Method and equipment for simultaneously detecting at least two target gases in exhaled gases
By combining differential absorption spectral feature space and multi-scale convolutional network, the problem of spectral overlap between ammonia and isoprene in exhaled gas is solved, achieving high-precision simultaneous detection, which is suitable for online detection of multiple components in exhaled gas.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-11
- Publication Date
- 2026-03-27
AI Technical Summary
Existing technologies struggle to simultaneously and accurately detect the concentrations of ammonia and isoprene in exhaled breath, especially in the ultraviolet band where there is spectral overlap and high noise interference.
By constructing a feature space for differential absorption spectra and using a multi-scale convolutional network for spectral decoupling, combined with optical and data units, the differential absorption spectra of ammonia and isoprene in exhaled gas can be separated and their concentrations measured.
Under complex composition and spectral overlap conditions, high-precision simultaneous detection of ammonia and isoprene was achieved, with measurement errors of less than 3.09% and 1.76%, respectively.
Smart Images
Figure CN121740784A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of gas detection, and in particular to a method and device for simultaneously detecting at least two target gases in exhaled breath. BACKGROUND
[0002] The concentration of various components in exhaled breath reflects the physiological state of the human body. By detecting the concentration of various disease markers in exhaled breath, the health status of the human body can be monitored, for example, detecting the concentration of ammonia in exhaled breath can assist in monitoring the state of patients with kidney disease, and detecting the concentration of isoprene in exhaled breath can help diagnose the physiological health status of lung cancer patients. The concentration of NH3 in the exhaled breath of patients with kidney disease is significantly higher than that of healthy people, usually reaching ppm level. The concentration of isoprene in the exhaled breath of lung cancer patients is usually less than 50 ppb, deviating from the normal exhaled isoprene concentration of more than 50 ppb. Therefore, it is of great significance to monitor the concentration of ammonia and isoprene in exhaled breath. At the same time, these two gases are also common pollution gases in the atmosphere, and the concentration of ammonia and isoprene in the atmosphere can reflect the air quality and environmental pollution.
[0003] Chromatography is a process that separates the components of a mixture using solubility, adsorption and other properties of substances. However, it is time-consuming and cannot be used for real-time and continuous measurements on site. Electrochemical methods measure concentration by generating electrical signals through reactions with gas molecules, but often have the disadvantages of low selectivity, short detection life, and the need to contact the measured gas. In contrast, optical detection methods have the advantages of high sensitivity and fast analysis speed, and can perform non-contact real-time measurements on site. In the ultraviolet band, ammonia and isoprene have significant characteristic absorption, which makes ultraviolet absorption spectroscopy a suitable method for measuring these component gases. The principle is that gas molecules absorb light energy at a specific wavelength, producing a unique band-shaped absorption spectrum. Jan van den Broek et al. proposed a non-specific nano-structured SnO2 sensor system doped with platinum, which can achieve a detection limit of 5 ppb for isoprene. Zhu Rui et al. proposed a model based on ultraviolet absorption spectroscopy and feature enhancement technology, which achieved a minimum ammonia measurement of 9.50 ppb. Li Qingyuan et al. measured the absorption characteristic peak of isoprene in the ultraviolet band based on cavity ring-down spectroscopy, achieving a detection limit of 0.47 ppb. Although ppb-level ammonia and isoprene measurements can be achieved, simultaneous detection of isoprene and ammonia in exhaled breath is still difficult to achieve. This is because the gas components in exhaled breath are complex, and there is a problem of spectral overlap between ammonia and isoprene in the ultraviolet band, which makes spectral decoupling and high-precision measurement at the ppb level still difficult. SUMMARY
[0004] The embodiment of the present application provides a method and equipment for simultaneously detecting at least two target gases in exhaled gas, so as to solve the problems of high noise interference caused by low concentration and spectral overlap caused by complex components when simultaneously detecting multiple target gases.
