High-concentration gas detection method and system based on direct absorption spectrum

Abstract

The invention discloses a high-concentration gas detection method and system based on a direct absorption spectrum. The method comprises the following steps: acquiring a direct absorption transmission spectrum signal of gas to be detected; performing baseline fitting processing on the direct absorption transmission spectrum signal of the to-be-detected gas by adopting a low-order polynomial to obtain an absorption spectrum signal of the to-be-detected gas; carrying out peak value normalization processing on the absorption spectrum signal of the gas to be detected; inputting the normalized absorption spectrum signal of the to-be-detected gas into a pre-trained deep learning concentration detection model for processing to obtain a predicted concentration value of the to-be-detected gas; according to the method, a deep learning technology of a convolutional neural network is combined with full-spectrum characteristics, so that the problem of insufficient gas detection precision of a direct absorption spectrum in a high-concentration range is solved, and the detection requirements of high precision, high resolution and the high-concentration range are met.

Description

Technical Field

[0001] The present invention relates to the field of gas detection technology, and more specifically, to a high-concentration gas detection method and system based on direct absorption spectroscopy. Background Art

[0002] In many application scenarios, high concentration and wide range gas detection provides more reliable safety protection compared to the detection limit of parts per billion (ppb). Especially in tunnel engineering, mining operations, transportation and use of natural gas, the explosion limits of some gases are relatively wide. Therefore, gas sensors that can cover a wide concentration range are not only of great significance in scientific research, but also meet the needs of practical applications. However, traditional electrochemical sensors have problems such as short service life, poor long-term stability and low detection accuracy. In contrast, fiber optic sensing technology has significant advantages in the field of gas detection with its high transmission speed, high sensitivity and safety and explosion-proof. In particular, tunable diode laser absorption spectroscopy (TDLAS) has been widely used in gas concentration detection due to its non-contact, narrow linewidth, strong anti-interference ability and real-time online detection.

[0003] At present, wavelength modulation spectroscopy (WMS) and direct absorption spectroscopy (DAS) are commonly used in the TDLAS field. However, it should be noted that when the absorbance is close to 0.05 or greater than 0.05, the second harmonic peak of the absorption signal measured by the WMS method is no longer linearly related to the gas concentration, and the gas concentration cannot be accurately inverted. When using the DAS method to detect gas concentrations in a high concentration range, although it has an advantage over the WMS method in measuring large absorbances, it also faces the problem of nonlinear absorption spectra under large absorbances, which makes it impossible to accurately invert gas concentrations.

[0004] In summary, at this stage, there is an urgent need to develop a high-concentration gas detection method and system based on direct absorption spectroscopy to solve the above-mentioned technical problems. Summary of the invention

[0005] An object of the present invention is to provide a new technical solution for a high concentration gas detection method and system based on direct absorption spectroscopy.

[0006] According to a first aspect of the present invention, a high-concentration gas detection method based on direct absorption spectroscopy is provided, the method comprising:

[0007] Step S1: obtaining direct absorption and transmission spectrum signals of the gas to be tested;

[0008] Step S2: using a low-order polynomial to perform baseline fitting processing on the direct absorption and transmission spectrum signal of the gas to be measured to obtain the absorption spectrum signal of the gas to be measured;

[0009] Step S3: performing peak normalization processing on the absorption spectrum signal of the gas to be tested;

[0010] Step S4: input the normalized absorption spectrum signal of the gas to be tested into a pre-trained deep learning concentration detection model for processing to obtain a predicted concentration value of the gas to be tested.

[0011] Optionally, the step S2 specifically includes:

[0012] Step S21: using a low-order polynomial to perform baseline fitting processing on the non-absorption data in the direct absorption transmission spectrum signal of the gas to be measured, so as to obtain a transmission spectrum signal of the gas to be measured after fitting;

[0013] Step S22: performing baseline removal processing on the fitted transmission spectrum signal of the gas to be measured to obtain the transmission spectrum signal of the gas to be measured with the baseline removed;

[0014] Step S23: performing an inversion operation on the transmission spectrum signal of the gas to be measured with the baseline removed to obtain the absorption spectrum signal of the gas to be measured.

[0015] Optionally, the low-order polynomial is a first-order polynomial.

