IoT-based air quality monitoring methods and systems

By using singular value decomposition and classification accuracy evaluation functions in air quality monitoring, combining the distribution differences and autocorrelation of eigenvalues ​​to accurately distinguish noise signals from normal signals, the noise interference problem is solved, and the accuracy of air quality monitoring is improved. The reliability of data provides a reliable basis for air quality early warning.

CN118585756BActive Publication Date: 2024-11-15XIAN LECHI TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411060397.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-08-05
Publication Date
2024-11-15
Estimated Expiration
2044-08-05

AI Technical Summary

Technical Problem

The prior art fails to effectively consider noise data interference in air quality monitoring, resulting in noise signals in the collected data, reducing the accuracy of air quality monitoring, and the traditional singular value decomposition algorithm retains characteristic values ​​during denoising. Impacts the denoising quality.

Method used

The Internet of Things air quality monitoring method is used to process the time sequence of harmful gas concentrations through the singular value decomposition algorithm, and a classification accuracy evaluation function is constructed, combining the distribution differences and autocorrelation of the characteristic values ​​to accurately distinguish noise signals from normal signals, and The noise is removed by zeroing the eigenvalue to improve monitoring accuracy.

Benefits of technology

It effectively removes noise signals in the time sequence of harmful gas concentrations, improves the accuracy of air quality monitoring and data reliability, and provides a reliable basis for subsequent air quality warning.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118585756B_ABST
    Figure CN118585756B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of data processing technology, and in particular to an air quality monitoring method and system based on the Internet of Things. The method includes the steps of: obtaining a time series sequence of harmful gas concentration of each type at each location; performing singular value decomposition on the time series sequence of harmful gas concentration to obtain an eigenvalue set, and taking the difference between the time series sequence of harmful gas concentration before and after a eigenvalue in the eigenvalue set is set to zero as the signal time series sequence of the eigenvalue; randomly combining the eigenvalues ​​in the eigenvalue set into two categories to obtain several classification results; constructing a classification accuracy evaluation function according to the classification results, and taking the classification result corresponding to the maximum value of the classification accuracy evaluation function as the target classification result; calculating the noise conformity of each category in the target classification result; and setting the eigenvalue of the category corresponding to the maximum value of the noise conformity to zero to remove the noise of the time series sequence of harmful gas concentration to assist in air quality monitoring. The accuracy of air detection is improved by accurate denoising.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to an air quality monitoring method and system based on the Internet of Things. Background Art

[0002] Air quality affects people's health, so it is necessary to monitor the air quality in real time. Air quality assessment is usually based on air monitoring data, but when air monitoring data is collected, it means that polluted gases have already affected the air quality. Therefore, in order to prevent the impact of harmful gases on air quality, it is necessary to detect harmful gases in time before they affect the air quality. Therefore, it is necessary to estimate the emission of harmful gases and issue early warning reminders based on the estimated results.

[0003] Before estimating harmful gases, it is necessary to use sensors to collect harmful gas concentration data. Since the sensors are affected by temperature, wind speed, etc., noise signals exist in the collected data. Such noise signals are easily misjudged as abnormal harmful gas concentration data, thereby reducing the accuracy of air quality monitoring.

[0004] The Chinese patent application document with application publication number CN116735807A shows an air quality monitoring and evaluation method based on multi-sensor data. The document mainly uses cluster analysis to screen out abnormal data to judge the air quality. The method in the document does not take into account the impact of noise data interference on abnormal data screening, so the method in the document cannot eliminate the impact of noise interference on air quality monitoring.

[0005] As a commonly used denoising method, the singular value decomposition algorithm generally retains eigenvalues ​​based on human experience during denoising analysis. The retention of eigenvalues ​​will affect the denoising quality, so the traditional method of retaining eigenvalues ​​cannot achieve a good denoising effect. How to retain appropriate eigenvalues ​​becomes the research focus of this invention. Summary of the invention

[0006] In order to solve the problem of how to retain suitable characteristic values, the present invention provides an air quality monitoring method and system based on the Internet of Things.

