A Denoising Method and Device for Laser Absorption Spectroscopy

Abstract

The present application discloses a noise reduction method and device for laser absorption spectroscopy. The method includes: determining a laser absorption spectrum formed when a laser irradiates a gas to be measured; determining a noise reduction spectrum under current environmental variables according to a noise reduction spectral function corresponding to the current pressure of the gas chamber where the gas to be measured is located and the relationship between the current temperature and the laser absorption spectrum, the noise reduction spectrum and the laser absorption spectrum being in the same coordinate system, and the current environmental variables including the current temperature and the current pressure; determining an update direction and an update displacement amount of the noise reduction spectrum in the coordinate system by comparing the laser absorption spectrum and the noise reduction spectrum; updating the position of the noise reduction spectrum in the coordinate system according to the update direction and the update displacement amount until the updated noise reduction spectrum meets a preset condition, and the noise reduction spectrum that meets the preset condition is the laser absorption spectrum after noise filtering. The present application can reduce the influence of various noises on the laser absorption spectrum of the gas to be measured, and thus is beneficial to improving the detection accuracy.

Description

Technical Field

[0001] The present application relates to the technical field of laser absorption spectrum analysis, and in particular to a noise reduction method and device for laser absorption spectrum. Background Art

[0002] At present, the main methods of gas concentration detection include electrochemical gas sensor detection, catalytic combustion gas sensor detection and laser absorption spectrum detection. Among them, laser absorption spectrum detection is more and more popular due to its advantages of high sensitivity, fast response time and low detection limit compared with traditional electrochemical gas sensor detection and catalytic combustion gas sensor detection. Laser absorption spectrum detection is mainly based on the principle that "specific gases only absorb lasers of specific wavelengths". The gas concentration is calculated by comparing the difference between the laser absorption spectra presented by the gas to be tested before and after absorbing the laser of a specific wavelength.

[0003] Conventional detection devices in laser absorption spectrum detection technology generally include a laser transmitter, an air chamber, a photoelectric detection module, and a signal processing module. Specifically, the laser beam emitted by the laser transmitter propagates in the air chamber and can be absorbed by the gas to be measured in the air chamber. After receiving the laser beam absorbed by the gas to be measured, the photoelectric detection module converts the optical signal into an electrical signal and sends it to the signal processing module. The signal processing module can output the laser absorption spectrum according to the received electrical signal. In the above detection process, due to the presence of various noises, the output laser absorption spectrum often has a low signal-to-noise ratio, thereby increasing the error of the subsequent gas concentration calculation process, which is not conducive to improving the accuracy of gas concentration detection.

[0004] In order to reduce the impact of various noises on gas concentration detection, the existing technology mainly performs noise reduction processing on the gas concentration detection process from two aspects: hardware improvement and software denoising. In terms of hardware improvement, for example, the laser emitter in the detection device can be improved, such as selecting a laser emitter that can emit lasers with better light absorption intensity to improve the quality of the laser, thereby reducing the light source intensity noise of the laser; or, the gas chamber can be improved to increase the optical path or improve the reflection structure of the gas chamber to reduce the light source phase noise of the laser; or, the back-end electrical module can be improved, such as designing a higher performance filter circuit to eliminate the noise generated by the circuit system. Although the above-mentioned various hardware improvement methods can reduce the impact of different types of noise on gas concentration detection, the hardware improvement method has higher requirements on the performance of each component of the detection device, so it often makes the structure of the detection device more complicated, and the use and maintenance methods are more cumbersome, which greatly increases the cost of the detection device.

[0005] Based on this, software denoising is gradually becoming a research hotspot in the field of laser absorption spectroscopy noise reduction technology due to its advantages such as easy debugging and low cost. In recent years, the commonly used software denoising algorithm is the window mean algorithm (AveragingWindow). When detecting gas concentration through the window mean algorithm, it is first necessary to collect the laser absorption spectra of the gas to be measured multiple times. The absorption spectra collected multiple times are used as sample data. By comparing numerous sample data, the differences between the sample data are determined, and the parts with larger differences are considered to be affected by noise. Then, the parts with larger differences are removed, and the gas concentration is determined through the remaining sample data to filter out the influence of noise on gas detection.

[0006] However, the window mean algorithm is extremely dependent on the number of sample data collected. If the sample data is too small, the noise in each laser absorption spectrum of the gas to be measured does not have universality, so it often affects the noise filtering effect in the laser absorption spectrum. If the sample data is too large, it will greatly increase the computational load of the algorithm and extend the noise filtering time in the laser absorption spectrum. Summary of the Invention

[0007] This application provides a noise reduction method and device for laser absorption spectra to solve at least one of the above technical problems.

[0008] The technical solution adopted in this application is as follows:

[0009] In the first aspect, this application provides a noise reduction method for laser absorption spectra. The noise reduction method includes the following steps:

[0010] Determine the laser absorption spectrum formed when the laser irradiates the gas to be measured;

[0011] According to the noise reduction spectral function corresponding to the current pressure of the gas chamber where the gas to be measured is located, and the relationship between the current temperature and the laser absorption spectrum, determine the noise reduction spectrum under the current environmental variables. The noise reduction spectrum and the laser absorption spectrum are in the same coordinate system. The current environmental variables include the current temperature and the current pressure;

[0012] By comparing the laser absorption spectrum and the noise reduction spectrum, determine the update direction and update displacement of the noise reduction spectrum in the coordinate system;

[0013] According to the update direction and the update displacement, update the position of the noise reduction spectrum in the coordinate system until the updated noise reduction spectrum meets the preset conditions. The noise reduction spectrum that meets the preset conditions is the laser absorption spectrum after noise filtering.

[0014] As a preferred embodiment of the present application, the determination of the laser absorption spectrum formed when the laser irradiates the gas to be measured includes: obtaining the original spectrum generated by the laser irradiating the gas to be measured, where the original spectrum corresponds to multiple consecutive periods; determining the extreme values of each period in the original spectrum through an extreme value search algorithm; using the determined multiple extreme values as nodes to cut the multiple consecutive periods into multiple discrete periods; performing mean processing on the multiple discrete periods superimposed together to obtain a mean spectrum, where the mean spectrum has a gas absorption peak, and the gas absorption peak is the wavelength absorption region corresponding to the gas to be measured in the laser absorption spectrum; fitting the part of the mean spectrum without the gas absorption peak to obtain a reference spectrum; and performing normalization processing on the mean spectrum and the reference spectrum to obtain the laser absorption spectrum.

