A method and system for auditing industrial emission pollution data based on feature recognition

Abstract

The present invention relates to the technical field of industrial pollution data auditing, and specifically to an industrial emission pollution data auditing method and system based on feature recognition. First, all emission spectrum curves are obtained, and all deviation frequencies in the most abnormal emission spectrum curves are screened out; then, the reference amplitude of each deviation frequency is obtained to obtain the standard amplitude of each deviation frequency; the initial standard emission spectrum curve is fitted according to all emission spectrum curves, and then the initial standard emission spectrum curve is adjusted according to the standard amplitude to obtain the standard emission spectrum curve, thereby performing a pollution audit. The present invention adjusts and corrects the amplitude of the deviation frequency based on the frequency characteristics of high-frequency sensor noise caused by electronic mechanical interference and the spectrum similarity characteristics of emission curves of different periods, combined with all periodic emission spectrum curves, to obtain a standard emission spectrum curve, thereby improving the accuracy of pollution auditing.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial pollution data auditing, and in particular to an industrial emission pollution data auditing method and system based on feature recognition. Background Art

[0002] Industrial emission pollution refers to the phenomenon that various pollutants and harmful substances discharged during industrial production cause harm to the ecosystem and human health. There are many types of industrial emission pollutants, including waste gas, wastewater and solid waste. Among them, the excessive pH value of industrial wastewater may cause water pollution, destroy the balance of the water ecosystem, and further affect the reproduction and survival of aquatic organisms; therefore, monitoring and auditing the pH value of discharged wastewater is crucial for sustainable development, compliance supervision of corporate departments and scientific research.

[0003] Since the discharge of industrial wastewater is usually affected by the cycle of production plans, industrial wastewater also has a certain discharge cycle; by comparing the pH value in each discharge cycle with the relevant emission standards, it is possible to evaluate whether there is any pollution behavior; but the pH data monitored in each discharge cycle may contain noise data due to interference with the sensor, and the noise data is usually caused by the electronic mechanical interference of the sensor, that is, there is a lot of high-frequency noise, so it is usually necessary to perform denoising on it to filter out the high-frequency noise; and the existing technology uses traditional filters, which often rely on the setting of the cutoff frequency. The unreasonable setting of the filter parameters may filter out the real and useful signal components or residual high-frequency noise components, thereby causing signal distortion, which in turn affects the accuracy of the industrial wastewater discharge pollution audit. Summary of the invention

[0004] In order to solve the technical problem that the filtering effect of periodic emission data in the prior art is poor, resulting in low accuracy of industrial wastewater emission pollution audit, the purpose of the present invention is to provide an industrial emission pollution data audit method based on feature recognition, and the technical solution adopted is as follows:

[0005] Obtain emission spectrum curves of all periodic emission curves in the industrial emission process during the audit period;

[0006] According to the spectrum difference between the emission spectrum curves, an abnormal emission spectrum curve is screened out from all the emission spectrum curves; according to the amplitude difference between the abnormal emission spectrum curve and other emission spectrum curves at the same frequency, all deviation frequencies in the abnormal emission spectrum curve are screened out;

[0007] When each of the deviation frequencies corresponds to the same frequency, the reference amplitude of each of the emission spectrum curves is obtained according to the amplitude characteristics of each of the emission spectrum curves and the influence of the corresponding frequency on the time domain characteristics of the emission spectrum curve; the standard amplitude of each of the deviation frequencies is obtained according to the reference amplitude of each of the deviation frequencies and the frequency amplitudes in all of the emission spectrum curves at the same frequency, combined with the frequency characteristics of the deviation frequency; the initial standard emission spectrum curve is fitted according to all of the emission spectrum curves, and the initial standard emission spectrum curve is adjusted according to the standard amplitude to obtain a standard emission spectrum curve;

[0008] A pollution audit is conducted on industrial emissions within the audit period according to the standard emission frequency curve.

[0009] Furthermore, the method for obtaining the abnormal emission spectrum curve includes:

[0010] Taking the frequency intersection of all the emission spectrum curves as the frequency range to be analyzed, and obtaining the integral area of ​​each emission spectrum curve within the frequency range to be analyzed;

[0011] Obtaining a deviation parameter of each emission spectrum curve according to a difference between the integral area corresponding to each emission spectrum curve and the other emission spectrum curves;

[0012] The emission spectrum curve with the largest deviation parameter is taken as an abnormal emission spectrum curve.

[0013] Furthermore, the method for obtaining the deviation frequency includes:

[0014] At the same frequency, according to the difference between the frequency amplitude in the abnormal emission spectrum curve and the frequency amplitude in each of the other emission spectrum curves, obtaining the offset index of each frequency in the abnormal emission spectrum curve;

[0015] In the abnormal emission spectrum curve, the frequency whose deviation index is greater than or equal to a preset threshold is taken as the deviation frequency.