[0007] In a first aspect, the present invention provides an air quality monitoring method based on the Internet of Things, which adopts the following technical solutions:

[0008] The air quality monitoring method based on the Internet of Things includes the following steps:

[0009] Obtain the time series of each type of harmful gas concentration at each location;

[0010] Perform singular value decomposition on the harmful gas concentration time series to obtain an eigenvalue set, and use the difference between the harmful gas concentration time series before and after an eigenvalue in the eigenvalue set is set to zero as the signal time series of the eigenvalue;

[0011] The eigenvalues ​​in the eigenvalue set are randomly combined into two categories to obtain several classification results; a classification accuracy evaluation function is constructed, wherein the classification accuracy evaluation function is positively correlated with the distribution difference of the signal time series of the eigenvalues ​​of the two categories in one classification result, and is also positively correlated with the correlation difference between the signal time series of the eigenvalues ​​of the two categories in one classification result and the time series of the concentration of harmful gases of the same type at the nearest position; and the classification result corresponding to the maximum value of the classification accuracy evaluation function is taken as the target classification result;

[0012] Calculating the noise conformity of each category in the target classification result, wherein the noise conformity is negatively correlated with the autocorrelation coefficient and the eigenvalue of the signal time series of the eigenvalue in the category;

[0013] The characteristic value of the category corresponding to the maximum noise compliance value is set to zero to remove the noise of the harmful gas concentration time series to assist in air quality monitoring.

[0014] When constructing the classification accuracy evaluation function, the present invention considers the distribution difference of the signal time series of the characteristic values ​​of the two categories in each classification result, and can better utilize the difference characteristics between the noise signal and the normal signal to evaluate the situation of separating the noise signal from the normal signal in the classification result, thereby providing a basis for obtaining a better segmentation result; further, when constructing the classification accuracy evaluation function, considering the correlation between the signal time series of the analysis characteristic value and the time series of the same type of harmful gas concentration at the nearest position, it can better utilize the characteristics that the correlation between the noise signal and the time series of the same type of harmful gas concentration at the nearest position is poor, and the correlation between the normal signal and the time series of the same type of harmful gas concentration at the nearest position is strong, to evaluate the situation of separating the noise signal from the normal signal in the classification result, thereby providing a basis for obtaining a better segmentation result; further, when calculating the noise conformity, the autocorrelation and the characteristic value are considered to accurately describe the signal regularity and content, and then distinguish the noise signal from the normal signal in the target classification result, providing a basis for subsequent accurate denoising; further, by accurately removing the noise information in the time series of harmful gas concentration, a basis is provided for subsequent air quality monitoring.

[0015] Preferably, the construction of the classification accuracy evaluation function includes:

[0016] ;

[0017] Among them, the eigenvalue in one category and the eigenvalue in another category in a classification result of the eigenvalue set are recorded as the first eigenvalue and the second eigenvalue respectively, Respectively represent The first eigenvalue and the The mean difference and variance difference of the signal time series of the second eigenvalue, denote the number of the first and second eigenvalues, respectively, Indicates The correlation between the signal time series of the first eigenvalue and the concentration time series of the same harmful gas at the nearest position, Indicates The correlation between the signal time series of the second eigenvalue and the concentration time series of the same harmful gas at the nearest position, Represents the classification accuracy evaluation function.

[0018] When constructing a classification accuracy evaluation function, the present invention introduces mean differences and variance differences to accurately reflect the distribution differences of the eigenvalue corresponding signals between the two categories in the classification results, and then accurately judges whether the classification results can reflect the difference between the noise signal and the normal signal, thereby judging the accuracy of the classification results; further, when constructing a classification accuracy evaluation function, by introducing the correlation between the eigenvalue corresponding signals and the harmful gas concentration signals at other locations, it is judged whether the classification results can accurately reflect the correlation differences between the noise signal and the normal signal and the harmful gas concentration time series at other locations, thereby judging the accuracy of the classification results.

[0019] Preferably, the calculating the noise conformity of each category in the target classification result includes:

[0020] ;

[0021] in, Indicates the first category of the target classification result. eigenvalues, Indicates the number of feature values ​​contained in this category in the target classification result. Indicates the first The autocorrelation coefficient of the signal time series of eigenvalues, Indicates the noise conformity of this category in the target classification result.

[0022] The present invention introduces autocorrelation coefficients and eigenvalues ​​to accurately reflect the regularity and signal content of signals corresponding to eigenvalues ​​in a category in the target classification result, so as to determine whether the signals corresponding to the eigenvalues ​​in the category conform to the noise.