[0015] As a preferred embodiment of the present application, the noise reduction spectral function corresponding to the current pressure includes:

[0016] When the current pressure is less than a preset first threshold, the noise reduction spectral function is:

[0017]

[0018] where, g G (v) represents the Gaussian line shape function g G , Δv G is the full width at half maximum of the noise reduction spectrum when the current pressure is less than the preset first threshold, with the unit of cm -1 , v0 is the wave number at the center of the spectral line, with the unit of cm -1 , v is the wave number of the laser, with the unit of cm -1 ;

[0019] When the current pressure is greater than a preset second threshold, the noise reduction spectral function is:

[0020]

[0021] where, g L (v) represents the Lorentz line shape function g L , Δv L is the full width at half maximum of the noise reduction spectrum when the current pressure is greater than the preset second threshold, with the unit of cm -1 ; v0 is the wave number at the center of the spectral line, with the unit of cm -1 ; v is the wave number of the laser, with the unit of cm -1 ;

[0022] When the current pressure is between the first threshold and the second threshold, the noise reduction spectral function is:

[0023]

[0024] where g v (v) represents the Voigt line shape function g v , and g v is the convolution of g G and g L .

[0025] As a preferred embodiment of the present application, determining the noise reduction spectrum under the current environmental variables according to the noise reduction spectrum function corresponding to the current pressure of the gas chamber where the gas to be measured is located and the relationship between the current temperature and the laser absorption spectrum includes: selecting the noise reduction spectrum function corresponding to the current pressure; constructing a compensation function based on the spectral line intensity of the current temperature and the laser absorption spectrum, and performing a product calculation on the noise reduction spectrum function and the compensation function to obtain the noise reduction spectrum under the current environmental variables.

[0026] As a preferred embodiment of the present application, before determining the update direction and update displacement of the noise reduction spectrum in the coordinate system by comparing the laser absorption spectrum and the noise reduction spectrum, the noise reduction method further includes:

[0027] Importing the data corresponding to the laser absorption spectrum and the noise reduction spectrum into an adaptively constructed particle swarm optimization algorithm respectively, where the adaptively constructed particle swarm optimization algorithm is configured to include i particles, each particle represents a data object, each data object is a spectral point in the laser absorption spectrum, the total population number is NP, and the maximum number of updates is N;

[0028] Correcting each particle of the laser absorption spectrum through a boundary processing function, and the boundary processing function is:

[0029]

[0030] where is the i-th particle of the corrected corresponding j-dimensional component, is the i-th particle of the corresponding j-dimensional component before correction, is the preset maximum upper limit value in j dimensions, is the preset minimum lower limit value in j dimensions;

[0031] Determining the update direction and update displacement of the noise reduction spectrum in the coordinate system by comparing the laser absorption spectrum and the noise reduction spectrum includes:

[0032] Obtain the laser absorption spectrum corresponding to the corrected particles, and determine the update direction and update displacement of the noise-reduced spectrum in the coordinate system by comparing the noise-reduced spectrum and the laser absorption spectrum corresponding to the corrected particles.

[0033] As a preferred embodiment of the present application, the determining the update direction and update displacement of the noise-reduced spectrum in the coordinate system by comparing the laser absorption spectrum and the noise-reduced spectrum includes: setting self-learning factors, population learning factors, and adjacent learning factors in the adaptive particle swarm algorithm respectively, where the self-learning factor is used to characterize the update trend of each particle in each dimension, the population learning factor is used to characterize the common update trend of any particle in different dimensions, and the adjacent learning factor is used to characterize the influence trend of adjacent particles on the current particle in each dimension; calculating the particle fitness value, population fitness value, and adjacent fitness value corresponding to the self-learning factor, population learning factor, and adjacent learning factor in the laser absorption spectrum respectively; and performing weighted calculations on the particle fitness value, population fitness value, and adjacent fitness value according to a preset weight strategy to determine the update direction and update displacement of the noise-reduced spectrum.

[0034] As a preferred embodiment of the present application, performing weighted calculations on the particle fitness value, population fitness value, and adjacent fitness value according to a preset weight strategy to determine the update direction and update displacement of the noise-reduced spectrum includes: calculating particle weight coefficients, population weight coefficients, and adjacent weight coefficients corresponding to the particle fitness value, population fitness value, and adjacent fitness value respectively according to a preset weight strategy; and performing weighted calculations on the particle fitness value, population fitness value, and adjacent fitness value according to the particle weight coefficient, population weight coefficient, and adjacent weight coefficient to obtain the update direction and update displacement of the noise-reduced spectrum.

[0035] As a preferred embodiment of the present application, the particle weight coefficient, population weight coefficient, and adjacent weight coefficient are determined in the following manner:

[0036]

[0037] Wherein, if the number of updates of the position of the noise-reduced spectrum in the coordinate system reaches the maximum number of updates N, it is determined that the updated noise-reduced spectrum meets the preset conditions, and n is the current number of updates; w fitness is the particle weight coefficient applied when performing weighted calculation on the particle fitness value, w NP is the population weight coefficient applied when performing weighted calculation on the population weight coefficient, w ADThe adjacent weight coefficient applied when calculating the weighted adjacent fitness values;

[0038] Determine the update direction and update displacement of the noise reduction spectrum according to the weighted calculation result.

[0039] As a preferred embodiment of the present application, the determining the update direction and update displacement of the noise reduction spectrum according to the weighted calculation result includes: after calculating the product of the particle fitness value, the population fitness value and the adjacent fitness value to obtain an update vector, determining the absolute value of the update vector as the update displacement of the noise reduction spectrum, and determining the positive or negative of the update vector as the update direction of the noise reduction spectrum, where a positive value indicates upward along the ordinate of the coordinate system, and a negative value indicates downward along the ordinate of the coordinate system.

[0040] In a second aspect, the present application further provides a noise reduction device for laser absorption spectroscopy, including:

[0041] A first determination module, configured to determine the laser absorption spectrum formed when a laser irradiates a gas to be measured;

[0042] A second determination module, configured to determine the noise reduction spectrum under the current environmental variables according to the noise reduction spectrum function corresponding to the current pressure of the gas chamber where the gas to be measured is located, and the relationship between the current temperature and the laser absorption spectrum, the noise reduction spectrum and the laser absorption spectrum are located in the same coordinate system, and the current environmental variables include the current temperature and the current pressure;

[0043] A comparison module, configured to determine the update direction and update displacement of the noise reduction spectrum in the coordinate system by comparing the laser absorption spectrum and the noise reduction spectrum;

[0044] An update module, configured to update the position of the noise reduction spectrum in the coordinate system according to the update direction and the update displacement until the updated noise reduction spectrum meets a preset condition, and the noise reduction spectrum that meets the preset condition is the laser absorption spectrum after noise filtering.

