An environmental monitoring method and system based on an environmental monitoring instrument

Abstract

This application relates to the field of environmental monitoring technology, specifically to an environmental monitoring method and system based on an environmental monitoring instrument. The method includes: acquiring a first concentration sequence of a target pollutant gas in the environment to be monitored; performing multi-scale decomposition on the portion of the first concentration sequence within a preset sliding window to obtain component sequences at multiple scales; determining the change pattern of the first concentration sequence based on the amplitude characteristics of each scale component, and determining the next sampling interval of the sensor based on the change pattern; determining a sparse sampling interval, reconstructing the concentration sequence within the sparse sampling interval to obtain a second concentration sequence, and determining the monitoring result for the target pollutant gas. This application identifies the pattern type of concentration change through multi-scale decomposition, matching the sampling interval with the essential characteristics of the change, and restores the temporal resolution of the sparse interval through reconstruction, thus reducing power consumption while ensuring the effectiveness of the monitoring data.

Description

Technical Field

[0001] This application relates to the field of environmental monitoring technology, and in particular to an environmental monitoring method and system based on an environmental monitoring instrument. Background Technology

[0002] Environmental monitoring instruments used to continuously monitor the concentration of target polluting gases are typically equipped with gas sensors. They periodically collect gas concentration data in the environment at a certain sampling frequency, and analyze and judge based on the collected concentration sequence to determine whether there is any pollution anomaly in the monitored environment.

[0003] In practical applications, the higher the sampling frequency, the higher the temporal resolution and the stronger the ability to capture rapid changes in the environment, but the corresponding power consumption is also greater; the lower the sampling frequency, the lower the power consumption, but critical events occurring between two samplings may be missed. This contradiction is particularly prominent for battery-powered portable or remotely deployed environmental monitoring instruments. Summary of the Invention

[0004] In view of this, the purpose of this application is to provide an environmental monitoring method and system based on an environmental monitoring instrument, solving the problem of the trade-off between power consumption and time resolution caused by fixed-frequency sampling in the prior art. The specific solution is as follows: In a first aspect, this application provides an environmental monitoring method based on an environmental monitoring instrument, comprising the following steps: The first concentration sequence of the target pollutant gas in the environment to be monitored is obtained through the sensors of the environmental monitoring instrument; The portion of the first concentration sequence within a preset sliding window is decomposed into multiple scales to obtain component sequences at multiple scales. Based on the amplitude characteristics of each scale component, the change pattern of the first concentration sequence is determined, and the next sampling interval of the sensor is determined based on the change pattern. The time interval between adjacent sampling points in the first concentration sequence where the sampling interval is greater than the preset standard interval is determined as the sparse sampling interval. The concentration sequence within the sparse sampling interval is reconstructed to obtain the second concentration sequence. Based on the second concentration sequence, the monitoring results for the target pollutant gas are determined.

[0005] Optionally, the multi-scale decomposition is wavelet decomposition; the component sequences at multiple scales include at least one high-frequency component sequence and at least one low-frequency component sequence.

[0006] Optionally, the variation pattern of the first concentration sequence is determined based on the amplitude characteristics of each scale component, including: Calculate the rate of change of the amplitude statistics of the high-frequency component sequence within the preset sliding window, and the rate of change of the amplitude statistics of the low-frequency component sequence within the preset sliding window; The change pattern is determined based on the comparison between the rate of change of the amplitude statistics of the high-frequency component sequence and the first preset threshold, and the comparison between the rate of change of the amplitude statistics of the low-frequency component sequence and the second preset threshold.

[0007] Optionally, the variation mode includes a stationary mode and a non-stationary mode; determining the next sampling interval of the sensor based on the variation mode includes: in the stationary mode, determining the next sampling interval as a basic sampling interval; in the non-stationary mode, determining the next sampling interval from a preset plurality of sampling intervals according to the rate of change of the amplitude statistics of the high-frequency component sequence.

[0008] Optionally, the upper and lower limits of the next sampling interval are determined based on the total number of samplings of the first concentration sequence and the number of samplings already performed.

[0009] Optionally, when the rate of change of the amplitude statistics of the high-frequency component sequence exceeds a preset trigger threshold, the next sampling interval is set to a preset minimum sampling interval, and the sensor is triggered to perform sampling.

[0010] Optionally, the first preset threshold and the second preset threshold are adjusted based on the deviation between the second concentration sequence and the actual sampling values ​​of the sensor at both ends of the sparse sampling interval.