[0005] In a first aspect, the embodiment of the present application provides a method for simultaneously detecting at least two target gases in exhaled gas, wherein the target gases at least include a first target gas and a second target gas, and the method comprises the following steps: constructing a first feature space according to a differential absorption spectrum of the first target gas under a standard state, and constructing a second feature space according to a differential absorption spectrum of the second target gas under the standard state; collecting an exhaled gas sample to obtain a differential absorption spectrum of the exhaled gas sample; projecting the differential absorption spectrum of the exhaled gas sample into the first feature space and the second feature space respectively to perform spectral decoupling, so as to obtain a first differential absorption spectrum of the first target gas and a second differential absorption spectrum of the second target gas in the exhaled gas sample; inputting the first differential absorption spectrum into a first multi-scale convolution network and inputting the second differential absorption spectrum into a second multi-scale convolution network, so as to respectively obtain a concentration of the first target gas and a concentration of the second target gas in the exhaled gas sample; wherein the first multi-scale convolution network is obtained by training a multi-scale convolution network model through a plurality of groups of differential absorption spectra of the first target gas with different concentrations, and the second multi-scale convolution network is obtained by training a multi-scale convolution network model through a plurality of groups of differential absorption spectra of the second target gas with different concentrations.
[0006] In a possible implementation, the step of constructing a first feature space according to a differential absorption spectrum of the first target gas under a standard state and constructing a second feature space according to a differential absorption spectrum of the second target gas under the standard state comprises: constructing a first projection matrix according to the differential absorption spectrum of the first target gas under the standard state, and taking the first projection matrix as the first feature space; constructing a second projection matrix according to the differential absorption spectrum of the second target gas under the standard state, and taking the second projection matrix as the second feature space.
[0007] In a possible implementation, the step of constructing a first projection matrix according to the differential absorption spectrum of the first target gas under the standard state comprises: constructing the first projection matrix according to ; The step of constructing a second projection matrix according to the differential absorption spectrum of the second target gas under the standard state comprises: According to , a second projection matrix is constructed; wherein, is the first projection matrix, is the second projection matrix, is a differential absorption spectrum of the first target gas in a standard state, is a differential absorption spectrum of the second target gas in a standard state.
[0008] In a possible implementation, the differential absorption spectrum of the exhaled gas sample is projected into the first feature space and the second feature space respectively for spectral decoupling, to obtain a first differential absorption spectrum of the first target gas and a second differential absorption spectrum of the second target gas in the exhaled gas sample, comprising: projecting the differential absorption spectrum of the exhaled gas sample into the first projection matrix to obtain the first differential absorption spectrum of the first target gas in the exhaled gas sample; projecting the differential absorption spectrum of the exhaled gas sample into the second projection matrix to obtain the second differential absorption spectrum of the second target gas in the exhaled gas sample.
[0009] In a possible implementation, projecting the differential absorption spectrum of the exhaled gas sample into the first projection matrix to obtain the first differential absorption spectrum of the first target gas in the exhaled gas sample, comprising: According to obtaining the first differential absorption spectrum of the first target gas in the exhaled gas sample; projecting the differential absorption spectrum of the exhaled gas sample into the second projection matrix to obtain the second differential absorption spectrum of the second target gas in the exhaled gas sample, comprising: According to obtaining the second differential absorption spectrum of the second target gas in the exhaled gas sample; wherein, is the first differential absorption spectrum of the first target gas in the exhaled gas sample, is the second differential absorption spectrum of the second target gas in the exhaled gas sample, is the differential absorption spectrum of the exhaled gas sample, is the first projection matrix, is the second projection matrix.
[0010] In a possible implementation, before constructing the first feature space according to the differential absorption spectrum of the first target gas in a standard state, further comprising: acquire the ultraviolet absorption spectrum of the first target gas in a standard state, process the ultraviolet absorption spectrum of the first target gas in a standard state by using a differential absorption spectrum technology, and obtain the differential absorption spectrum of the first target gas in a standard state; Before constructing the second feature space according to the differential absorption spectrum of the second target gas in a standard state, the method further comprises: acquire the ultraviolet absorption spectrum of the second target gas in a standard state, process the ultraviolet absorption spectrum of the second target gas in a standard state by using a differential absorption spectrum technology, and obtain the differential absorption spectrum of the second target gas in a standard state.
[0011] In a possible implementation, the first target gas is ammonia, and the second target gas is isoprene gas.
[0012] In a possible implementation, the standard state is a state in which the signal-to-noise ratio of the ultraviolet absorption spectrum of the first target gas and the ultraviolet absorption spectrum of the second target gas is greater than 20 dB.