[0016] Optionally, in step S3, the calculation formula for the peak normalization process is expressed as:

[0017]

[0018] Among them, MAX represents the peak normalized value, X i represents the original peak value of the i-th absorption spectrum signal, cofe i Represents the coefficient after the normalized peak of the i-th absorption spectrum signal.

[0019] Optionally, in step S4, the process of acquiring the training data set required by the deep learning concentration detection model is:

[0020] Step 41: obtaining direct absorption and transmission spectrum signals of sample gas at different concentrations through simulation;

[0021] Step 42: performing baseline fitting processing on the direct absorption and transmission spectrum signals of the sample gas at different concentrations using a low-order polynomial to obtain absorption spectrum signals of the sample gas at different concentrations;

[0022] Step 43: performing peak normalization processing on the absorption spectrum signals at different concentrations of the sample gas;

[0023] Step 44: selecting a preset number of concentration points at equal intervals from the nonlinear region of the normalized absorption spectrum signal of the sample gas at different concentrations, and determining corresponding absorption spectrum signals according to the selected concentration points;

[0024] Step 45: construct a training data set with the selected preset number of concentration points and the corresponding absorption spectrum signals to train the deep learning concentration detection model.

[0025] Optionally, in step S44, the process of selecting the preset number of concentration points is:

[0026] Determine a contour line based on the peak normalized value;

[0027] Determine the initial position and the end position where the contour line intersects with the nonlinear region of the absorption spectrum signal at different concentrations of the sample gas;

[0028] Determining a concentration interval based on a concentration point corresponding to the initial position and a concentration point corresponding to the end position;

[0029] A preset number of concentration points are selected from the concentration range at equal intervals.

[0030] Optionally, the step S41 is specifically as follows: under standard atmospheric pressure and room temperature conditions, simulation is performed based on the parameters of the sample gas at the set wavenumber corresponding to the spectral database and the preset optical path to obtain direct absorption and transmission spectral signals of the sample gas at different concentrations.

[0031] According to a second aspect of the present invention, there is provided a high-concentration gas detection system based on direct absorption spectroscopy, the system comprising:

[0032] A first processing module is configured to obtain a direct absorption and transmission spectrum signal of the gas to be tested;

[0033] The second processing module is configured to perform baseline fitting processing on the direct absorption and transmission spectrum signal of the gas to be measured using a low-order polynomial to obtain the absorption spectrum signal of the gas to be measured;

[0034] A third processing module is configured to perform peak normalization processing on the absorption spectrum signal of the gas to be measured;

[0035] The fourth processing module is configured to input the normalized absorption spectrum signal of the gas to be tested into a pre-trained deep learning concentration detection model for processing to obtain a predicted concentration value of the gas to be tested.

[0036] Optionally, the second processing module specifically includes:

[0037] A fitting unit is configured to perform baseline fitting processing on non-absorption data in the direct absorption transmission spectrum signal of the gas to be measured using a low-order polynomial to obtain a transmission spectrum signal of the gas to be measured after fitting;

[0038] A baseline removal unit is configured to perform baseline removal processing on the fitted transmission spectrum signal of the gas to be measured to obtain a transmission spectrum signal of the gas to be measured with the baseline removed;

[0039] The inversion unit is configured to perform an inversion operation on the transmission spectrum signal of the gas to be measured with the baseline removed, so as to obtain the absorption spectrum signal of the gas to be measured.

[0040] Optionally, the low-order polynomial is a first-order polynomial.

[0041] Optionally, the calculation formula for the peak normalization process is expressed as:

[0042]

[0043] Among them, MAX represents the peak normalized value, X i represents the original peak value of the i-th absorption spectrum signal, cofe i Represents the coefficient after the normalized peak of the i-th absorption spectrum signal.

[0044] Optionally, the process of acquiring the training data set required for the deep learning concentration detection model is:

[0045] The direct absorption and transmission spectrum signals of the sample gas at different concentrations are obtained through simulation;

[0046] A low-order polynomial is used to perform baseline fitting processing on the direct absorption and transmission spectrum signals of the sample gas at different concentrations to obtain the absorption spectrum signals of the sample gas at different concentrations;

[0047] Perform peak normalization processing on the absorption spectrum signals of sample gas at different concentrations;

[0048] Selecting a preset number of concentration points at equal intervals from the nonlinear region of the absorption spectrum signal of the normalized sample gas at different concentrations, and determining the corresponding absorption spectrum signal according to the selected concentration points;

[0049] A preset number of selected concentration points and corresponding absorption spectrum signals are constructed into a training data set to train the deep learning concentration detection model.

