarm notification record, historical detection data and the comprehensive detection data after the secure transmission, combined with real-time detection data, a machine learning algorithm is used to construct a biological contamination prediction model, and a biological contamination prediction report is generated;

[0179] In this step, based on existing data and real-time data, a machine learning algorithm is used to construct a biological contamination prediction model, and a biological contamination prediction report is generated. This model can help predict future biological contamination events.

[0180] First, collect alarm notification records, historical detection data, environmental parameters and real-time detection data, and perform preprocessing and feature selection. Second, select a suitable machine learning algorithm and train the model. Use cross-validation method to optimize model performance, then use the optimized model to predict future data.

[0181] In the examples of the present application, it is assumed that in the water quality monitoring system, the alarm notification records, historical detection data and environmental parameters of the past year are collected. Through preprocessing and feature selection of these data, a vector machine algorithm is used to construct a biological contamination prediction model. After cross-validation and optimization of the model, it is used to predict possible biological contamination events in the next month and generate a detailed prediction report.

[0182] Optionally, in step 104, the environmental parameters in the alarm notification record, the historical detection data, and the comprehensive detection data after the secure transmission are combined with real-time detection data to construct a biological pollution prediction model using a machine learning algorithm, and a biological pollution prediction report is generated, including: collecting and preprocessing the environmental parameters in the alarm notification record, the historical detection data, the comprehensive detection data after the secure transmission, and the real-time detection data to obtain a high-quality data set; according to the high-quality data set, feature selection and conversion are performed to determine key features affecting biological pollution prediction, and some features are converted or encoded as necessary to obtain feature data after feature selection and conversion; using the feature data after feature selection and conversion, support vector machines are selected as machine learning algorithms to construct a biological pollution prediction model, and appropriate kernel functions are selected and related parameters are adjusted to optimize model performance; based on the constructed biological pollution prediction model, the model is trained and verified using a cross-validation method to ensure that the model has good generalization ability and prediction accuracy, and an optimized biological pollution prediction model is obtained; using the optimized biological pollution prediction model, new or future data is predicted to generate a detailed biological pollution prediction report, and the biological pollution prediction report content includes the predicted time range, the prediction result, the potential high-risk area, and the corresponding recommended measures.

[0183] Optionally, in step 104, the environmental parameters in the alarm notification record, the historical detection data, and the comprehensive detection data after the secure transmission are combined with real-time detection data to construct a biological pollution prediction model using a machine learning algorithm, and a biological pollution prediction report is generated, including: collecting and preprocessing the environmental parameters in the alarm notification record, the historical detection data, the comprehensive detection data after the secure transmission, and the real-time detection data to obtain a high-quality data set; according to the high-quality data set, feature selection and conversion are performed to determine key features affecting biological pollution prediction, and some features are converted or encoded as necessary to obtain feature data after feature selection and conversion; using the feature data after feature selection and conversion, support vector machines are selected as machine learning algorithms to construct a biological pollution prediction model, and appropriate kernel functions are selected and related parameters are adjusted to optimize model performance; based on the constructed biological pollution prediction model, the model is trained and verified using a cross-validation method to ensure that the model has good generalization ability and prediction accuracy, and an optimized biological pollution prediction model is obtained; using the optimized biological pollution prediction model, new or future data is predicted to generate a detailed biological pollution prediction report, and the biological pollution prediction report content includes the predicted time range, the prediction result, the potential high-risk area, and the corresponding recommended measures.

[0184] In this step, the alarm notification records contain past biological contamination events and their handling, the historical detection data include past fluorescence intensity values and environmental parameters, and the integrated detection data after secure transmission are the current data after processing and integration. These data are used to build a biological contamination prediction model to predict future possible biological contamination events. Feature selection and transformation refer to extracting the most helpful features for the prediction task from the original data and performing necessary transformations or encodings to improve model performance. Support vector machine is a commonly used machine learning algorithm that can effectively handle nonlinear problems and optimize model performance by selecting appropriate kernel functions.

[0185] First, collect the alarm notification records, historical detection data, environmental parameters in the integrated detection data after secure transmission, and real-time detection data, and preprocess them to obtain a high-quality dataset. Then, based on the high-quality dataset, perform feature selection and transformation to determine the key features affecting biological contamination prediction, and perform necessary transformations or encodings on some features to obtain feature data after feature selection and transformation. Next, use the feature data after feature selection and transformation, select support vector machine as the machine learning algorithm, and build a preliminary vector machine model. According to the requirements of the biological contamination prediction task, select an appropriate kernel function to optimize the nonlinear mapping ability and generalization ability of the model. Then, use the cross-validation method to train and validate the model to ensure that the model has good generalization ability and prediction accuracy, and obtain the optimized biological contamination prediction model. Finally, use the optimized biological contamination prediction model to predict new or future data and generate a detailed biological contamination prediction report, including the predicted time range, prediction results, potential high-risk areas, and corresponding recommended measures.