[0013] In a possible implementation, the concentration range of the several groups of different concentrations of the first target gas is 20-2000 ppb, and the concentration range of the several groups of different concentrations of the second target gas is 20-1800 ppb.
[0014] In a second aspect, an embodiment of the present application provides a device for simultaneously detecting at least two target gases in exhaled gas, comprising a light path unit, a gas path unit, and a data unit. The gas path unit comprises a first gas path provided with a gas mixing instrument and a second gas path provided with a flow control air pump; the gas mixing instrument is connected with a first target gas storage tank, a second target gas storage tank, and a nitrogen gas storage tank, respectively, and is used to configure the first target gas in a standard state, the second target gas in a standard state, the several groups of different concentrations of the first target gas, and the several groups of different concentrations of the second target gas; the flow control air pump is connected with a gas sample collection bag for storing the exhaled gas sample; The light path unit comprises a deuterium lamp light source, a convex lens, a gas cell, a light shield, a fiber collimator, an optical fiber, and a spectrometer arranged in sequence; the front end of the gas cell is connected with the first gas path and the second gas path, respectively, and the rear end of the gas cell is connected with a waste gas cell; the light path unit is used to obtain the ultraviolet absorption spectrum of a to-be-measured gas; the to-be-measured gas comprises the first target gas in a standard state, the second target gas in a standard state, the several groups of different concentrations of the first target gas, the several groups of different concentrations of the second target gas, and the exhaled gas sample. The data unit comprises a computer; the computer is connected with the spectrometer; the computer comprises a memory and a processor; the memory stores a computer program; and the processor implements the method of the first aspect or any possible implementation manner of the first aspect when executing the computer program.
[0015] In the embodiment of the present application, by establishing a feature space (including a first feature space and a second feature space), the differential absorption spectrum of the exhaled gas sample is directly projected into two different feature spaces to realize direct decoupling of the spectrum; and a multi-scale convolution network (including a first multi-scale convolution network and a second multi-scale convolution network) is established to obtain the corresponding relationship between the component concentration and the differential absorption spectrum, and then the multi-scale convolution network is used to realize simultaneous detection of multiple components in the exhaled gas sample, thereby realizing online detection of multiple components of the exhaled gas under simple equipment conditions. BRIEF DESCRIPTION OF DRAWINGS
[0016] Figure 1 is an implementation flowchart of the method for simultaneously detecting at least two target gases in the exhaled gas provided by the embodiment of the present application; Figure 2 is an implementation flowchart of the method for simultaneously detecting ammonia and isoprene in the exhaled gas provided by the embodiment of the present application; Figure 3 is a structural schematic diagram of the device for simultaneously detecting at least two target gases in the exhaled gas provided by the embodiment of the present application; 1, deuterium lamp light source, 2, convex lens, 3, gas cell, 4, light shield, 5, optical fiber collimator, 6, optical fiber, 7, spectrometer, 8, computer, 9, waste gas cell, 10, gas mixing instrument, 11, first target gas storage tank, 12, second target gas storage tank, 13, nitrogen storage tank, 14, gas sample collection bag, 15, flow control air pump. DETAILED DESCRIPTION
[0017] The embodiment of the present application will be described in detail below with reference to the accompanying drawings.
[0018] Figure 1 The implementation flowchart of the method for simultaneously detecting at least two target gases in the exhaled gas provided by the embodiment of the present application is described in detail as follows: In step 101, a first feature space is constructed according to the differential absorption spectrum of the first target gas under standard conditions, and a second feature space is constructed according to the differential absorption spectrum of the second target gas under standard conditions.
[0019] In the embodiment, the ultraviolet absorption spectrum of the first target gas under standard conditions is obtained, and the differential absorption spectrum technology is used to process the ultraviolet absorption spectrum of the first target gas under standard conditions to obtain the differential absorption spectrum of the first target gas under standard conditions.
[0020] According to the differential absorption spectrum of the first target gas in the standard state, a first projection matrix is constructed, and the first projection matrix is taken as a first feature space, comprising: According to , the first projection matrix is constructed.