[0016] Furthermore, the method for obtaining the reference amplitude includes:

[0017] When each of the deviation frequencies corresponds to the same frequency, the relative amplitude level of the corresponding amplitude in each of the emission spectrum curves is used as the first reference weight of the amplitude in each of the emission spectrum curves corresponding to the same frequency;

[0018] When each of the deviation frequencies corresponds to the same frequency, a negative correlation mapping of the difference between the corresponding amplitude in each of the emission spectrum curves and the amplitude mean of the corresponding amplitudes in all other emission spectrum curves is used as a second reference weight for the amplitude in each of the emission spectrum curves corresponding to the same frequency;

[0019] Adjust the frequency amplitude corresponding to the same frequency for each of the deviation frequencies in each of the emission spectrum curves, and obtain the spectrum influence degree of each of the deviation frequencies corresponding to the same frequency according to the difference between the emission spectrum curves before and after the adjustment;

[0020] According to the first reference weight, the second reference weight and the degree of spectrum influence, obtaining a reference weight of the amplitude corresponding to the same frequency in each of the emission spectrum curves, wherein the first reference weight and the second reference weight are both positively correlated with the reference weight, and the degree of spectrum influence is negatively correlated with the reference weight;

[0021] The corresponding amplitudes are weighted and summed according to the reference weights, and the weighted summation result is used as the reference amplitude corresponding to the deviation frequency.

[0022] Furthermore, the method for obtaining the spectrum impact degree includes:

[0023] When each of the deviation frequencies corresponds to the same frequency, the amplitude corresponding to the same frequency in each of the emission spectrum curves is replaced in sequence with the amplitude corresponding to the same frequency in each of the remaining emission spectrum curves to obtain all the adjusted spectrum curves of each of the emission spectrum curves;

[0024] All the adjusted spectrum curves and the corresponding emission spectrum curves are respectively subjected to time-frequency inverse transformation, and the reconstruction error between each adjusted spectrum curve and the emission spectrum curve after the time-frequency inverse transformation is obtained, and the sum of all reconstruction errors is used as the spectrum influence degree of the deviation frequency corresponding to the same frequency.

[0025] Furthermore, the method for obtaining the standard amplitude includes:

[0026] Normalizing the frequency value of each of the deviation frequencies as a first weight, obtaining a second weight according to the first weight, wherein the sum of the second weight and the first weight is a constant of 1;

[0027] When each of the deviation frequencies corresponds to the same frequency, obtaining the amplitude mean of the corresponding amplitudes in all the emission spectrum curves;

[0028] The reference amplitude is weighted with the first weight, the amplitude mean is weighted with the second weight, and the weighted sum result is used as the standard amplitude corresponding to the deviation frequency.

[0029] Furthermore, the initial standard emission spectrum curve and the method for obtaining the standard emission spectrum curve include:

[0030] Constructing the abscissa axis of the standard emission spectrum curve according to the union of all frequencies in all the emission spectrum curves, obtaining the initial standard amplitude at each frequency in the standard emission spectrum curve according to the frequencies and corresponding amplitudes of all the emission spectrum curves, and obtaining the initial standard emission spectrum curve;

[0031] The initial standard amplitude at the same frequency corresponding to the deviation frequency in the initial standard emission spectrum curve is replaced by the corresponding standard amplitude to obtain a standard emission spectrum curve.

[0032] Furthermore, the audit method for conducting pollution audit on industrial emissions within the audit period includes:

[0033] The standard emission spectrum curve is subjected to an inverse time-frequency transformation to obtain a standard emission curve; when a preset number of continuous emission data in the standard emission curve exceeds the set standard, it is determined that there is a pollution risk in the industrial emissions during the audited time period.

[0034] Furthermore, the time period to be audited includes a preset number of cycle emission processes.

[0035] The present invention also proposes an industrial emission pollution data auditing system based on feature recognition, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the industrial emission pollution data auditing method based on feature recognition are implemented.

[0036] The present invention has the following beneficial effects:

[0037] The present invention obtains emission spectrum curves of all periodic emission curves in the industrial emission process within the audited time period, so as to facilitate the subsequent spectrum feature analysis of all periodic emission processes; according to the spectrum differences between the emission spectrum curves, the abnormal emission spectrum curves are screened out from all emission spectrum curves, so as to screen out all deviation frequencies in the abnormal emission spectrum curve according to the amplitude differences between the abnormal emission spectrum curve and other emission spectrum curves at the same frequency, and the deviation frequency is the frequency component that may be interfered by the sensor noise, so as to facilitate the subsequent adjustment, and obtain the deviation frequency according to the difference between the emission spectrum curves with similar fluctuation characteristics, so as to avoid the poor filtering effect caused by the unreasonable setting of the filter parameters, and improve the analysis accuracy of the noise component and the accuracy of the subsequent standard emission spectrum curve; under the condition that each deviation frequency corresponds to the same frequency, according to each emission spectrum curve The amplitude characteristics and the corresponding frequency affect the time domain characteristics of the emission spectrum curve, and the reference amplitude of each deviation frequency is obtained; according to the reference amplitude of each deviation frequency and the frequency amplitude in all emission spectrum curves at the same frequency, combined with the frequency characteristics of the deviation frequency, the standard amplitude of each deviation frequency is obtained, and the frequency amplitude in all emission spectrum curves can provide an average amplitude reference under ideal conditions. The reference amplitude is an amplitude reference for weighted evaluation considering the noise spectrum characteristics, and then the reference amplitude and the average amplitude are adjusted in combination with the high-frequency characteristics of the noise to consider the weights, and the standard amplitude under each deviation frequency can be evaluated; the initial standard emission spectrum curve is fitted according to all emission spectrum curves, and the initial standard emission spectrum curve is adjusted according to the standard amplitude to obtain the standard emission spectrum curve; pollution audits are conducted on industrial emissions within the audit time period according to the standard emission frequency curve. According to the frequency characteristics of high-frequency sensor noise caused by electronic mechanical interference and the spectrum similarity characteristics of emission curves of different periods, the present invention adjusts and corrects the amplitude of the deviation frequency in combination with all period emission spectrum curves to obtain the standard emission spectrum curve, avoids the poor filtering effect of traditional filters, and improves the accuracy of pollution audits. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0039] Figure 1 A flowchart of an industrial emission pollution data audit method based on feature recognition provided by an embodiment of the present invention;