[0023] Preferably, the step of taking the difference between the time series of the concentration of the harmful gas before and after the zeroing process of an eigenvalue in the eigenvalue set as the signal time series of the eigenvalue comprises:

[0024] After setting an eigenvalue in the eigenvalue set to zero, a singular value decomposition inverse transformation is performed based on the diagonal singular matrix, the left singular matrix and the right singular matrix formed by all the eigenvalues ​​to obtain a harmful gas concentration matrix lacking the eigenvalue;

[0025] The difference between the harmful gas concentration matrix and the harmful gas concentration matrix lacking the eigenvalue is recorded as the signal matrix of the eigenvalue;

[0026] A sequence obtained by splicing together each row of the signal matrix of the eigenvalue is recorded as the signal timing sequence of the eigenvalue.

[0027] The present invention obtains the signal time series of each characteristic value by analyzing the difference between the harmful gas concentration time series before and after zeroing, thereby decomposing the signals with different characteristics in the harmful gas concentration time series, and providing a basis for distinguishing noise signals from normal signals.

[0028] Preferably, the eigenvalues ​​in the eigenvalue set are randomly combined into two categories to obtain several classification results, including:

[0029] Set the number of variables , any eigenvalue in the set eigenvalues ​​as one class, and the remaining eigenvalues ​​as one class. Take 1 to Get an integer between Each value of corresponds to the classification result, Represents the number of eigenvalues.

[0030] The present invention classifies the characteristic values ​​without controlling the quantity and combination mode, so as to obtain a more comprehensive classification result, and prevent the classification result from being incomplete and the noise signal from being unable to be accurately distinguished from the normal signal.

[0031] Preferably, the step of performing singular value decomposition on the harmful gas concentration time series to obtain a set of eigenvalues ​​includes:

[0032] The harmful gas concentration time series is evenly divided into data segments of length L, where L represents a preset segment length, and each data segment is used as each row of the matrix to obtain a harmful gas concentration matrix;

[0033] The singular value decomposition algorithm is used to perform singular value decomposition on the harmful gas concentration matrix to obtain a diagonal singular matrix; all eigenvalues ​​on the diagonal are obtained in the diagonal singular matrix, and the set composed of all eigenvalues ​​is recorded as an eigenvalue set.

[0034] Preferably, the process of setting the characteristic value of the category corresponding to the maximum value of noise compliance to zero to remove the noise of the harmful gas concentration time series includes:

[0035] After the eigenvalues ​​of the category corresponding to the maximum noise compliance value are set to zero, the inverse transformation of singular value decomposition is performed based on the diagonal singular matrix composed of all eigenvalues ​​to obtain the denoised harmful gas concentration matrix;

[0036] Each row of the denoised harmful gas concentration matrix is ​​spliced ​​together to obtain the denoised harmful gas concentration time series.

[0037] Preferably, the auxiliary air quality monitoring comprises:

[0038] The Kalman filter algorithm is used to predict and analyze all the data in the denoised harmful gas concentration time series to obtain the estimated harmful gas concentration at the future moment;

[0039] When the estimated concentration of harmful gases is greater than the preset concentration threshold, an early warning is issued.

[0040] In a second aspect, the present invention provides an air quality monitoring system based on the Internet of Things, which adopts the following technical solutions:

[0041] The air quality monitoring system based on the Internet of Things includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the air quality monitoring method based on the Internet of Things is implemented.

[0042] By adopting the above technical solution, the above-mentioned air quality monitoring method based on the Internet of Things is generated into a computer program and stored in a memory to be loaded and executed by a processor, so that a terminal device is made according to the memory and the processor for easy use.

[0043] The present invention has the following technical effects:

[0044] When constructing the classification accuracy evaluation function, the present invention considers the distribution difference of the signal time series of the characteristic values ​​of the two categories in each classification result, and can better use the difference characteristics between the noise signal and the normal signal to evaluate the situation of segmenting the noise signal from the normal signal in the classification result, thereby providing a basis for obtaining a better segmentation result;

[0045] Furthermore, when constructing the classification accuracy evaluation function, the correlation between the signal time series of the analysis feature value and the time series of the concentration of the same type of harmful gas at the nearest position is taken into account. The characteristics that the correlation between the noise signal and the time series of the concentration of the same type of harmful gas at the nearest position is poor, while the correlation between the normal signal and the time series of the concentration of the same type of harmful gas at the nearest position is strong, can be used to evaluate the situation of segmenting the noise signal and the normal signal in the classification result, thereby providing a basis for obtaining a better segmentation result;

[0046] Furthermore, when calculating noise compliance, autocorrelation and eigenvalues ​​are taken into account to accurately describe the signal regularity and content, thereby distinguishing the noise signal from the normal signal in the target classification result, providing a basis for subsequent accurate denoising; further, by accurately removing the noise information in the time series of harmful gas concentrations, a basis is provided for subsequent air quality monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.

