Micro water analysis data preprocessing and transmission method and system based on edge computing

Abstract

The application discloses a micro water analysis data preprocessing and transmission method and system based on edge computing, relates to the technical field of data analysis and processing, and comprises the following steps: deploying a sampler at a first edge end, guiding a multi-source sensor array to collect multi-source water body signals, performing first value classification processing based on signal morphology comparison, performing second screening and compression processing based on a value density negotiation protocol, and performing third water body systematized entropy reduction processing based on an entropy reduction gateway, generating a water body preprocessing result, packaging and determining data transmission conditions at a sending interface of the first edge end, and performing communication backhaul based on a second middle station, so as to solve the technical problems that the intelligence of a water body multi-source signal processing system is insufficient and it is difficult to reasonably balance the dynamic adaptation of adaptive sampling and preprocessing transmission in the prior art, and realize intelligent water body analysis management. The real-time performance and response speed of micro water analysis are improved.

Description

Technical Field

[0001] This invention relates to the field of data analysis and processing technology, specifically to a method and system for preprocessing and transmitting trace water analysis data based on edge computing. Background Technology

[0002] Currently, water quality monitoring technologies generally rely on central nodes for centralized data processing. However, limited by network bandwidth, communication power consumption, and transmission latency, these technologies struggle to meet the requirements of efficiency and reliability in real-time monitoring and anomaly early warning for trace water bodies. Furthermore, due to the complexity of water body signal sources, the sheer volume of data, and the inherent uncertainty of the data, directly transmitting raw signals not only wastes resources but also easily leads to critical information being buried in redundant data, affecting the accuracy and timeliness of monitoring results.

[0003] In existing technologies, some technologies add a preprocessing module at the sampling end to compress or filter some data, but most methods only stay at the level of simple filtering or compression, and cannot effectively balance data value and communication bandwidth in dynamic acquisition scenarios.

[0004] In summary, optimizing the processing system for multi-source water signals, balancing the dynamic adaptation of adaptive sampling and preprocessing transmission, and improving the intelligence of water analysis and processing are urgent technical problems that need to be solved. Summary of the Invention

[0005] This application provides a method and system for preprocessing and transmitting trace water analysis data based on edge computing, which is used to address the technical problem of insufficient intelligence in existing water multi-source signal processing systems, making it difficult to reasonably balance the dynamic adaptation of adaptive sampling and preprocessing transmission.

[0006] In view of the above problems, this application provides a method and system for preprocessing and transmitting trace water analysis data based on edge computing.

[0007] In a first aspect, this application provides a method for preprocessing and transmitting trace water analysis data based on edge computing. The method includes: deploying a sampler at a first edge end to guide a multi-source sensor array to collect multi-source water signals, wherein the sampler guides multi-source water sampling with minimum uncertainty and risk control guidance; preprocessing the multi-source water signals to generate water preprocessing results, wherein the preprocessing steps include a first value classification processing based on signal morphology comparison, a second screening and compression processing based on a value density negotiation protocol, and a third systematic entropy reduction processing based on an entropy reduction gateway; encapsulating the water preprocessing results at a transmission interface at the first edge end, determining data transmission conditions according to the encapsulation mode, and performing communication backhaul based on a second middle platform.

[0008] Secondly, this application provides a micro-water analysis data preprocessing and transmission system based on edge computing. The system includes: a water sampling unit for deploying a sampler at a first edge end to guide a multi-source sensor array to collect multi-source water signals, wherein the sampler guides multi-source water sampling with minimum uncertainty and risk control guidance; a preprocessing unit for preprocessing the multi-source water signals to generate water preprocessing results, wherein the preprocessing steps include a first value classification processing based on signal morphology comparison, a second screening and compression processing based on a value density negotiation protocol, and a third systematic entropy reduction processing based on an entropy reduction gateway; and a transmission unit for encapsulating the water preprocessing results at the sending interface at the first edge end, determining data transmission conditions according to the encapsulation mode, and performing communication backhaul based on a second middle platform.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] The edge computing-based data preprocessing and transmission method for trace water analysis provided in this application deploys a sampler at a first edge to guide a multi-source sensor array to collect multi-source water signals. The method preprocesses these signals to generate preprocessed water results, including a first value grading process based on signal morphology comparison, a second screening and compression process based on a value density negotiation protocol, and a third systematic entropy reduction process based on an entropy reduction gateway. The preprocessed water results are encapsulated at the transmission interface at the first edge. Data transmission conditions are determined based on the encapsulation mode, and communication backhaul is performed based on a second platform. This method addresses the insufficient intelligence of existing multi-source water signal processing systems and the difficulty in reasonably balancing the dynamic adaptation of adaptive sampling and preprocessing transmission, thereby achieving intelligent water analysis management and improving the real-time performance and response speed of trace water analysis. Attached Figure Description

[0011] Figure 1 This application provides a schematic diagram of the data preprocessing and transmission method for trace water analysis based on edge computing;

[0012] Figure 2 This application provides a schematic diagram of the structure of a trace water analysis data preprocessing and transmission system based on edge computing.

[0013] Explanation of reference numerals in the attached figures: Water sampling unit 11, pretreatment unit 12, transmission unit 13. Detailed Implementation

[0014] This application provides a method and system for preprocessing and transmitting trace water analysis data based on edge computing, which addresses the technical problem in the existing technology of insufficient intelligence in multi-source signal processing systems for water bodies and the difficulty in reasonably balancing the dynamic adaptation of adaptive sampling and preprocessing transmission.

[0015] Example 1: As Figure 1 As shown, this application provides a method for preprocessing and transmitting trace water analysis data based on edge computing, the method comprising:

[0016] S1: Deploy a sampler at the first edge to guide the multi-source sensor array to collect multi-source water signals, wherein the sampler guides multi-source water sampling with minimal uncertainty and risk control guidance.