[0050] Optionally, the process of selecting the preset number of concentration points is:

[0051] Determine a contour line based on the peak normalized value;

[0052] Determine the initial position and the end position where the contour line intersects with the nonlinear region of the absorption spectrum signal at different concentrations of the sample gas;

[0053] Determining a concentration interval based on a concentration point corresponding to the initial position and a concentration point corresponding to the end position;

[0054] A preset number of concentration points are selected from the concentration range at equal intervals.

[0055] Optionally, under standard atmospheric pressure and room temperature conditions, simulation is performed based on the parameters of the sample gas at a set wavenumber corresponding to the spectral database and a preset optical path to obtain direct absorption and transmission spectral signals of the sample gas at different concentrations.

[0056] According to a third aspect of the present invention, an electronic device is provided, comprising a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of a high-concentration gas detection method based on direct absorption spectroscopy as described in the first aspect of the present invention are implemented.

[0057] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of a high-concentration gas detection method based on direct absorption spectroscopy as described in the first aspect of the present invention are implemented.

[0058] According to an embodiment disclosed in the present invention, a high-concentration gas detection method and system based on direct absorption spectroscopy of the present invention has the following beneficial effects:

[0059] The high-concentration gas detection method based on direct absorption spectroscopy of the present invention first obtains the direct absorption transmission spectral signal of the gas to be tested; then uses a low-order polynomial to perform baseline fitting processing on the direct absorption transmission spectral signal of the gas to be tested to obtain the absorption spectral signal of the gas to be tested; and performs peak normalization processing on the absorption spectral signal of the gas to be tested; finally, the normalized absorption spectral signal of the gas to be tested is input into a pre-trained deep learning concentration detection model for processing to obtain the predicted concentration value of the gas to be tested. Compared with the traditional linear polynomial concentration inversion method, the method of the embodiment of the present invention utilizes the deep learning technology of convolutional neural networks combined with full spectrum features to solve the problem of insufficient gas detection accuracy of direct absorption spectroscopy in a high concentration range, and realizes high-precision detection when the peak of the absorption spectrum of high-concentration gas disappears or is saturated, thereby meeting the detection requirements of high precision, high resolution and high concentration range.

[0060] Further features and advantages of the present invention will become apparent from the following detailed description of exemplary embodiments of the present invention with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.

[0062] Figure 1 A schematic flow chart of a high-concentration gas detection method based on direct absorption spectroscopy provided according to an embodiment;

[0063] Figure 2 A schematic diagram of removing the baseline of a direct absorption transmission spectrum signal provided according to an embodiment;

[0064] Figure 3 A schematic diagram of a peak normalization process provided according to an embodiment;

[0065] Figure 4 A schematic diagram of polynomial fitting according to an embodiment;

[0066] Figure 5 A schematic diagram showing a comparison of concentration inversion results using different methods provided in the embodiments;

[0067] Figure 6 A schematic structural diagram of a high-concentration gas detection system based on direct absorption spectroscopy provided according to an embodiment;

[0068] Figure 7 A schematic diagram of an electronic device. DETAILED DESCRIPTION

[0069] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present invention unless otherwise specifically stated.

[0070] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

[0071] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered as part of the specification.

[0072] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0073] Embodiment 1:

[0074] See also Figure 1 As shown, an embodiment of the present invention provides a high-concentration gas detection method based on direct absorption spectroscopy, the method comprising:

[0075] Step S1: Obtain the direct absorption and transmission spectrum signal of the gas to be tested.

[0076] It should be noted that after the light signal emitted by the light source passes through the gas to be tested, part of the light signal is absorbed, and the remaining light signal passes through the gas to be tested and is received by the spectrometer. The spectrometer can obtain the direct absorption and transmission spectral signal of the gas to be tested by analyzing the intensity and distribution of the transmitted light signal.