[0186] In the water quality monitoring project of a city's water supply system in the embodiments of the present application, ATP fluorescence detection sensors are deployed at multiple key nodes, and fluorescence intensity values and environmental parameters of water samples are automatically collected every hour. These data are processed and transmitted to the cloud data center. In the cloud, the system collects alarm notification records, historical detection data, environmental parameters in the integrated detection data after secure transmission, and real-time detection data for the past year, and preprocesses them to remove noise and outliers to obtain a high-quality dataset.

[0187] Next, perform feature selection and transformation to determine the key features affecting biological contamination prediction, such as fluorescence intensity values, temperature, pH, etc. Standardize some continuous features and one-hot encode categorical features to obtain feature data after feature selection and transformation.

[0188] With these feature data, the support vector machine is selected as the machine learning algorithm to build a preliminary vector machine model. According to the requirements of the biological pollution prediction task, the radial basis function kernel is selected because the radial basis function kernel performs well in handling nonlinear problems and can provide strong nonlinear mapping ability and generalization ability. Through the grid search method, the best parameter combination is found in the pre-defined parameter range to optimize the model performance.

[0189] The model is trained and verified using the 5-fold cross-validation method to ensure that the model has good generalization ability and prediction accuracy on unseen data. The optimized biological pollution prediction model is finally obtained.

[0190] Using the optimized biological pollution prediction model, new or future data is predicted to generate detailed biological pollution prediction reports. The report content includes the predicted time range, prediction results, potential high-risk areas, and corresponding recommended measures. In this way, management personnel can take measures in advance according to the prediction report to effectively prevent and control the occurrence of biological pollution events.

[0191] The present application considers that the support vector machine is a powerful supervised learning algorithm, especially suitable for classification and regression tasks. In biological pollution prediction, the support vector machine distinguishes different categories of data points by constructing a decision boundary. The kernel function is a key component of the support vector machine, which allows the model to find a nonlinear decision boundary in high-dimensional space, thereby improving the generalization ability of the model.

[0192] Optionally, in step 104, the feature data after feature selection and conversion is used to select a support vector machine as a machine learning algorithm to build a biological pollution prediction model, and a suitable kernel function is selected to optimize the model performance, including:

[0193] Using the feature data after feature selection and conversion, a support vector machine is selected as a machine learning algorithm to model the biological pollution prediction task to obtain a preliminary vector machine model;

[0194] The basic form of the vector machine model f(x) is represented as:

[0195]

[0196] Where N represents the number of support vectors, α i is the Lagrange multiplier, y i is the label of each support vector (+1 or -1), K(x, x i ) is the kernel function, and b is the bias term.

[0197] According to the requirements of the biological pollution prediction task, a suitable kernel function is selected for the vector machine model, the selection of the kernel function is based on the nonlinear mapping ability and generalization ability of the model, and a vector machine model with strong nonlinear mapping ability is obtained; Commonly used kernel functions include linear kernel, polynomial kernel and radial basis function kernel; The mathematical expression of the radial basis function kernel K(x, x i ) is:

[0198] K(x, x i ) = exp(-γ∥x-x i ∥ 2 )

[0199] Where γ is the kernel function parameter, and ∥x-x i ∥ represents the Euclidean distance between two vectors.

[0200] Support vector machines maximize the separation between different classes by selecting the optimal hyperplane, thereby achieving good classification performance. Kernel functions map the original feature space to a high-dimensional space, making it easier to find a linearly separable decision boundary in the high-dimensional space. Commonly used kernel functions include linear kernel, polynomial kernel and radial basis function kernel.

[0201] The following briefly introduces the acquisition method of each parameter of the formula:

[0202] Where the number of support vectors N is automatically determined through the training process; The Lagrange multiplier α i is obtained by solving the optimization problem; The label y i is directly obtained from the training data set; The kernel function K(x, x i ) selects the appropriate kernel function type according to the requirements, and optimizes the parameters through cross-validation; The bias term b is obtained by solving the optimization problem; The kernel function parameter γ is optimized and selected through cross-validation and other methods.

