A method and system for analyzing marine ecological environment based on big data

Abstract

The present application relates to the technical field of marine ecological environment analysis, and discloses a marine ecological environment analysis method and system based on big data, which comprises the following steps: obtaining environment monitoring and management object working condition time series data, determining a transmission lag time window based on flow field transmission logic; calculating cross-correlation coefficients in the window, and if they fall into a gray verification interval, generating a pseudo-random working condition control sequence orthogonal to the background main frequency; driving the management object to perform irregular actions through an administrative interface, calculating the local waveform matching degree of the environment data and the sequence; and determining responsibility and generating a punishment instruction when the threshold is exceeded. The system comprises a data acquisition and space-time mapping module, a cross-correlation calculation module, an active attribution module, and a fingerprint verification and decision module. Through the active pseudo-random disturbance and frequency domain orthogonal verification mechanism, the present application realizes exclusive locking of the responsible subject under complex hydrological environment, and solves the attribution problem that passive monitoring is difficult to distinguish between natural background fluctuations and human emissions.

Description

Technical Field

[0001] This invention relates to a method and system for analyzing marine ecological environment based on big data, belonging to the field of marine ecological environment technology. Background Technology

[0002] Currently, marine ecological environment administrative supervision and management relies on a three-dimensional monitoring network composed of satellite remote sensing, fixed-point buoys, and shore-based stations to collect water quality environmental parameters in real time and compare them with the values ​​of statutory environmental quality standards as the basis for judging environmental anomalies and triggering administrative enforcement procedures. The marine environment has strong dynamic coupling characteristics and complex system attributes with high background noise. Tidal rise and fall, thermocline fluctuations, and seasonal ocean current changes cause periodic or random oscillations in environmental monitoring indicators. Natural factors cause background fluctuations, the temporal characteristics and numerical amplitude of which are highly confused with environmental anomalies caused by human emissions. The existing static judgment logic based on absolute numerical thresholds is difficult to separate and determine causal relationships from aliased monitoring data when facing non-steady-state hydrological environments. This leads to natural fluctuations being misjudged as human responsibility, or the inability to identify specific violators under the interference of multi-source emissions.

[0003] To improve the accuracy of regulatory attribution, increasing sensor deployment density or constructing high-precision three-dimensional hydrodynamic numerical models for pollution source tracing is often employed. However, large-scale applications have limitations: the construction and maintenance costs of high-density hardware arrays increase exponentially, and it is difficult to overcome the nonlinear drift and failure of sensor data caused by marine organism attachment and seawater corrosion; complex fluid physics models suffer from difficulties in obtaining boundary conditions and lengthy computation times, failing to meet the timeliness requirements of administrative supervision in real-time blocking and rapid evidence collection for illegal emissions; in complex scenarios where multiple emission sources are spatially adjacent and have similar emission characteristics, it is difficult to construct a legally exclusive chain of evidence based solely on physical diffusion simulation conclusions, leading to evidentiary difficulties for administrative penalties. Existing methods use data analysis logic to address attribution blind spots in complex hydrological contexts. For example, Chinese invention patent application CN114659503A discloses an artificial intelligence-based marine ecological environment monitoring method that combines meteorological data with historical comparisons to identify environmental change trends and establish data correlations. However, this method is a passive monitoring method that relies on the statistical correlation of monitoring indicators rather than physical causality. Under the cover of strong periodic background noise, if the emission behavior of the controlled object overlaps with the natural hydrological cycle frequency or multiple emission sources coexist, passive logic based solely on observation and comparison cannot distinguish between natural fluctuations and human emissions from a physical perspective, making it difficult to identify the responsible party.

[0004] Therefore, the technical problem to be solved by this invention is how to utilize existing heterogeneous data resources, reduce the scale of hardware investment, decouple natural background noise and anthropogenic emission characteristics through data processing logic, and achieve accurate identification and real-time monitoring of the responsible parties for environmental anomalies under complex hydrological conditions. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A method for marine ecological environment analysis based on big data, comprising the following steps:

[0006] Acquire environmental monitoring time-series data of the target sea area and operational time-series data of the controlled objects, and determine the transmission lag time window from the controlled objects to the environmental monitoring points based on the preset flow field transmission logic;

[0007] Within the transmission lag time window, the cross-correlation coefficient between the operating condition time series data and the environmental monitoring time series data is calculated; in response to the cross-correlation coefficient falling into the preset grayscale verification interval, a pseudo-random operating condition control sequence is generated, and the frequency domain characteristics of the pseudo-random operating condition control sequence are constructed to be orthogonal to the background dominant frequency of the current environmental monitoring time series data.

[0008] The administrative management interface sends scheduling instructions containing pseudo-random operating condition control sequences to the controlled objects, and executes irregular actions in accordance with the pseudo-random operating condition control sequences so as to change the operating status of the controlled objects accordingly.

[0009] Acquire the verification environment timing data after the execution of irregular actions, and calculate the local waveform matching degree between the verification environment timing data and the pseudo-random operating condition control sequence;

[0010] When the local waveform matching degree exceeds the confirmation threshold, it is determined that the abnormal fluctuation of the environmental monitoring time series data originates from the controlled object, a responsibility association mapping for the controlled object is established and a regulatory penalty instruction is generated; among them, the pseudo-random operating condition control sequence is used to demodulate the exclusive responsible entity fingerprint from the environmental background noise in a multi-source interference environment.

[0011] Preferably, before generating regulatory penalty instructions, the method further includes performing a side-channel-based data logic consistency verification step: acquiring energy consumption time-series data of the controlled object during irregular actions, the energy consumption time-series data originating from a third-party utility metering system; calculating the logical synchronization rate between the energy consumption time-series data and the pseudo-random operating condition control sequence based on a pre-set production and pollution coupling model; and blocking the generation of regulatory penalty instructions and generating a data audit work order for the controlled object in response to the logical synchronization rate being lower than a preset self-consistency threshold, the calculation of the logical synchronization rate is based on the cross-correlation analysis between the normalized waveform of the energy consumption time-series data and the pseudo-random operating condition control sequence.