[0045] Due to the adoption of the above technical solutions, the beneficial effects obtained by the present application at least include:

[0046] The noise reduction method for laser absorption spectroscopy provided by this application mainly improves the software algorithm in the detection device to reduce the influence of various noises on the laser absorption spectroscopy, thereby improving the accuracy of gas concentration detection. Specifically, this application mainly determines the update direction and update displacement of the noise reduction spectrum by comparing the obtained laser absorption spectrum of the gas to be measured with the fitted noise reduction spectrum, and finally outputs the noise reduction spectrum of the gas to be measured after filtering out the noise through continuous iterative processing of the noise reduction spectrum. On the one hand, since this method does not require modification of the hardware part in the detection device, it can greatly reduce the cost of the detection device and improve the convenience of use and maintenance of the detection device; on the other hand, this application filters out various noises in the detection device by fitting the noise reduction spectrum, which can not only eliminate the extreme deviations in the laser absorption spectrum, but also eliminate the small noises in the laser absorption spectrum due to the smooth characteristics of the line type of the fitted noise reduction spectrum, thereby greatly reducing the influence of various noises on the laser absorption spectrum of the gas to be measured, and further facilitating the improvement of the accuracy in the subsequent gas concentration detection process of the gas to be measured.

[0047] In addition, after fitting the noise reduction spectrum, this application only needs to continuously evolve the initially fitted noise reduction spectrum towards the laser spectrum of the real gas to be measured according to the update direction, update displacement and continuous iterative processing. And since the update direction and update displacement are determined based on the laser absorption spectrum of the gas to be measured, the accuracy of the above evolution process can also be well guaranteed. Compared with the window mean algorithm in the prior art, this method obviously does not require multiple sample data acquisitions of the gas to be measured, so it can avoid the defects of unsatisfactory filtering effect or too long algorithm processing time in the window mean algorithm. This can enable this application to reduce the computational load of the algorithm and shorten the noise reduction processing time of the laser absorption spectrum while ensuring the noise reduction effect on the laser absorption spectrum. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The drawings described herein are used to provide a further understanding of this application and form a part of this application. The schematic embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation to this application. In the drawings:

[0049] Figure 1 is a schematic diagram of a laser absorption spectrum provided by an embodiment of this application;

[0050] Figure 2 is a schematic flowchart of a noise reduction method for a laser absorption spectrum provided by an embodiment of this application;

[0051] Figure 3 is a schematic flowchart of a process for determining a laser absorption spectrum provided by an embodiment of this application;

[0052] Figure 4 Schematic diagram of an original spectrum provided by an embodiment of the present application;

[0053] Figure 5 Schematic diagram of an averaged spectrum provided by an embodiment of the present application;

[0054] Figure 6 Schematic diagram of a reference spectrum provided by an embodiment of the present application;

[0055] Figure 7 Schematic flowchart of another noise reduction method for laser absorption spectrum provided by an embodiment of the present application;

[0056] Figure 8 Schematic diagram of the structure of a noise reduction device for laser absorption spectrum provided by an embodiment of the present application. Detailed implementation manners

[0057] To make the objectives, technical solutions and advantages of the exemplary embodiments of the present application clearer, the technical solutions in the exemplary embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the exemplary embodiments of the present application. Apparently, the described exemplary embodiments are only a part rather than all of the embodiments of the present application.

[0058] For the convenience of understanding the technical solutions of the application, some concepts related to the present application will be described below first.

[0059] Laser absorption spectroscopy detection technology, also known as tunable diode laser absorption spectroscopy (TDLAS for short), mainly uses the narrow linewidth of the tunable semiconductor laser and the characteristic that the wavelength changes with the injection current to measure single or several absorption lines of molecules that are very close and difficult to distinguish.

[0060] The light intensity attenuation of the laser passing through the gas to be measured is based on the Lambert-Beer law, that is, the gas to be measured has an absorption characteristic for light of a specific wavelength, and the absorption intensity is proportional to the concentration of the gas component. The gas concentration is measured by measuring the attenuation of the laser by the gas. If the laser absorption spectroscopy detection technology is applied to gas concentration detection, it is possible to detect the concentration of the gas to be measured by using the wavelength tuning characteristic of the semiconductor laser and the selective absorption of the gas to be measured for the laser.

[0061] Specifically, the detection principle is that the tunable semiconductor laser emits a laser of a specific wavelength under the modulation of the driving current. With the modulation of the injected periodic current, the laser wavelength changes periodically, so that the central wavelength of the laser is adjusted to the absorption spectrum of the gas to be measured, and selective absorption occurs. The concentration of the gas to be measured is then inverted using the spectral intensity signal obtained after gas absorption. Because of the high monochromaticity of the semiconductor laser, an isolated absorption spectrum line of the gas molecule to be measured can be used for measurement, avoiding the cross-interference of the absorption spectrum lines of different gas molecules, thereby accurately identifying the gas to be measured.

[0062] Absorption spectrum refers to the spectrum produced when a substance absorbs photons and transitions from a low energy level to a high energy level; the laser absorption spectrum in this application specifically refers to the spectrum formed when the gas to be measured absorbs photons when a laser is used as a light source.

[0063] In the actual gas concentration detection process, the laser absorption spectrum obtained may be affected to some extent by random changes in the laser amplitude, phase, frequency, etc., defects in the gas chamber structure, or poor stability of the circuit system. Figure 1 A schematic diagram of a laser absorption spectrum provided in this application. Figure 1 The messy lines in the image are spectral lines containing noise, and the smooth curves in the messy lines are ideal spectral lines of the laser absorption spectrum without noise interference, or pure signals. In order to obtain pure signals, the influence of noise on the actual laser absorption spectrum is reduced as much as possible. The prior art not only improves the various devices involved in the gas concentration detection process from the hardware improvement aspect, but also proposes a variety of different noise reduction algorithms from the software aspect to filter out the noise in the laser absorption spectrum. In addition to the window average algorithm, there are also wavelet denoising algorithms and deep learning neural networks.

[0064] Among them, the principle of the wavelet denoising algorithm is: first set a critical threshold λ, if the wavelet coefficient is less than λ, it is considered that the coefficient is mainly caused by noise, and this part of the coefficient is removed; if the wavelet coefficient is greater than λ, it is considered that this coefficient is mainly caused by the signal, and this part of the coefficient is retained, and then the processed wavelet coefficient is subjected to inverse wavelet transform to obtain the denoised signal. Although the wavelet denoising algorithm has good adaptability, its shortcomings are also obvious. It often requires the selection of a suitable wavelet basis function and the frequent adjustment of the threshold parameters. If the wavelet basis function and threshold parameters are not selected properly, it will have a great impact on the denoising effect of the laser absorption spectrum.