[0011] Optionally, the concentration sequence within the sparse sampling interval is reconstructed to obtain a second concentration sequence, including: The concentration sequence within the sparse sampling interval is sparsely represented in a preset sparse transform domain to obtain sparse representation coefficients. The sparse representation coefficients are optimized under the norm minimization constraint to obtain the concentration reconstruction values ​​at each time point within the sparse sampling interval, forming the second concentration sequence.

[0012] Optionally, determining the monitoring results for the target pollutant gas based on the second concentration sequence includes: Based on the concentration values ​​at each time point in the second concentration sequence, the cumulative exposure to the target pollutant gas within the sparse sampling interval is calculated. When the cumulative exposure exceeds a preset exposure threshold, an abnormality alert is output for the monitored environment within the time period corresponding to the sparse sampling interval.

[0013] Secondly, this application provides an environmental monitoring system based on an environmental monitoring instrument to implement the above method, including the following modules: The sensor module is used to acquire the first concentration sequence of the target pollutant gas in the environment to be monitored; The sampling control module is used to perform multi-scale decomposition on the portion of the first concentration sequence within a preset sliding window to obtain component sequences at multiple scales, determine the change pattern of the first concentration sequence based on the amplitude characteristics of each scale component, and determine the next sampling interval of the sensor module based on the change pattern. The reconstruction module is used to determine the time interval between adjacent sampling points in the first concentration sequence where the sampling interval is greater than a preset standard interval as a sparse sampling interval, and to reconstruct the concentration sequence within the sparse sampling interval to obtain a second concentration sequence. The output module is used to determine the monitoring results of the target pollutant gas based on the second concentration sequence.

[0014] Compared with the prior art, this application has the following advantages: Multi-scale decomposition distinguishes changes in concentration sequences into components at different time scales, allowing the judgment of change patterns to be based on the compositional structure of changes at different scales, rather than a general overall change amplitude. High-frequency components correspond to rapid fluctuations at short time scales, while low-frequency components correspond to trend evolution at long time scales. Different combinations of their amplitude characteristics correspond to different environmental change patterns, enabling the sampling interval decision to match the essential characteristics of the change, avoiding over-response to non-substantial fluctuations and lag in response to trend changes.

[0015] By reconstructing data from sparse sampling intervals, a second concentration sequence with full temporal resolution is obtained. This ensures that the power savings from reducing the sampling frequency do not come at the expense of data continuity, thus achieving an effective balance between power consumption and monitoring accuracy. Based on the second concentration sequence, monitoring indicators dependent on continuous concentration time histories can be further calculated, expanding the dimensionality and practicality of the monitoring results. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of this application. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 A flowchart illustrating an environmental monitoring method based on an environmental monitoring instrument, provided for an embodiment of this application; Figure 2 This is a schematic diagram of an environmental monitoring system based on an environmental monitoring instrument, provided as an embodiment of this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] Reference Figure 1 This application provides an environmental monitoring method based on an environmental monitoring instrument, which includes the following steps: S1) Obtain the first concentration sequence of the target pollutant gas in the environment to be monitored through the sensors of the environmental monitoring instrument; S2) Perform multi-scale decomposition on the portion of the first concentration sequence within a preset sliding window to obtain component sequences at multiple scales; S3) Based on the amplitude characteristics of each scale component, determine the change pattern of the first concentration sequence, and determine the next sampling interval of the sensor based on the change pattern; S4) The time interval between adjacent sampling points in the first concentration sequence where the sampling interval is greater than the preset standard interval is determined as the sparse sampling interval. The concentration sequence in the sparse sampling interval is reconstructed to obtain the second concentration sequence. S5) Based on the second concentration sequence, determine the monitoring results for the target pollutant gas.

[0020] Specifically, in step S1, the sensors of the environmental monitoring instrument continuously collect the concentration of the target pollutant gas in the monitored environment, forming a first concentration sequence arranged chronologically. The sensors may include electrochemical sensors, semiconductor sensors, or optical sensors, etc., with the specific type selected based on the properties of the target pollutant gas. For ease of subsequent processing, the first concentration sequence is denoted as... ,in Indicates the first Concentration values ​​at each sampling time.