[0203] Suppose in a water quality monitoring project of a city water supply system, there are the following data:

[0204] The feature data X includes fluorescence intensity, temperature, pH value, etc.

[0205] The label data Y indicates whether biological pollution occurs (+1 indicates pollution, -1 indicates no pollution).

[0206] The following is a data example:

[0207] Fluorescence intensity Temperature pH value Label 10 20 7.2 -1 12 22 7.1 -1 14 24 7.0 +1 16 25 6.9 +1 18 26 6.8 +1

[0208] Feature selection and conversion:

[0209] Suppose after feature selection and conversion, the fluorescence intensity, temperature and pH value are retained as features.

[0210] Modeling is performed using support vector machines, with radial basis function kernels selected.

[0211] Training the support vector machine model:

[0212] Suppose the following parameters are obtained through training:

[0213] Number of support vectors N=3

[0214] The Lagrange multipliers are α1 = 0.5, α2 = 0.2, and α3 = 0.3.

[0215] Labels y1 = -1, y2 = -1, y3 = +1

[0216] Kernel function parameter γ = 0.1

[0217] Bias term b = -1.5

[0218] Support Vector Machine Model:

[0219] f(x)=0.5·(-1)·exp(-0.1·∥x-x1∥ 2 )+0.2·(-1)·exp(-0.1·∥x-x2∥ 2 )

[0220] +0.3·(+1)·exp(-0.1·∥x-x3∥ 2 -1.5

[0221] Predicting new data points:

[0222] Suppose we have a new data point x = [15, 23, 7.1], calculate its predicted value:

[0223] Support vectors x1 = [10, 20, 7.2];

[0224] Support vectors x2 = [12, 22, 7.1];

[0225] Support vectors x3 = [14, 24, 7.0];

[0226] Calculate the Euclidean distance:

[0227]

[0228] Calculate the kernel function value:

[0229] K(x, x1) = exp(-0.1·5.83) 2 )=exp(-3.3989)≈0.033

[0230] K(x, x2) = exp(-0.1·3.16) 2)=exp(-0.9986)≈0.367

[0231] K(x, x3) = exp(-0.1·1.42) 2 )=exp(-0.2016)≈0.817

[0232] Calculate the predicted value:

[0233] f(x)=0.5·(-1)·0.033+0.2·(-1)·0.367+0.3·(+1)·0.817-1.5

[0234] f(x)=-0.0165-0.0734+0.2451-1.5

[0235] f(x)≈-1.3448

[0236] The predicted value f(x) ≈ -1.3448. Since the predicted value is negative, according to the classification rules of support vector machine, this data point is predicted as no biological pollution has occurred (label is -1).

[0237] Based on the above calculations, it can be concluded that, given the characteristic data, the new data point x = [15, 23, 7.1] is predicted to be free of biological pollution. This indicates that the current water quality conditions have not yet reached the threshold for biological pollution, but continuous monitoring is still necessary to ensure water quality safety.

[0238] 105. Using the biological pollution prediction report and the problem area information reflected in the alarm notification record, adjust the ATP fluorescence detection frequency setting of each monitoring point, increase the detection frequency for monitoring points that are predicted to be high-risk areas, and decrease the detection frequency for low-risk areas, thereby generating an optimized monitoring strategy to improve the response speed and resource utilization efficiency of the monitoring system.

[0239] In this step, the detection frequency at each monitoring point is adjusted based on the biocontamination prediction report and alarm notification records. The detection frequency is increased in high-risk areas and decreased in low-risk areas, thereby optimizing the monitoring strategy and improving the system's response speed and resource utilization efficiency.

[0240] First, based on biocontamination prediction reports, high-risk areas are identified. Second, for these areas, the frequency of ATP fluorescence detection is increased; for low-risk areas, the detection frequency is appropriately reduced. This allows for a more rational allocation of detection resources and improves the overall efficiency of the system.

[0241] In the example of the present application, in the water quality monitoring system, according to the prediction report, it is found that some areas have a higher biological pollution risk in the future. Therefore, the ATP fluorescence detection frequency of these areas is increased from once an hour to once every half hour. While for other low-risk areas, the detection frequency is reduced from once an hour to once every two hours. This adjustment not only improves the response speed to potential pollution events, but also optimizes the resource allocation of the entire system.