[0077] In this embodiment, the gas to be measured can be methane, and of course it can also be other gases, which are not listed here one by one. In this embodiment, the absorbance of the gas to be measured is close to 0.05 or greater than 0.05. At this time, there is a nonlinear absorption spectrum problem, and the existing technology cannot accurately invert the gas concentration.

[0078] Step S2: using a low-order polynomial to perform baseline fitting processing on the direct absorption and transmission spectrum signal of the gas to be measured, so as to obtain the absorption spectrum signal of the gas to be measured.

[0079] Optionally, step S2 in the high-concentration gas detection method based on direct absorption spectroscopy of this embodiment specifically includes:

[0080] Step S21: using a low-order polynomial to perform baseline fitting processing on the non-absorption data in the direct absorption transmission spectrum signal of the gas to be measured, so as to obtain a transmission spectrum signal of the gas to be measured after fitting;

[0081] It should be noted that in this embodiment, a first-order polynomial, a second-order polynomial or a third-order polynomial can be used for baseline fitting processing. The baseline fitting processing can remove the noise of the measuring instrument and the background interference, thereby improving the accuracy of the spectral data analysis.

[0082] Preferably, in order to improve the detection effect, a third-order polynomial is used to perform baseline fitting processing in this embodiment.

[0083] Step S22: performing baseline removal processing on the fitted transmission spectrum signal of the gas to be measured to obtain the transmission spectrum signal of the gas to be measured with the baseline removed;

[0084] It should be noted that, in this embodiment, the baseline is subtracted from the fitted transmission spectrum signal of the gas to be measured to obtain the transmission spectrum signal of the gas to be measured with the baseline removed.

[0085] Step S23: performing an inversion operation on the transmission spectrum signal of the gas to be measured with the baseline removed to obtain the absorption spectrum signal of the gas to be measured.

[0086] It should be noted that, in this embodiment, the absorption spectrum signal of the gas to be measured can be obtained by inverting the transmission spectrum signal of the gas to be measured with the baseline removed.

[0087] Step S3: performing peak normalization processing on the absorption spectrum signal of the gas to be tested.

[0088] Optionally, in the high-concentration gas detection method based on direct absorption spectroscopy of this embodiment, in step S3, the calculation formula for peak normalization processing is expressed as:

[0089]

[0090] Among them, MAX represents the peak normalized value, X i represents the original peak value of the i-th absorption spectrum signal, cofe i Represents the coefficient after the normalized peak of the i-th absorption spectrum signal.

[0091] Step S4: input the normalized absorption spectrum signal of the gas to be tested into a pre-trained deep learning concentration detection model for processing to obtain a predicted concentration value of the gas to be tested.

[0092] It should be noted that the deep learning concentration detection model in this embodiment is obtained by pre-training the training data set. In this embodiment, the deep learning concentration detection model can be a CNN network model, and of course other network models can also be used, which will not be illustrated one by one here.

[0093] Optionally, the process of acquiring the training data set required for the deep learning concentration detection model in the high concentration gas detection method based on direct absorption spectroscopy in this embodiment is:

[0094] The direct absorption and transmission spectrum signals of the sample gas at different concentrations are obtained through simulation;

[0095] A low-order polynomial is used to perform baseline fitting processing on the direct absorption and transmission spectrum signals of the sample gas at different concentrations to obtain the absorption spectrum signals of the sample gas at different concentrations;

[0096] Perform peak normalization processing on the absorption spectrum signals of sample gas at different concentrations;

[0097] Selecting a preset number of concentration points at equal intervals from the nonlinear region of the absorption spectrum signal of the normalized sample gas at different concentrations, and determining the corresponding absorption spectrum signal according to the selected concentration points;

[0098] A preset number of selected concentration points and the corresponding absorption spectrum signals are constructed into a training data set to train the deep learning concentration detection model.

[0099] It should be noted that, in this embodiment, the sample gas may be methane.

[0100] Optionally, the process of selecting a preset number of concentration points in the high-concentration gas detection method based on direct absorption spectroscopy in this embodiment is:

[0101] Determine a contour line based on the peak normalized value;

[0102] Determine the initial position and the end position where the contour line intersects the nonlinear region of the absorption spectrum signal at different concentrations of the sample gas;

[0103] Determine the concentration interval based on the concentration point corresponding to the initial position and the concentration point corresponding to the end position;

[0104] A preset number of concentration points are selected at equal intervals from the concentration range.