[0065] Deep learning neural networks are signal processing algorithms developed in recent years. They extract features from laser absorption spectra using computers with powerful computing capabilities, and then reconstruct the laser absorption spectra to achieve noise reduction processing for the laser absorption spectra. However, deep learning neural networks often require a large number of parameters, such as the network topology structure, initial values of weights and thresholds, etc., resulting in too long learning time and even possibly failing to achieve the learning purpose. Especially for laser absorption spectra that overall conform to the spectral line shape theory, the deep learning neural network searches for features without direction, which is very likely to increase the detection time of gas concentration.

[0066] In addition, laser absorption spectra are affected by various noises. Moreover, due to environmental variables in the gas cell (such as temperature, pressure, etc.) having a great impact on the spectral line shape of the laser absorption spectra, the noise can also include environmental variables. Specifically, due to the limitation of the internal structure of the gas cell, temperature sensors for detecting the internal temperature of the gas cell and / or pressure sensors for detecting the internal pressure of the gas cell, etc., are usually set outside the gas cell, and the detected temperature and pressure are not actual values but measured values with errors from the actual values. Based on this, existing technologies mostly perform calibration measurements of known-concentration gases at different temperatures and pressures through calibration compensation, and calculate and compensate the final gas concentration to reduce the errors caused by the setting methods of the temperature sensors and / or pressure sensors. However, the conventional calibration compensation method not only has a cumbersome calculation process, additionally increasing the computational load of gas concentration calculation, but also is limited by the performance of the sensors, thus unable to guarantee the calibration accuracy.

[0067] In view of this, in order to avoid the influence of environmental variables and noises on laser absorption spectra to the greatest extent and improve the accuracy and detection efficiency of gas concentration detection. This application provides a noise reduction method for laser absorption spectra. Referring to Figure 2 as shown, the noise reduction method includes the following steps:

[0068] S100. Determine the laser absorption spectrum formed when the laser irradiates the gas to be measured.

[0069] It should be noted that the laser absorption spectrum in this step may include the original spectrum obtained by irradiating the gas to be measured with the laser, where the original spectrum may contain a large amount of noise.

[0070] Alternatively, to improve the accuracy of noise reduction, the laser absorption spectrum in this step is not the original spectrum formed by directly irradiating the gas to be measured with a laser, nor is it a pure spectrum (pure signal) with all noise completely filtered out. The laser absorption spectrum here is more like a special spectrum between the original spectrum and the pure spectrum. The amount of noise contained in this special spectrum is much smaller than that of the original spectrum, but not all noise can be completely filtered out. The reason for determining the laser absorption spectrum is to provide a reference for the subsequent determined noise reduction spectrum, thereby facilitating the update of the noise reduction spectrum.

[0071] S200. Determine the noise reduction spectrum under the current environmental variables according to the noise reduction spectrum function corresponding to the current pressure of the gas chamber where the gas to be measured is located, and the relationship between the current temperature and the laser absorption spectrum.

[0072] Among them, the noise reduction spectrum and the laser absorption spectrum are in the same coordinate system, and the current environmental variables include the current temperature and the current pressure. This coordinate system can be a two-dimensional coordinate system as shown in Figure 1 with the number of points (Points) as the abscissa and the absorbance as the ordinate. Among them, the number of points corresponds to the wavelength of the laser, or rather, the number of points is the data points measured based on the wavelength of the laser.

[0073] S300. Determine the update direction and update displacement of the noise reduction spectrum in the coordinate system by comparing the laser absorption spectrum and the noise reduction spectrum.

[0074] S400. Update the position of the noise reduction spectrum in the coordinate system according to the update direction and update displacement, and determine whether the updated noise reduction spectrum meets the preset conditions; if so, execute step S500; if not, return to step S300.

[0075] That is to say, if the noise reduction spectrum after one update does not meet the preset conditions, it is necessary to compare the laser absorption spectrum with the noise reduction spectrum after one update again, and re-determine the update direction and update displacement of the noise reduction spectrum after one update in the coordinate system; then, according to the re-determined update direction and update displacement, perform a secondary update on the position of the noise reduction spectrum after one update in the coordinate system, and determine whether the noise reduction spectrum after the secondary update meets the preset conditions. This cycle continues until the noise reduction spectrum finally meets the preset conditions.

[0076] S500. Determine the laser absorption spectrum with noise filtered out as the noise reduction spectrum that meets the preset conditions.

[0077] Among them, the preset condition may be the maximum number of updates reached by the number of updates of the position of the noise reduction spectrum in the coordinate system, such as 80 times, 100 times, etc. In this case, if the number of updates of the position of the noise reduction spectrum in the coordinate system reaches the maximum number of updates, it can be determined that the updated noise reduction spectrum meets the preset condition; it can also be that the difference between the coordinates of the noise reduction spectrum and the laser absorption spectrum in the coordinate system is within a preset difference range. The difference in coordinates can be the difference in the ordinate between the noise reduction spectrum and the laser absorption spectrum when the abscissa is the same, or it can be the difference in the abscissa between the noise reduction spectrum and the laser absorption spectrum when the ordinate is the same. In this case, when the difference between the coordinates of the noise reduction spectrum and the laser absorption spectrum in the coordinate system is within the preset difference range, it is considered that the updated noise reduction spectrum meets the preset condition. The present application does not limit the specific setting method of the preset condition.

[0078] The noise reduction method for laser absorption spectrum provided by the present application mainly reduces the influence of various noises on the laser absorption spectrum by improving the software algorithm in the detection device, thereby improving the accuracy of gas concentration detection. Specifically, the present application mainly determines the update direction and update displacement of the noise reduction spectrum through the comparison between the laser absorption spectrum of the gas to be measured obtained and the fitted noise reduction spectrum, and finally outputs the noise reduction spectrum of the gas to be measured after filtering out the noise through continuous iterative update processing of the noise reduction spectrum. On the one hand, since this method does not require modification of the hardware part in the detection device, it can greatly reduce the cost of the detection device and improve the convenience of use and maintenance of the detection device; on the other hand, the present application filters out various noises in the detection device by fitting the noise reduction spectrum, which can not only eliminate the extreme deviations in the laser absorption spectrum, but also eliminate the small noises in the laser absorption spectrum due to the smooth characteristics of the line type of the fitted noise reduction spectrum itself, thereby greatly reducing the influence of various noises on the laser absorption spectrum of the gas to be measured, and thus facilitating the improvement of the accuracy in the subsequent gas concentration detection process.