[0242] Optionally, in step 105, the ATP fluorescence detection frequency of each monitoring point is adjusted based on the problem area information reflected in the alarm notification record, and the detection frequency of the monitoring points in the high-risk areas is increased, and the detection frequency of the monitoring points in the low-risk areas is reduced, to generate an optimized monitoring strategy, so as to improve the response speed and resource utilization efficiency of the monitoring system, including: using the prediction results, prediction confidence and potential impact analysis in the biological pollution prediction report to evaluate the future biological pollution risk of each monitoring point to obtain a risk area evaluation result; based on the risk area evaluation result, combining the problem area information reflected in the alarm notification record, comprehensively analyzing the actual pollution risk level of each monitoring point, and generating a risk evaluation report of each monitoring point; according to the risk evaluation report of each monitoring point, for the monitoring points in the high-risk areas, the ATP fluorescence detection frequency is adjusted to increase the detection frequency, so as to ensure that the biological pollution event can be discovered and responded in time, and a monitoring strategy for the high-risk areas is generated; for the monitoring points in the low-risk areas, based on the risk evaluation report of each monitoring point, the detection frequency is appropriately reduced to reduce resource waste and improve the operation efficiency of the monitoring system, and a monitoring strategy for the low-risk areas is generated; the monitoring strategy for the high-risk areas and the monitoring strategy for the low-risk areas are integrated to form a comprehensive and optimized monitoring strategy, so as to improve the response speed and resource utilization efficiency of the monitoring system and ensure effective monitoring and management of biological pollution.

[0243] In this step, the biological pollution prediction report includes prediction results, prediction confidence and potential impact analysis, which are used to evaluate the biological pollution risk of each monitoring point in the future. The alarm notification record contains the biological pollution events and their handling in the past, reflecting the information of the actual problem areas. The risk evaluation report combines the prediction results and historical data to determine the actual pollution risk level of each monitoring point. By adjusting the ATP fluorescence detection frequency, the monitoring density of the high-risk areas can be improved to ensure timely discovery and response to biological pollution events, while the detection frequency in the low-risk areas is reduced to save resources and improve the overall efficiency of the monitoring system.

[0244] Firstly, the future biological contamination risk of each monitoring point is evaluated by using the prediction results, prediction confidence and potential impact analysis in the biological contamination prediction report, to obtain the risk area evaluation results. Then, based on these risk area evaluation results, combined with the problem area information reflected in the alarm notification record, the actual contamination risk level of each monitoring point is comprehensively analyzed to generate a risk assessment report for each monitoring point. According to the risk assessment report, for the monitoring points predicted to be high-risk areas, the detection frequency of the ATP fluorescence detection is adjusted to increase the detection frequency, so as to ensure that the biological contamination events can be discovered and responded in time, and the monitoring strategy for the high-risk area is generated. For the monitoring points predicted to be low-risk areas, the detection frequency is appropriately reduced to reduce resource waste and improve the operation efficiency of the monitoring system, and the monitoring strategy for the low-risk area is generated. Finally, the monitoring strategies for the high-risk area and the low-risk area are integrated to form a comprehensive and optimized monitoring strategy, so as to improve the response speed and resource utilization efficiency of the monitoring system, and ensure effective monitoring and management of biological contamination.

[0245] In the embodiment of the present application, in a water quality monitoring project of a city water supply system, ATP fluorescence detection sensors are deployed at multiple key nodes, and the fluorescence intensity value and environmental parameters of water samples are automatically collected once an hour. The system generates a biological contamination prediction report, which includes the prediction results, prediction confidence and potential impact analysis of each monitoring point in the next week. For example, the prediction report shows that several areas have a higher risk of biological contamination in the next week, while other areas have a lower risk.

[0246] Based on these prediction results, combined with the problem area information reflected in the alarm notification record in the past year, the system comprehensively analyzes the actual contamination risk level of each monitoring point to generate a risk assessment report for each monitoring point. For example, a certain area has had biological contamination events multiple times in the past year, and the prediction result shows that the area still has a high risk in the next week, so it is marked as a high-risk area.

[0247] Figure 2 A structural schematic diagram of a remote monitoring and data analysis system based on ATP fluorescence detection and Internet of Things is provided for the embodiment of the present application, as shown in Figure 2 The system comprises:

[0248] A collection and processing module 21 is configured to collect original detection data generated by the ATP fluorescence detection sensor, wherein the original detection data includes fluorescence intensity values and accompanying environmental parameters, and to perform real-time denoising processing on the original detection data to obtain processed detection data with stable and high signal-to-noise ratio.