[0012] Preferably, before calculating the cross-correlation coefficient, a background noise differential removal step based on flow field topology is performed: based on flow field transport logic, a background reference point unaffected by emissions from controlled objects is selected in the spatial topology of the target sea area; background environmental data of the background reference point in the same period is obtained, and a background baseline sequence is generated by corresponding flow field transport correction; the difference sequence between the environmental monitoring time series data and the background baseline sequence is calculated; the difference sequence is used as input data for calculating the cross-correlation coefficient to filter out regional common-mode environmental noise.

[0013] Preferably, the step of establishing the responsibility association mapping adopts an iterative residual stripping logic, specifically including: based on environmental monitoring time-series data, determining the primary control object with the highest cross-correlation coefficient, and establishing a first-level responsibility association mapping; based on the operating condition time-series data of the primary control object and the corresponding cross-correlation coefficient, generating the theoretical environmental contribution sequence of the primary control object; calculating the difference between the environmental monitoring time-series data and the theoretical environmental contribution sequence, and generating an environmental anomaly residual sequence. Its computational logic satisfies: ,in, For environmental monitoring time series data, The operating sequence data for the primary controlled object. The response coefficients are determined based on the cross-correlation coefficient. As the baseline constant; the environmental anomaly residual sequence As a new input benchmark, the cross-correlation coefficient calculation steps are re-executed for all controllable objects other than the primary controllable object to identify secondary responsible entities.

[0014] Preferably, the lower limit of the grayscale verification interval is dynamically configured based on a risk-weighted sensitivity drift mechanism, specifically including: establishing a risk accumulation account for the controlled object; extracting historical cross-correlation coefficients of the controlled object when it has not triggered regulatory penalty instructions within a preset historical period, filtering out values ​​that fall into a preset observation interval and calculating the corresponding risk accumulation index; adding the risk accumulation index to the risk accumulation account, and dynamically lowering the lower limit of the grayscale verification interval based on the current balance of the risk accumulation account; wherein, the higher the current balance of the risk accumulation account, the lower the lower limit of the grayscale verification interval.

[0015] Preferably, the steps for generating a pseudo-random operating condition control sequence include: performing spectral analysis on environmental monitoring time series data to extract the main environmental background periodic frequencies; constructing a notch filter logic in the frequency domain to suppress the frequency bands corresponding to the environmental background periodic frequencies; generating a wideband pseudo-random noise sequence and applying the notch filter logic to filter the wideband pseudo-random noise sequence to obtain the pseudo-random operating condition control sequence.

[0016] Preferably, the step of determining the transmission lag time window specifically includes: acquiring gridded average flow field data of the target sea area; extracting streamline paths connecting the emission points of the controlled object and the environmental monitoring points; performing path integration on the velocity vector of the gridded average flow field data along the streamline path to calculate the theoretical transmission time; and constructing a transmission lag time window including the start time and the end time based on the theoretical transmission time and a preset diffusion coefficient.

[0017] The determination of abnormal fluctuations in environmental monitoring time series data originating from the controlled object also relies on topological consistency verification logic, specifically including: determining the verification monitoring point located downstream of the environmental monitoring point based on the geographical information of the target sea area; verifying the actual monitoring data of the verification monitoring point within the secondary lag time to determine whether it exhibits a secondary characteristic response that conforms to the theoretical attenuation amplitude; if the verification result is negative, determining that the data of the environmental monitoring point is abnormal, blocking the establishment of the responsibility association mapping, and generating an equipment calibration work order for the environmental monitoring point.

[0018] Preferably, the steps for generating regulatory penalty instructions include: obtaining the specific value of the local waveform matching degree; when the local waveform matching degree is within the first confidence interval, the generated regulatory penalty instruction includes encrypted monitoring requirements and on-site verification tasks for the controlled object; when the local waveform matching degree is within the second confidence interval higher than the first confidence interval, the generated regulatory penalty instruction includes an immediate shutdown signal or administrative penalty data packet for the controlled object.

[0019] Preferably, the method further includes defensive detection of data tampering behavior of the controlled object, specifically including: monitoring the statistical characteristics of the operating condition sequence data; in response to the operating condition sequence data showing a variance or information entropy lower than a preset threshold during the execution of irregular actions, determining that the controlled object has data modification behavior; directly generating the highest priority on-site enforcement instructions and locking the electronic supervision file of the controlled object.

[0020] A big data-based marine ecological environment analysis system includes:

[0021] The data acquisition and spatiotemporal mapping module is used to acquire environmental monitoring time-series data of the target sea area and operational time-series data of the controlled objects, and to determine the transmission lag time window from the controlled objects to the environmental monitoring points based on the preset flow field transmission logic.

[0022] The cross-correlation calculation module is used to calculate the cross-correlation coefficient between the operating condition time series data and the environmental monitoring time series data within the transmission lag time window;

[0023] The proactive attribution and rights confirmation module is used to generate a pseudo-random operating condition control sequence in response to the cross-correlation number falling into the preset grayscale verification range. The frequency domain characteristics of the pseudo-random operating condition control sequence are constructed to be orthogonal to the background main frequency of the current environmental monitoring time series data. The module also sends the scheduling instructions containing the pseudo-random operating condition control sequence to the controlled objects through the administrative management interface. The controlled objects then perform irregular actions according to the pseudo-random operating condition control sequence to change the operating status of the controlled objects accordingly.

[0024] The fingerprint verification and decision-making module is used to acquire the verification environment time-series data after the execution of irregular actions, and to calculate the local waveform matching degree between the verification environment time-series data and the pseudo-random operating condition control sequence. When the local waveform matching degree exceeds the confirmation threshold, it is determined that the abnormal fluctuation of the environmental monitoring time-series data originates from the controlled object, establishes a responsibility association mapping for the controlled object, and generates regulatory penalty instructions. Among them, the pseudo-random operating condition control sequence is used to demodulate the exclusive responsible entity fingerprint from the environmental background noise in a multi-source interference environment.