[0021] Specifically, after obtaining the first concentration sequence, in step S2, the portion of the first concentration sequence within a preset sliding window is decomposed into multiple scales. The length of the preset sliding window is... That is, each time the nearest concentration is selected from the first concentration sequence. The analysis is performed on actual sampled data points. It should be noted that the preset sliding window only uses the actual sampled values ​​from the first concentration sequence, excluding the reconstructed concentration values ​​from the second concentration sequence, to avoid interference from potential artificial signal features in the reconstructed values ​​that could affect the results of multi-scale decomposition and the judgment of subsequent change patterns. The preset sliding window's sliding step size is one actual sampled data point; that is, for each new actual sampled data point obtained, the window slides forward one data point and triggers one multi-scale decomposition. During periods in stationary mode where the sampling interval is set to the basic sampling interval, no new actual sampled data points are generated between two samples; therefore, multi-scale decomposition is only performed once after each new sample arrives.

[0022] Since the preset sliding window always contains a fixed number of actual sampled data points, the physical time length covered by the window changes as the sampling interval changes. In stationary mode, the sampling interval is larger, and the physical time covered by the window is longer, allowing multi-scale decomposition to capture the trend changes in concentration over a longer time range. In non-stationary mode, the sampling interval is shortened, and the physical time covered by the window is correspondingly shortened, allowing multi-scale decomposition to focus on recent rapid changes. This adaptive change in the physical time length of the window is consistent with the focus of the sampling strategy in different modes. Wavelet decomposition is performed on the index dimension of the data points, and the frequency corresponding to each scale component is the index domain frequency. When the sampling interval is non-uniform, the mapping relationship between the index domain frequency and the physical time domain frequency changes with the sampling interval, but this change does not affect the determination of the change mode. The determination of the change mode is based on the relative rate of change of the amplitude statistics of each scale component between the current window and the previous window. The sampling intervals of two adjacent windows are basically the same (the windows differ by only one new sampling point), so the calculation of the rate of change is performed under the same physical scale reference.

[0023] The purpose of multi-scale decomposition is to separate the components of change at different time scales in a concentration sequence. This is because some changes occur at shorter time scales, such as rapid fluctuations caused by local turbulence or transient sensor noise; while other changes occur at longer time scales, such as trend changes caused by variations in pollution source intensity or evolution of meteorological conditions. By separating the components at different scales, we can grasp the dynamic characteristics of the current environment from the compositional structure of the changes rather than the overall magnitude.

[0024] In this embodiment, wavelet decomposition is used for multi-scale decomposition. Wavelet decomposition projects the original sequence onto different scales using a set of wavelet basis functions to obtain component sequences corresponding to each scale. The db4 wavelet (Daubechies4 wavelet) is selected as the wavelet basis function, which achieves a good balance between smoothness and tight support, making it suitable for separating components with varying degrees of smoothness in concentration sequences. Number of decomposition levels Determine by the following formula: After wavelet decomposition, the high-frequency component sequence of the first layer is taken as the high-frequency component sequence. Take the first The low-frequency component sequence of the highest layer is used as the low-frequency component sequence. By using a fixed number of layers (layer 1) to select high-frequency components, it is ensured that the high-frequency component sequences correspond to the same time scale range at different times, thus making subsequent calculations based on the rate of change of their amplitude statistics more accurate. The pattern determination and sampling interval matching have a unified physical basis. In other embodiments, the high-frequency component sequences of the second or third layer can also be taken, but they should be consistent with the preset sliding window length. and preset standard interval Matching ensures that the frequency range corresponding to the selected high-frequency component can effectively reflect the rapidly fluctuating components in the concentration sequence that need to be responded to in a timely manner.

[0025] Finally, the low-frequency component sequence was obtained. and high-frequency component sequences .in, and The values ​​represent the lengths of the low-frequency component sequence and the high-frequency component sequence, respectively. The low-frequency component sequence reflects the overall trend of concentration changes within the preset sliding window, and its fluctuation period is relatively long. The high-frequency component sequence reflects the detailed fluctuations of concentration within the preset sliding window, and its fluctuation period is relatively short.

[0026] Specifically, in step S3, the change pattern of the first concentration sequence is determined based on the amplitude characteristics of each scale component, and the next sampling interval of the sensor is determined based on the change pattern. This can be further subdivided into the following processing steps.