[0040] Figure 2 A flowchart of a method for obtaining a reference amplitude provided by an embodiment of the present invention;

[0041] Figure 3 A schematic diagram of the structure of an industrial emission pollution data audit system based on feature recognition provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the industrial emission pollution data audit method and system based on feature recognition proposed by the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0043] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0044] The following is a detailed description of a method and system for auditing industrial emission pollution data based on feature recognition provided by the present invention in conjunction with the accompanying drawings.

[0045] See also Figure 1 , which shows a flow chart of an industrial emission pollution data audit method based on feature recognition provided by an embodiment of the present invention, specifically comprising:

[0046] Step S1, obtaining emission spectrum curves of all periodic emission curves in the industrial emission process within the audit time period.

[0047] Taking into account that the emission of industrial pollutants has a certain periodicity, in one embodiment of the present invention, taking industrial wastewater as an example, a pH sensor is first installed at the factory wastewater discharge pipe, and the pH value of the industrial wastewater is collected in real time at a sampling frequency of once per minute, and all periodic emission curves of the industrial emission process within the audited time period are obtained, wherein the audited time period includes a preset number of periodic emission processes, and the periodic emission curve is the pH value curve of each periodic emission process; to facilitate the subsequent feature identification of all periodic emission processes, the embodiment of the present invention further performs Fourier transform on each periodic emission curve to obtain all emission spectrum curves, thereby analyzing the periodic emission curves interfered by noise for adjustment and correction.

[0048] It should be noted that the preset number of cycles is 15. In other embodiments of the present invention, the implementer may also set other values, or determine the preset number of cycles according to the number of discharges in the daily production plan or monthly production plan of the factory.

[0049] In other embodiments of the present invention, implementers may also install other types of sensors to monitor the periodic emission curves of other exhaust gases or solid wastes such as sulfur dioxide, nitrogen oxides, etc. according to audit requirements; in other embodiments of the present invention, implementers may also use other time-frequency transformation methods such as wavelet transform to obtain emission spectrum curves, which and Fourier transform are existing technologies well known to those skilled in the art and will not be elaborated here.

[0050] Step S2, based on the spectrum differences between the emission spectrum curves, filter out abnormal emission spectrum curves from all emission spectrum curves; based on the amplitude differences between the abnormal emission spectrum curve and other emission spectrum curves at the same frequency, filter out all deviation frequencies in the abnormal emission spectrum curve.

[0051] Taking into account that the periodic emission curve of industrial wastewater has a certain periodicity, it should have similar fluctuation characteristics and similar spectral characteristics. Taking into account that under normal circumstances, the periodic emission curve may also have certain fluctuations, which may also appear as high-frequency components in the corresponding emission spectrum curve. While hoping to filter out high-frequency noise, inappropriate filter parameters may directly smooth out the actual fluctuations, or residual noise, thereby causing signal distortion and affecting subsequent audit results; at the same time, the noise interference is random, and there is a large difference in spectral characteristics between the emission spectrum curve affected by the noise and the normal emission spectrum curve. Therefore, the embodiment of the present invention selects abnormal emission spectrum curves from all emission spectrum curves based on the spectral differences between the emission spectrum curves. Compared with the normal emission spectrum curve, there is a certain amount of noise interference in the abnormal emission spectrum curve, and the noise components in the abnormal emission spectrum curve can be further analyzed to avoid emission signal distortion caused by unreasonable filter parameter settings.

[0052] Preferably, in one embodiment of the present invention, considering that similar emission spectrum curves should have similar signal energy distribution, and the curve integral within the spectrum range can evaluate the signal energy, there may be slight differences in the frequency ranges of different emission spectrum curves, so the frequency intersection is taken; and considering that when an emission spectrum curve has a large difference in integral area from all other emission spectrum curves, it can be considered that it contains noise; based on this, the method for obtaining an abnormal emission spectrum curve includes:

[0053] The frequency intersection of all emission spectrum curves is taken as the frequency range to be analyzed, and the integral area of ​​each emission spectrum curve in the frequency range to be analyzed is obtained;

[0054] Obtaining a deviation parameter of each emission spectrum curve according to a difference between the corresponding integral areas of each emission spectrum curve and each of the other emission spectrum curves;

[0055] The emission spectrum curve with the largest deviation parameter is taken as the abnormal emission spectrum curve.