[0079] In addition, after the noise reduction spectrum is fitted in the present application, only based on the update direction, update displacement and continuous iterative update processing, the initially fitted noise reduction spectrum can be continuously evolved towards the laser spectrum of the real gas to be measured. And since the update direction and update displacement are determined based on the laser absorption spectrum of the gas to be measured, the accuracy of the above evolution process can also be well guaranteed. Compared with the window mean algorithm in the prior art, this method obviously does not require multiple sample data acquisitions of the gas to be measured, so it can avoid the defects of unsatisfactory filtering effect or too long algorithm processing time in the window mean algorithm, which enables the present application to reduce the operation load of the algorithm and shorten the noise reduction processing time of the laser absorption spectrum while ensuring the noise reduction effect on the laser absorption spectrum.

[0080] Further, with reference to Figure 3 as shown, when determining the laser absorption spectrum, it can be determined through the following steps:

[0081] S110. Obtain the original spectrum generated by the laser irradiating the gas to be measured.

[0082] It should be noted that the original spectrum in this application corresponds to multiple consecutive cycles, and the original spectrum can be collected through an amplification acquisition circuit. However, since the collected original spectrum is continuous and infinitely looped (as shown in Figure 4 ), in actual calculation, it is often not necessary and not possible to directly process the original spectrum. Because in the actual detection process, the acquisition time of the original spectrum is very short (such as 1 s, 0.5 s, etc.), but even in a very short acquisition time, hundreds of consecutive cycles of the original spectrum can be collected. Since each cycle actually contains the concentration information of the gas to be measured, and the concentration information contained in each cycle has very little difference, it is obvious that it is not necessary to process all the collected cycles to finally calculate the gas concentration. Therefore, on the basis of obtaining the original spectrum, it is usually also necessary to process the original spectrum as described in the following steps S120 - S160, that is, to perform cropping, mean value processing, and normalization processing on the original spectrum respectively.

[0083] S120. Determine the extreme values of each cycle in the original spectrum through an extreme value search algorithm.

[0084] S130. Take the determined multiple extreme values as nodes, and crop multiple consecutive cycles into multiple discrete cycles.

[0085] When cropping the original spectrum, the part between the minimum value or maximum value of adjacent two cycles can be saved with the minimum value or maximum value of each cycle as the node, so as to crop out multiple discrete cycles. It can be understood that there is no limitation on the cropping method of the original spectrum in this application. It can be cropped with extreme values as nodes as shown in this step S130, or it can be cropped with any user - set value as the node, as long as it is ensured that each discrete cycle cropped contains complete gas concentration information.

[0086] S140. Perform mean value processing on multiple superimposed discrete cycles to obtain a mean - value spectrum.

[0087] Among them, with reference to Figure 5 as shown, the mean - value spectrum in this step has a gas absorption peak (i.e., the concave region in Figure 5 ), and the gas absorption peak is the wavelength absorption region corresponding to the gas to be measured in the laser absorption spectrum.

[0088] In addition, although multiple discrete periods are obtained through the above steps, if each discrete period is magnified many times, it can be found that in fact each discrete period contains a large amount of random noise. In this step, the influence of random noise on the laser absorption spectrum is reduced by averaging multiple discrete periods. The averaging method can especially filter out the noise that differs greatly from the laser absorption spectrum in each period.

[0089] S150. Fit the part of the averaged spectrum that does not contain gas absorption peaks to obtain a reference spectrum.

[0090] Since in the averaged spectrum, except for the gas absorption peaks caused by gas absorption, there is no gas absorption anywhere else, and the averaged spectrum is approximately a sawtooth wave signal. Therefore, as shown in Figure 6 , the spectra on both sides of the gas absorption peak can be fitted to a straight line by the quadratic fitting method, that is, fitted to a straight line in the form of y = ax + b, and this straight line is considered to be the spectrum without gas absorption. It should be noted that in actual situations, the averaged spectrum is not a perfect sawtooth wave signal, and it often bends due to gas absorption, that is, it can also be fitted in the form of the quadratic function y = ax 2 + bx + c.

[0091] S160. Normalize the averaged spectrum and the reference spectrum to obtain the laser absorption spectrum.

[0092] The reference spectrum obtained by fitting in step S150 is a spectrum without gas concentration information, and the averaged spectrum obtained in step S140 is a spectrum with large deviation noise filtered out. In this step S160, the normalized spectrum, that is, the laser absorption spectrum, is obtained by ratio processing of the reference spectrum and the averaged spectrum. This laser absorption spectrum is smoother than the averaged spectrum and contains less noise; compared with the reference spectrum, it contains gas concentration information, thus providing a basis for the subsequent gas concentration calculation process.

[0093] As an optional embodiment of the present application, in step S200, the noise reduction spectral function corresponding to the current pressure can be determined in the following manner:

[0094] For example, when the current pressure is less than a preset first threshold, the noise reduction spectral function can be:

[0095]

[0096] where g G (v) represents the Gaussian line shape function g G , Δv GWhen the current pressure is less than the preset first threshold, the full width at half maximum of the noise-reduced spectrum, with the unit of cm -1 , v0 is the wave number at the center of the spectral line, with the unit of cm -1 , v is the wave number of the laser, with the unit of cm -1 .

[0097] For example, when the current pressure is greater than the preset second threshold, the noise-reduced spectral function can be:

[0098]

[0099] where g L (v) represents the Lorentz line shape function g L , Δv L is the full width at half maximum of the noise-reduced spectrum when the current pressure is greater than the preset second threshold, with the unit of cm -1 ; v0 is the wave number at the center of the spectral line, with the unit of cm -1 ; v is the wave number of the laser, with the unit of cm -1 .

[0100] For example, when the current pressure is between the first threshold and the second threshold, the noise-reduced spectral function can be:

[0101]

[0102] where g v (v) represents the Voigt line shape function g v , g v is the convolution of g G and g L ,

[0103] It should be noted that the present application does not limit the specific values of the first threshold and the second threshold, which can be selected by the user according to actual needs. Exemplarily, the first threshold can be 0.01 atm, and the second threshold can be 0.5 atm.

[0104] Furthermore, considering the influence of temperature on the laser absorption spectral line intensity (or spectral line strength), and the influence of line strength on the line shape of the laser absorption spectrum, line strength is a physical quantity representing the energy distribution of the spectral line, which describes the total energy absorbed by atoms per unit volume per unit time. In practical applications, the influence of the environmental variable of temperature on the laser absorption spectrum cannot be ignored.

[0105] In the above step S200, when determining the noise reduction spectrum under the current environmental variables, after selecting the noise reduction spectrum function corresponding to the current pressure, a compensation function based on the temperature of the gas to be measured and the spectral line intensity of the laser absorption spectrum can also be constructed, and the noise reduction spectrum can be finally obtained by multiplying the noise reduction spectrum function and the compensation function.