[0027] First, for the high-frequency and low-frequency component sequences obtained through multi-scale decomposition, their amplitude statistics are calculated respectively. Amplitude statistics characterize the overall amplitude level of the component sequence within a preset sliding window; in this embodiment, the root mean square value is used. Amplitude statistics of the high-frequency component sequence... Amplitude statistics of low-frequency component sequences Calculate them separately according to the following formulas, where The first high-frequency component sequence The value of each data point The first low-frequency component in the sequence Values ​​of each data point: Then, the rate of change of the amplitude statistics of the high-frequency component sequence within a preset sliding window is calculated. And the rate of change of the amplitude statistics of the low-frequency component sequence within a preset sliding window. The rate of change is used to measure the degree of change in the amplitude statistics between the current preset sliding window and the previous preset sliding window. Let... and These are the high-frequency amplitude statistics and low-frequency amplitude statistics corresponding to the current preset sliding window, respectively. and Given the corresponding values ​​of the previous preset sliding window, the rate of change is calculated using the following formula: After obtaining the above rate of change, the high-frequency rate of change will be... With the first preset threshold Comparison, using low-frequency change rates With the second preset threshold A comparison is made to determine the change pattern to which the current concentration sequence belongs. In this embodiment, the change patterns include stationary and non-stationary patterns. When and When the current concentration sequence is in a stable state, both the low-frequency trend and high-frequency fluctuations of concentration changes are relatively stable, and the environmental conditions have not changed substantially; otherwise, i.e. or When the current concentration sequence is determined to be in a non-stationary mode, it indicates that the environment is changing, such as changes in pollution sources or sudden changes in meteorological conditions.

[0028] Among them, the first preset threshold Second preset threshold The initial value can be determined in the following way. During the initial operation phase after the environmental monitoring instrument is deployed, the sensor continuously samples at a preset standard interval. After an initial sampling period (the length of this initial sampling period is not less than the preset sliding window length), the initial value is determined. (twice as the initial acquisition period), calculate the value corresponding to each window slide within this initial acquisition period. Value and The values ​​were obtained respectively. sequence sum sequence. Will All values ​​greater than 0 in the sequence are sorted in ascending order, and the 90th percentile is taken as... Candidate values; All values ​​greater than 0 in the sequence are sorted in ascending order, and the 90th percentile is taken as... The candidate values. If The number of values ​​greater than 0 in the sequence is less than 10, or the selected candidate value is less than the preset minimum threshold. Then As The initial value; Similarly, a preset minimum threshold is set. A small positive number, set empirically, is used to prevent the initial threshold from being too small and causing the system to be overly sensitive. Using the 90th percentile means that in the initial stage, only the approximately 10% of cases with the highest rate of change are considered non-stationary, balancing the system's initial sensitivity to environmental changes and avoiding frequent entry into non-stationary modes due to normal fluctuations. After obtaining the initial value, as the system continues to run, and It will be dynamically updated according to the threshold adaptive adjustment mechanism.

[0029] Based on the identified change pattern, the next sampling interval of the sensor is determined according to the following strategy. In stationary mode, the environmental conditions are stable, and high-frequency sampling is unnecessary; therefore, the next sampling interval is determined as the base sampling interval. The base sampling interval is a relatively large sampling interval, which can be set to several times the preset standard interval to significantly reduce the sampling frequency and power consumption during stable periods. In non-stationary mode, the environment is changing, and the sampling interval needs to be determined according to the degree of change.

[0030] Preset a set consisting of multiple sampling intervals. ,in ; Will With preset trigger threshold The intervals between them are divided into geometric sequences. There are 10 sub-intervals, among which , The common ratio; The boundary is defined as ,in , ; thus, The sub-intervals are as follows: , respectively corresponding to sampling intervals .

[0031] when When, the sampling interval is taken ;when When, the sampling interval is taken ; And so on, The larger the value, the smaller the corresponding sampling interval. Using this method, low-frequency components are used to determine whether the environment has entered a non-stationary mode, while high-frequency components are used to further quantify the severity of changes in the non-stationary mode and match the corresponding sampling interval. This ensures that the sampling strategy remains sensitive to trend changes while providing differentiated responses to changes of varying severity.

[0032] After determining the next sampling interval, the sensor performs subsequent sampling at this interval, which may change the time interval between adjacent sampling points in the first concentration sequence. When the sampling interval between two adjacent samples is greater than the preset standard interval, it indicates that the sensor did not perform sampling at the normal frequency during this period, and there is missing concentration data for that period.

[0033] Specifically, in step S4, the time interval between adjacent sampling points in the first concentration sequence where the sampling interval is greater than a preset standard interval is determined as a sparse sampling interval. For example, the preset standard interval is... In the The second sampling and the first Between each sampling, the sampling interval is .like Then take the first The sampling time is the starting point, the... The time period defined by the sampling time point as the endpoint constitutes a sparse sampling interval. Within this sparse sampling interval, actual sampled values ​​are only present at the two ends; concentration values ​​at intermediate time points are not collected by the sensor. To recover these missing concentration data, the concentration sequence within the sparse sampling interval is reconstructed.