[0017] In this embodiment of the invention, a sampler is first deployed at a first edge end. The first edge end refers to an edge computing node close to the source of water data generation, which has the ability to directly interact with the multi-source sensor array. By deploying a sampler at this edge end, the multi-source sensor array is guided to collect multi-source water signals, ensuring that the sampling logic can be dynamically adjusted to closely match the actual characteristics of the water body on site.

[0018] In this application, the multi-source sensor array refers to a collection set composed of multiple types of sensor units. This array can simultaneously sense multi-dimensional parameters in the water body, such as temperature, pH value, dissolved oxygen, conductivity, and the concentration of specific pollutant ions. It acquires multi-dimensional and multi-level signals within the same spatiotemporal range, thereby avoiding the information limitations that may result from a single sensing channel.

[0019] During the guided sampling process in the technical solution of this application, the sampler conducts multi-source water sampling guidance based on the principles of minimizing uncertainty and risk control, and collects multi-source water signals of the target area.

[0020] Specifically, the minimum uncertainty refers to minimizing the ambiguity and noise of the collected data in a statistical sense through probability analysis or uncertainty quantification during sampling decision-making. For example, when preliminary environmental monitoring shows rapid fluctuations in dissolved oxygen levels in water, the sampler can instruct relevant sensors to increase the sampling frequency and simultaneously trigger the synchronous collection of other related indicators.

[0021] The risk control guidance refers to incorporating risk control logic into the data acquisition strategy, optimizing sampling decisions in advance by predicting potential water quality anomalies. For example, when a region is likely to experience rapid accumulation of pollutants, the sensor array is guided to prioritize acquiring signals related to that pollutant and correspondingly reduce the acquisition frequency of low-risk indicators, ensuring that limited resources are concentrated on serving high-risk areas.

[0022] In summary, by combining minimum uncertainty with risk control guidance, the sampler can achieve efficient data value prioritization and risk perception during the acquisition phase, thereby significantly improving the scientific rigor and reliability of multi-source water body signal acquisition.

[0023] Furthermore, prior to deploying the sampler at the first edge, the construction of the sampler, in step S1 of this application, includes:

[0024] Based on the first uncertainty positioning, a first sampling decision node is deployed, wherein the first sampling decision node outputs a first uncertainty probability field; based on the second risk control trend prediction, a second sampling decision node is deployed, wherein the second sampling decision node outputs a second risk control probability field; a probability weighting channel is established between the first sampling decision node and the second sampling decision node to generate the sampler, and the sampler is embedded in the first edge terminal.

[0025] In this embodiment of the invention, a first sampling decision node is deployed based on a first uncertainty positioning. Specifically, the uncertainty positioning refers to establishing an uncertainty quantification probability in a water signal sampling scenario. Specifically, signal noise, sensor drift, and data volatility in the sampling environment are analyzed to determine the data reliability range; unstable signals have larger fluctuations, and their uncertainty probability is higher than that of stable signals, etc.

[0026] The first sampling decision node transforms the signals acquired by the multi-source sensor array into a probability distribution, which serves as the first uncertain probability field. This first uncertain probability field is a spatialized or vectorized uncertainty mapping used to characterize the confidence level of different water body signals within a confidence interval. For example, when dissolved oxygen data acquired by a certain sensor fluctuates abnormally, the node can reduce the weight of that signal through uncertainty calculation, thereby reducing its impact on the overall sampling decision.

[0027] Subsequently, a second sampling decision node is deployed based on the second risk control trend prediction. Specifically, the risk control trend prediction refers to using statistical modeling or machine learning methods to identify risks and make forward-looking inferences about the changing trends of water quality parameters, especially predicting possible limit exceedances or sudden changes. The second sampling decision node deployed based on this prediction logic is mainly used to mark and focus on potential risk signals in advance, and its output is the second risk control probability field.

[0028] The second risk control probability field reflects the risk probability distribution that various signals may bring in the future. For example, when the concentration of a certain pollutant ion in the water body shows an accelerating upward trend, the node will assign a higher risk weight to its output risk control probability field, thereby guiding subsequent sampling strategies to pay more attention to this type of signal.

[0029] Subsequently, a probability weighting channel is established between the first sampling decision node and the second sampling decision node. Specifically, the probability weighting channel refers to a computational path that integrates information from different probability fields. By weighting and combining the first uncertain probability field and the second risk control probability field, a composite evaluation system that takes into account both data uncertainty and risk perception is generated.

[0030] In a preferred embodiment of this application, the relative importance between uncertainty reduction and risk prediction can be balanced according to a preset weighting coefficient or a dynamic adjustment method.

[0031] For example, when the aquatic environment is relatively stable, the weight of the uncertainty probability field can be increased to ensure data reliability; while when the risk of sudden changes in the aquatic environment is high, the weight of the risk control probability field can be increased to achieve rapid response. Through this weighting channel, a sampling control mechanism with dual guidance capabilities is ultimately formed.

[0032] In summary, the sampler is generated based on the first sampling decision node, the second sampling decision node, and the probability weighting channel, and is used to dynamically guide the acquisition behavior of the multi-source sensor array in real-time operation.

[0033] Finally, the sampler is embedded within the first edge end, allowing it to reside and operate as part of the first edge end long-term. This not only enables local real-time data reliability assessment and risk prediction but also avoids the latency and communication overhead caused by remote computing, thereby ensuring a comprehensive improvement in accuracy, sensitivity, and real-time performance throughout the entire water signal acquisition process.