[0025] Compared with the prior art, the beneficial effects of the present invention are:

[0026] 1. In the application of big data supervision of marine ecological environment, administrative supervision orders introduce pseudo-random operating condition control sequences orthogonal to the dominant frequency of the environmental background, construct an active system identification mechanism, and utilize the physical characteristic that natural environmental fluctuations do not respond to administrative orders to transform the operating condition adjustment of the controlled object into an excitation signal with a specific feature code. The waveform matching degree between environmental monitoring data and feature code sequence within the transmission lag window is calculated, and the anthropogenic emission response component is extracted from marine environmental data with strong periodic noise. Based on excitation-response logic verification, the false correlation caused by the overlap of emission cycle and natural cycle frequency in passive monitoring is eliminated, ensuring the uniqueness of responsibility determination logic under complex hydrological conditions.

[0027] 2. To address the masking effect of strong signals on weak signals in areas with multiple emission sources, an iterative signal separation and analysis process is established. Based on cross-correlation calculation, the dominant responsible party is identified. The theoretical environmental contribution sequence is synthesized using emission behavior characteristics. The theoretical contribution component is removed from the original environmental monitoring data to generate an environmental anomaly residual sequence. The environmental anomaly residual sequence is used as a new input benchmark. Cross-correlation analysis is recursively performed to identify secondary responsible parties. By drawing on the data processing logic of serial interference elimination, the mixed and superimposed environmental anomaly signals are decoupled layer by layer into independent components corresponding to different parties, thereby enabling the identification and quantification of secondary violations hidden behind visible pollution sources.

[0028] 3. Introduce a side-channel verification mechanism based on utility metering data. Utilize the physical coupling relationship between industrial production material flow and energy flow to construct a data credibility firewall. Obtain the time-series energy consumption data of the controlled objects during the same period. Construct a theoretical emission envelope based on the production and pollution discharge coupling model. Compare it with the emission behavior characteristic sequence reported by the controlled objects. Since the energy consumption data comes from a third-party system, its information path is independent of the emission data and it corresponds physically and logically. To address the problem of logical mutual exclusion between the time-series waveforms or statistical characteristics caused by unilateral tampering of emission data, the inherent constraints of data logic can be used to eliminate the interference of untrustworthy data on subsequent analysis. Attached Figure Description

[0029] Figure 1 This invention provides a flowchart for marine environmental analysis that integrates active attribution and side-channel verification.

[0030] Figure 2 The time-series waveform evolution diagram of environmental monitoring time-series data and background baseline sequence differential removal provided by the present invention;

[0031] Figure 3 A functional architecture diagram of a marine environmental analysis system for achieving exclusive liability locking provided by the present invention;

[0032] Figure 4 This invention provides a logic decision flow diagram based on cross-correlation coefficients and waveform matching degree;

[0033] Figure 5 This invention provides a schematic diagram illustrating the data interaction and signal transmission principle between functional modules in a big data-based marine ecological environment analysis system. Detailed Implementation

[0034] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0035] This invention discloses a big data-based method for marine ecological environment analysis. By establishing a spatiotemporal mapping logic between natural environmental flow fields and human activities, and combining passive time-series cross-correlation analysis with an active frequency-domain orthogonal pseudo-random perturbation mechanism, it achieves exclusive identification of responsible parties for environmental anomalies and automated generation of regulatory instructions in complex hydrological environments. The electronic monitoring platform executes data acquisition and spatiotemporal alignment procedures, acquiring time-series environmental monitoring data for the target sea area. and the operating sequence data of each controlled object. Among them, environmental monitoring time series data Sampling sequences containing dissolved oxygen, chemical oxygen demand, or specific pollutant concentrations, with sampling frequencies set to standard frequencies that satisfy the Nyquist sampling theorem; operating condition sequence data. The platform simultaneously retrieves gridded average flow field data for the sea area, including industrial parameters characterizing emission intensity such as valve opening, pump speed, or instantaneous flow meter readings, and includes the flow velocity vectors at each spatial grid point within the sea area. Based on the flow field transport logic, the platform calculates the emission points of the controlled objects. To environmental monitoring points Transmission lag time window Specifically, the platform extracts and connects emission points. Environmental monitoring sites streamline path and along that streamline path For velocity vector Perform path integration to calculate the theoretical transmission time. Its computational logic satisfies ,in, For path differential elements, The magnitude of the velocity vector is based on the calculated theoretical transmission time. and the preset diffusion coefficient The platform constructs a transmission lag time window. Among them, the time grace period The value is based on a preset diffusion coefficient. Confirmed, and and They show a positive correlation and are used to characterize the projection width of the spatial diffusion effect of pollutants in the flow field during transport in the time domain.

[0036] To eliminate the interference of regional common-mode environmental noise on single causal determination, the platform performs a background noise differential elimination procedure based on flow field topology before performing correlation calculations. Based on the flow field direction logic, the platform selects emission points located within the controlled target area's spatial topology. Background reference points upstream or laterally, and unaffected by the emission plumes of the controlled object. The platform obtains the background environmental data of the reference point during the same period, and applies the aforementioned flow field transmission correction logic to translate and align it to the monitoring point. Time axis, generating background baseline sequence Platform computing environment monitoring time series data Compared with background baseline sequence Difference sequences , The difference sequence Filter out regional background fluctuations caused by seasonal ocean currents, large-scale red tides, or thermocline fluctuations, while retaining potential local anthropogenic emission increments; determine the transport lag time window. And obtain the differential sequence of net emission impact. Subsequently, the platform executes the delayed cross-correlation fingerprint calculation procedure, and the platform operates within the transmission delay time window. Within a limited time frame, calculate the operating condition and runtime sequence data of the controlled object. with difference sequence cross-relationships between .

[0037] During the calculation process, the platform uses runtime sequence data for operating conditions. Perform time translation, translation amount Traversing the transmission lag time window Select cross-correlation numbers for all time steps covered. The maximum value is taken as the final cross-correlation coefficient. To address the signal masking problem caused by the coexistence of multiple emission sources, the platform employs an iterative residual stripping logic to progressively identify the responsible parties. The platform is based on the original differential sequence. Determine the cross-correlation number The highest-level primary control object establishes a first-level responsibility association mapping, and the platform uses the operational sequence data of this primary control object as the basis for this mapping. Based on the corresponding cross-correlation coefficient and response lag time, the theoretical environmental contribution sequence of the primary controlled object is generated. The platform calculates the difference sequence. Contribution sequence of theoretical environment The difference is used to generate an environmental anomaly residual sequence. Its computational logic satisfies ,in, The response coefficients are determined based on the cross-correlation coefficient. As the baseline constant, this environmental anomaly residual sequence The platform represents the remaining environmental anomaly components after removing the influence of the primary responsible party, and will use the environmental anomaly residual sequence. As a new input baseline, the cross-correlation coefficient calculation steps are re-executed for all control objects except the primary control object until the secondary responsible parties are identified or the residual sequence approaches white noise.