[0106] Among them, the compensation function can refer to the expression: Q(T) = a + bT + cT 2 + dT 3 for determination. Q(T) is the partition function, T represents the temperature of the gas to be measured, and a, b, c, and d are all constants and can be obtained from a pre-established database. It should be noted that regarding the multiplication calculation method between the noise reduction spectrum function and the compensation function, since the actual form of the noise reduction spectrum function of this application in the coordinate system is not a continuous curve but consists of many points, each point can be called a spectral point, and each point represents a particle in the adaptive particle swarm algorithm. Therefore, when multiplying the noise reduction spectrum function and the compensation function, the value of the compensation function Q(T) can be determined according to the corresponding database query and the measurement of the current temperature, and then the vectors corresponding to each particle in the noise reduction spectrum function are multiplied by Q(T) to obtain multiple new vectors, that is, the ordinates of each point in the noise reduction spectrum function are multiplied by Q(T) respectively to obtain the new ordinates of each point, and then the multiple points corresponding to the new ordinates are connected in the coordinate system to obtain the final noise reduction spectrum.

[0107] In addition, it should also be noted that this application can also evaluate the accuracy of the temperature parameter in the current environmental variables by constructing a calibration equation. The reason for setting the calibration equation for the temperature parameter is also due to the influence brought by the setting position of the aforementioned temperature sensor. Due to various reasons, the temperature measured inside the gas chamber by the temperature sensor is a measured value, and there are more or less differences between it and the actual value. The setting of the calibration equation can calibrate the measured value of the temperature parameter, so as to determine whether the temperature sensor is faulty according to the calibration result.

[0108] Among them, the expression of the calibration equation can be: r L = r0(T0 / T) n p, r L is the full width at half maximum, with the unit cm -1 ; r0 is the atmospheric broadening coefficient at temperature T0, T0 is the reference temperature (open Celsius), p is the pressure, and n is the temperature coefficient.

[0109] Furthermore, this application can also calculate the noise reduction spectrum and the laser absorption spectrum according to the root mean square error to determine the optimal noise reduction spectrum.

[0110] Furthermore, referring toFigure 7 As shown in Figure 7 , step S300 in the present application may include:

[0111] S310. Import the data corresponding to the laser absorption spectrum and the noise reduction spectrum into the pre-constructed adaptive particle swarm optimization algorithm respectively.

[0112] Among them, the adaptive particle swarm optimization algorithm is configured to include i particles, each particle represents a data object, each data object is a spectral point in the laser absorption spectrum, the total population is NP, and the maximum number of updates is N. Moreover, by virtue of the advantages of the adaptive particle swarm optimization algorithm, such as fast convergence speed, few parameters, and simple and easy implementation, when applying this algorithm to the gas concentration detection process, on the one hand, it can increase the temperature and pressure range of the gas monitored by the system, and on the other hand, it can improve the detection efficiency and shorten the detection time. In addition, it should be noted that when determining the laser absorption spectrum formed by laser irradiating the gas to be measured, the obtained laser absorption spectrum actually contains multiple different spectral lines, and each spectral line contains multiple spectral points; when importing this laser absorption spectrum into the adaptive particle swarm optimization algorithm, the number of spectral lines included in the laser absorption spectrum is configured as the dimension of the algorithm. In other words, the j dimension in the following corresponds to the number of spectral lines in the laser absorption spectrum. And the population in the algorithm is defined as the whole composed of the i-th particles in different spectral lines, so the population size actually also corresponds to the j dimension.

[0113] S320. Correct each particle in the adaptive particle swarm optimization algorithm through the boundary processing function.

[0114] Among them, the boundary processing function is:

[0115]

[0116] Among them, is the i-th particle corresponding to the corrected j-dimensional component, is the i-th particle corresponding to the j-dimensional component before correction, is the preset maximum upper limit value in the j dimension, is the preset minimum lower limit value in the j dimension.

[0117] Specifically, when correcting each particle through the boundary processing function, for example, when or , it is considered that the particle belongs to the particle with extreme deviation, and the value of the particle needs to be corrected. Among them, when , the value of the particle is calculated according to the formula to obtain the corrected value of the particle; when , then according to the formula Calculating the value of the particle yields the corrected value of the particle.

[0118] The setting of the boundary processing function can correct the particles that may have extreme deviations in the laser absorption spectrum before the noise reduction spectrum is updated, so that each particle in the algorithm can be updated in the best direction and position, thereby ensuring that the noise reduction spectrum has a better fitting effect.

[0119] In addition, in step S300, when determining the update direction and update displacement of the noise reduction spectrum in the coordinate system by comparing the laser absorption spectrum and the noise reduction spectrum, it includes: obtaining the laser absorption spectrum corresponding to the corrected particle, and determining the update direction and update displacement of the noise reduction spectrum in the coordinate system by comparing the noise reduction spectrum and the laser absorption spectrum corresponding to the corrected particle.

[0120] Correcting the particles with obvious deviations in the laser absorption spectrum through the boundary function, thereby reducing the influence of extreme noise on the laser absorption spectrum, and determining the update direction and update displacement of the noise reduction spectrum in the coordinate system by comparing the noise reduction spectrum and the corrected laser absorption spectrum, not only makes the fitted noise reduction spectrum have a more realistic line shape, but also can greatly improve the noise reduction effect of the noise reduction spectrum.

[0121] In one embodiment, the specific steps for determining the update direction and update displacement of the noise reduction spectrum in the coordinate system by comparing the laser absorption spectrum and the noise reduction spectrum can refer to the following description:

[0122] First, the self-learning factor, population learning factor, and adjacent learning factor can be set for the laser absorption spectrum in the adaptive particle swarm algorithm respectively. Among them, the self-learning factor is used to characterize the update trend of each particle in each dimension, which can extract all the information in the laser absorption spectrum, thereby greatly reducing the influence of the technician's subjective purpose on the algorithm; the population learning factor is used to characterize the common update trend of any particle in different dimensions, and this learning factor can quickly extract the gas concentration information contained in the laser absorption spectrum; the adjacent learning factor is used to characterize the influence trend of adjacent particles on the current particle in each dimension, especially the influence trend of interference fringes on the laser absorption spectrum. Since the influence of interference fringes on the noise reduction spectrum will not be eliminated by the preprocessing of the above steps S110-S160, but will increase with the increase of the number of updates; the setting of the adjacent learning factor can not only effectively reduce the influence of interference fringes, but also increase the ability of the algorithm to locally search for the optimal solution, thereby accelerating the update convergence speed of the algorithm.