[0034] Furthermore, to guide the multi-source sensor array in acquiring multi-source water signals, step S1 of this application includes:

[0035] According to the preset sampling period, the multi-source sensor array is driven to perform one-step standardized sampling to determine the multi-source water body signal; the multi-source water body signal is imported into the sampler, and the uncertain probability field analysis based on the first sampling decision node and the risk control probability field analysis based on the second sampling decision node are performed in parallel. By performing probability weighting calculation, the re-sampled signal sources with a probability threshold are screened; for the re-sampled signal sources, the multi-source sensor array is directionally driven and re-sampled to compensate for the multi-source water body signal.

[0036] In this embodiment of the invention, a multi-source sensor array is first driven to perform a standardized sampling step according to a preset sampling period. The preset sampling period refers to a fixed time interval set based on monitoring requirements or environmental characteristics, used to uniformly trigger the acquisition operations of various sensors. Standardized sampling refers to a signal acquisition step completed under a unified timestamp, a unified parameter range, and a unified sampling accuracy. Its purpose is to ensure that the data output by different sensors within the same time period are comparable and consistent.

[0037] For example, within a ten-minute sampling period, the temperature sensor, dissolved oxygen sensor, and conductivity sensor will be activated simultaneously and collect data, resulting in a preliminary set of multi-source water body signals.

[0038] Subsequently, after acquiring the aforementioned multi-source water body signals, the multi-source water body signals are imported into the sampler, and the uncertain probability field analysis based on the first sampling decision node and the risk control probability field analysis based on the second sampling decision node are executed in parallel.

[0039] Specifically, through uncertainty probability field analysis—that is, by performing probability analysis on the uncertainty and confidence intervals of each signal source through the first sampling decision node—potential drifts and outliers in the data can be identified. Simultaneously, through risk control probability field analysis—that is, by using the second sampling decision node to predict the risk of signal change trends—potential indicators that may lead to pollution spread or water quality exceeding limits can be identified. Two types of probability fields can be obtained.

[0040] Subsequently, after completing the analysis of the two types of probability fields mentioned above, probability weighting calculations are performed to screen out repeatable signal sources exceeding the probability threshold. Specifically, through probability weighting calculations, that is, using a pre-set or dynamically adjusted weight distribution, the two types of probability fields, including the uncertain probability field and the risk control probability field, are merged to generate a comprehensive probability evaluation value. Further, a threshold-based judgment is performed, whereby signals exceeding the probability threshold are listed as key repeatable targets.

[0041] For example, if the heavy metal ion concentration data collected by a certain sensor is stable in uncertainty analysis, but shows a rapid upward trend in risk control prediction, then the comprehensive weighted probability of the signal source will inevitably exceed the threshold, and thus it will be identified as a re-collection signal source.

[0042] Subsequently, for the resampled signal source, the multi-source sensor array is directionally driven and resampled to compensate for the multi-source water signal. Specifically, the directional driving refers to the sampler issuing instructions to specific sensor units to perform additional acquisition tasks at a specific parameter range or a higher sampling frequency; the resampled refers to sampling the same signal source multiple times to further verify the stability and risk characteristics of its values.

[0043] Finally, by fusing the resampled data with the initial sampling results, the proportion of high-value data in the overall signal set can be increased, thus providing a more complete and reliable data foundation for subsequent preprocessing steps.

[0044] S2: Preprocess the multi-source water signals to generate water preprocessing results. The preprocessing steps include a first value classification process based on signal morphology comparison, a second screening and compression process based on value density negotiation protocol, and a third systematic entropy reduction process based on entropy reduction gateway.

[0045] In this embodiment of the invention, the multi-source water signals undergo preprocessing. Specifically, before the raw sampled data enters the central analysis platform, a multi-level data optimization operation is performed at the first edge. The goal is to remove redundancy, highlight valuable data, and reduce transmission burden, while ensuring that key feature information is not lost. Through this preprocessing step, the original complex data stream is transformed into a more compact processing result with higher information value, thus laying the foundation for subsequent transmission and in-depth analysis.

[0046] In the specific steps, the preprocessing includes the first value classification process based on signal morphology comparison. That is, by establishing baseline curve templates for various target water body types, the actual collected signals are compared with the trend of standard signals to identify their morphological characteristics, such as trend changes, peak fluctuations and duration, and different signals can be classified into different value levels.

[0047] For example, if the output data of a sensor highly overlaps with the baseline template, it can be identified as a stable signal with a relatively low value level, and can be compressed at a high ratio in subsequent processing; conversely, if a signal shows a significant trend shift or sudden anomaly, its value level is high, and it should be retained and given special processing in the compression stage.

[0048] Subsequently, after completing the value classification, a second screening and compression process based on the value density negotiation protocol is further executed. That is, the dynamic parameter negotiation mechanism agreed upon between the first edge end and the second middle platform is used to determine the priority of data screening and compression according to the value density of the data in different scenarios.

[0049] In this application, the specific negotiation method is as follows: based on the real-time generated value assessment summary, the corresponding value density distribution is calculated, and the signals in the first classification result are filtered accordingly. The different value signals retained after filtering will be processed using different compression modes according to their classification level.

[0050] For example, trending data may be adaptively compressed, while outlier data remains lossless, thus achieving an optimization effect that balances efficiency and integrity.

[0051] Subsequently, after completing the above screening and compression processes, the data structure needs to be further optimized through the systematic entropy reduction processing of the third water body system based on the entropy reduction gateway. That is, the signal is subjected to multi-dimensional projection and structured analysis, and the data volume is reduced and the information value is improved through entropy reduction operation.

[0052] Specifically, this step involves systematically projecting the second processing result to determine the correlation between different data points, and then using reverse deduction to obtain a state vector representing the characteristics of the underlying water body. In the implementation of this application, redundant information is systematically removed, while key features are retained and given higher weights. Finally, the underlying state vector output by the entropy reduction gateway serves as the water body preprocessing result, providing high-value, high-density input for subsequent data encapsulation and transmission.