[0056] In one embodiment of the present invention, the calculation formula of the deviation parameter is: ; In the formula, For the Deviation parameters of emission spectrum curves; is the maximum and minimum normalization function; For the The serial numbers of other emission spectrum curves other than the emission spectrum curve; is the total number of emission spectrum curves, that is, the number of preset cycles, which is 15 in the embodiment of the present invention; For the The integral area of ​​the emission spectrum curve within the frequency range to be analyzed is The absolute value of the difference between the integrated areas of the emission spectrum curves within the frequency range to be analyzed.

[0057] In the calculation formula of the deviation parameter, the difference between the integral area corresponding to each emission spectrum curve and each other emission spectrum curve is specifically summed, and the sum result is normalized; in other embodiments of the present invention, the implementer may also use the difference mean and the difference weighted average result to evaluate the deviation parameter, which is not described here. It should be noted that the sum result is normalized by the maximum and minimum normalization method. The larger the deviation parameter, the greater the possibility of being an abnormal emission spectrum curve, so the emission spectrum curve with the largest deviation parameter is taken as the abnormal emission spectrum curve; in other embodiments of the present invention, the implementer may also adopt other normalization methods, which are existing technologies and are not described here.

[0058] In another embodiment of the present invention, the implementer may also obtain the main frequency and the total number of peak points of each emission spectrum curve based on feature extraction, and further obtain and evaluate the spectral feature differences of different emission spectrum curves in combination with energy density; obtain the spectral feature parameters of each emission spectrum curve by multiplying the frequency, the total number of peak points and the energy density, and then sum the differences between each emission spectrum curve and the corresponding spectral feature parameters of each other emission spectrum curve, normalize the sum result, and obtain the deviation parameters of each emission spectrum curve.

[0059] It should be noted that the method for obtaining the main frequency and the total number of peak points is: obtain all peak points through the peak detection algorithm, and take the frequency corresponding to the maximum peak point as the main frequency; the energy density is the result of dividing the integral area of ​​the emission spectrum curve by the length of the frequency interval; the implementer can also use other spectral characteristic parameters to obtain the spectral characteristic parameters corresponding to each emission spectrum curve, and then obtain the deviation parameters; they are all existing technologies well known to those skilled in the art and will not be repeated here.

[0060] After obtaining the abnormal emission spectrum curve, the frequency components affected by noise interference can be further analyzed in the abnormal emission spectrum curve. Considering that the embodiment of the present invention is mainly aimed at the correction and adjustment of high-frequency noise, when there is a large difference in the corresponding amplitude of the frequency components of an emission spectrum curve and all other emission spectrum curves at the same frequency, it can be considered that it contains noise; the embodiment of the present invention screens out all deviation frequencies in the abnormal emission spectrum curve according to the amplitude difference between the abnormal emission spectrum curve and other emission spectrum curves at the same frequency.

[0061] Preferably, in one embodiment of the present invention, the method for obtaining the deviation frequency includes:

[0062] At the same frequency, according to the difference between the frequency amplitude in the abnormal emission spectrum curve and the frequency amplitude in each of the other emission spectrum curves, the deviation index of each frequency in the abnormal emission spectrum curve is obtained;

[0063] In the abnormal emission spectrum curve, the frequency whose deviation index is greater than or equal to a preset threshold is regarded as the deviation frequency.

[0064] In one embodiment of the present invention, the calculation formula of the offset index is: ; In the formula, is the sequence number of the frequency in the abnormal emission spectrum curve; The abnormal emission spectrum curve The frequency deviation index; is the maximum and minimum normalization function; is the serial number of the abnormal emission spectrum curve; is the serial number of other emission spectrum curves except the abnormal emission spectrum curve; is the total number of emission spectrum curves, that is, the number of preset cycles, which is 15 in the embodiment of the present invention; The abnormal emission spectrum curve The frequency amplitude of the frequency; For the In the emission spectrum curve The frequency amplitude of the frequency; It should be noted that, in one embodiment of the present invention, the preset threshold is 0.3, and the implementer may also set other values; in other embodiments of the present invention, the implementer may also adopt other normalization methods, which will not be described in detail here.

[0065] It should be noted that, in other embodiments of the present invention, considering that there may be more than one abnormal emission spectrum curve, the implementer may select multiple curves, such as 2-5, and perform frequency deviation analysis on each abnormal emission spectrum curve to determine the deviation frequency to be adjusted.

[0066] Step S3, when each deviation frequency corresponds to the same frequency, obtain the reference amplitude of each deviation frequency according to the amplitude characteristics of each emission spectrum curve and the influence of the corresponding frequency on the time domain characteristics of the emission spectrum curve; obtain the standard amplitude of each deviation frequency according to the reference amplitude of each deviation frequency and the frequency amplitude in all emission spectrum curves at the same frequency, combined with the frequency characteristics of the deviation frequency; fit the initial standard emission spectrum curve according to all emission spectrum curves, adjust the initial standard emission spectrum curve according to the standard amplitude, and obtain the standard emission spectrum curve.