[0021] An ultraviolet absorption spectrum of a second target gas in the standard state is obtained, and the ultraviolet absorption spectrum of the second target gas in the standard state is processed using a differential absorption spectrum technology to obtain a differential absorption spectrum of the second target gas in the standard state.
[0022] According to the differential absorption spectrum of the second target gas in the standard state, a second projection matrix is constructed, and the second projection matrix is taken as a second feature space, comprising: According to , the second projection matrix is constructed.
[0023] Wherein, the first projection matrix is the second projection matrix is the differential absorption spectrum of the first target gas in the standard state is the differential absorption spectrum of the second target gas in the standard state is.
[0024] In step 102, the exhaled gas sample is collected, and the differential absorption spectrum of the exhaled gas sample is obtained.
[0025] In step 103, the differential absorption spectrum of the exhaled gas sample is projected into the first feature space and the second feature space respectively to decouple the spectrum, and the first differential absorption spectrum of the first target gas and the second differential absorption spectrum of the second target gas in the exhaled gas sample are obtained.
[0026] In this embodiment, the differential absorption spectrum of the exhaled gas sample is projected into the first projection matrix to obtain the first differential absorption spectrum of the first target gas in the exhaled gas sample, comprising: According to the first differential absorption spectrum of the first target gas in the exhaled gas sample is obtained.
[0027] The differential absorption spectrum of the exhaled gas sample is projected into the second projection matrix to obtain the second differential absorption spectrum of the second target gas in the exhaled gas sample, comprising: According to the second differential absorption spectrum of the second target gas in the exhaled gas sample is obtained.
[0028] Wherein, the first differential absorption spectrum of the first target gas in the exhaled gas sample is a second differential absorption spectrum of a second target gas in the exhaled gas sample, a differential absorption spectrum of the exhaled gas sample, a first projection matrix, a second projection matrix.
[0029] In step 104, the first differential absorption spectrum is input into the first multi-scale convolution network, and the second differential absorption spectrum is input into the second multi-scale convolution network to obtain the concentration of the first target gas and the concentration of the second target gas in the exhaled gas sample respectively, wherein the first multi-scale convolution network is obtained by training a multi-scale convolution network model through a plurality of groups of differential absorption spectra of the first target gas with different concentrations, and the second multi-scale convolution network is obtained by training a multi-scale convolution network model through a plurality of groups of differential absorption spectra of the second target gas with different concentrations.
[0030] The method for simultaneously detecting at least two target gases in exhaled gas provided by the embodiment of the present application is described below through specific examples: Referring to Figure 2 , which shows the implementation flowchart of the method for simultaneously detecting ammonia and isoprene in exhaled gas based on absorption spectrum provided by the embodiment of the present application, which is described in detail as follows: The first concentration ammonia gas and the second concentration isoprene gas are configured, the first feature space is constructed according to the differential absorption spectrum of the first concentration ammonia gas, and the second feature space is constructed according to the differential absorption spectrum of the second concentration isoprene gas.
[0031] In this embodiment, the signal-to-noise ratio of the ultraviolet absorption spectrum of the first concentration ammonia gas and the second concentration isoprene gas is greater than 20 dB.
[0032] In this embodiment, the ultraviolet absorption spectrum of the first concentration ammonia gas is collected, and the differential absorption spectrum of the first concentration ammonia gas is obtained by using the differential absorption spectrum method, including: The ultraviolet absorption spectrum of the first concentration ammonia gas and the ultraviolet absorption spectrum of the second concentration isoprene gas are respectively subjected to polynomial fitting to obtain the slow-varying absorption part of ammonia and the slow-varying absorption part of isoprene .
[0033] After filtering out the slow-varying absorption part of ammonia from the ultraviolet absorption spectrum of the first concentration ammonia gas , the obtained fast-varying absorption part of ammonia is taken as the differential absorption spectrum of the first concentration ammonia gas, that is, the differential absorption spectrum of the first target gas :
[0034] Obtain the first projection matrix projected onto the first feature space. :
[0035] The ultraviolet absorption spectrum of the second concentration of isoprene gas was collected, and the differential absorption spectrum of the second concentration of isoprene gas was obtained using differential absorption spectroscopy, including: UV absorption spectrum of isoprene gas at the second concentration Filtering out the slow-change absorption portion of isoprene gas Subsequently, the rapidly varying absorption portion of the obtained isoprene gas was used as the differential absorption spectrum of the second concentration of isoprene gas, i.e., the differential absorption spectrum of the second target gas. :
[0036] Obtain the second projection matrix projected onto the second feature space. :
[0037] Exhaled gas samples were collected, and differential absorption spectra of the exhaled gas samples were obtained.