[0105] It should be noted that, in this embodiment, the contour line can be 1 / 5, 1 / 4 or 1 / 3 of the peak normalized value MAX, etc., determined according to actual conditions, and no specific limitation is made here. In this embodiment, the preset number can be 100, 200, 400, 500, 600, 800 or 1000, etc., determined according to actual conditions, and no specific limitation is made here.

[0106] Optionally, in the high-concentration gas detection method based on direct absorption spectroscopy of the present embodiment, simulation is performed under standard atmospheric pressure and normal temperature conditions based on the parameters of the sample gas at the corresponding set wavenumber in the spectral database and the preset optical path to obtain direct absorption transmission spectral signals of the sample gas at different concentrations.

[0107] The following is a detailed description of the process of obtaining a training data set in the high-concentration gas detection method based on direct absorption spectroscopy of the present invention using a specific embodiment:

[0108] 1) Using the HITRAN2020 spectral database, methane has a wave number of 6046.95 cm -1 The relevant parameters are set under standard atmospheric pressure (pressure is 1atm) and temperature is 25℃, the optical path is set to 1m, and the direct absorption and transmission spectral signals of different concentrations are simulated;

[0109] 2) According to the non-absorption data on both wings of the absorption profile of the spectral signal, a first-order polynomial is directly used for baseline fitting, such as Figure 2 (a), the baseline is then subtracted from the transmission spectrum and inverted to obtain Figure 2 (b), i.e., the corresponding absorption spectrum signal (i.e., raw data);

[0110] 3)) Before the raw data is input into the convolutional neural network model (CNN), it must be processed by appropriate conversion methods, otherwise the direct application of the raw data may reduce the performance of the model. Therefore, the peak value of the absorption spectrum signal is normalized, and the specific expression is shown in formula (1):

[0111]

[0112] According to the peak saturation of the absorption spectrum signal, the peak normalization value MAX=10 is defined. 8 , X i represents the original peak value of the i-th absorption spectrum, cofe i Represents the coefficient of the normalized peak value of the i-th absorption spectrum line. The peak value of the absorption spectrum signal was normalized for a total of 16 concentration points from 1000ppm to 10wppm. The results are as follows: Figure 2 At the same time, it should be noted that after the normalized data is sent to the CNN model for training, a peak denormalization operation is required to restore the original absorption spectrum signal.

[0113] 4) The concentration range selected in the training data set is 1000 ppm to 10w ppm, such as Figure 3 As shown, a contour line y=1 / 4MAX is selected. The contour line has 4 intersections with two concentration points. Only the two intersections on the left semi-axis of the absorption spectrum are concerned, and their horizontal coordinates are 1315ppm and 1369ppm, i.e., the initial position and the end position. By determining the initial and end positions of the absorption spectrum intersecting with the contour line in the concentration range, the concentration interval is obtained as [1315,1369]. 16 "equally spaced" concentration points are selected regularly, and 1000 additional concentration points are selected between every two equally spaced concentration points, so that the scale of the standard data set reaches 15*1000.

[0114] 5. The numerical values ​​of the concentration points obtained above and the corresponding absorption spectrum signals are constructed into a training data set, and are input into the above-established convolutional neural network (CNN) model for training to obtain a deep learning concentration detection model for accurate concentration prediction.

[0115] The following will select the gas absorption spectrum peak inversion concentration method in the prior art and the concentration detection model obtained by training the CNN network model with an unprocessed data set as a comparison. Figure 4 As shown in (a), a first-order polynomial is used to linearly fit the absorption spectrum signal peak of 0-1000ppm methane and the corresponding concentration. Figure 4 (b)-(d) show the 5Fitting of the methane absorption spectrum signal peak and concentration using second, third, and fourth order polynomials in the ppm range.