[0034] The reconstruction process begins with sparse representation. Let the total number of time points within the sparse sampling interval, divided according to a preset standard interval, be . Its value is determined as follows Construct the vector to be reconstructed. ,in and The endpoints represent known actual sampled values, while the remaining elements are unknowns to be determined. Sparse representation is performed in a preset sparse transform domain to obtain sparse representation coefficients. In this embodiment, the preset sparse transform domain is the discrete cosine transform domain, and the dictionary matrix composed of DCT basis functions is used as the synthesis matrix, denoted as . , its first Line 1 Column elements Defined by the following formula: when When, the coefficient is .

[0035] The sparse representation coefficient vector is denoted as Vector to be reconstructed With sparse representation coefficients They satisfy the composition relation .

[0036] Correspondingly, the forward discrete cosine transform (analysis) is derived from... Transpose implementation: dictionary matrix Each column corresponds to a DCT basis function of a different frequency, exhibiting good energy value clustering characteristics. It can concentrate most of the energy values ​​of the natural signal on a few low-frequency coefficients, making the transformed coefficient vector... It has good sparsity.

[0037] Then, optimization is performed. An observation vector is established. Known within the sparse sampling interval It consists of several actual sampled values. For example, that is Measurement matrix Used from the vector to be reconstructed Extracting the elements corresponding to known sampling points is constructed as follows: The first row has a value of 1 in the first column and 0 in all other columns; the second row has a value of 0 in the first column. One column has a value of 1, and the rest have values ​​of 0. Therefore, the relationship between the observed vector and the vector to be reconstructed is: Will Substituting into the above equation, we get . This is an underdetermined equation with infinitely many solutions. However, due to the unavoidable measurement noise in the actual sampled values ​​from the sensor, strict equality constraints would be necessary. Requiring the reconstruction result to be exactly equal to the actual sampled value at the endpoints would force measurement noise into the reconstruction result, affecting the reconstruction accuracy. Therefore, this embodiment relaxes the equality constraint into an inequality constraint and introduces a noise margin. The optimization problem can be expressed as: in , Determined based on the sensor's nominal measurement error, for example, taking... ,in Let be the standard deviation of the measurement noise of the sensor under calibration conditions. This relaxation constraint allows for a controlled deviation between the reconstruction results and the actual sampled values ​​at the endpoints, enabling the reconstruction process to smooth out measurement noise to some extent, rather than precisely fitting noisy data.

[0038] The above optimization problem is a basis pursuit denoising model, which can be solved using one of the following two approaches.

[0039] Orthogonal matching pursuit algorithm. The above... The norm-constrained problem is transformed into a solution under sparsity constraints. First, the residuals are initialized as observation vectors. The support set is initialized to an empty set. In each iteration, the current residual and the perception matrix are calculated. The absolute value of the inner product of each column is used to select the index corresponding to the column with the largest absolute value of the inner product and add it to the support set. Then, the least squares method is used to update the sparse coefficient estimates in the subspace spanned by the support set, and the residuals are updated. The iteration terminates when either of the following conditions is met: the size of the support set reaches its upper limit. In this embodiment, we take That is, the support set size should not exceed half the number of known observations and the length of the vector to be reconstructed, to prevent overfitting; or the current residual... Norm less than or equal to noise margin This indicates that the reconstruction results have met the accuracy requirements. The optimal coefficient vector is obtained after the iteration terminates. .

[0040] The LASSO algorithm transforms the constrained optimization problem into an unconstrained LASSO form: in This is a regularization parameter that controls the trade-off between data fidelity and sparsity. Selection and noise tolerance Relatedly, based on the correspondence in compressed sensing theory, take... In practical solutions, the coordinate descent method or the alternating direction multiplier method can be used to iteratively solve this unconstrained optimization problem to obtain the optimal coefficient vector. .

[0041] It should be noted that Path 1 (OMP) is suitable for a large number of observations. Smaller (as in this embodiment) In scenarios where the first path is sufficient, the second path (LASSO) is computationally efficient and easy to implement; the third path (LASSO) is suitable for scenarios with a large number of observations or higher requirements for reconstruction accuracy. In other embodiments, the basis pursuit algorithm can also be used for solving the problem. Then, the estimated value of the vector to be reconstructed is obtained through the composition relation: Each element in the sequence represents the reconstructed concentration value at each time point within the sparse sampling interval. These reconstructed concentration values ​​are then merged with the portion of the first concentration sequence not identified as a sparse sampling interval to form the second concentration sequence. During periods sampled at a preset standard interval or smaller, the second concentration sequence equals the actual sampled values ​​in the first concentration sequence; within sparse sampling intervals, the second concentration sequence is filled with the reconstructed concentration values ​​at each time point. Thus, the second concentration sequence maintains the overall temporal resolution corresponding to the preset standard interval.