[0123] Secondly, the particle fitness value, population fitness value, and adjacent fitness value corresponding to the self-learning factor, population learning factor, and adjacent learning factor in the laser absorption spectrum can also be calculated respectively.

[0124] Among them, the expression of the particle fitness value is:

[0125]

[0126] The expression of the population fitness value is:

[0127]

[0128] The expression of the adjacent fitness value is:

[0129]

[0130] Among them, x i is the i-th particle corresponding to the j-th dimensional component in the laser absorption spectrum, is the i-th particle in the noise-reduced spectrum, f(x i ) is the value corresponding to x i , is corresponding value, fitness n (x) is the value corresponding to the x-th particle in the j-th population; x i-1 is the (i - 1)-th particle corresponding to the j-th dimensional component in the laser absorption spectrum, x i+1 is the value of the (i + 1)-th particle corresponding to the j-th dimensional component in the laser absorption spectrum, f(x i-1 ) is the value corresponding to x i-1 , f(x i+1 ) is the value corresponding to x i+1 .

[0131] Finally, according to the preset weight strategy, the particle fitness value, the population fitness value, and the adjacent fitness value are weighted and calculated respectively to determine the update direction and the update displacement of the noise-reduced spectrum.

[0132] Here, when weighting and calculating the particle fitness value, the population fitness value, and the adjacent fitness value, the particle weight coefficient, the population weight coefficient, and the adjacent weight coefficient corresponding to the particle fitness value, the population fitness value, and the adjacent fitness value can be calculated respectively according to the preset weight strategy. Among them, the preset weight strategy can be to keep the particle weight coefficient and the population weight coefficient unchanged and gradually decrease the adjacent weight coefficient as the number of updates increases; or keep the particle weight coefficient unchanged and gradually decrease the population weight coefficient and the adjacent weight coefficient. This application has no limitation on this.

[0133] Then, according to the particle weight coefficient, the population weight coefficient, and the adjacent weight coefficient, perform a product calculation on the particle fitness value, the population fitness value, and the adjacent fitness value to obtain an update vector; and determine the absolute value of the update vector as the update displacement of the noise reduction spectrum, and determine the positive or negative of the update vector as the update direction of the noise reduction spectrum. Wherein, a positive value of the update vector indicates that the update direction is upward along the ordinate of the coordinate system, and a negative value of the update vector indicates that the update direction is downward along the ordinate of the coordinate system.

[0134] Preferably, the particle weight coefficient, the population weight coefficient, and the adjacent weight coefficient can be determined according to the following method:

[0135]

[0136] Wherein, if the number of updates of the position of the noise reduction spectrum in the coordinate system reaches the maximum number of updates N, it is determined that the updated noise reduction spectrum meets the preset conditions, and n is the current number of updates; w fitness is the particle weight coefficient, w NP is the population weight coefficient, w AD is the adjacent weight coefficient.

[0137] In addition, as shown in Figure 8 , the present application also provides a noise reduction device for a laser absorption spectrum, including:

[0138] A first determination module 410, configured to determine a laser absorption spectrum formed when a laser irradiates a gas to be measured;

[0139] A second determination module 420, configured to determine a noise reduction spectrum under the current environmental variables according to the noise reduction spectrum function corresponding to the current pressure of the gas chamber where the gas to be measured is located and the relationship between the current temperature and the laser absorption spectrum. The noise reduction spectrum and the laser absorption spectrum are located in the same coordinate system, and the current environmental variables include the current temperature and the current pressure;

[0140] A comparison module 430, configured to determine the update direction and the update displacement of the noise reduction spectrum in the coordinate system by comparing the laser absorption spectrum and the noise reduction spectrum;

[0141] An update module 440, configured to update the position of the noise reduction spectrum in the coordinate system according to the update direction and the update displacement until the updated noise reduction spectrum meets the preset conditions. The noise reduction spectrum that meets the preset conditions is the laser absorption spectrum after noise filtering.

[0142] Based on the exemplary embodiments shown in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present application. In addition, although the disclosure in the present application is introduced according to one or several exemplary instances, it should be understood that each aspect of these disclosures may also constitute a complete technical solution alone.

[0143] It should be understood that the terms "first", "second", "third", etc. in the description, claims and the above drawings of the present application are used to distinguish similar objects and do not necessarily need to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, for example, it can be implemented in an order other than those given in the illustration or description of the embodiments of the present application.

[0144] In addition, the terms "comprising" and "having" and any variations thereof are intended to cover but not exclude inclusion. For example, a product or device comprising a series of components does not necessarily have to be limited to those components clearly listed, but may include other components not clearly listed or inherent to these products or devices.

[0145] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not intended to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A noise reduction method for laser absorption spectroscopy, characterized in that The noise reduction method includes the following steps: Determine the laser absorption spectrum formed when the laser irradiates the gas to be measured; Select a noise reduction spectral function corresponding to the current pressure of the gas chamber where the gas to be measured is located, construct a spectral line intensity compensation function based on the current temperature and the laser absorption spectrum, and perform a product calculation on the noise reduction spectral function and the compensation function to determine the noise reduction spectrum under the current environmental variables. The noise reduction spectrum and the laser absorption spectrum are located in the same coordinate system. The current environmental variables include the current temperature and the current pressure. Wherein, when the current pressure is less than a preset first threshold, the noise reduction spectral function adopts a Gaussian linear function; when the current pressure is greater than a preset second threshold, the noise reduction spectral function adopts a Lorentz line shape function; when the current pressure is between the first threshold and the second threshold, the noise reduction spectral function adopts a Voigt line shape function; By comparing the laser absorption spectrum and the noise reduction spectrum, determine the update direction and update displacement of the noise reduction spectrum in the coordinate system; According to the update direction and the update displacement, update the position of the noise reduction spectrum in the coordinate system until the updated noise reduction spectrum meets the preset conditions. The noise reduction spectrum that meets the preset conditions is the laser absorption spectrum after noise filtering.

2. The noise reduction method for laser absorption spectroscopy according to claim 1, characterized in that, The determination of the laser absorption spectrum formed when the laser irradiates the gas to be measured includes: Obtain the original spectrum generated when the laser irradiates the gas to be measured. The original spectrum corresponds to multiple consecutive periods; Determine the extreme values of each period in the original spectrum through an extreme value search algorithm; Using the determined multiple extreme values as nodes, cut the multiple consecutive periods into multiple discrete periods; Perform an averaging process on the multiple superimposed discrete periods to obtain an averaged spectrum. The averaged spectrum has a gas absorption peak, and the gas absorption peak is the wavelength absorption region corresponding to the gas to be measured in the laser absorption spectrum; Fit the part of the averaged spectrum that does not contain the gas absorption peak to obtain a reference spectrum; Perform a normalization process on the averaged spectrum and the reference spectrum to obtain the laser absorption spectrum.