[0067] Since the periodic emission curves of industrial wastewater should have similar spectral characteristics, it is possible to fit a standard emission spectrum curve for all emission processes within the audited time period through comprehensive comparative analysis of the spectral characteristics of all emission spectrum curves, and then audit the standard emission spectrum curve to assess whether there is a pollution risk in the emission process.

[0068] Considering that the standard emission spectrum curve may have a large deviation due to noise interference, the weight of considering the deviation frequency in the abnormal emission spectrum curve should be reduced during fitting; and the noise is usually randomly distributed and not concentrated in a specific frequency, so the frequency amplitude of the noise is relatively low relative to the real useful frequency components, and has a special amplitude characteristic. Considering that in similar emission spectrum curves, the influence of the same frequency component on the time domain curve characteristics of each emission spectrum curve should also be similar, but some frequency components may be affected by noise, resulting in a greater impact on the time domain characteristics; therefore, in the embodiment of the present invention, when each deviation frequency corresponds to the same frequency, the reference amplitude of each deviation frequency is obtained according to the amplitude characteristics of each emission spectrum curve and the influence of the corresponding frequency on the time domain characteristics of the emission spectrum curve.

[0069] Preferably, in one embodiment of the present invention, considering that the noise frequency amplitude is relatively low, if the relative amplitude level of the corresponding amplitude at the same frequency corresponding to the deviation frequency in each emission spectrum curve is higher, the corresponding reference weight is larger; considering that the deviation frequency at the same frequency corresponding to the average level in each emission spectrum curve is more, it means that it may be affected by noise, and the corresponding reference weight is smaller; considering that if the same frequency components in different emission spectrum curves are replaced and adjusted, the time domain characteristics of the original emission spectrum curve change greatly, it means that the corresponding frequency components may be affected by noise, and the corresponding reference weight should be smaller; based on this, the method for obtaining the reference amplitude includes:

[0070] See also Figure 2 , which shows a flow chart of a method for obtaining a reference amplitude provided by an embodiment of the present invention:

[0071] Step S201 : when each deviation frequency corresponds to the same frequency, the relative amplitude level of the corresponding amplitude in each emission spectrum curve is used as the first reference weight of the amplitude in each emission spectrum curve corresponding to the same frequency.

[0072] Considering that the ratio of the amplitude to the total amplitude can reflect the relative level of the amplitude to a certain extent, in one embodiment of the present invention, the calculation formula of the first reference weight is: ;in, is the frequency number of each emission spectrum curve that is the same as the deviation frequency; is the serial number of the emission spectrum curve; For the In the emission spectrum curve The first reference weight of the amplitude at a frequency; For the In the emission spectrum curve The amplitude at the frequency; For all emission spectrum curves The sum of the amplitudes corresponding to the frequencies.

[0073] Step S202, when each deviation frequency corresponds to the same frequency, a negative correlation mapping of the difference between the corresponding amplitude in each emission spectrum curve and the amplitude mean of the corresponding amplitude in all other emission spectrum curves is used as a second reference weight for the amplitude at the same frequency in each emission spectrum curve.

[0074] Considering that at the same frequency, if the corresponding amplitude in one emission spectrum curve deviates more from the corresponding amplitude in the emission spectrum curve, the possibility of abnormality is greater, so in one embodiment of the present invention, the calculation formula of the second reference weight is: ;in, is the frequency number of each emission spectrum curve that is the same as the deviation frequency; is the serial number of the emission spectrum curve; For the In the emission spectrum curve A second reference weight for the amplitude at a frequency; For the In the emission spectrum curve The amplitude at the frequency; For all emission spectrum curves The mean amplitude at each frequency; is the absolute value symbol; The natural constant The exponential function of base .

[0075] Specific The negative correlation is mapped to the exponential function to adjust the logic and normalize it to obtain the second reference weight.

[0076] Step S203, adjusting the frequency amplitude corresponding to the same frequency for each deviation frequency in each emission spectrum curve, and obtaining the spectrum influence degree of each deviation frequency corresponding to the same frequency according to the difference between the emission spectrum curves before and after the adjustment.

[0077] In one embodiment of the present invention, a method for obtaining the degree of influence of the deviation includes:

[0078] When each deviation frequency corresponds to the same frequency, the amplitude corresponding to the same frequency in each emission spectrum curve is replaced with the amplitude corresponding to the same frequency in each of the remaining emission spectrum curves in turn, so as to obtain all the adjusted spectrum curves of each emission spectrum curve;

[0079] Perform inverse Fourier transform on all adjustment spectrum curves and corresponding emission spectrum curves respectively, obtain the reconstruction error between each adjustment spectrum curve and the emission spectrum curve after inverse Fourier transform, and take the sum of all reconstruction errors as the spectrum influence degree of the deviation frequency corresponding to the same frequency.

[0080] It should be noted that, since the time-frequency transformation method of Fourier transform is adopted in step S1, the corresponding time-frequency inverse transformation method is inverse Fourier transform; the method for obtaining the reconstruction error is specifically the mean square error between all signal points between each adjusted spectrum curve and the emission spectrum curve after the inverse Fourier transform, which is an existing technology well known to technicians in this field and will not be elaborated here.