[0042] Specifically, after data reconstruction is completed, in step S5, the monitoring results for the target pollutant gas are determined based on the second concentration sequence. The cumulative exposure to the target pollutant gas within the sparse sampling interval is calculated using the second concentration sequence. Let... and These represent the start and end times of the sparse sampling interval, respectively. The time in the second concentration sequence Corresponding concentration value, cumulative exposure Calculate using the following formula: Cumulative exposure reflects the total dose of contaminants received by personnel or equipment in the monitored environment within the time period corresponding to a sparse sampling interval. Even if the concentration value at a single time point does not exceed a preset concentration threshold, cumulative exposure over a longer period may still pose a health hazard. When the calculated cumulative exposure... Exceeding the preset exposure threshold When an anomaly occurs, an alert is output for the time period corresponding to the sparse sampling interval. The alert may include the start and end times of the time period and the cumulative exposure value. The calculation of this monitoring indicator depends on the continuous time resolution provided by the second concentration sequence. If only the actual sampled values ​​at both ends of the sparse sampling interval are available, the concentration integral at each time point within the interval cannot be accurately calculated.

[0043] In some specific implementations, a sampling count constraint mechanism is also included. This mechanism is used to impose boundary constraints on the next sampling interval determined in step S3, ensuring that the adaptive sampling decision driven by the changing mode does not exceed the range allowed by the overall power consumption budget of the monitoring task.

[0044] If the total duration of the monitoring task is The maximum number of samples that the sensor can perform when fully charged is As of now, the number of samples that have been performed is recorded as follows. The elapsed task time is recorded as Then the remaining number of samples can be obtained. and remaining task time ; Based on the above variables, determine the upper limit of the next sampling interval. and lower limit value Upper limit The sampling interval should not be too large. Its purpose is to prevent the number of data points collected in the remaining task time from being too small due to the interval being too wide, which would affect the continuity of the monitoring data. It is determined by both the remaining task time and the remaining number of samples: In the formula, This represents the minimum number of sampling attempts that should be performed within the preset remaining task time, and can take the value of [value missing]. , Set the preset sampling retention ratio. For example, take... This means that even if the system tends to use a larger sampling interval, at least 50% of the remaining sampling attempts must still be performed within the remaining time. It is already very small (e.g.) )hour, Degenerates to 1, at this time Ensure that at least one more sample is taken.

[0045] lower limit value The sampling interval must not be too small. Its purpose is to prevent sampling from being too frequent, so that the remaining sampling times are exhausted too early before the task ends. The remaining task time and the remaining number of samples are determined by distributing them equally.

[0046] With this setting, if the remaining task time is spent on... If sampling is done at intervals, then the remaining number of samples will be used up exactly. When When smaller, The corresponding increase reflects the tightening of sampling resources, forcing the system to lengthen the sampling interval to maintain monitoring coverage. When the sampling count is exhausted, the system can switch to the lowest power consumption mode or issue a prompt to replace the battery. The specific settings for this situation depend on the environmental monitoring equipment used in the actual implementation.

[0047] After determining the next sampling interval in step S3, it is compared with... and Comparison: If the sampling interval is greater than Then adjust it to If the sampling interval is less than Then adjust it to If the sampling interval falls within this range, the original value remains unchanged. Thus, the final next sampling interval does not exceed the upper limit and is not less than the lower limit, achieving a balance between adaptive decision-making for changing modes and the overall power consumption budget.

[0048] In some implementations, an emergency triggering mechanism is also included to handle sudden and drastic changes in the environment that require the sensor to respond as quickly as possible, without being limited by the aforementioned normal adaptive decision-making process.

[0049] When the rate of change of the amplitude statistics of the high-frequency component sequence Exceeding the preset trigger threshold Regardless of whether the current mode is stationary or non-stationary, the next sampling interval is directly set to the preset shortest sampling interval. This immediately triggers the sensor to perform sampling. A preset trigger threshold is set. Greater than the first preset threshold And as mentioned above, it satisfies Preset minimum sampling interval Smaller than the set of sampling intervals minimum value This is to distinguish it from the shortest sampling interval in non-stationary mode, reflecting the highest response priority for sudden events.