[0048] Figure 1 It is a flow chart of a method in an air quality monitoring method based on the Internet of Things in an embodiment of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0050] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0051] The embodiment of the present invention discloses an air quality monitoring method based on the Internet of Things, referring to Figure 1 , comprising steps S1 to S8:

[0052] S1: Obtain the time series of each type of harmful gas concentration at each location.

[0053] Specifically, a group of sensors for collecting harmful gas concentrations are installed at every m meters. The concentration of each type of harmful gas is collected once per second, and the collection is N times. The types of harmful gases include but are not limited to the following aspects: sulfur dioxide, nitrogen dioxide, carbon monoxide, and hydrogen sulfide. The concentration of each type of harmful gas collected by the sensor at each position at all times is arranged in time sequence to obtain a time series of the concentration of each type of harmful gas at each position. m represents the preset interval distance, n represents the preset interval time, and N represents the preset number of collections. This embodiment is described by taking m as 100, n as 10, and N as 1000 as an example. Other embodiments may take other values, and this embodiment does not make specific restrictions.

[0054] S2: Perform singular value decomposition on the harmful gas concentration time series to obtain the eigenvalue set.

[0055] It should be noted that, in order to perform singular value decomposition processing, the harmful gas concentration time series sequence needs to be converted into a matrix, and in order to convert into a matrix, the harmful concentration time series sequence needs to be segmented, and then all the segments form a harmful gas concentration matrix.

[0056] S20: Segment the harmful gas concentration time series to obtain a number of data segments, and construct a harmful gas concentration matrix from all the data segments.

[0057] Optionally, as an example, the harmful gas concentration time series is segmented to obtain a number of data segments, and all the data segments are used to form a harmful gas concentration matrix, including:

[0058] The harmful gas concentration time series is evenly divided into data segments of length L. L represents a preset segment length. This embodiment takes L as 50 as an example for description. Other embodiments may take other values, and this embodiment does not make any specific limitation.

[0059] Each data segment is taken as each row of the matrix to obtain the harmful gas concentration matrix.

[0060] S21: Perform singular value decomposition on the harmful gas concentration matrix to obtain an eigenvalue set.

[0061] Optionally, as an example, singular value decomposition is performed on the harmful gas concentration matrix to obtain an eigenvalue set, including:

[0062] The singular value decomposition algorithm is used to perform singular value decomposition on the harmful gas concentration matrix to obtain a left singular matrix, a diagonal singular matrix and a right singular matrix. All eigenvalues ​​on the diagonal are obtained in the diagonal singular matrix, and the set composed of all eigenvalues ​​is recorded as the eigenvalue set.

[0063] S3: taking the difference between the time series of the harmful gas concentration before and after a eigenvalue in the eigenvalue set is set to zero as the signal time series of the eigenvalue.

[0064] It should be noted that in order to separate the noise signal from the normal signal, it is necessary to decompose the signals representing different characteristics in the harmful gas concentration time series, and then determine which signals corresponding to which characteristics are noise signals and which are normal signals.

[0065] Preferably, as an example, the difference between the time series of the concentration of harmful gas before and after the zeroing process of a feature value in the feature value set is used as the signal time series of the feature value, including:

[0066] After setting one eigenvalue in the eigenvalue set to zero, a singular value decomposition inverse transform is performed based on the diagonal singular matrix, the left singular matrix and the right singular matrix composed of all eigenvalues ​​to obtain a harmful gas concentration matrix lacking the eigenvalue.

[0067] The difference between the harmful gas concentration matrix and the harmful gas concentration matrix lacking the eigenvalue is recorded as the signal matrix of the eigenvalue.

[0068] A sequence obtained by splicing together each row of the signal matrix of the eigenvalue is recorded as the signal timing sequence of the eigenvalue.