[0081] Step S204, based on the first reference weight, the second reference weight and the degree of spectrum impact, obtain the reference weight of the amplitude corresponding to the same frequency in each emission spectrum curve, the first reference weight and the second reference weight are both positively correlated with the reference weight, and the degree of spectrum impact is negatively correlated with the reference weight.

[0082] In one embodiment of the present invention, the calculation formula of the reference weight is: ;in, is the frequency number of each emission spectrum curve that is the same as the deviation frequency; is the serial number of the emission spectrum curve; For the In the emission spectrum curve The reference weight of the amplitude at each frequency; For the In the emission spectrum curve The first reference weight of the amplitude at a frequency; For the In the emission spectrum curve A second reference weight for the amplitude at a frequency; The natural constant The exponential function with base ; For the In the emission spectrum curve The degree of spectrum impact of each frequency; is the normalization function.

[0083] The negative correlation of the spectrum impact degree is mapped to the exponential function to adjust the logical relationship, and then the first reference weight, the second reference weight and the spectrum impact degree after the negative correlation mapping are multiplied and combined, so that the first reference weight and the second reference weight are both positively correlated with the reference weight, and the spectrum impact degree is negatively correlated with the reference weight.

[0084] It should be noted that The normalization method is as follows: calculate the product of the frequency at the same frequency corresponding to each deviation frequency in all emission spectrum curves and sum them up, that is, , the sum is normalized as the denominator so that the sum of the reference weights of the amplitudes at the same frequency in different emission spectrum curves is 1; in other embodiments of the present invention, the implementer may also use other basic mathematical operations and related mapping methods, which are all existing technologies and will not be described here.

[0085] Step S205: weighted sum the corresponding amplitudes according to the reference weights, and use the weighted sum result as the reference amplitude of the corresponding deviation frequency. The calculation formula of the reference amplitude is: ; In the formula, is the frequency number of each emission spectrum curve that is the same as the deviation frequency; is the serial number of the emission spectrum curve; For the The reference amplitude of the frequency; is the total number of emission spectrum curves, that is, the number of preset cycles, which is 15 in the embodiment of the present invention; For the In the emission spectrum curve The amplitude at a frequency.

[0086] Taking into account that under normal circumstances, a standard emission spectrum curve can be fitted by averaging the spectral characteristics of all emission spectrum curves, but in the case of noise interference, directly averaging may lead to low accuracy of the standard emission spectrum curve; and considering that sensor noise is usually a high-frequency component caused by mechanical and electronic interference, the average amplitude is the amplitude reference under ideal circumstances, and the reference amplitude is the amplitude reference for weighted evaluation considering the noise spectrum characteristics, and then the weights of the reference amplitude and the average amplitude are adjusted in combination with the high-frequency characteristics of the noise, so that the standard amplitude at each deviation frequency can be evaluated; therefore, the embodiment of the present invention obtains the standard amplitude of each deviation frequency based on the reference amplitude of each deviation frequency and the frequency amplitudes in all emission spectrum curves at the same frequency, combined with the frequency characteristics of the deviation frequency.

[0087] Preferably, in one embodiment of the present invention, considering that the higher the frequency value of the deviation frequency is, the greater the possibility of being sensor noise is, the greater the weight of considering the reference amplitude is; if the lower the frequency value of the deviation frequency is, the smaller the possibility of being sensor noise is, the greater the weight of considering the average amplitude is; based on this, the method for obtaining the standard amplitude includes:

[0088] The frequency value of each deviation frequency is normalized as a first weight, and a second weight is obtained according to the first weight, and the sum of the second weight and the first weight is a constant 1;

[0089] In all emission spectrum curves, obtain the amplitude mean of all amplitudes corresponding to the frequency that is the same as each deviation frequency;

[0090] The reference amplitude is weighted with a first weight, the amplitude mean is weighted with a second weight, and the weighted sum result is used as the standard amplitude corresponding to the deviation frequency.

[0091] The calculation formula of standard amplitude is: ; In the formula, For the The standard amplitude of the deviation frequency; is the serial number of the deviation frequency; For the The frequency value corresponding to the deviation frequency; It is the maximum value of the frequency in the abnormal emission spectrum curve; For the The first weight of the deviation frequency; For the The reference amplitude of the frequency; For the The second weight of the deviation frequency; For all emission spectrum curves, The mean amplitude of all amplitudes corresponding to the same frequency of the deviation frequency.

[0092] In the calculation formula of the standard amplitude, the deviation frequency is specifically divided by the maximum value of the frequency in the abnormal emission spectrum curve for normalization. In other embodiments of the present invention, other normalization means may also be used, which will not be described in detail here.

[0093] After obtaining the standard amplitude of each deviation frequency, the initial standard emission spectrum curve can be fitted according to all emission spectrum curves, and then the initial standard emission spectrum curve can be adjusted according to the standard amplitude to obtain the standard emission spectrum curve.