[0249] The transmission module 22 is configured to transmit the processed detection data to a cloud data center via an Internet of Things technology, and generate comprehensive detection data after secure transmission together with stored historical detection data.

[0250] The analysis activation module 23 is configured to analyze the detection data by using a time series analysis algorithm in combination with the environmental parameters in the comprehensive detection data after secure transmission, identify a change pattern of the biological contamination level, and activate a multi-level alarm response system when the detection data exceeds a preset threshold dynamically adjusted based on the historical detection data.

[0251] The construction module 24 is configured to construct a biological contamination prediction model by using a machine learning algorithm based on the alarm notification record, historical detection data, environmental parameters in the comprehensive detection data after secure transmission, and real-time detection data, and generate a biological contamination prediction report.

[0252] The adjustment generation module 25 is configured to adjust the ATP fluorescence detection frequency settings of each monitoring point by using the biological contamination prediction report in combination with the problem area information reflected in the alarm notification record, increase the detection frequency of monitoring points in a high-risk area, and reduce the detection frequency of monitoring points in a low-risk area, and generate an optimized monitoring strategy to improve the response speed and resource utilization efficiency of the monitoring system.

[0253] Figure 2 The remote monitoring and data analysis system based on ATP fluorescence detection and the Internet of Things can perform Figure 1 The remote monitoring and data analysis method based on ATP fluorescence detection and the Internet of Things of the embodiments described above has the same implementation principles and technical effects. The specific operation modes of each module and unit of the remote monitoring and data analysis system based on ATP fluorescence detection and the Internet of Things in the above embodiments have been described in detail in the embodiments related to the method, and will not be described in detail here.

[0254] In one possible design, Figure 2 The remote monitoring and data analysis system based on ATP fluorescence detection and the Internet of Things of the embodiments described above can be implemented as a computing device, such as Figure 3 As shown, the computing device can include a storage component 31 and a processing component 32.

[0255] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32.

[0256] The processing component 32 is configured to collect raw detection data generated by the ATP fluorescence detection sensor, the raw detection data including fluorescence intensity values and their accompanying environmental parameters, and to perform real-time denoising processing on the raw detection data to obtain stable and high signal-to-noise ratio processed detection data; according to the processed detection data, the processed detection data is efficiently transmitted to a cloud data center via Internet of Things technology, and the processed detection data is generated together with the stored historical detection data to generate comprehensive detection data after safe transmission; using a time series analysis algorithm in combination with the environmental parameters in the comprehensive detection data after safe transmission, the detection data is comprehensively analyzed to identify the change pattern of the biological contamination level, and when the detection data exceeds the preset threshold dynamically adjusted based on the historical detection data, a multi-level alarm response system is activated, the multi-level alarm response system can select a suitable communication channel to send customized alarm information to designated management personnel according to different levels of alarm types, and generate alarm notification records; based on the alarm notification records, the historical detection data and the environmental parameters in the comprehensive detection data after safe transmission, in combination with real-time detection data, a biological contamination prediction model is constructed using a machine learning algorithm to generate a biological contamination prediction report; using the biological contamination prediction report, in combination with the problem area information reflected in the alarm notification records, the ATP fluorescence detection frequency settings of each monitoring point are adjusted, the detection frequency of the monitoring point predicted to be a high-risk area is increased, and the detection frequency of a low-risk area is reduced, and an optimized monitoring strategy is generated to improve the response speed and resource utilization efficiency of the monitoring system.

[0257] The processing component 32 can include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component can also be an application specific integrated circuit (ASIC), a digital data processor (DSP), a digital data processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a controller, a microcontroller, a microprocessor or other electronic elements for executing the above method.

[0258] The storage component 31 is configured to store various types of data to support the operation of the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0259] Of course, the computing device can also include other components, such as an input / output interface, a display component, a communication component, etc.

[0260] The input / output interface provides an interface between the processing component and peripheral interface modules, which can be output devices, input devices, and the like.

[0261] The communication component is configured to facilitate wired or wireless communication between the computing device and other devices, and the like.

[0262] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform, and the processing component, the storage component, and the like can be basic server resources rented or purchased from the cloud computing platform.

[0263] The embodiment of the application further provides a computer storage medium storing a computer program, and the computer program can implement the above-mentioned Figure 1 The remote monitoring and data analysis method based on ATP fluorescence detection and the Internet of Things.

[0264] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-mentioned system, device and unit can refer to the corresponding processes in the foregoing method embodiments, and will not be described here.