[0053] Furthermore, the multi-source water body signals are preprocessed. The preprocessing step includes a first value classification process based on signal morphological comparison. Step S2 of this application includes:

[0054] For the target water body type, a baseline curve template is prepared, wherein the baseline curve template contains the standard signal trend state of each sensor source; based on the baseline curve template, the multi-source water body signals are morphologically compared, value classification is performed, first-processed water body data is generated, and a value assessment summary is generated, wherein the morphological features include at least trend, peak value, and duration.

[0055] The value grading levels include at least a Level 1 based on stable baseline data, a Level 2 conforming to specific trend changes, and a Level 3 for out-of-limit abnormal data; wherein Level 1 performs high-distortion compression, Level 2 performs adaptive distortion compression, and Level 3 is lossless.

[0056] In this embodiment of the invention, a baseline curve template is first established for the target water body type. Specifically, the target water body type refers to the category of water body that needs to be monitored in different application scenarios, such as surface water, industrial circulating water, or drinking water. Different water body types have different physicochemical properties and monitoring indicators, so corresponding baseline curve templates need to be established for each type of water body before implementation. The baseline curve template contains the standard signal trend state of each sensor source, that is, the numerical range and change pattern that each monitoring parameter should exhibit under normal or acceptable conditions.

[0057] For example, temperature signals typically exhibit periodic fluctuations, pH values ​​should be maintained within a narrow range, and the concentrations of certain pollutants should remain close to zero for extended periods. By establishing these standard trend states, comparative references can be created to identify deviations between the collected data and the ideal conditions.

[0058] Subsequently, based on the baseline curve template, the multi-source water body signals are subjected to morphological comparison to perform value classification.

[0059] In the embodiments of this application, the morphological comparison refers to the corresponding analysis of the actual collected multi-source water body signals with the baseline curve template in multiple dimensions such as trend, peak value, and duration. For example, if the actual curve closely matches the baseline curve, it indicates that the data is in a stable state; if a slight but continuous trend deviation occurs, it indicates that there is a potential risk; if a sudden peak value exceeding the limit occurs, it indicates that the water body may have been abnormally polluted.

[0060] In this analysis, the signals are value-graded based on the comparison results, and first-stage water body data is generated accordingly. Simultaneously, a value assessment summary is generated for the first-stage water body data to briefly describe the degree of deviation and importance of different signals in terms of characteristics such as trend, peak value, and duration.

[0061] In the specific grading system of this application, the value grading level includes at least three levels. Level 1 is a signal based on stable baseline data, indicating that the signal is highly consistent with the standard state, with low risk and limited information increment; Level 2 is a signal that conforms to a specific trend change, indicating that although the data has not exceeded the limit, it shows a certain deviation and needs to be continuously monitored; Level 3 is abnormal data that exceeds the limit, indicating that the signal has seriously deviated from the standard and needs to be fully preserved for further analysis.

[0062] Synchronously, in terms of corresponding processing strategies, the first-level signal undergoes high-distortion compression to significantly reduce the data volume; the second-level signal undergoes adaptive distortion compression to ensure both the preservation of key trend information and transmission efficiency; and the third-level signal remains lossless to ensure that its key anomaly information is not weakened or lost. Through this hierarchical and corresponding compression strategy, the system can achieve a dynamic balance between data scale and information fidelity, thereby providing an optimized data foundation for subsequent filtering and transmission stages.

[0063] Furthermore, the multi-source water signals are preprocessed. The preprocessing step includes a second screening and compression process based on a value density negotiation protocol. Step S2 of this application includes:

[0064] By using a value density negotiation protocol maintained at the first edge and the second middle platform, dynamic value negotiation is performed on the value assessment summary to generate a value probability threshold; based on the value probability threshold, value-oriented screening and adaptive compression processing based on value grading are performed on the first processed water body data to determine the second processed water body data.

[0065] In this embodiment of the invention, the value assessment summary is first dynamically negotiated through a value density negotiation protocol maintained between the first edge terminal and the second middle platform. Specifically, the value density negotiation protocol refers to a bidirectional interaction mechanism between the edge terminal and the middle platform, used to evaluate the value density of different water body signals in the overall data stream in real time, that is, the amount of information value carried by a unit of data.

[0066] In one exemplary process, the value density negotiation protocol can dynamically adjust the importance ranking of different signals by comparing the value assessment summary generated in real time at the first edge with the historical monitoring model stored in the second middle platform, thereby allocating value density under limited resource constraints.

[0067] For example, when the edge detects that the numerical trends of certain sensor sources are highly correlated with historical anomaly patterns, the negotiation protocol will immediately increase the value weight of such signals. Ultimately, through this dynamic negotiation process, a quantifiable value probability threshold is generated for subsequent data filtering.

[0068] Subsequently, based on the value probability threshold, the first processed water data is subjected to value-oriented screening and adaptive compression processing based on value grading levels.

[0069] Specifically, the value-oriented filtering refers to using a value probability threshold as a dividing line, discarding data below the threshold as redundant or low-contribution information, while retaining data above the threshold for further processing. This ensures that high-value portions of the data stream are highlighted, while low-value portions no longer consume transmission and storage resources.

[0070] Based on this, corresponding adaptive compression strategies are implemented for data of different value levels: for Level 1 stable data, high-distortion compression is performed to significantly reduce the data volume; for Level 2 trend change data, adaptive distortion compression is adopted to reduce the storage and transmission burden while ensuring the integrity of trend information; for Level 3 abnormal data, lossless compression is maintained to ensure that key abnormal information is completely preserved.