[0094] Preferably, in one embodiment of the present invention, the horizontal axis of the standard emission spectrum curve is constructed according to the union of all frequencies in all emission spectrum curves, and the initial standard amplitude at each frequency in the standard emission spectrum curve is obtained according to the frequencies and corresponding amplitudes of all emission spectrum curves to obtain the initial standard emission spectrum curve; the initial standard amplitude at the same frequency corresponding to the deviation frequency in the initial standard emission spectrum curve is replaced by the corresponding standard amplitude to obtain the standard emission spectrum curve.

[0095] In the embodiment of the present invention, considering that there may be slight differences in the frequency ranges of all emission spectrum curves, the frequency union is taken to obtain all frequencies, and the frequencies are sorted in ascending order to construct the horizontal axis of the standard emission spectrum curve. Then, at each frequency, the frequency amplitude mean of all emission spectrum curves corresponding to the frequency is calculated to obtain the initial standard amplitude at each frequency in the standard emission spectrum curve. Considering that the existence of abnormal emission spectrum curves may cause deviations in the initial standard amplitude, thereby affecting the accuracy and audit effect, the initial standard amplitude at the same frequency corresponding to the deviation frequency in the initial standard emission spectrum curve is replaced with the standard amplitude corresponding to the deviation frequency obtained in the above step, thereby obtaining the standard emission spectrum curve.

[0096] Step S4, performing pollution audit on industrial emissions within the audit period according to the standard emission frequency curve.

[0097] In one embodiment of the present invention, the audit method for performing pollution audit on industrial emissions within the audit period includes:

[0098] The standard emission spectrum curve is inversely transformed by Fourier transform to obtain the standard emission curve; when there are a preset number of consecutive emission data in the standard emission curve that exceed the set standard, it is considered that there is a pollution risk in the industrial emissions during the audit period.

[0099] It should be noted that the preset number is set to 10, that is, when 10 consecutive PH data exceed the preset standard, it is considered that there is a pollution risk; the set standard is set to 6.0-9.0, and the implementer can also determine the set standard according to the PH emission environmental protection standards formulated by relevant departments and the industrial field, or set other preset number values; inverse Fourier transform is a prior art well known to those skilled in the art and will not be elaborated here.

[0100] In summary, the embodiment of the present invention first obtains the emission spectrum curves of all periodic emission curves in the industrial emission process within the audit time period; further screens out the abnormal emission spectrum curves, and screens out all deviation frequencies in the abnormal emission spectrum curves; then, when each deviation frequency corresponds to the same frequency, according to the amplitude characteristics of each emission spectrum curve and the influence of the corresponding frequency on the time domain characteristics of the emission spectrum curve, obtain the reference amplitude of each deviation frequency; according to the reference amplitude of each deviation frequency and the frequency amplitude in all emission spectrum curves at the same frequency, combined with the frequency characteristics of the deviation frequency, obtain the standard amplitude of each deviation frequency; according to all emission spectrum curves, fit the initial standard emission spectrum curve, adjust the initial standard emission spectrum curve according to the standard amplitude, and obtain the standard emission spectrum curve; according to the standard emission frequency curve, perform pollution audit on industrial emissions within the audit time period. According to the frequency characteristics of high-frequency sensor noise caused by electronic and mechanical interference and the spectrum similarity characteristics of emission curves of different periods, the present invention adjusts and corrects the amplitude of the deviation frequency in combination with all periodic emission spectrum curves, obtains the standard emission spectrum curve, and improves the accuracy of pollution audit.

[0101] Based on the same inventive concept, the embodiment of the present invention also proposes an industrial emission pollution data audit system based on feature recognition, please refer to Figure 3 The audit system includes a memory 301, a processor 302, and a computer program 303 stored in the memory 301 and executable on the processor 302. When the processor 302 executes the computer program 303, the above-mentioned industrial emission pollution data audit method based on feature recognition is implemented.