[0265] The device embodiments described above are only schematic, and the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment. Those skilled in the art can understand and implement without creative labor.

[0266] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and necessary general hardware platforms, and of course, can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of software products, which can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, and the like, and include a plurality of instructions for making a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0267] Finally, it should be noted that the above examples are only used to illustrate the technical solutions of the present application, and are not intended to limit the same; although the present application has been described in detail with reference to the foregoing examples, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing examples, or make equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A remote monitoring and data analysis method based on ATP fluorescence detection and the Internet of Things, characterized in that, include: Raw detection data generated by the ATP fluorescence detection sensor is collected. The raw detection data includes fluorescence intensity values ​​and their associated environmental parameters. The raw detection data is then subjected to real-time noise reduction to obtain processed detection data. Based on the processed detection data, it is efficiently transmitted to the cloud data center via IoT technology, and together with the stored historical detection data, a comprehensive detection data after secure transmission is generated; By using time series analysis algorithms and combining environmental parameters in the comprehensive detection data after secure transmission, the detection data is comprehensively analyzed to identify the changing patterns of biological pollution levels. When the detection data exceeds a preset threshold that is dynamically adjusted based on historical detection data, a multi-level alarm response system is activated. The multi-level alarm response system can select appropriate communication channels to send customized alarm information to designated management personnel according to different alarm types and generate alarm notification records. The process of using time series analysis algorithms to comprehensively analyze the detection data in conjunction with environmental parameters in the securely transmitted integrated detection data to identify the changing patterns of biopollution levels includes: analyzing the fluorescence intensity values ​​and their accompanying environmental parameters in the securely transmitted integrated detection data using time series analysis algorithms, extracting key indicators reflecting biopollution levels and their changing trends over time, and obtaining time series analysis results of biopollution levels, including: Using time series decomposition technology, the fluorescence intensity value F(t) and its accompanying environmental parameters in the comprehensive detection data after secure transmission are decomposed to separate three components: long-term trend T(t), seasonal fluctuation S(t), and random fluctuation R(t). The preliminary decomposition result is obtained as F(t) = T(t) + S(t) + R(t), where t represents time, F(t) represents the fluorescence intensity value at time t, T(t) represents the long-term trend, indicating the long-term variation of the fluorescence intensity value, S(t) represents the seasonal fluctuation, indicating the periodic variation of the fluorescence intensity value, and R(t) represents the random fluctuation, indicating the random variation of the fluorescence intensity value. Based on the preliminary decomposition results, statistical regression analysis was applied to identify the main environmental factors E affecting the change in fluorescence intensity. i (t), establish the fluorescence intensity value F(t) and environmental parameter E i The following is a nonlinear mathematical model between (t), and the model for the influence of environmental factors: Among them; E i (t) represents the environmental parameter, indicating the i-th environmental parameter at time t; β0 is the intercept term, β i γ i and δ i These are the linear, quadratic, and cubic regression coefficients of the i-th environmental parameter, respectively. α1 and α2 are the regression coefficients of the long-term trend and seasonal fluctuation, respectively. n is the number of environmental parameters, and ∈ is the error term, thus obtaining the environmental factor impact model. Using the environmental factor impact model, combined with the long-term trend T(t) and seasonal fluctuation S(t) in the preliminary decomposition results, the level of biological pollution is predicted, key indicators reflecting the level of biological pollution are extracted, and a set of key indicators is generated. The average fluorescence intensity is calculated using the following formula. The fluorescence intensity change rate ΔF(t) can be calculated using the following formula: Where Δt represents the time interval, and represents the time interval for calculating the rate of change of fluorescence intensity; Weighted moving average fluorescence intensity WMA(F(t)): Among them, w i Here, k represents the weight, k is the length of the time window, and i represents the index within the time window. The exponentially smoothed fluorescence intensity ES(F(t)) is calculated using the following formula: ES(F(t))=αF(t)+(1-α)ES(F(t-1)) Where α is the smoothing factor, and its value is between 0 and 1; Based on the set of key indicators, the changing trend of biological pollution levels over time is analyzed to identify the rising, falling, or stable states of biological pollution levels, and time series analysis results of biological pollution levels are generated. Based on the environmental parameters in the alarm notification records, historical detection data, and comprehensive detection data after secure transmission, combined with real-time detection data, a biological pollution prediction model is constructed using machine learning algorithms to generate a biological pollution prediction report. Using the biological pollution prediction report and the problem area information reflected in the alarm notification record, the ATP fluorescence detection frequency setting of each monitoring point is adjusted. The detection frequency is increased for monitoring points that are predicted to be high-risk areas, and the detection frequency is decreased for low-risk areas, thus generating an optimized monitoring strategy.