[0071] In summary, after the above filtering and adaptive compression processes, the second processed water body data is obtained. This dataset is smaller in size than the original data, but its information density is significantly improved. This avoids redundant data consuming limited resources and ensures that high-risk and high-value information can be completely transmitted to the central platform for further analysis and decision-making.

[0072] Thus, a balance was achieved in overall performance, reducing the amount of data and enhancing the value of information, thereby improving the effectiveness of water quality monitoring in terms of real-time performance and reliability.

[0073] Furthermore, the multi-source water body signals are preprocessed. The preprocessing step includes a third-systematic entropy reduction process based on an entropy reduction gateway. Step S2 of this application includes:

[0074] Through the entropy reduction gateway, multi-dimensional projection based on the water system is performed on the second processed water data to determine the water projection state; for the water projection state, the underlying state vector is deduced by reverse inference, wherein the entropy reduction processing target is the direction of data volume reduction and the direction of information value increase; the underlying state vector is used as the water preprocessing result.

[0075] In this embodiment of the invention, the water body data is first processed by an entropy reduction gateway to perform a multi-dimensional projection based on the water body system, thereby determining the water body projection state.

[0076] The entropy reduction gateway proposed in this application refers to a processing unit that performs entropy reduction operations at the data structure level. Its function is to transform complex raw data streams into projection results with lower redundancy and higher aggregation by remapping multi-source water body signals across multiple feature dimensions. Specifically, the multi-dimensional projection refers to mapping multiple dimensions such as trend features, peak features, duration features, and cross-parameter correlation features on the same signal set to generate a systematic data state representation as the water body projection state.

[0077] For example, when dissolved oxygen, pH value, and conductivity show a high correlation over a certain period of time, the entropy reduction gateway can aggregate this correlation into a more compact projected state vector through multi-dimensional projection, thereby reducing the existence of redundant information.

[0078] Subsequently, after obtaining the water body projection state, the underlying state vector is further deduced in reverse based on the water body projection state.

[0079] Specifically, the reverse deduction refers to, based on the projection results, gradually tracing back to the minimum state vector that can represent the core characteristics of the water system, following the logic of data compression and feature aggregation. Preferably, the underlying state vector emphasized here is not a simple set of all original data, but a set of core feature indicators extracted through mathematical modeling, capable of expressing maximum information value with minimal data volume.

[0080] In this simulation, the entropy reduction processing aims to reduce the amount of data and increase the value of information. That is, by eliminating redundant signals and weakly correlated features, the overall data size is reduced, while the dataset's ability to describe the actual state of the water body is improved.

[0081] For example, when a large number of similar stable signals are integrated into a trend indicator, the amount of data is significantly reduced, but the trend indicator can more directly reveal the nature of water body evolution.

[0082] Finally, the underlying state vector obtained from the above reverse deduction is used as the preprocessing result of the water body. This underlying state vector not only has high value density and low redundancy, but also can be directly interfaced with the analysis model of the subsequent second platform, thus ensuring transmission efficiency without losing key monitoring information.

[0083] In this embodiment of the application, a feasible way to construct the entropy reduction gateway is as follows: using the multi-dimensional projection of the water system as the first processing logic and the reverse inference of the underlying state vector as the second processing logic, a gateway logic architecture is constructed. Furthermore, through sample-supervised training, the gateway logic architecture is trained to convergence, the constructed entropy reduction gateway is obtained, and it is embedded and deployed on the first edge.

[0084] In summary, through this entropy reduction process, the present invention can achieve in-depth optimization of water signals, enabling the entire preprocessing system to complete data slimming and value extraction at the edge, laying an efficient and reliable foundation for subsequent encapsulation and transmission.

[0085] S3: The water pretreatment results are encapsulated at the sending interface at the first edge, and the data transmission conditions are determined according to the encapsulation mode to perform communication backhaul based on the second middle platform.

[0086] In this embodiment of the invention, the water pretreatment results are first encapsulated at the sending interface at the first edge end, i.e., the communication outlet at the first edge end used for external data transmission. Specifically, the sending interface includes not only a physical transmission port but also a logic module for implementing data packaging, encryption, and protocol adaptation.

[0087] During the encapsulation process, the underlying state vector, optimized for entropy reduction, is transformed into standardized transmission units according to a predetermined encapsulation format to ensure the integrity and compatibility of data in subsequent communication. This ensures consistency across different batches of data and provides structured support for parsing at the receiving end.

[0088] Subsequently, the data transmission conditions are determined based on the encapsulation mode. The encapsulation mode refers to the different packaging strategies used in the encapsulation process, including but not limited to the heartbeat packet mode and the information ingot mode.

[0089] Specifically, the heartbeat packet mode is typically used to transmit low-value or routine data. It is characterized by small data volume and high transmission frequency, and is suitable for maintaining the middle platform's real-time awareness of the edge status. The information ingot mode is used to carry data containing high-value or abnormal signals. Especially when there is data with level 3 value, it is automatically encapsulated into the information ingot mode to ensure that the data receives higher priority and stronger security during transmission.

[0090] Subsequently, based on the different encapsulation modes, data security levels, communication power consumption, and transmission bandwidth are further combined to generate corresponding data transmission conditions, thereby dynamically adjusting transmission priorities and resource allocation.

[0091] Furthermore, based on the data transmission conditions, communication backhaul is performed using the second middle platform. That is, the first edge terminal sends the encapsulated water pretreatment results to the second middle platform via an interactive network according to predetermined data transmission conditions. The second middle platform not only receives and stores the data but also performs further global analysis and risk assessment on the returned data. Through this backhaul process, a closed-loop data interaction mechanism is formed between the edge terminal and the middle platform, enabling high-value signals to be rapidly transmitted to the central layer while ensuring efficiency and security, thus realizing a complete technical chain from edge acquisition and local preprocessing to global analysis.