[0038] To address the spurious correlations and high-risk critical game problems caused by the frequency overlap between natural and emission cycles, the platform introduces an active attribution and weighting mechanism based on frequency domain orthogonality. This mechanism is applied when the calculated cross-correlation coefficients... When the data falls within the preset grayscale verification range, the platform initiates active perturbation verification logic, which verifies the current environmental monitoring time series data. Perform spectrum analysis algorithms such as Fast Fourier Transform to extract the main environmental background periodic frequencies. For frequencies such as tidal frequencies or biological circadian rhythm frequencies, the platform constructs notch filter logic in the frequency domain, with its stopband center frequency set to [value missing]. The platform generates a wideband pseudo-random noise sequence and applies the notch filter logic to filter it, obtaining a pseudo-random operating condition control sequence. The pseudo-random operating condition control sequence The power spectral density at the ambient background period frequency It exhibits a minimum value at a certain point, thus achieving orthogonality with the fluctuations of the natural environment in the frequency domain.

[0039] Pseudo-random operating condition control sequence The generation follows a deterministic spectral masking procedure: the processor processes environmental monitoring time-series data. Perform a Fast Fourier Transform to extract the highest frequency point of the energy amplitude in the power spectral density map, which is then used as the environmental background period frequency. Construct an infinite impulse response notch filter, with the stopband center frequency set to... The stopband width is set to The linear feedback shift register is used to generate an m-sequence, which is then used as the original wideband noise input notch filter. The output filtered sequence is converted by a linear mapping algorithm into a 4-20mA analog current signal or Modbus bus control command that meets the input impedance requirements of the controlled actuator, forming a pseudo-random operating condition control sequence. The platform, through its administrative management interface, will include the pseudo-random operating condition control sequence. The scheduling instructions are issued to the controlled objects, driving them to execute control sequences corresponding to pseudo-random operating conditions. Irregular actions are used to cause the operating status of the controlled object to fluctuate accordingly according to the sequence. During the verification period following the execution of these irregular actions, the platform acquires time-series data of the verification environment. And calculate the time series data of the verification environment. With pseudo-random operating condition control sequence In the transmission lag time window The platform determines that the environmental anomaly originates from the controlled object when the local waveform matching degree exceeds the confirmation threshold, and generates a regulatory penalty instruction.

[0040] To prevent monitored entities from tampering with operating data to evade supervision, the platform executes a side-channel-based data logic consistency verification procedure in parallel. The platform obtains the energy consumption time-series data of the monitored entities for the same period from a third-party utility metering system. For example, regarding industrial electricity load curves, the platform calculates time-series energy consumption data based on a pre-set production and pollution discharge coupling model. Operating condition sequence data reported by the controlled objects The logical synchronization rate between the two is calculated based on the cross-correlation analysis of their normalized waveforms. The logical synchronization rate calculation follows a sliding window-based normalized cross-correlation analysis procedure: the platform acquires energy consumption time-series data. Operating condition sequence data Z-score standardization was applied to eliminate dimensional differences and map the data to a distribution interval with a mean of 0 and a variance of 1; within a window of length equal to the transmission lag time... On the time axis, calculate the Pearson correlation coefficient between the two standardized sequences; the processor traverses the transmission lag time window. For all phase offset points within the system, the maximum correlation coefficient is extracted as the logic synchronization rate. When the phase offset corresponding to the logic synchronization rate falls within the range of the pre-stored electromechanical response delay constant of the device, the logic synchronization rate is considered valid. If the logic synchronization rate is lower than the preset self-consistency threshold, the platform determines that the operating condition is not valid. If there is a risk of breach of trust, the generation of regulatory penalty instructions based on this data will be blocked, and a data audit work order for the controlled object will be automatically generated.

[0041] The platform is also equipped with a dynamic threshold configuration mechanism based on risk-weighted sensitivity drift. The platform establishes a risk accumulation account for the monitored objects. Within a preset historical period, it extracts the historical cross-correlation coefficients of the monitored objects when no regulatory penalty instructions have been triggered. If the coefficient falls within a preset observation interval, the platform calculates the corresponding risk accumulation index and adds it to the account. Based on the current balance of the risk accumulation account, the platform dynamically lowers the lower limit of the above-mentioned gray-scale verification interval and the fingerprint lock threshold. The higher the balance of the risk accumulation account, the lower the threshold, thereby improving the platform's regulatory sensitivity to long-term critical game behavior. Based on the above judgment results, the platform generates graded administrative regulatory instructions. When the judgment results such as cross-correlation coefficients or local waveform matching degree are within the first confidence interval, the platform generates an instruction data packet containing encrypted monitoring requirements and on-site verification tasks. When the judgment results are within the second confidence interval, which is higher than the first confidence interval, the platform generates an instruction containing an immediate shutdown signal or an administrative penalty data packet, and sends it to the law enforcement terminal through the e-government network.

[0042] Example 1

[0043] In monitoring scenarios of nearshore industrial parks with strong semi-diurnal tidal hydrological characteristics and multiple spatially adjacent controlled objects, electronic monitoring platforms face the challenge of attribution due to the high frequency domain overlap between natural tidal cycles and the periodic production and emission behaviors of the controlled objects. Environmental monitoring time series data When the chemical oxygen demand (COD) index shows periodic fluctuations and its peak value approaches the legal limit, the platform uses retrieved gridded average flow field data and velocity vectors. Calculate the emission points of the controlled objects To the monitoring point Transmission lag time window And simultaneously select locations at emission points Lateral background reference point unaffected by plume To generate background baseline sequences The platform performs differential operations. To filter out common-mode environmental noise caused by wide-area tidal effects, and based on this, calculate the operating condition sequence data. with difference sequence cross-correlation coefficient .