2. The remote monitoring and data analysis method based on ATP fluorescence detection and the Internet of Things according to claim 1, characterized in that, When the detected data exceeds a preset threshold dynamically adjusted based on historical detected data, a multi-level alarm response system is activated. This system can select appropriate communication channels to send customized alarm information to designated management personnel according to different alarm types, and generate alarm notification records, including: Based on the time series analysis results of the biopollution level, compared with the stored historical detection data, the preset warning threshold is dynamically adjusted using statistical methods to obtain a dynamic warning threshold that adapts to changes in biopollution levels under different environmental conditions. Based on the dynamic warning threshold, the time series analysis results of the biological pollution level are monitored. When the detected fluorescence intensity value exceeds the dynamic warning threshold, a multi-level alarm response mechanism is immediately activated to obtain an alarm trigger signal. Based on the alarm trigger signal, the alarm level is automatically determined according to the severity of the alarm, and the most appropriate communication channel is selected to send customized alarm information, including alarm level, specific pollution location, cause analysis and preliminary suggested response measures, to the preset management personnel or emergency team, and an alarm notification record is generated.

3. The remote monitoring and data analysis method based on ATP fluorescence detection and the Internet of Things according to claim 1, characterized in that, The environmental parameters, based on the alarm notification records, historical detection data, and the comprehensive detection data after secure transmission, combined with real-time detection data, are used to construct a biological pollution prediction model using machine learning algorithms, generating a biological pollution prediction report, including: By collecting and preprocessing environmental parameters from alarm notification records, historical detection data, and comprehensive detection data after secure transmission, as well as real-time detection data, a high-quality dataset is obtained. Based on high-quality datasets, feature selection and transformation are performed to identify key features affecting the prediction of biological pollution. These features are then transformed or encoded to obtain feature data after feature selection and transformation. Using the feature data after feature selection and transformation, support vector machine is selected as the machine learning algorithm to build a biopollution prediction model. At the same time, kernel function is selected and relevant parameters are adjusted to optimize model performance. Based on the constructed biopollution prediction model, cross-validation was used to train and validate the model to ensure that the model has good generalization ability and prediction accuracy, and an optimized biopollution prediction model was obtained. Using an optimized biopollution prediction model, predictions are made on new or future data to generate a detailed biopollution prediction report. The biopollution prediction report includes the prediction time range, prediction results, potential high-risk areas, and corresponding recommended measures.

4. The remote monitoring and data analysis method based on ATP fluorescence detection and the Internet of Things according to claim 3, characterized in that, The process involves using feature data that has undergone feature selection and transformation, selecting Support Vector Machine (SVM) as the machine learning algorithm to construct a biopollution prediction model, and simultaneously selecting a kernel function and adjusting relevant parameters to optimize model performance, including: Using the feature data after feature selection and transformation, support vector machine was selected as the machine learning algorithm to model the task of predicting biological pollution, and a preliminary vector machine model was obtained. Based on the requirements of the biological pollution prediction task, a kernel function is selected for the vector machine model; Using a preliminary vector machine model, the penalty parameters and kernel function parameters are adjusted. The adjustment process employs a parameter optimization method to find the optimal parameter combination within a predefined parameter range, thereby obtaining the optimized parameter combination.

5. The remote monitoring and data analysis method based on ATP fluorescence detection and the Internet of Things according to claim 4, characterized in that, The process involves using feature data that has undergone feature selection and transformation, selecting Support Vector Machine (SVM) as the machine learning algorithm to construct a biopollution prediction model, and simultaneously selecting a kernel function and adjusting relevant parameters to optimize model performance, including: Using the feature data after feature selection and transformation, support vector machine was selected as the machine learning algorithm to model the task of predicting biological pollution, and a preliminary vector machine model was obtained. The basic form of the vector machine model f(x) is expressed as: Where N represents the number of support vectors, α i It is a Lagrange multiplier, y i The label of each support vector is either +1 or -1, K(x, x) i ) is the kernel function, and b is the bias term; Based on the requirements of the biological pollution prediction task, kernel functions are selected for the vector machine model, including linear kernels, polynomial kernels, and radial basis function kernels; the mathematical expression of the radial basis function kernel is K(x, x). i )for: K(x,x i )=exp(-γ||x-x i || 2 ) Where γ is the kernel function parameter, ||xx i || represents the Euclidean distance between two vectors.