[0038] The differential absorption spectra of the exhaled gas sample are projected onto the first feature space and the second feature space, respectively, to perform spectral decoupling and obtain the single differential absorption spectra of ammonia and isoprene in the exhaled gas sample.
[0039] In this embodiment, the differential absorption spectral data of the exhaled gas sample Projection to the first matrix Projection was used to obtain the differential absorption spectrum of ammonia in the exhaled gas sample; the differential absorption spectrum data of the exhaled gas sample were then processed. Projection to the second matrix Projection was used to obtain the differential absorption spectrum of isoprene gas in the exhaled gas sample.
[0040] The purpose of establishing a projection matrix is to decouple and obtain the absorption spectra of individual components in the detection of mixed gases. This is based on the principle that the absorption spectral characteristics of each component gas are unique. Therefore, the absorption spectra of the mixed gas can be directionally projected onto the characteristic space spanned by the absorption spectra of each individual component gas to obtain the absorption spectra of the individual components. An application scenario is that the absorption spectra of exhaled gas components overlap in the ultraviolet band; the projection matrix is used to obtain the absorption spectra of the ammonia / isoprene individual components in exhaled gas.
[0041] The ammonia differential absorption spectrum is input into a first multi-scale convolution network, and the differential absorption spectrum of the isoprene gas is input into a second multi-scale convolution network to obtain the concentration of ammonia and the concentration of isoprene gas in the exhaled gas sample, respectively, wherein the first multi-scale convolution network is obtained by training a multi-scale convolution network model through a plurality of groups of differential absorption spectra of ammonia with different concentrations, and the second multi-scale convolution network is obtained by training a multi-scale convolution network model through a plurality of groups of differential absorption spectra of isoprene gas with different concentrations.
[0042] In this embodiment, the multi-scale convolution network model comprises: A first convolution-pooling-activation module is configured to extract differential absorption spectrum features and narrow-band absorption peaks.
[0043] A second convolution-pooling-activation module is configured to extract local shape information of the differential absorption spectrum features.
[0044] A third convolution-pooling-activation module is configured to extract differential absorption structures of a wide band of the differential absorption spectrum.
[0045] A splicing module is configured to splice the spectral feature vectors output by the first convolution-pooling-activation module, the second convolution-pooling-activation module, and the third convolution-pooling-activation module to form a fusion spectral feature covering the details and the overall structure.
[0046] A fully connected regression layer is configured to inversely calculate the target gas concentration from the fusion spectral feature.
[0047] A plurality of groups of ammonia with different concentrations and a plurality of groups of isoprene gas with different concentrations are configured, and the gas concentrations are shown in Table 1: Table 1
[0048] The obtained ultraviolet absorption spectrum is processed using the differential absorption spectrum technology (DOAS) to obtain differential absorption spectrum data of ammonia and differential absorption spectrum data of isoprene .
[0049] In this embodiment, the first training set is obtained by extending the differential absorption spectra of a plurality of groups of ammonia with different concentrations through a linear superposition method, and the first multi-scale convolution network is obtained by training a multi-scale convolution network model through the first training set.
[0050] The second training set is obtained by extending the differential absorption spectra of a plurality of groups of isoprene gas with different concentrations through a linear superposition method, and the second multi-scale convolution network is obtained by training a multi-scale convolution network model through the second training set.
[0051] It should be understood that the size of the serial number of each step in the above embodiment does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiment of the application.
[0052] The following is an equipment embodiment of the application, and for details not described in detail, reference can be made to the corresponding method embodiments described above.
[0053] The embodiment of the application also provides an apparatus for simultaneously detecting at least two target gases in exhaled gas, comprising a memory and a processor, the memory storing a computer program, and the processor realizing the method in the above method embodiment when executing the computer program.