[0044] Given that the regular production cycle frequency of the controlled object at this time has a risk of phase aliasing with the residual tidal background frequency, leading to a cross-correlation coefficient... If the value falls within the grayscale verification range, the platform immediately initiates an active attribution and rights confirmation mechanism, by constructing a stopband center frequency of... The notch filter logic generates a pseudo-random operating condition control sequence orthogonal to the ambient background dominant frequency. The sequence is encapsulated as scheduling instructions and issued through the administrative management interface, driving the sewage lift pumps of the controlled objects to perform irregular frequency conversion actions. During the verification period, the platform calculates the time-series data of the verification environment. With pseudo-random operating condition control sequence The local waveform matching degree, combined with energy consumption time-series data obtained from third-party utility systems. Operating condition sequence data The logical synchronization rate verification results, after eliminating natural background interference and data tampering risks, establish a responsibility association mapping for the controlled object and generate hierarchical regulatory penalty instructions. This process transforms the original signal demixing problem that relied on passive statistical separation into an incentive response verification process based on active channel detection.

[0045] Example 2

[0046] This embodiment verifies the attribution accuracy and anti-interference capability of the analytical method of the present invention under scenarios of multi-source emission interference and data misrepresentation. The experimental platform is constructed based on a hydrodynamic numerical model and a working condition simulation unit. The hydrodynamic model uses an unstructured grid to simulate the nearshore flow field including semi-diurnal tides and random wind fields. The flow velocity is set to 0.2 m / s to 1.5 m / s. The emission sources include target source A (with random fluctuation characteristics), co-frequency interference source B (close to the tidal frequency), and random interference source C. Environmental monitoring data... The emissions from various sources are superimposed through transmission and diffusion, and Gaussian white noise with a signal-to-noise ratio of 10 dB is added. A baseline control test was performed. In the control group, which relied solely on passive cross-correlation analysis, the cross-correlation coefficient between the emission period of source B (with the same frequency) and the tidal background frequency highly overlapped with the environmental data, reaching 0.85, while that of target source A was only 0.62. This led to a misjudgment of the responsible party, triggering the active attribution and determination mechanism of this invention. The system identified the dominant background frequency as... Hz, and generate pseudo-random operating condition control sequences orthogonal in the frequency domain. Driven by source A, irregular emissions were carried out for 48 hours, and environmental time-series data were collected and verified in the experiment. The local waveform matching degree was calculated, and Table 1 lists the attribution performance comparison data under different signal-to-noise ratios and interference conditions.

[0047] Table 1: Performance Comparison Data of Attribution and Rights Confirmation Mechanisms

[0048]

[0049] As shown in Table 1, when the signal-to-noise ratio is 10dB, the confidence level of the target source attribution of the sample group of the present invention is increased from 0.62 in the unmodulated control group to 0.94, and the false alarm rate is reduced from 85.4% to 2.3%. Even in the extreme noise environment of 0dB, the confidence level is still maintained at 0.81, which verifies the anti-interference effectiveness of pseudo-random sequences in frequency domain orthogonality.

[0050] In addition, the experiment conducted side-channel verification for data misconduct scenarios. When target source A tampered with emission data, the system introduced independently measured energy consumption time-series data. The results showed that the logical synchronization rate between the two dropped from the normal 0.92 to 0.35, which was lower than the preset threshold of 0.60, successfully triggering the breach of trust blocking.

[0051] Example 3

[0052] This embodiment combines Figures 1 to 5 This section describes a method and system for analyzing the marine ecological environment based on big data, such as... Figure 1As shown, the system performs data acquisition and spatiotemporal mapping steps. Based on the flow field transmission logic, it determines the transmission lag time window and enters the background noise differential elimination stage to filter out regional common-mode environmental noise and baseline drift. On this basis, it calculates the cross-correlation coefficient between the operating conditions and environmental data within the lag time window. If the cross-correlation coefficient falls into the grayscale verification range, the system immediately starts the active attribution process and generates a pseudo-random operating condition control sequence orthogonal to the dominant frequency of the environmental background. Through the administrative interface, it drives the controlled objects to perform irregular actions and issues scheduling instructions and action execution. Then, it acquires and verifies the environmental time series data and performs local waveform matching degree calculation to verify the matching degree between the environmental data and the pseudo-random sequence. When the matching degree exceeds the threshold and the logic consistency verification of the side channel data based on the production and discharge model is passed, i.e., the logic synchronization rate is greater than the self-consistency threshold, the responsibility association mapping is finally established and the hierarchical disposal is executed to generate regulatory penalty instructions.

[0053] like Figure 2 As shown, this time-series waveform plot uses time step as the horizontal axis and concentration or difference as the vertical axis to depict the evolution trend of three key sequences. The solid line represents environmental monitoring time-series data. The dashed line represents the background baseline sequence after flow field transport correction. The two exhibit highly consistent periodic background fluctuation characteristics, while the dotted lines below represent the difference sequences generated through difference operations. It removes common-mode background noise and highlights potential anomalous incremental features; such as Figure 3 As shown, the system aims to achieve exclusive locking of responsibility for marine ecological environment. It unfolds six functional modules through a fishbone-shaped topology. The data acquisition and perception module is responsible for collecting environmental monitoring time-series data and operating condition time-series data. The flow field and noise processing module performs flow field transmission lag time window calculation and background noise differential removal. The passive correlation calculation module covers cross-correlation coefficient calculation and iterative residual stripping. The active attribution and rights confirmation module includes frequency domain orthogonal sequence generation, pseudo-random operating condition control sequence construction and irregular action execution. The defense and verification module relies on the production and discharge coupling model to perform side-channel logic consistency verification. The decision and disposal module finally performs local waveform matching degree calculation and completes the generation of regulatory penalty instructions.