[0069] It should be noted that the signal time series of the characteristic value is a signal that describes a characteristic. The following needs to analyze which signals corresponding to which characteristics are noise signals and which are normal signals.

[0070] S4: Randomly combine the eigenvalues ​​in the eigenvalue set into two categories to obtain several classification results; construct a classification accuracy evaluation function, which is positively correlated with the distribution difference of the signal time series sequence of the eigenvalues ​​of the two categories in one classification result, and is also positively correlated with the correlation difference between the signal time series sequence of the eigenvalues ​​of the two categories in one classification result and the time series sequence of the concentration of harmful gases of the same type at the nearest position; the classification result corresponding to the maximum value of the classification accuracy evaluation function is used as the target classification result.

[0071] S40: randomly combining the feature values ​​in the feature value set into two categories to obtain several classification results.

[0072] It should be noted that, since it is currently unknown which features correspond to noise signals and which features correspond to normal signals, we can only randomly classify the signals corresponding to the feature values ​​to obtain several classification results.

[0073] Preferably, as an example, the feature values ​​in the feature value set are randomly combined into two categories to obtain several classification results, including:

[0074] Set the number of variables , any eigenvalue in the set eigenvalues ​​as one class, and the remaining eigenvalues ​​as one class. Take 1 to Get an integer between Each value of corresponds to the classification result, The corresponding value Classification results, Represents the number of eigenvalues.

[0075] S41: Construct a classification accuracy evaluation function.

[0076] It should be noted that it is not known which classification result can separate the noise signal from the normal signal. The accuracy of each classification result needs to be judged based on the difference characteristics between the noise and normal signals and the correlation characteristics of the harmful gas concentration sequence collected at adjacent positions. In this embodiment, the classification accuracy evaluation function is used to judge the accuracy of each classification result.

[0077] Preferably, as an example, constructing a classification accuracy evaluation function includes:

[0078]

[0079] Among them, the characteristic value in one category and the characteristic value in another category in a classification result of the characteristic value set of a type of harmful gas concentration time series at a location are recorded as the first characteristic value and the second characteristic value of the classification result of the type at the location, respectively. Indicates the first classification result of this type at this location The first eigenvalue and the The absolute value of the mean difference of the signal time series of the second eigenvalues, Indicates the first classification result of this type at this location The first eigenvalue and the The absolute value of the variance difference of the signal time series of the second eigenvalue, Indicates the first classification result of this type at this location The first eigenvalue and the The distribution difference of the signal time series of the second eigenvalue is large. Since the difference between the noise signal and the normal signal is large, the distribution characteristics of the noise signal and the normal signal are also large. Therefore, the classification result can be used to evaluate the accuracy of the classification of the noise signal and the normal signal. The number of first feature values ​​representing the classification result of this type at this location, The number of second eigenvalues ​​representing the classification result of this type at this location, Indicates the first classification result of this type at this location The correlation between the signal time series of the first eigenvalue and the concentration time series of the same type of harmful gas at the nearest position, Indicates the first classification result of this type at this location The correlation between the signal time series of the second eigenvalue and the time series of the harmful gas concentration of the same type at the nearest position. Since the harmful gas will flow, the harmful gas concentration in the closer area will be more correlated. Therefore, the normal signal will be more correlated with the harmful gas concentration at the nearest position, while the noise signal has a poor correlation with the harmful gas concentration at the nearest position. Therefore, the classification result can be judged by analyzing the difference in correlation with the harmful gas concentration at other positions to determine whether the noise signal is separated from the normal signal. It reflects the difference in correlation between the two types of signals in the classification results of this type at this location and the harmful gases at the nearest location. Represents the classification accuracy evaluation function.

[0080] S5: The classification result corresponding to the maximum value of the classification accuracy evaluation function is taken as the target classification result.

[0081] Preferably, as an example, the classification result corresponding to the maximum value of the classification accuracy evaluation function is used as the target classification result, including:

[0082] The data corresponding to each classification result of a position is input into the classification accuracy evaluation function to obtain the classification accuracy evaluation function value of each classification result, and the classification result corresponding to the maximum value of the classification accuracy evaluation function is taken as the target classification result.

[0083] S6: Calculate the noise conformity of each category in the target classification result, where the noise conformity is negatively correlated with the autocorrelation coefficient and eigenvalue of the signal time series of the eigenvalue in the category.