[0054] Exemplarily, Figure 3 A structure diagram of the apparatus for simultaneously detecting ammonia and isoprene in exhaled gas based on absorption spectrum is shown, only parts related to the embodiment of the application are shown for the convenience of description, and the details are as follows: The apparatus for simultaneously detecting ammonia and isoprene in exhaled gas based on absorption spectrum comprises a light path unit, a gas path unit and a data unit.
[0055] The gas path unit comprises a first gas path provided with a gas mixing instrument 10 and a second gas path provided with a flow control gas pump 15; the gas mixing instrument is connected with a first target gas storage tank 11, a second target gas storage tank 12 and a nitrogen storage tank 13 respectively, and is used for configuring the first target gas in a standard state, the second target gas in a standard state, a plurality of groups of first target gases with different concentrations and a plurality of groups of second target gases with different concentrations; the flow control gas pump 15 is connected with a gas sample collection bag 14 for storing exhaled gas samples.
[0056] In the embodiment, exemplarily, the first target gas storage tank 11 is an ammonia storage tank, and the second target gas storage tank 12 is an isoprene storage tank.
[0057] Exemplarily, a plurality of groups of ammonia gases with different concentrations are configured by the gas mixing instrument 10 based on 15.6 ppm ammonia gas and pure nitrogen gas, and a plurality of groups of isoprene gases with different concentrations are configured by the gas mixing instrument 10 based on 11.4 ppm isoprene gas and pure nitrogen gas, at this time, the flow control gas pump 15 is closed, and the configured gas is output to the gas cell 3 by the gas mixing instrument 10.
[0058] The light path unit comprises a deuterium lamp light source 1, a convex lens 2, a gas cell 3, a light shield 4, a fiber collimator 5, an optical fiber 6 and a spectrometer 7 arranged in sequence; the front end of the gas cell 3 is connected with the first gas path and the second gas path respectively, and the rear end of the gas cell 3 is connected with a waste gas cell 9; the light path unit is used for obtaining the ultraviolet absorption spectrum of the gas to be measured; the gas to be measured includes the first target gas under standard state, the second target gas under standard state, a plurality of groups of different concentrations of the first target gas, a plurality of groups of different concentrations of the second target gas and the exhaled gas sample.
[0059] The data unit comprises a computer 8; the computer 8 is connected with the spectrometer 7; the computer comprises a memory and a processor; the memory stores a computer program, and the processor executes the method for simultaneously detecting at least two target gases in exhaled gas.
[0060] For example, the flow control air pump 15 is closed, the light emitted by the deuterium lamp light source 1 is converted into parallel light by the quartz convex lens 2, then passes through the 50 cm long gas absorption cell 3 filled with ammonia gas or isoprene gas under standard state respectively, the transmitted light passes through the light shield 4 and is guided into the optical fiber 6 by the optical fiber coupler 5, the ultraviolet absorption spectrum of the ammonia gas or isoprene gas under standard state is collected by the spectrometer 7 and transmitted to the computer 8, the computer 8 obtains the differential absorption spectrum of the ammonia gas and isoprene gas under standard state, and establishes a first characteristic space and a second characteristic space.
[0061] The light emitted by the deuterium lamp light source 1 is converted into parallel light by the quartz convex lens 2, then passes through the 50 cm long gas absorption cell 3 filled with a plurality of groups of different concentrations of ammonia gas or isoprene gas respectively, the transmitted light passes through the light shield 4 and is guided into the optical fiber 6 by the optical fiber coupler 5, the ultraviolet absorption spectrum of the plurality of groups of different concentrations of ammonia gas or isoprene gas is collected by the spectrometer 7 and transmitted to the computer 8, the computer 8 obtains the first differential absorption spectrum of the ammonia gas and the second differential absorption spectrum of the isoprene gas, and obtains the first multi-scale convolution network and the second multi-scale convolution network by training the multi-scale convolution network model in the computer 8.