3. The noise reduction method for laser absorption spectroscopy according to claim 1, characterized in that, The noise reduction spectral function corresponding to the current pressure includes: When the current pressure is less than a preset first threshold, the noise reduction spectral function is: ； Among them, represents the Gaussian line shape function , is the full width at half maximum of the noise reduction spectrum when the current pressure is less than the preset first threshold, with the unit of , is the wave number at the center of the spectral line, with the unit of , is the wave number of the laser, with the unit of ; When the current pressure is greater than a preset second threshold, the noise reduction spectral function is: ； Among them, represents the Lorentz line shape function , is the full width at half maximum of the noise reduction spectrum when the current pressure is greater than the preset second threshold, with the unit of ; is the wave number at the center of the spectral line, with the unit of ; is the wave number of the laser, with the unit of ; When the current pressure is between the first threshold and the second threshold, the noise reduction spectral function is: ； Among them, represents the Volkert line-type function , is and 's convolution, , , .

4. The noise reduction method for laser absorption spectroscopy according to claim 1, characterized in that Before determining the update direction and update displacement of the noise reduction spectrum in the coordinate system by comparing the laser absorption spectrum and the noise reduction spectrum, the noise reduction method further includes: Import the data corresponding to the laser absorption spectrum and the noise reduction spectrum into a pre-constructed adaptive particle swarm optimization algorithm respectively, where the adaptive particle swarm optimization algorithm is configured to include particles, each particle represents a data object, each data object is a spectral point in the laser absorption spectrum, and the total population is , and the maximum number of updates is N; Correct each particle of the laser absorption spectrum through a boundary processing function. The boundary processing function is: Among them, is the th particle of the -dimensional component after correction, is the corresponding th particle of the -dimensional component before correction, is the maximum upper limit value preset in the dimension, is the minimum lower limit value preset in the dimension; The determination of the update direction and update displacement of the noise reduction spectrum in the coordinate system by comparing the laser absorption spectrum and the noise reduction spectrum includes: Obtain the laser absorption spectrum corresponding to the corrected particles, and determine the update direction and update displacement of the noise-reduced spectrum in the coordinate system by comparing the noise-reduced spectrum and the laser absorption spectrum corresponding to the corrected particles.

5. The noise reduction method for laser absorption spectroscopy according to claim 4, wherein The determining of the update direction and update displacement of the noise-reduced spectrum in the coordinate system by comparing the laser absorption spectrum and the noise-reduced spectrum includes: Set the self-learning factor, population learning factor, and adjacent learning factor respectively in the adaptive particle swarm algorithm, where the self-learning factor is used to characterize the update trend of each particle in each dimension, the population learning factor is used to characterize the common update trend of any particle under different dimensions, and the adjacent learning factor is used to characterize the influence trend of adjacent particles on the current particle in each dimension; Calculate the particle fitness value, population fitness value, and adjacent fitness value corresponding to the self-learning factor, population learning factor, and adjacent learning factor in the laser absorption spectrum respectively; Perform weighted calculations on the particle fitness value, population fitness value, and adjacent fitness value respectively according to a preset weight strategy to determine the update direction and update displacement of the noise-reduced spectrum.

6. The noise reduction method for laser absorption spectroscopy according to claim 5, wherein Performing weighted calculations on the particle fitness value, population fitness value, and adjacent fitness value respectively according to a preset weight strategy to determine the update direction and update displacement of the noise-reduced spectrum includes: Calculate the particle weight coefficient, population weight coefficient, and adjacent weight coefficient corresponding to the particle fitness value, population fitness value, and adjacent fitness value respectively according to a preset weight strategy; Perform weighted calculations on the particle fitness value, population fitness value, and adjacent fitness value according to the particle weight coefficient, population weight coefficient, and adjacent weight coefficient to obtain the update direction and update displacement of the noise-reduced spectrum.

7. A noise reduction method for laser absorption spectroscopy according to claim 6, characterized in that, The particle weight coefficient, population weight coefficient, and adjacent weight coefficient are determined according to the following method: Among them, if the number of updates of the position of the noise reduction spectrum in the coordinate system reaches the maximum number of updates , it is determined that the updated noise reduction spectrum meets the preset conditions, which is the current number of updates; is the particle weight coefficient applied when calculating the weighted value of the particle fitness value, is the population weight coefficient applied when calculating the weighted value of the population weight coefficient, is the adjacent weight coefficient applied when calculating the weighted value of the adjacent fitness values; Determine the update direction and update displacement of the noise-reduced spectrum according to the weighted calculation result.

8. A noise reduction method for laser absorption spectroscopy according to claim 7, characterized in that, The determining of the update direction and update displacement of the noise-reduced spectrum according to the weighted calculation result includes: After performing a product calculation on the particle fitness value, population fitness value, and adjacent fitness value to obtain an update vector, determine the absolute value of the update vector as the update displacement of the noise-reduced spectrum, and determine the positive or negative of the update vector as the update direction of the noise-reduced spectrum, where a positive value indicates upward along the ordinate of the coordinate system, and a negative value indicates downward along the ordinate of the coordinate system.

9. A noise reduction device for laser absorption spectroscopy, characterized in that, Includes: A first determination module for determining the laser absorption spectrum formed when a laser irradiates a gas to be measured; A second determination module, configured to select a noise reduction spectral function corresponding to the current pressure of the gas chamber where the gas to be measured is located, construct a spectral line intensity compensation function based on the current temperature and the laser absorption spectrum, and perform a multiplication calculation on the noise reduction spectral function and the compensation function to determine a noise reduction spectrum under the current environmental variables, where the noise reduction spectrum and the laser absorption spectrum are located in the same coordinate system, and the current environmental variables include the current temperature and the current pressure; wherein, when the current pressure is less than a preset first threshold, the noise reduction spectral function adopts a Gaussian linear function; when the current pressure is greater than a preset second threshold, the noise reduction spectral function adopts a Lorentzian line shape function; when the current pressure is between the first threshold and the second threshold, the noise reduction spectral function adopts a Voigt line shape function; A comparison module, configured to determine an update direction and an update displacement amount of the noise reduction spectrum in the coordinate system by comparing the laser absorption spectrum and the noise reduction spectrum; An update module, configured to update the position of the noise reduction spectrum in the coordinate system according to the update direction and the update displacement amount until the updated noise reduction spectrum meets a preset condition, and the noise reduction spectrum that meets the preset condition is the laser absorption spectrum after noise filtering.

Images