[0092] Furthermore, determining the data transmission conditions, step S3 of this application includes:

[0093] The water pretreatment results are encapsulated, and a first priority condition is generated according to the encapsulation mode, wherein the encapsulation mode is either a heartbeat packet mode or an information ingot mode, and when there is three levels of value data, it is encapsulated as an information ingot mode; a second channel condition is generated based on the data security level, communication power consumption and transmission bandwidth; and data transmission conditions are determined according to the first priority condition and the second channel condition.

[0094] In this embodiment of the invention, the water pretreatment result is first encapsulated, that is, the underlying state vector obtained by the first edge preprocessing is structured according to a predetermined rule, so as to transform it into a data unit that can be directly entered into the transmission channel.

[0095] In one specific implementation, during the encapsulation process, the data is numbered, segmented, verified, and formatted to ensure that the data can be correctly identified and restored in different transmission paths. Through this encapsulation operation, the originally scattered preprocessing results are unified into standardized transmission objects, laying the foundation for subsequent condition settings.

[0096] After encapsulation is completed, the first priority condition is generated according to the encapsulation mode.

[0097] Specifically, the encapsulation mode refers to the encapsulation strategy adopted for different data value characteristics, mainly including two forms: heartbeat packet mode and information ingot mode. Heartbeat packet mode is suitable for low-value or routine monitoring data, aiming to maintain communication stability between the edge and the middleware with low load. Information ingot mode is used to carry high-value or abnormal data, especially in the presence of level 3 value data, where information ingot mode will be forcibly activated to ensure that such critical information receives the highest transmission priority. Therefore, the first priority condition is determined by the selected encapsulation mode, guiding the priority order of data in the transmission link.

[0098] Subsequently, after determining the first priority condition, the second channel condition is generated based on data security level, communication power consumption, and transmission bandwidth.

[0099] Specifically, the data security level refers to the requirements for data encryption, integrity verification, and anti-attack capabilities during transmission; the communication power consumption refers to the energy level consumed under different transmission methods, especially applicable to battery-powered edge nodes; and the transmission bandwidth is the maximum data rate that currently available communication links can support. By comprehensively evaluating the magnitude of the above three factors, a suitable transmission channel strategy is allocated to the data. For example, a fully encrypted high-speed link is selected when high security requirements are needed and bandwidth is sufficient, while compression and low-energy transmission methods are selected in low-power scenarios.

[0100] Finally, based on the first priority condition and the second channel condition, the final data transmission conditions are determined. This constitutes the overall set of strategies allocated to the currently encapsulated data in the transmission link, including priority, channel selection, power consumption constraints, and security requirements. Through joint matching of priority and channel conditions, the optimal transmission path can be dynamically adapted, ensuring that high-value data is transmitted with the highest priority and strongest security, while regular data is transmitted in a resource-efficient manner.

[0101] Thus, this invention achieves differentiated management of data of different values ​​while ensuring transmission efficiency and reliability.

[0102] Furthermore, step S3 of this application includes:

[0103] The interaction network between the first edge terminal and the second middle platform is determined; based on the data transmission conditions, matching based on real-time channel status is performed in the interaction network to determine the target interaction thread; based on the target interaction thread, the water pretreatment result is transmitted back from the first edge terminal to the second middle platform.

[0104] In this embodiment of the invention, the interaction network between the first edge terminal and the second middle platform is first determined. The interaction network refers to the communication architecture that enables data backhaul between the first edge terminal and the second middle platform, which can include different forms such as cellular communication networks, low-power wide area networks, satellite links, or wired dedicated networks. The process of determining the interaction network not only considers the reachability of the physical link but also takes into account the bandwidth, latency, and security requirements of the on-site deployment environment, thereby providing a stable foundation for subsequent transmission.

[0105] For example, in remote monitoring scenarios, low-power wide-area communication methods can be prioritized to ensure long-term equipment operation; while in high-risk waters, dedicated networks with encryption and high bandwidth capabilities can be selected to ensure data security and real-time performance.

[0106] Subsequently, based on the data transmission conditions, matching based on real-time channel status is performed in the interactive network to determine the target interactive thread. The real-time channel status refers to the instantaneous parameter performance of the current network link, including bandwidth utilization, packet loss rate, latency level, and interference intensity. Before transmission, the network status is dynamically monitored, and combined with the data transmission conditions generated in the previous step, specifically including priority requirements, security levels, and power consumption limits, a matching algorithm selects the most suitable transmission path and thread in the interactive network as the target interactive thread. This is a dedicated logical channel allocated within the given network for executing this data transmission, ensuring that data of different value levels do not interfere with each other during parallel transmission and obtain corresponding quality of service guarantees.

[0107] For example, high-value Level 3 anomaly data will be assigned to low-latency, high-bandwidth, and encrypted threads, while low-value routine monitoring data will go into low-power threads to save energy.

[0108] Subsequently, according to the target interaction thread, the water pretreatment results are transmitted back from the first edge terminal to the second middle platform, enabling the second middle platform to receive the high-value pretreatment results from the edge terminal in the shortest possible time. Furthermore, after receiving the results, the second middle platform can not only store and back them up, but also perform global analysis and prediction based on large-scale models and historical data.

[0109] In summary, by adopting a hierarchical backhaul mechanism based on interactive networks, transmission conditions, and thread matching, this invention achieves an efficient data closed loop between the edge and the middle platform, ensuring the priority transmission of high-risk signals while also balancing communication energy consumption and network resources.

[0110] The edge computing-based method for preprocessing and transmitting trace water analysis data provided in this application has the following technical advantages:

[0111] 1. The sampler guides sampling with minimal uncertainty and risk control, and combines a resampling mechanism to compensate for the signal, thereby improving the accuracy and integrity of multi-source water body signals and providing a high-quality data foundation for subsequent analysis.