[0054] like Figure 4 As shown, the process is to obtain time-series environmental monitoring data for the target marine area. and the operating sequence data of the controlled objects Starting from the input, the system first determines the transmission lag time window based on the flow field transmission logic, and calculates the cross-correlation coefficient between the two within this time window. Then, it enters the key logic branch judgment. If the cross-correlation coefficient falls into the grayscale verification range, a pseudo-random operating condition control sequence orthogonal to the main frequency of the environmental background is generated, and instructions are issued through the administrative interface to drive the execution of irregular actions. Then, the local waveform matching degree between the verification environmental data and the pseudo-random sequence is calculated. Only when the matching degree exceeds the confirmation threshold, the system determines the responsibility and generates a penalty instruction to achieve exclusive locking of the responsible party. Otherwise, it determines that the party is not responsible or terminates the process.

[0055] like Figure 5 As shown, this invention provides a marine ecological environment analysis system based on big data. In this system, the data acquisition and spatiotemporal mapping module receives time-series data input from environmental monitoring points and industrial control terminals. Based on the flow field logic, it outputs an aligned data stream to the cross-correlation calculation module. If the cross-correlation coefficient calculated within the transmission lag time window falls into the grayscale verification interval, it will trigger the active attribution and rights confirmation module to generate a frequency domain orthogonal pseudo-random sequence. This active attribution and rights confirmation module sends scheduling instructions to the execution terminal of the controlled object through the administrative management interface. The irregular action responses fed back by the execution terminal of the controlled object are captured by the fingerprint verification and decision-making module. This fingerprint verification and decision-making module calculates the local waveform matching degree through the characteristics of the transmitted pseudo-random sequence, establishes a responsibility mapping when the threshold is exceeded, and finally outputs a regulatory penalty instruction.

[0056] Example 4

[0057] In the environmental monitoring system of deep-sea aquaculture platforms, the maintenance of conventional flow field monitoring equipment is difficult and the data transmission bandwidth is limited due to the harsh sea conditions. To address the problem of calculating the transmission lag time window caused by this data blind spot, this embodiment introduces an adaptive parameter calibration procedure based on the fusion of historical flow field knowledge base and sparse observation data. The system establishes a historical flow field knowledge base for the target sea area and collects satellite remote sensing flow field inversion data, long-term buoy observation data and numerical simulation reanalysis data of the area over the past five years. The data is then indexed and classified in multiple dimensions according to season, tidal phase and wind field characteristics. When real-time flow field data is missing or insufficient in accuracy, the system acquires the current sparse observation data, including the instantaneous readings of single-point current meters and sea surface wind field remote sensing data.

[0058] Next, the system performs flow field reconstruction based on feature matching. Using the current sparse observation data as the query vector, it retrieves the historical flow field pattern with the highest matching degree from the historical flow field knowledge base. The matching algorithm employs a weighted Euclidean distance metric, where the weight of the velocity vector direction is set higher than the velocity magnitude to ensure the consistency of the flow field topology. The system selects the pattern with the highest matching degree from the previous data. A historical flow field sample, such as The system calculates a weighted average value as the reconstructed flow field data for the current moment. Based on the reconstructed flow field data, the system calibrates the transmission lag time window. The system extracts the streamline path connecting the aquaculture platform's discharge point and the downstream monitoring point, and integrates the velocity vector of the reconstructed flow field along this path to calculate the theoretical transmission time. To compensate for the uncertainty of the reconstructed flow field, the system introduces a dynamic time grace factor. Adaptive adjustment is made based on the variance of the deviation between the current sparse observation data and the reconstructed flow field: the larger the deviation, the better. The larger the value, the wider the coverage of the transmission lag time window, ensuring that the actual pollutant transmission process is included. Finally, the system uses the calibrated transmission lag time window to perform cross-correlation analysis and responsibility locking.

[0059] Example 5

[0060] In open marine monitoring scenarios where multiple emission sources are spatially densely distributed and environmental baselines are significantly affected by seasonal ocean currents, electronic monitoring platforms face the challenge of attribution difficulties, where single cross-correlation analysis is highly susceptible to false positives under common-mode noise interference. This is particularly relevant when monitoring points... When the dissolved oxygen index shows an abnormal downward trend over a certain period of time, the system, based on the flow field topology, at the monitoring points... A background reference point was selected 2 kilometers upstream of the target emission plume, which is not directly affected by the target controlled object. The system obtains the background reference point. The environmental monitoring data from the same location was used, and the time axis of the data was shifted and aligned to the monitoring point using flow field transport correction logic. At this point, a background baseline sequence is generated. The system performs difference operations. This filters out common-mode environmental noise caused by large-scale ocean current changes or regional eutrophication.

[0061] To address the signal masking issue caused by the presence of multiple controlled entities with similar emission characteristics in the sea area, the system further initiates iterative residual stripping logic, calculating the operational timing data of each controlled entity. with difference sequence The cross-correlation coefficient is used to identify the primary controlled object with the highest coefficient, and its operating condition sequence data is used as a basis. and response coefficient generation theory environmental contribution sequence The system calculates the difference sequence. Contribution sequence of theoretical environment The difference is used to obtain the environmental anomaly residual sequence. The residual sequence is then used as a new input benchmark to recalculate the cross-correlation of the remaining controlled objects. This process is repeated until the statistical characteristics of the residual sequence approach white noise, thereby achieving step-by-step locking of secondary responsible entities. In this embodiment, a simulation environment with three simulated emission sources is constructed: source A is the primary emission source, source B is the secondary emission source, and source C is the background interference source. The experimental results show that when the differential and stripping logic is not enabled, the emission signal of source B is completely masked by source A and background noise, and the attribution confidence is only 0.42, failing to trigger an alarm. However, after enabling the differential background removal and iterative stripping mechanism of this invention, the attribution confidence of source B increases to 0.89, and it is successfully identified and locked by the system. This result confirms that the method of this invention improves the ability to identify concealed illegal emission behavior in complex multi-source interference environments through signal separation and purification at the logic level.

[0062] Example 6

[0063] During the initial deployment phase of the system targeting the new sea area, to ensure the universality and accuracy of the environmental anomaly judgment logic under different hydrological backgrounds, the platform executes baseline calibration and threshold tuning procedures based on historical data. The system retrieves historical environmental monitoring data for the past complete hydrological years of the sea area and combines it with the flow field reanalysis data of the same period to construct a background noise statistical distribution model unique to the sea area. Based on this model, the system filters out pure background time periods with no known industrial emission events and calculates the operating condition sequence data within those time periods. Environmental monitoring time series data The cross-correlation coefficient sequence is used to establish the noise correlation distribution under the null hypothesis. Based on the statistical significance test logic, the system sets the lower limit of the grayscale validation interval to the noise correlation distribution. The 95th percentile sets the fingerprint locking threshold to a noise-correlation distribution. The 99th percentile of this procedure ensures that the system threshold is set not based on empirical estimates, but on the objective statistical characteristics of the local hydrological environment, thus avoiding the risk of systemic false alarms or omissions due to differences in background noise intensity.