6. The remote monitoring and data analysis method based on ATP fluorescence detection and the Internet of Things according to claim 1, characterized in that, The process involves using the biocontamination prediction report, combined with the problem area information reflected in the alarm notification record, to adjust the ATP fluorescence detection frequency settings at each monitoring point. The detection frequency is increased for monitoring points predicted to be high-risk areas, and decreased for low-risk areas, generating an optimized monitoring strategy, including: Using the prediction results, prediction confidence levels, and potential impact analysis in the aforementioned biological pollution prediction report, the future biological pollution risk of each monitoring point is assessed, and the risk area assessment results are obtained. Based on the risk area assessment results and the problem area information reflected in the alarm notification records, the actual pollution risk level of each monitoring point is comprehensively analyzed, and a risk assessment report for each monitoring point is generated. Based on the risk assessment reports of each monitoring point, for monitoring points that are predicted to be high-risk areas, the detection frequency is increased by adjusting the ATP fluorescence detection frequency setting to ensure timely detection and response to biocontamination events, and a monitoring strategy for high-risk areas is generated. For monitoring points that are predicted to be low-risk areas, the detection frequency is appropriately reduced based on the risk assessment reports of each monitoring point, and a monitoring strategy for low-risk areas is generated. The monitoring strategies for high-risk areas and low-risk areas are integrated to form a comprehensive and optimized monitoring strategy.

7. The remote monitoring and data analysis method based on ATP fluorescence detection and the Internet of Things according to claim 1, characterized in that, The processed detection data is efficiently transmitted to a cloud data center via IoT technology, and together with the stored historical detection data, securely transmitted comprehensive detection data is generated, including: Using the processed detection data, a structured detection data packet is formed, resulting in a structured detection data packet; Based on the structured detection data packet, a security encryption protocol in Internet of Things technology is used to encrypt the detection data packet to generate an encrypted detection data packet. The encrypted detection data packet is efficiently transmitted to the cloud data center via the Internet of Things network. During the transmission process, data compression technology is used to reduce the amount of data transmitted, while error control technology is used to ensure the integrity of the data transmission, so as to obtain the detection data after secure transmission. In the cloud data center, the securely transmitted detection data is integrated with the stored historical detection data to generate comprehensive detection data after secure transmission.

8. A remote monitoring and data analysis system based on ATP fluorescence detection and the Internet of Things, used to execute the remote monitoring and data analysis method based on ATP fluorescence detection and the Internet of Things as described in any one of claims 1-7, characterized in that, include: The collection and processing module is used to collect the raw detection data generated by the ATP fluorescence detection sensor. The raw detection data includes fluorescence intensity values ​​and their associated environmental parameters. The module performs real-time noise reduction on the raw detection data to obtain processed detection data. The transmission module is used to efficiently transmit the processed detection data to the cloud data center via Internet of Things technology, and generate comprehensive detection data after secure transmission together with the stored historical detection data. The analysis activation module is used to comprehensively analyze the detection data using time series analysis algorithms combined with environmental parameters in the securely transmitted integrated detection data, identify the changing patterns of biological pollution levels, and activate a multi-level alarm response system when the detection data exceeds a preset threshold dynamically adjusted based on historical detection data. This multi-level alarm response system can select appropriate communication channels to send customized alarm information to designated management personnel according to different alarm types, generating alarm notification records. The process of comprehensively analyzing the detection data using time series analysis algorithms combined with environmental parameters in the securely transmitted integrated detection data to identify the changing patterns of biological pollution levels includes: analyzing the fluorescence intensity values ​​and accompanying environmental parameters in the securely transmitted integrated detection data using time series analysis algorithms, extracting key indicators reflecting biological pollution levels and their changing trends over time, and obtaining time series analysis results of biological pollution levels. The construction module is used to build a biological pollution prediction model based on the environmental parameters in the alarm notification records, historical detection data and the comprehensive detection data after secure transmission, combined with real-time detection data, and to generate a biological pollution prediction report by using machine learning algorithms. The adjustment generation module is used to adjust the ATP fluorescence detection frequency settings of each monitoring point by combining the biological pollution prediction report with the problem area information reflected in the alarm notification record. The detection frequency is increased for monitoring points that are predicted to be high-risk areas, and the detection frequency is decreased for low-risk areas, thereby generating an optimized monitoring strategy.

Images