[0112] 2. The three-tiered preprocessing steps work synergistically: morphological comparison enables value grading, value density protocol optimizes filtering and compression, and entropy reduction gateway reduces data redundancy and enhances information value, significantly improving data processing efficiency and effectiveness. Based on a value-grading-based differentiated compression strategy, high-distortion, adaptive distortion, and lossless processing are applied to data of different levels, reducing data volume while ensuring the integrity of key information and saving storage and transmission resources. The value density negotiation protocol supports dynamic negotiation between the edge and the middle platform, making data filtering more aligned with actual needs, avoiding invalid data transmission, and improving the targeting and flexibility of preprocessing. The entropy reduction gateway achieves the dual goals of reducing data volume and enhancing information value through multi-dimensional projection and back-inference from the underlying state vector, strengthening the systematization and usability of data.

[0113] 3. Transmission conditions are determined based on the encapsulation mode and channel conditions. Target threads are matched in conjunction with real-time channel status to ensure timely and reliable data transmission, adapting to complex network environments. The information ingot mode prioritizes the transmission of three levels of value data, ensuring high-value anomaly information is processed first. The heartbeat packet mode adapts to regular data transmission, optimizing transmission priority management. Preprocessing is completed at the edge, reducing the amount of uploaded data, decreasing reliance on transmission bandwidth, alleviating the computational pressure on the central platform, and improving the real-time performance and response speed of trace water analysis.

[0114] Example 2: Based on the same inventive concept as the edge computing-based micro-water analysis data preprocessing and transmission method in the aforementioned examples, such as... Figure 2 As shown, this application provides a micro-water analysis data preprocessing and transmission system based on edge computing, the system comprising:

[0115] The water sampling unit 11 is used to deploy a sampler at the first edge end to guide the multi-source sensor array to collect multi-source water signals, wherein the sampler guides the multi-source water sampling with minimal uncertainty and risk control guidance.

[0116] Preprocessing unit 12 is used to preprocess the multi-source water body signals and generate water body preprocessing results. The preprocessing steps include a first value classification processing based on signal morphology comparison, a second screening and compression processing based on value density negotiation protocol, and a third systematic entropy reduction processing based on entropy reduction gateway.

[0117] The transmission unit 13 is used to encapsulate the water pretreatment results at the sending interface of the first edge end, determine the data transmission conditions according to the encapsulation mode, and perform communication backhaul based on the second middle platform.

[0118] Furthermore, the water sampling unit 11 performs the following steps: deploying a first sampling decision node based on a first uncertainty positioning, wherein the first sampling decision node outputs a first uncertainty probability field; deploying a second sampling decision node based on a second risk control trend prediction, wherein the second sampling decision node outputs a second risk control probability field; establishing a probability weighting channel between the first sampling decision node and the second sampling decision node, generating the sampler, and embedding the sampler into the first edge end.

[0119] Furthermore, the water sampling unit 11 performs the following steps: according to a preset sampling period, it drives the multi-source sensor array to perform one-step standardized sampling to determine the multi-source water signals; it imports the multi-source water signals into the sampler and performs in parallel uncertain probability field analysis based on the first sampling decision node and risk control probability field analysis based on the second sampling decision node; by performing probability weighting calculation, it filters out re-sampled signal sources with a probability threshold; for the re-sampled signal sources, it performs directional driving and re-sampling of the multi-source sensor array to compensate for the multi-source water signals.

[0120] Furthermore, the preprocessing unit 12 performs the following steps: for the target water body type, it organizes a baseline curve template, wherein the baseline curve template contains the standard signal trend state of each sensor source; based on the baseline curve template, it performs morphological comparison of the multi-source water body signals, performs value classification, generates first processed water body data, and generates a value assessment summary, wherein the morphological features include at least trend, peak value, and duration.

[0121] The value grading levels include at least a Level 1 based on stable baseline data, a Level 2 conforming to specific trend changes, and a Level 3 for out-of-limit abnormal data; wherein Level 1 performs high-distortion compression, Level 2 performs adaptive distortion compression, and Level 3 is lossless.

[0122] Furthermore, the preprocessing unit 12 performs the following steps: dynamically negotiates the value assessment summary through a value density negotiation protocol maintained at the first edge and the second middle platform to generate a value probability threshold; and performs value-oriented screening and adaptive compression processing based on value grading on the first processed water data according to the value probability threshold to determine the second processed water data.

[0123] Furthermore, the preprocessing unit 12 performs the following steps:

[0124] Through the entropy reduction gateway, multi-dimensional projection based on the water system is performed on the second processed water data to determine the water projection state; for the water projection state, the underlying state vector is deduced by reverse inference, wherein the entropy reduction processing target is the direction of data volume reduction and the direction of information value increase; the underlying state vector is used as the water preprocessing result.

[0125] Furthermore, the transmission unit 13 performs the following steps: encapsulates the water pretreatment result, generates a first priority condition according to the encapsulation mode, wherein the encapsulation mode is a heartbeat packet mode or an information ingot mode, and when there is three-level value data, it is encapsulated as an information ingot mode; generates a second channel condition based on data security level, communication power consumption and transmission bandwidth; and determines data transmission conditions according to the first priority condition and the second channel condition.

[0126] Furthermore, the transmission unit 13 performs the following steps: determining the interaction network between the first edge terminal and the second middle platform; performing matching based on real-time channel status in the interaction network according to the data transmission conditions to determine the target interaction thread; and transmitting the water pretreatment result back from the first edge terminal to the second middle platform according to the target interaction thread.

[0127] Through the foregoing detailed description of the edge computing-based micro-water analysis data preprocessing and transmission method, those skilled in the art can clearly understand the edge computing-based micro-water analysis data preprocessing and transmission method and system in this embodiment. As for the apparatus disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and relevant parts can be referred to in the method section.