[0050] In executing one After an emergency sampling at intervals, the system resumes its normal adaptive decision-making process. The first concentration sequence is updated with the new data points obtained from this emergency sampling. The system then performs the multi-scale decomposition in step S2 and the change pattern judgment in step S3 according to the normal procedure, and redetermines the next sampling interval. If... Still more If so, the emergency mechanism will be triggered again; if Falling back to The subsequent sampling interval is then determined according to the conventional strategy for stationary or non-stationary modes.

[0051] In some specific implementations, a threshold adjustment mechanism is also included. This mechanism is used to automatically optimize the first preset threshold during long-term operation. Second preset threshold This allows the sensitivity of the change pattern judgment to gradually match the actual dynamic characteristics of the environment to be monitored.

[0052] The adjustment is based on the deviation between the reconstructed value of the second concentration sequence within the sparse sampling interval and the linear interpolation formed by the actual sampled values ​​from the sensor at both ends of the sparse sampling interval. This deviation reflects the following fact: if the system classifies a certain period as a stationary mode at a certain sampling interval (and therefore uses a larger sampling interval), but the reconstructed concentration sequence afterward shows that the concentration actually changed significantly during that period, then the current... or The settings are too lenient and need to be tightened.

[0053] If the actual sampled value at the endpoint of a certain sparse sampling interval is and The reconstructed value of the second concentration sequence within this interval is (in , The linear interpolation sequence between endpoints is denoted as... ,in: Root mean square error between reconstructed values ​​and linear interpolation for: The smaller the value, the closer the concentration change within the sparse sampling interval is to linear (i.e., the change is gradual), and it is reasonable to classify this interval as a stationary mode. The larger the value, the greater the deviation of the concentration from the linear trend within that range, indicating that there may be changes that have not been captured in time.

[0054] For the recent sparse sampling intervals Take the average value, and denote it as A preset lower limit value is set. And an adjustment upper limit value ( ).when If the system's stable mode assessment has been conservative recently, maintaining a large sampling interval despite multiple instances of gradual actual changes while achieving good reconstruction results, then the threshold can be appropriately relaxed to further reduce power consumption. The first preset threshold should be set accordingly. Second preset threshold Multiply by an adjustment factor greater than 1. .when This indicates that the system's assessment of a stable mode has been too lenient recently, and an excessively large sampling interval was incorrectly used in cases of significant actual changes. This resulted in a large deviation between the reconstructed values ​​and the actual trend. In this situation, the threshold needs to be tightened to improve sensitivity to changes. The first preset threshold should be adjusted accordingly. Second preset threshold Multiply by an adjustment factor less than 1. .

[0055] The value of determines the smoothness of the adjustment. The larger the value, the smoother the adjustment, but the slower the response. A smaller threshold value results in more sensitive adjustments but may be affected by single abnormal intervals. In practical deployments, an appropriate value can be selected based on the duration of the monitoring task and the stability of the environment. Through this mechanism, the system can continuously optimize the threshold setting during long-term operation, gradually matching the sensitivity of change pattern judgment and sampling strategy with the actual dynamic characteristics of the monitored environment.

[0056] Furthermore, refer to Figure 2 This application also provides an environmental monitoring system based on an environmental monitor, including a sensor module 11, a sampling control module 12, a reconstruction module 13, and an output module 14.

[0057] The sensor module 11 is used to acquire the first concentration sequence of the target pollutant gas in the environment to be monitored. It includes at least one gas sensor and performs sampling operations according to the sampling interval determined by the sampling control module 12.

[0058] The sampling control module 12 is used to perform multi-scale decomposition on the portion of the first concentration sequence within a preset sliding window to obtain component sequences at multiple scales, determine the change pattern of the first concentration sequence based on the amplitude characteristics of each scale component, and determine the next sampling interval of the sensor module 11 based on the change pattern.

[0059] The reconstruction module 13 is used to determine the time interval between adjacent sampling points in the first concentration sequence where the sampling interval is greater than the preset standard interval as a sparse sampling interval, and to reconstruct the concentration sequence within the sparse sampling interval to obtain the second concentration sequence.

[0060] Output module 14 is used to determine the monitoring results of the target pollutant gas based on the second concentration sequence. The data flow between the above modules is as follows: sensor module 11 sends the collected first concentration sequence to sampling control module 12; sampling control module 12 feeds back the determined sampling interval to sensor module 11 to guide the next sampling, and at the same time sends the first concentration sequence to reconstruction module 13; reconstruction module 13 sends the reconstructed second concentration sequence to output module 14; output module 14 outputs the monitoring results.