[0064] After baseline calibration, the system further performs on-site adaptive calibration for the production and pollution discharge coupling model. The system collects historical operating time series data of the controlled objects during stable production cycles and energy consumption time series data from third-party utility systems. The least squares method is used to perform regression fitting on the two parameters, thereby extracting the characteristic coefficients of unit energy consumption and pollution generation of the specific controlled object and the baseline constant. The system calculates the self-consistency threshold based on the variance of the fitting residuals, and specifically sets the self-consistency threshold as the goodness-of-fit determination coefficient. The calibration process refines the general industry production and discharge model into a special model for specific production processes and equipment characteristics, which is 0.8 times that of the general model. This avoids logical verification errors caused by equipment aging or process differences.

[0065] Example 7

[0066] This embodiment constructs a system adaptive calibration and online fault-tolerant procedure based on digital twins. The system introduces virtual trial operation calibration logic. Before formally accessing the measured data, a high-fidelity digital twin sea area model is constructed based on the historical hydrological data and shoreline geographic information of the target sea area. In this model, the system randomly generates thousands of virtual operating condition samples with different emission characteristics such as emission intensity, frequency, duration, and environmental background such as different tidal phases and wind field conditions using the Monte Carlo method. The system uses these samples to conduct offline training and stress testing on the cross-correlation analysis module. Through the gradient descent algorithm, it automatically optimizes and determines the optimal diffusion coefficient and fingerprint locking threshold for the specific sea area, thereby completing the initial parameter calibration of the system.

[0067] Secondly, to address potential performance degradation or environmental baseline drift that may occur during long-term equipment operation, the system integrates an online health self-check mechanism. The system periodically, such as once a week, selects a quiet period without operational conditions and automatically performs a reassessment of the statistical characteristics of ambient background noise. If a noise correlation distribution of the background noise is detected... If the initial calibration value deviates, such as the p-value of the Kolmogorov-Smirnov test being less than 0.05, the system will automatically trigger the parameter retuning process, dynamically updating the upper and lower limits of the grayscale verification interval based on the latest background data. In addition, the system also monitors the signal-to-noise ratio and outlier ratio of each sensor data in real time. Once the data quality of a certain measuring point is found to be lower than the preset health threshold, the weight of that measuring point in the cross-correlation calculation will be automatically reduced or it will be temporarily removed from the monitoring network until maintenance intervention is carried out.

[0068] Example 8

[0069] This embodiment establishes a digital twin model covering the target sea area and integrates high-precision seabed topographic data, historical meteorological and hydrological data, and simulated pollution source location information. In the digital twin environment, the system runs a Monte Carlo simulation experiment, randomly generating thousands of virtual operating condition samples with different emission intensities, durations, and environmental background noise. Furthermore, for the transmission lag time window calculation module, the system automatically optimizes and determines the optimal diffusion coefficient through simulation experiments. In each simulation iteration, the system is based on the current... The predicted lag time window is calculated and compared with the actual pollutant arrival time in the simulation model. The prediction error is calculated and continuously adjusted using the gradient descent algorithm. The value is calculated until the root mean square value of the prediction error converges to a minimum.

[0070] Meanwhile, for the cross-correlation analysis module, the system uses generated virtual samples to conduct stress tests to calibrate the fingerprint locking threshold. Under different signal-to-noise ratio conditions, the system statistically analyzes the distribution differences of cross-correlation coefficients in real and false correlation samples. Based on a preset false alarm rate upper limit such as 0.1%, the system uses receiver operating characteristic (ROC) curve analysis to determine the fingerprint locking threshold that can maximize the recognition accuracy.

[0071] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0072] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for analyzing marine ecological environment based on big data, characterized in that, Includes the following steps: Acquire environmental monitoring time-series data of the target sea area and operational time-series data of the controlled objects, and determine the transmission lag time window from the controlled objects to the environmental monitoring points based on the preset flow field transmission logic; Calculate the cross-correlation coefficients between operating condition time series data and environmental monitoring time series data within the transmission lag time window; In response to the cross-correlation coefficient falling into a preset grayscale verification interval, a pseudo-random operating condition control sequence is generated. The frequency domain feature of the pseudo-random operating condition control sequence is constructed to be orthogonal to the background dominant frequency of the current environmental monitoring time series data. The lower limit of the grayscale verification interval is dynamically configured based on a risk-weighted sensitivity drift mechanism, specifically including: establishing a risk accumulation account for the controlled object; extracting the historical cross-correlation coefficients of the controlled object when no regulatory penalty instruction is triggered within a preset historical period, filtering out the values ​​that fall into a preset observation interval and calculating the corresponding risk accumulation index; adding the risk accumulation index to the risk accumulation account, and dynamically lowering the lower limit of the grayscale verification interval based on the current balance of the risk accumulation account; wherein, the higher the current balance of the risk accumulation account, the lower the lower limit of the grayscale verification interval. The administrative management interface sends scheduling instructions containing pseudo-random operating condition control sequences to the controlled objects. Irregular actions are executed according to these sequences to change the operating status of the controlled objects accordingly. The steps for generating the pseudo-random operating condition control sequences include: performing spectral analysis on environmental monitoring time-series data to extract the main environmental background periodic frequencies; constructing a notch filter logic in the frequency domain to suppress the frequency bands corresponding to the environmental background periodic frequencies; generating a wideband pseudo-random noise sequence and applying the notch filter logic to filter the wideband pseudo-random noise sequence to obtain the pseudo-random operating condition control sequence. Acquire the verification environment timing data after the execution of irregular actions, and calculate the local waveform matching degree between the verification environment timing data and the pseudo-random operating condition control sequence; When the local waveform matching degree exceeds the weighting threshold, the abnormal fluctuations in the environmental monitoring time series data are determined to originate from the controlled object. A responsibility association mapping is then established for the controlled object, and regulatory penalty instructions are generated. Specifically, a pseudo-random operating condition control sequence is used to demodulate the exclusive responsible entity fingerprint from environmental background noise under multi-source interference conditions. The step of establishing the responsibility association mapping employs an iterative residual stripping logic, including: based on the environmental monitoring time series data, determining the top controlled object with the highest cross-correlation coefficient, and establishing a first-level responsibility association mapping; based on the operating condition time series data of the top controlled object and its corresponding cross-correlation coefficient, generating the theoretical environmental contribution sequence of the top controlled object; calculating the difference between the environmental monitoring time series data and the theoretical environmental contribution sequence, and generating an environmental anomaly residual sequence. Its computational logic satisfies: ,in, For environmental monitoring time series data, The operating sequence data for the primary controlled object. The response coefficients are determined based on the cross-correlation coefficient. As the baseline constant; the environmental anomaly residual sequence As a new input benchmark, the cross-correlation coefficient calculation steps are re-executed for all controllable objects other than the primary controllable object to identify secondary responsible entities.