[0128] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for preprocessing and transmitting trace water analysis data based on edge computing, characterized in that, The method includes: A sampler is deployed at the first edge to guide a multi-source sensor array to collect multi-source water signals, wherein the sampler guides multi-source water sampling with minimal uncertainty and risk control guidance; The multi-source water body signals are preprocessed to generate water body preprocessing results. The preprocessing steps include a first value classification processing based on signal morphology comparison, a second screening and compression processing based on a value density negotiation protocol, and a third systematic entropy reduction processing based on an entropy reduction gateway. The value density negotiation protocol refers to a two-way interaction mechanism between the edge and the middle platform, used to evaluate the value density of different water body signals in the overall data stream in real time. The value density is the amount of information value carried by a unit of data. The water pretreatment results are encapsulated at the sending interface at the first edge, and the data transmission conditions are determined according to the encapsulation mode to perform communication backhaul based on the second middle platform.

2. The method as described in claim 1, characterized in that, The construction of the sampler, prior to its deployment at the first edge, includes: Based on the first uncertainty, a first sampling decision node is deployed, wherein the first sampling decision node outputs a first uncertainty probability field; Based on the second risk control trend prediction, a second sampling decision node is deployed, wherein the second sampling decision node outputs a second risk control probability field; Establish a probability weighting channel between the first sampling decision node and the second sampling decision node, generate the sampler, and embed the sampler into the first edge terminal.

3. The method as described in claim 2, characterized in that, Instructing a multi-source sensor array to acquire multi-source water body signals, including: According to the preset sampling period, drive the multi-source sensor array to perform one step of standardized sampling to determine the multi-source water body signal; The multi-source water signals are imported into the sampler, and the uncertain probability field analysis based on the first sampling decision node and the risk control probability field analysis based on the second sampling decision node are performed in parallel. By performing probability weighting calculation, the re-collected signal sources with a probability threshold are screened. For the re-collected signal source, the multi-source sensor array is directionally driven and re-collected to compensate for the multi-source water body signal.

4. The method as described in claim 1, characterized in that, The preprocessing of the multi-source water body signals includes a first value classification process based on signal morphological comparison, comprising: For the target water body type, a baseline curve template is prepared, wherein the baseline curve template contains the standard signal trend state of each sensor source; Based on the baseline curve template, the multi-source water body signals are morphologically compared, value is graded, first-processed water body data is generated, and a value assessment summary is generated, wherein the morphological features include at least trend, peak value, and duration.

5. The method as described in claim 4, characterized in that, A value grading system is established, wherein the value grading system includes at least a Level 1 based on stable baseline data, a Level 2 conforming to specific trend changes, and a Level 3 for out-of-limit abnormal data, wherein the specific trend changes indicate that the data, although not exceeding the limit, exhibits a certain deviation. The first level performs high-distortion compression, the second level performs adaptive distortion compression, and the third level is lossless.

6. The method as described in claim 5, characterized in that, The second screening and compression process based on the value density negotiation protocol includes: By using the value density negotiation protocol maintained at the first edge and the second middle platform, dynamic value negotiation is performed on the value assessment summary to generate a value probability threshold. Based on the value probability threshold, the first processed water body data is subjected to value-oriented screening and adaptive compression based on value grading to determine the second processed water body data.

7. The method as described in claim 6, characterized in that, The preprocessing of the multi-source water body signals includes a systematic entropy reduction process for a third water body system based on an entropy reduction gateway, comprising: By using an entropy reduction gateway, a multi-dimensional projection based on the water system is performed on the second processed water data to determine the water projection state. For the water body projection state, the underlying state vector is deduced by reverse inference, with the entropy reduction processing target being the reduction of data volume and the increase of information value. The underlying state vector is used as the preprocessing result of the water body.

8. The method as described in claim 1, characterized in that, Determine the data transmission conditions, including: The water pretreatment results are encapsulated, and a first priority condition is generated according to the encapsulation mode. The encapsulation mode is either a heartbeat packet mode or an information ingot mode. When there is tertiary value data, it is encapsulated as an information ingot mode. The second channel conditions are generated based on data security level, communication power consumption, and transmission bandwidth. The data transmission conditions are determined based on the first priority condition and the second channel condition.

9. The method as described in claim 8, characterized in that, Determine the interaction network between the first edge terminal and the second middle platform; Based on the data transmission conditions, a matching process based on real-time channel status is performed in the interactive network to determine the target interactive thread; According to the target interaction thread, the water pretreatment results are transmitted back from the first edge terminal to the second middle platform.

10. A data preprocessing and transmission system for trace water analysis based on edge computing, characterized in that, The system is used to perform the edge computing-based micro-water analysis data preprocessing and transmission method as described in any one of claims 1-9, the system comprising: A water sampling unit is used to deploy a sampler at the first edge end to guide a multi-source sensor array to collect multi-source water signals, wherein the sampler guides multi-source water sampling with minimal uncertainty and risk control guidance. The preprocessing unit is used to preprocess the multi-source water body signals and generate water body preprocessing results. The preprocessing steps include a first value classification processing based on signal morphology comparison, a second screening and compression processing based on a value density negotiation protocol, and a third systematic entropy reduction processing based on an entropy reduction gateway. The value density negotiation protocol refers to a two-way interaction mechanism between the edge terminal and the middle platform, which is used to evaluate the value density of different water body signals in the overall data stream in real time. The value density is the amount of information value carried by a unit of data. The transmission unit is used to encapsulate the water pretreatment results at the sending interface of the first edge end, determine the data transmission conditions according to the encapsulation mode, and perform communication backhaul based on the second middle platform.

Images