[0061] The monitoring system is used to implement the aforementioned environmental monitoring method. For the specific steps of the method, please refer to the relevant content disclosed in the foregoing embodiments, which will not be repeated here.

[0062] Finally, it should be noted that in the above content, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0063] The terms “comprising,” “including,” or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase “comprising one…” does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0064] The technical solutions provided in this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. An environmental monitoring method based on an environmental monitoring instrument, characterized in that, Includes the following steps: The first concentration sequence of the target pollutant gas in the environment to be monitored is obtained through the sensors of the environmental monitoring instrument; The portion of the first concentration sequence within a preset sliding window is decomposed into multiple scales to obtain component sequences at multiple scales. Based on the amplitude characteristics of each scale component, the change pattern of the first concentration sequence is determined, and the next sampling interval of the sensor is determined based on the change pattern. The time interval between adjacent sampling points in the first concentration sequence where the sampling interval is greater than the preset standard interval is determined as the sparse sampling interval. The concentration sequence within the sparse sampling interval is reconstructed to obtain the second concentration sequence. Based on the second concentration sequence, the monitoring results for the target pollutant gas are determined.

2. The method according to claim 1, characterized in that: The multi-scale decomposition is wavelet decomposition; the component sequences at multiple scales include at least one high-frequency component sequence and at least one low-frequency component sequence.

3. The method according to claim 2, characterized in that, Determining the variation pattern of the first concentration sequence based on the amplitude characteristics of each scale component includes: Calculate the rate of change of the amplitude statistics of the high-frequency component sequence within the preset sliding window, and the rate of change of the amplitude statistics of the low-frequency component sequence within the preset sliding window; The change pattern is determined based on the comparison between the rate of change of the amplitude statistics of the high-frequency component sequence and the first preset threshold, and the comparison between the rate of change of the amplitude statistics of the low-frequency component sequence and the second preset threshold.

4. The method according to claim 3, characterized in that: The change patterns include stationary and non-stationary patterns; Determining the next sampling interval of the sensor based on the change pattern includes: In the stable mode, the next sampling interval is determined as the basic sampling interval; In the non-stationary mode, the next sampling interval is determined from a set of preset sampling intervals based on the rate of change of the amplitude statistics of the high-frequency component sequence.

5. The method according to claim 4, characterized in that, Also includes: Based on the total number of samplings of the first concentration sequence and the number of samplings already performed, the upper and lower limits of the next sampling interval are determined.

6. The method according to claim 3, characterized in that, Also includes: When the rate of change of the amplitude statistics of the high-frequency component sequence exceeds the preset trigger threshold, the next sampling interval is set to the preset shortest sampling interval, and the sensor is triggered to perform sampling.

7. The method according to claim 3, characterized in that: The first preset threshold and the second preset threshold are adjusted based on the deviation between the second concentration sequence and the actual sampling values ​​of the sensor at both ends of the sparse sampling interval.

8. The method according to claim 1, characterized in that, The concentration sequence within the sparse sampling interval is reconstructed to obtain a second concentration sequence, including: The concentration sequence within the sparse sampling interval is sparsely represented in a preset sparse transform domain to obtain sparse representation coefficients. The sparse representation coefficients are optimized under the norm minimization constraint to obtain the concentration reconstruction values ​​at each time point within the sparse sampling interval, forming the second concentration sequence.

9. The method according to claim 1, characterized in that, Determining the monitoring results for the target pollutant gas based on the second concentration sequence includes: Based on the concentration values ​​at each time point in the second concentration sequence, the cumulative exposure to the target pollutant gas within the sparse sampling interval is calculated. When the cumulative exposure exceeds a preset exposure threshold, an abnormality alert is output for the monitored environment within the time period corresponding to the sparse sampling interval.

10. An environmental monitoring system based on an environmental monitoring instrument, characterized in that, include: The sensor module is used to acquire the first concentration sequence of the target pollutant gas in the environment to be monitored; The sampling control module is used to perform multi-scale decomposition on the portion of the first concentration sequence within a preset sliding window to obtain component sequences at multiple scales, determine the change pattern of the first concentration sequence based on the amplitude characteristics of each scale component, and determine the next sampling interval of the sensor module based on the change pattern. The reconstruction module is used to determine the time interval between adjacent sampling points in the first concentration sequence where the sampling interval is greater than a preset standard interval as a sparse sampling interval, and to reconstruct the concentration sequence within the sparse sampling interval to obtain a second concentration sequence. The output module is used to determine the monitoring results of the target pollutant gas based on the second concentration sequence.

Images