[0102] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0103] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for auditing industrial emission pollution data based on feature recognition. It is characterized in that The method comprises: Obtain emission spectrum curves of all periodic emission curves in the industrial emission process during the audit period; According to the spectrum difference between the emission spectrum curves, an abnormal emission spectrum curve is screened out from all the emission spectrum curves; according to the amplitude difference between the abnormal emission spectrum curve and other emission spectrum curves at the same frequency, all deviation frequencies in the abnormal emission spectrum curve are screened out; When each of the deviation frequencies corresponds to the same frequency, the reference amplitude of each of the emission spectrum curves is obtained according to the amplitude characteristics of each of the emission spectrum curves and the influence of the corresponding frequency on the time domain characteristics of the emission spectrum curve; the standard amplitude of each of the deviation frequencies is obtained according to the reference amplitude of each of the deviation frequencies and the frequency amplitudes in all of the emission spectrum curves at the same frequency, combined with the frequency characteristics of the deviation frequency; the initial standard emission spectrum curve is fitted according to all of the emission spectrum curves, and the initial standard emission spectrum curve is adjusted according to the standard amplitude to obtain a standard emission spectrum curve; Conduct pollution audits on industrial emissions during the audit period according to the standard emission frequency curve; The audit methods for conducting pollution audits on industrial emissions within the audit period include: Performing an inverse time-frequency transformation on the standard emission spectrum curve to obtain a standard emission curve; when a preset number of emission data continuously exceed the set standard in the standard emission curve, it is determined that there is a pollution risk in the industrial emissions during the audit period; The method for obtaining the reference amplitude includes: When each of the deviation frequencies corresponds to the same frequency, the corresponding amplitude in each of the emission spectrum curves is used as the numerator, the sum of the corresponding amplitudes in all the emission spectrum curves is used as the denominator, and the ratio of the numerator to the denominator is used as the first reference weight of the amplitude in each of the emission spectrum curves corresponding to the same frequency; When each of the deviation frequencies corresponds to the same frequency, a negative correlation mapping of the difference between the corresponding amplitude in each of the emission spectrum curves and the amplitude mean of the corresponding amplitudes in all other emission spectrum curves is used as a second reference weight for the amplitude in each of the emission spectrum curves corresponding to the same frequency; Adjust the frequency amplitude corresponding to the same frequency for each of the deviation frequencies in each of the emission spectrum curves, and obtain the spectrum influence degree of each of the deviation frequencies corresponding to the same frequency according to the difference between the emission spectrum curves before and after the adjustment; According to the first reference weight, the second reference weight and the degree of spectrum influence, obtaining a reference weight of the amplitude corresponding to the same frequency in each of the emission spectrum curves, wherein the first reference weight and the second reference weight are both positively correlated with the reference weight, and the degree of spectrum influence is negatively correlated with the reference weight; The corresponding amplitudes are weighted and summed according to the reference weights, and the weighted summation result is used as the reference amplitude corresponding to the deviation frequency.

2. According to claim 1, a method for auditing industrial emission pollution data based on feature recognition, It is characterized in that The method for obtaining the abnormal emission spectrum curve includes: Taking the frequency intersection of all the emission spectrum curves as the frequency range to be analyzed, and obtaining the integral area of ​​each emission spectrum curve within the frequency range to be analyzed; Obtaining a deviation parameter of each emission spectrum curve according to a difference between the integral area corresponding to each emission spectrum curve and the other emission spectrum curves; The emission spectrum curve with the largest deviation parameter is taken as an abnormal emission spectrum curve.

3. According to claim 1, a method for auditing industrial emission pollution data based on feature recognition, It is characterized in that The method for obtaining the deviation frequency includes: At the same frequency, according to the difference between the frequency amplitude in the abnormal emission spectrum curve and the frequency amplitude in each of the other emission spectrum curves, obtaining the offset index of each frequency in the abnormal emission spectrum curve; In the abnormal emission spectrum curve, the frequency whose deviation index is greater than or equal to a preset threshold is taken as the deviation frequency.

4. According to claim 1, a method for auditing industrial emission pollution data based on feature recognition, It is characterized in that The method for obtaining the spectrum impact degree includes: When each of the deviation frequencies corresponds to the same frequency, the amplitude corresponding to the same frequency in each of the emission spectrum curves is replaced in sequence with the amplitude corresponding to the same frequency in each of the remaining emission spectrum curves to obtain all the adjusted spectrum curves of each of the emission spectrum curves; All the adjusted spectrum curves and the corresponding emission spectrum curves are respectively subjected to time-frequency inverse transformation, and the reconstruction error between each adjusted spectrum curve and the emission spectrum curve after the time-frequency inverse transformation is obtained, and the sum of all reconstruction errors is used as the spectrum influence degree of the deviation frequency corresponding to the same frequency.

5. According to claim 1, a method for auditing industrial emission pollution data based on feature recognition, It is characterized in that The method for obtaining the standard amplitude includes: Normalizing the frequency value of each of the deviation frequencies as a first weight, obtaining a second weight according to the first weight, wherein the sum of the second weight and the first weight is a constant of 1; When each of the deviation frequencies corresponds to the same frequency, obtaining the amplitude mean of the corresponding amplitudes in all the emission spectrum curves; The reference amplitude is weighted with the first weight, the amplitude mean is weighted with the second weight, and the weighted sum result is used as the standard amplitude corresponding to the deviation frequency.

6. According to claim 1, a method for auditing industrial emission pollution data based on feature recognition, It is characterized in that The initial standard emission spectrum curve and the method for obtaining the standard emission spectrum curve include: Constructing the abscissa axis of the standard emission spectrum curve according to the union of all frequencies in all the emission spectrum curves, obtaining the initial standard amplitude at each frequency in the standard emission spectrum curve according to the frequencies and corresponding amplitudes of all the emission spectrum curves, and obtaining the initial standard emission spectrum curve; The initial standard amplitude at the same frequency corresponding to the deviation frequency in the initial standard emission spectrum curve is replaced by the corresponding standard amplitude to obtain a standard emission spectrum curve.

7. According to claim 1, a method for auditing industrial emission pollution data based on feature recognition, It is characterized in that The time period to be audited includes a preset number of cycle emission processes.

8. An industrial emission pollution data audit system based on feature recognition, It is characterized in that It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of an industrial emission pollution data audit method based on feature recognition as described in any one of claims 1 to 7 are implemented.

Images