[0084] It should be noted that although the noise signal and the normal signal are separated in the above target classification results, it is still not possible to determine which category in the target classification results represents the noise signal. The following needs to analyze the situation in which the data in each category conforms to the noise signal to obtain the category corresponding to the noise signal.

[0085] Preferably, as an example, calculating the noise conformity of each category in the target classification result includes:

[0086]

[0087] in, Indicates the first category of the target classification result. The larger the value, the more signals containing this feature are in the harmful gas concentration time series, while the noise signal is generally relatively small. Therefore, the possibility that the signal corresponding to the feature value in this category is a noise signal is small. Indicates the number of feature values ​​contained in this category in the target classification result. Indicates the first The autocorrelation coefficient of the signal time series of the characteristic value is larger, indicating that the The signal time series of the eigenvalues ​​has a strong regularity, while the noise signal generally has a poor regularity. Therefore, the possibility that the signal corresponding to the eigenvalue in this category is a noise signal is small. Indicates the noise conformity of this category in the target classification result.

[0088] The above embodiments involve autocorrelation coefficients. The following is a process for obtaining the autocorrelation coefficients, including:

[0089] The autocorrelation monitoring algorithm is used to perform autocorrelation monitoring on the signal time series to obtain the autocorrelation coefficient. The autocorrelation monitoring algorithm is used to perform autocorrelation monitoring on the signal time series, which is a prior art and will not be described in detail here.

[0090] S7: The characteristic value of the category corresponding to the maximum value of noise compliance is set to zero to remove the noise of the harmful gas concentration time series.

[0091] Optionally, as an example, the characteristic value of the category corresponding to the maximum noise compliance value is set to zero to remove the noise of the harmful gas concentration time series, including:

[0092] After the eigenvalues ​​of the category corresponding to the maximum value of noise compliance are set to zero, the inverse transformation of singular value decomposition is performed based on the diagonal singular matrix, left singular matrix and right singular matrix composed of all eigenvalues ​​to obtain the denoised harmful gas concentration matrix.

[0093] Each row of the denoised harmful gas concentration matrix is ​​spliced ​​together to obtain the denoised harmful gas concentration time series.

[0094] S8: To assist in air quality monitoring.

[0095] Optionally, as an example, to assist with air quality monitoring, including:

[0096] The Kalman filter algorithm is used to predict and analyze all the data in the denoised harmful gas concentration time series to obtain the estimated harmful gas concentration at future moments.

[0097] When the estimated value of the harmful gas concentration is greater than the preset concentration threshold, an early warning is issued. In this embodiment, the preset concentration threshold of each type of harmful gas concentration is different, where the preset concentration thresholds of sulfur dioxide, nitrogen dioxide, carbon monoxide, and hydrogen sulfide are 0.0005%, 0.00025%, 0.0024%, and 0.00066%, respectively. Other embodiments may take other values, and this embodiment does not specifically limit them.

[0098] An embodiment of the present invention further discloses an air quality monitoring system based on the Internet of Things, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an air quality monitoring method based on the Internet of Things according to the present invention is implemented.

[0099] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0100] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory, a dynamic random access memory, a static random access memory, an enhanced dynamic random access memory, a high bandwidth memory, a hybrid storage cube, etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of the device or accessible or connectable to the device.