[0062] The flow control air pump 15 is opened, the light emitted by the deuterium lamp light source 1 is converted into parallel light by the quartz convex lens 2, then passes through the 50 cm long gas absorption cell 3 filled with the exhaled gas sample, the transmitted light passes through the light shield 4 and is guided into the optical fiber 6 by the optical fiber coupler 5, the ultraviolet absorption spectrum of the exhaled gas sample is collected by the spectrometer 7 and transmitted to the computer 8, the computer 8 obtains the first differential absorption spectrum of the ammonia gas and the second differential absorption spectrum of the isoprene gas, inputs the first differential absorption spectrum into the first multi-scale convolution network and inputs the second differential absorption spectrum into the second multi-scale convolution network, and obtains the concentration of the ammonia gas and the concentration of the isoprene gas in the exhaled gas sample respectively.
[0063] The comparison between the detection results of the method for simultaneously detecting ammonia and isoprene in exhaled breath based on absorption spectrum and the actual concentration of the gas is shown in Table 2, which shows the accuracy analysis of the measurement results of 5 groups of mixed gas components with different concentrations of ammonia and isoprene. In this embodiment, the maximum relative measurement error of ammonia obtained by measuring 5 groups of mixed gas with different concentrations by the new method proposed in the application is 3.09%, and the maximum relative measurement error of isoprene is 1.76%. The experimental results prove that the method for measuring trace ammonia and isoprene proposed in the application based on absorption spectrum and spectral projection decoupling method combined with neural network can decouple the absorption spectrum of single component gas under the condition of serious spectral overlap and noise interference, and realize high-precision and stable measurement.
[0064] Table 2
[0065] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in a certain embodiment can be referred to the related description of other embodiments. If there is no special description and logical conflict, the terms and / or descriptions of different embodiments are consistent and can be mutually referred to, and the technical features in different embodiments can be combined to form new embodiments according to their inherent logical relationship.
[0066] The above embodiments are only used to illustrate the technical solutions of the present application, but not limit it; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.
Claims
1. A method for simultaneously detecting at least two target gases in exhaled breath, characterized in that, The target gas includes at least a first target gas and a second target gas, and the method includes: A first feature space is constructed based on the differential absorption spectrum of the first target gas under standard conditions, and a second feature space is constructed based on the differential absorption spectrum of the second target gas under standard conditions. Collect exhaled gas samples and obtain the differential absorption spectrum of the exhaled gas samples; The differential absorption spectrum of the exhaled gas sample is projected onto the first feature space and the second feature space respectively to perform spectral decoupling, thereby obtaining the first differential absorption spectrum of the first target gas and the second differential absorption spectrum of the second target gas in the exhaled gas sample. The first differential absorption spectrum is input into a first multi-scale convolutional network, and the second differential absorption spectrum is input into a second multi-scale convolutional network to obtain the concentrations of the first target gas and the second target gas in the exhaled gas sample, respectively. The first multi-scale convolutional network is obtained by training a multi-scale convolutional network model with several sets of differential absorption spectra of the first target gas at different concentrations, and the second multi-scale convolutional network is obtained by training a multi-scale convolutional network model with several sets of differential absorption spectra of the second target gas at different concentrations.
2. The method for simultaneous detection of at least two target gases in exhaled gas according to claim 1, characterized in that, A first feature space is constructed based on the differential absorption spectrum of the first target gas under standard conditions, and a second feature space is constructed based on the differential absorption spectrum of the second target gas under standard conditions, including: Based on the differential absorption spectrum of the first target gas under standard conditions, a first projection matrix is constructed, and the first projection matrix is used as the first feature space. Based on the differential absorption spectrum of the second target gas under standard conditions, a second projection matrix is constructed, and the second projection matrix is used as the second feature space.
3. The method for simultaneous detection of at least two target gases in exhaled gas according to claim 2, characterized in that, The step of constructing a first projection matrix based on the differential absorption spectrum of the first target gas under standard conditions includes: according to Construct the first projection matrix; The step of constructing a second projection matrix based on the differential absorption spectrum of the second target gas under standard conditions includes: according to Construct the second projection matrix; in, The first projection matrix, This is the second projection matrix. The differential absorption spectrum of the first target gas under standard conditions. This is the differential absorption spectrum of the second target gas under standard conditions.