2. The method for analyzing marine ecological environment based on big data according to claim 1, characterized in that, Before generating regulatory penalty instructions, the process includes a side-channel-based data logic consistency verification step: acquiring energy consumption time-series data of the controlled object during irregular actions, with the energy consumption time-series data sourced from a third-party utility metering system; calculating the logical synchronization rate between the energy consumption time-series data and the pseudo-random operating condition control sequence based on a pre-set production and pollution coupling model; and blocking the generation of regulatory penalty instructions and generating a data audit work order for the controlled object in response to the logical synchronization rate falling below a preset self-consistency threshold. The calculation of the logical synchronization rate is based on the cross-correlation analysis between the normalized waveform of the energy consumption time-series data and the pseudo-random operating condition control sequence.

3. The method for analyzing marine ecological environment based on big data according to claim 1, characterized in that, Before calculating the cross-correlation coefficient, a background noise differential removal step based on the flow field topology is also included: based on the flow field transport logic, background reference points that are not affected by the emissions of the controlled objects are selected in the spatial topology of the target sea area; Background environmental data of the background reference point during the same period are obtained and corrected accordingly through flow field transmission to generate a background baseline sequence; Calculate the difference sequence between environmental monitoring time-series data and background baseline sequence; The difference sequence is used as input data for calculating the cross-correlation coefficient to filter out regional common-mode environmental noise.

4. The method for marine ecological environment analysis based on big data according to claim 1, characterized in that, The steps for determining the transmission lag time window specifically include: acquiring gridded average flow field data of the target sea area; extracting streamline paths connecting the emission points of the controlled object and the environmental monitoring points; performing path integration on the velocity vector of the gridded average flow field data along the streamline path to calculate the theoretical transmission time; and constructing a transmission lag time window that includes the start and end times based on the theoretical transmission time and a preset diffusion coefficient. The determination of abnormal fluctuations in environmental monitoring time series data originating from the controlled object also relies on topological consistency verification logic, specifically including: determining the verification monitoring point located downstream of the environmental monitoring point based on the geographical information of the target sea area; verifying the actual monitoring data of the verification monitoring point within the secondary lag time to determine whether it exhibits a secondary characteristic response that conforms to the theoretical attenuation amplitude; if the verification result is negative, determining that the data of the environmental monitoring point is abnormal, blocking the establishment of the responsibility association mapping, and generating an equipment calibration work order for the environmental monitoring point.

5. The method for marine ecological environment analysis based on big data according to claim 1, characterized in that, The steps for generating regulatory penalty instructions include: obtaining the specific value of the local waveform matching degree; when the local waveform matching degree is within the first confidence interval, the generated regulatory penalty instruction includes encrypted monitoring requirements and on-site verification tasks for the controlled object; when the local waveform matching degree is within the second confidence interval, which is higher than the first confidence interval, the generated regulatory penalty instruction includes an immediate shutdown signal or administrative penalty data packet for the controlled object.

6. The method for marine ecological environment analysis based on big data according to claim 1, characterized in that, The method also includes defensive detection of data tampering behavior of the controlled object, specifically including: monitoring the statistical characteristics of the operating condition sequence data; in response to the operating condition sequence data showing a variance or information entropy lower than a preset threshold during the execution of irregular actions, determining that the controlled object has data modification behavior; directly generating the highest priority on-site enforcement instructions and locking the electronic supervision file of the controlled object.

7. A marine ecological environment analysis system based on big data, characterized in that, The system is used to implement the method according to any one of claims 1 to 6, comprising: The data acquisition and spatiotemporal mapping module is used to acquire environmental monitoring time-series data of the target sea area and operational time-series data of the controlled objects, and to determine the transmission lag time window from the controlled objects to the environmental monitoring points based on the preset flow field transmission logic. The cross-correlation calculation module is used to calculate the cross-correlation coefficient between the operating condition time series data and the environmental monitoring time series data within the transmission lag time window; The proactive attribution and rights confirmation module is used to generate a pseudo-random operating condition control sequence in response to the cross-correlation number falling into the preset grayscale verification range. The frequency domain characteristics of the pseudo-random operating condition control sequence are constructed to be orthogonal to the background main frequency of the current environmental monitoring time series data. The module also sends the scheduling instructions containing the pseudo-random operating condition control sequence to the controlled objects through the administrative management interface. The controlled objects then perform irregular actions according to the pseudo-random operating condition control sequence to change the operating status of the controlled objects accordingly. The fingerprint verification and decision-making module is used to acquire the verification environment time-series data after the execution of irregular actions, and to calculate the local waveform matching degree between the verification environment time-series data and the pseudo-random operating condition control sequence. When the local waveform matching degree exceeds the confirmation threshold, it is determined that the abnormal fluctuation of the environmental monitoring time-series data originates from the controlled object, establishes a responsibility association mapping for the controlled object, and generates regulatory penalty instructions. Among them, the pseudo-random operating condition control sequence is used to demodulate the exclusive responsible entity fingerprint from the environmental background noise in a multi-source interference environment.

Images