[0101] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

[0102] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. The air quality monitoring method based on the Internet of Things is characterized by: Includes steps: Obtain the time series of each type of harmful gas concentration of the sensor at each location; Performing singular value decomposition on the harmful gas concentration time series sequence to obtain an eigenvalue set, including: uniformly dividing the harmful gas concentration time series sequence into data segments of length L, where L represents a preset segment length, and taking each data segment as each row of the matrix to obtain a harmful gas concentration matrix; performing singular value decomposition processing on the harmful gas concentration matrix using a singular value decomposition algorithm to obtain a diagonal singular matrix; obtaining all eigenvalues ​​on the diagonal in the diagonal singular matrix, recording a set consisting of all eigenvalues ​​as an eigenvalue set, and taking the difference between the harmful gas concentration time series sequence before and after an eigenvalue in the eigenvalue set is set to zero as the signal time series sequence of the eigenvalue, including: after setting an eigenvalue in the eigenvalue set to zero, performing singular value decomposition inverse transformation processing based on the diagonal singular matrix, the left singular matrix and the right singular matrix consisting of all eigenvalues ​​to obtain a harmful gas concentration matrix lacking the eigenvalue; recording the difference between the harmful gas concentration matrix and the harmful gas concentration matrix lacking the eigenvalue as the signal matrix of the eigenvalue; and recording a sequence obtained by splicing each row of the signal matrix of the eigenvalue as the signal time series sequence of the eigenvalue; Randomly combine the eigenvalues ​​in the eigenvalue set into two categories to obtain several classification results; construct a classification accuracy evaluation function, including: ; wherein, the characteristic value in one category and the characteristic value in another category in a classification result of the characteristic value set of a type of harmful gas concentration time series at a location are recorded as the first characteristic value and the second characteristic value of the classification result of the type at the location, respectively, Indicates the first classification result of this type at this location The first eigenvalue and the The absolute value of the mean difference of the signal time series of the second eigenvalues, Indicates the first classification result of this type at this location The first eigenvalue and the The absolute value of the variance difference of the signal time series of the second eigenvalue, Indicates the first classification result of this type at this location The first eigenvalue and the The distribution difference of the signal time series of the second eigenvalue, The number of first feature values ​​representing the classification result of this type at this location, The number of second eigenvalues ​​representing the classification result of this type at this location, Indicates the first classification result of this type at this location The correlation between the signal time series of the first eigenvalue and the concentration time series of the same type of harmful gas at the nearest position, Indicates the first classification result of this type at this location The correlation between the signal time series of the second eigenvalue and the concentration time series of the same type of harmful gas at the nearest position, It reflects the difference in correlation between the two types of signals in the classification results of this type at this location and the harmful gases at the nearest location. represents a classification accuracy evaluation function; the classification accuracy evaluation function is positively correlated with the distribution difference of the signal time series of the characteristic values ​​of two categories in a classification result, and is also positively correlated with the correlation difference between the signal time series of the characteristic values ​​of two categories in a classification result and the time series of the concentration of harmful gases of the same type at the nearest position; the classification result corresponding to the maximum value of the classification accuracy evaluation function is taken as the target classification result; Calculating the noise conformity of each category in the target classification result includes: ;in, Indicates the first category of the target classification result. eigenvalues, Indicates the number of feature values ​​contained in this category in the target classification result. Indicates the first The autocorrelation coefficient of the signal time series of eigenvalues, Indicates the noise conformity of the category in the target classification result, and the noise conformity is negatively correlated with the autocorrelation coefficient and eigenvalue of the signal time series of the eigenvalue in the category; The characteristic value of the category corresponding to the maximum noise compliance value is set to zero to remove the noise of the harmful gas concentration time series to assist in air quality monitoring.

2. The air quality monitoring method based on the Internet of Things according to claim 1, characterized in that: The eigenvalues ​​in the eigenvalue set are randomly combined into two categories to obtain several classification results, including: Set the number of variables , any eigenvalue in the set eigenvalues ​​as one class, and the remaining eigenvalues ​​as one class. Take 1 to Get an integer between Each value of corresponds to the classification result, Represents the number of eigenvalues.

3. The air quality monitoring method based on the Internet of Things according to claim 1, characterized in that: The process of setting the characteristic value of the category corresponding to the maximum value of noise compliance to zero to remove the noise of the harmful gas concentration time series includes: After the eigenvalues ​​of the category corresponding to the maximum noise compliance value are set to zero, the inverse transformation of singular value decomposition is performed based on the diagonal singular matrix composed of all eigenvalues ​​to obtain the denoised harmful gas concentration matrix; Each row of the denoised harmful gas concentration matrix is ​​spliced ​​together to obtain the denoised harmful gas concentration time series.

4. The air quality monitoring method based on the Internet of Things according to claim 1, characterized in that: The auxiliary air quality monitoring includes: The Kalman filter algorithm is used to predict and analyze all the data in the denoised harmful gas concentration time series to obtain the estimated harmful gas concentration at the future moment; When the estimated concentration of harmful gases is greater than the preset concentration threshold, an early warning is issued.

5. The air quality monitoring system based on the Internet of Things is characterized by: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the air quality monitoring method based on the Internet of Things according to any one of claims 1 to 4 is implemented.