4. The method for simultaneous detection of at least two target gases in exhaled gas according to claim 2, characterized in that, The differential absorption spectrum of the exhaled gas sample is projected onto the first feature space and the second feature space respectively to perform spectral decoupling, thereby obtaining the first differential absorption spectrum of the first target gas and the second differential absorption spectrum of the second target gas in the exhaled gas sample, including: The differential absorption spectrum of the exhaled gas sample is projected onto the first projection matrix to obtain the first differential absorption spectrum of the first target gas in the exhaled gas sample. The differential absorption spectrum of the exhaled gas sample is projected onto the second projection matrix to obtain the second differential absorption spectrum of the second target gas in the exhaled gas sample.
5. The method for simultaneous detection of at least two target gases in exhaled gas according to claim 4, characterized in that, Projecting the differential absorption spectrum of the exhaled gas sample onto the first projection matrix to obtain the first differential absorption spectrum of the first target gas in the exhaled gas sample includes: according to Obtain the first differential absorption spectrum of the first target gas in the exhaled gas sample; Projecting the differential absorption spectrum of the exhaled gas sample onto the second projection matrix to obtain the second differential absorption spectrum of the second target gas in the exhaled gas sample includes: according to Obtain the second differential absorption spectrum of the second target gas in the exhaled gas sample; in, The first differential absorption spectrum of the first target gas in the exhaled gas sample. The second differential absorption spectrum of the second target gas in the exhaled gas sample. The differential absorption spectrum of the exhaled gas sample. The first projection matrix, This is the second projection matrix.
6. The method for simultaneous detection of at least two target gases in exhaled gas according to claim 1, characterized in that, Before constructing the first feature space based on the differential absorption spectrum of the first target gas under standard conditions, the method further includes: The ultraviolet absorption spectrum of the first target gas under standard conditions is obtained, and the ultraviolet absorption spectrum of the first target gas under standard conditions is processed using differential absorption spectroscopy to obtain the differential absorption spectrum of the first target gas under standard conditions. Before constructing the second characteristic space based on the differential absorption spectrum of the second target gas under standard conditions, the method further includes: The ultraviolet absorption spectrum of the second target gas under standard conditions is obtained, and the ultraviolet absorption spectrum of the second target gas under standard conditions is processed using differential absorption spectroscopy to obtain the differential absorption spectrum of the second target gas under standard conditions.
7. The method for simultaneous detection of at least two target gases in exhaled gas according to claim 1, characterized in that, The first target gas is ammonia, and the second target gas is isoprene.
8. The method for simultaneous detection of at least two target gases in exhaled gas according to claim 1, characterized in that, The standard state is a state in which the signal-to-noise ratio of the ultraviolet absorption spectrum of the first target gas and the ultraviolet absorption spectrum of the second target gas are both greater than 20dB.
9. The method for simultaneous detection of at least two target gases in exhaled gas according to claim 1, characterized in that, The concentration range of the first target gas in several groups of different concentrations is 20 to 2000 ppb, and the concentration range of the second target gas in several groups of different concentrations is 20 to 1800 ppb.
10. A device for simultaneously detecting at least two target gases in exhaled breath, characterized in that, It includes an optical path unit, an air path unit, and a data unit; The gas path unit includes a first gas path equipped with a gas mixing device and a second gas path equipped with a flow-controlled gas pump; the gas mixing device is connected to a first target gas storage tank, a second target gas storage tank and a nitrogen storage tank respectively, and is used to configure the first target gas under standard conditions, the second target gas under standard conditions, several sets of the first target gas at different concentrations and several sets of the second target gas at different concentrations; the flow-controlled gas pump is connected to a gas sample collection bag storing the exhaled gas sample; The optical path unit includes a deuterium lamp light source, a convex lens, a gas cell, a light shield, a fiber collimating lens, an optical fiber, and a spectrometer arranged sequentially. The front end of the gas cell is connected to the first gas path and the second gas path, respectively, and the rear end of the gas cell is connected to the exhaust gas cell. The optical path unit is used to obtain the ultraviolet absorption spectrum of the gas to be tested. The gas to be tested includes the first target gas under standard conditions, the second target gas under standard conditions, several sets of the first target gas at different concentrations, several sets of the second target gas at different concentrations, and the exhaled gas sample. The data unit includes a computer; the computer is connected to the spectrometer; the computer includes a memory and a processor; the memory stores a computer program, and the processor executes the computer program to implement the method as described in any one of claims 1 to 9.