[0116] like Figure 5 As shown, the blue line represents the first-order polynomial from 1000ppm to 10 5 Concentration inversion results in the ppm range. It can be observed that in the range of 0ppm to 1000ppm, the first-order polynomial fitting shows a higher concentration inversion accuracy. As the concentration increases, the prediction results gradually show a nonlinear trend. The cyan line represents the concentration inversion results using the fourth-order polynomial fitting, which shows a wider concentration inversion range than the first-order polynomial fitting. However, when the concentration exceeds 2×10 4 When it reaches ppm, the prediction results of the fourth-order polynomial fitting also lose the linear relationship, which limits its application in high-concentration gas detection. The changes in these two curves once again verify the problem of failure of concentration inversion in high-concentration gas detection due to the saturation phenomenon of the absorption spectrum peak in direct absorption spectroscopy technology. The black line represents the prediction result of the concentration by the concentration detection model (i.e., Traditional-CNN) obtained by training the CNN network model with an unprocessed data set. It can be observed from the enlarged figure that its concentration inversion accuracy in the low concentration range is low. This is because when the step size is fixed, the concentration range has a large span from low to high, and there are fewer concentration point samples at low concentrations, resulting in insufficient low-concentration resolution. However, using the deep learning concentration detection model (i.e., Standard-CNN) of this embodiment for concentration prediction can effectively solve this problem, such as Figure 5 The observation results show that the concentration prediction of the deep learning concentration detection model of this embodiment has an extremely high goodness of fit, reaching 0.99996, showing that there is a significant linear relationship between the calibrated concentration and the model-predicted concentration.

[0117] In summary, in order to solve the problem that both the WMS method and the DAS method cannot achieve high-concentration range gas measurement, an embodiment of the present invention proposes a high-concentration gas detection method based on direct absorption spectroscopy, which first obtains the direct absorption transmission spectral signal of the gas to be measured; then uses a low-order polynomial to perform baseline fitting processing on the direct absorption transmission spectral signal of the gas to be measured to obtain the absorption spectral signal of the gas to be measured; and performs peak normalization processing on the absorption spectral signal of the gas to be measured; finally, the normalized absorption spectral signal of the gas to be measured is input into a pre-trained deep learning concentration detection model for processing to obtain the predicted concentration value of the gas to be measured. Compared with the traditional linear polynomial concentration inversion method, the method of the embodiment of the present invention uses the deep learning technology of convolutional neural networks combined with full spectrum features to solve the problem of insufficient gas detection accuracy of direct absorption spectroscopy in a high concentration range, and realizes high-precision detection when the peak of the absorption spectrum of the high-concentration gas disappears or is saturated, thereby meeting the detection requirements of high precision, high resolution and high concentration range.

[0118] Embodiment 2:

[0119] The embodiment of the present invention provides a high concentration gas detection system based on direct absorption spectroscopy, based on the high concentration gas detection method based on direct absorption spectroscopy described in Example 1, see Figure 6 As shown, the system 100 includes:

[0120] The first processing module 10 is configured to obtain a direct absorption and transmission spectrum signal of the gas to be tested;

[0121] The second processing module 20 is configured to perform baseline fitting processing on the direct absorption and transmission spectrum signal of the gas to be measured using a low-order polynomial to obtain the absorption spectrum signal of the gas to be measured;

[0122] The third processing module 30 is configured to perform peak normalization processing on the absorption spectrum signal of the gas to be measured;

[0123] The fourth processing module 40 is configured to input the normalized absorption spectrum signal of the gas to be measured into a pre-trained deep learning concentration detection model for processing to obtain a predicted concentration value of the gas to be measured.

[0124] Optionally, the second processing module 20 in the high-concentration gas detection system based on direct absorption spectroscopy of this embodiment specifically includes:

[0125] The fitting unit 21 is configured to perform baseline fitting processing on the non-absorption data in the direct absorption transmission spectrum signal of the gas to be measured using a low-order polynomial to obtain a transmission spectrum signal of the gas to be measured after fitting;

[0126] A baseline removal unit 22 is configured to perform baseline removal processing on the fitted transmission spectrum signal of the gas to be measured to obtain a transmission spectrum signal of the gas to be measured with the baseline removed;

[0127] The negation unit 23 is configured to perform a negation operation on the transmission spectrum signal of the gas to be measured with the baseline removed, so as to obtain the absorption spectrum signal of the gas to be measured.

[0128] Optionally, the low-order polynomial in the high-concentration gas detection system based on direct absorption spectroscopy in this embodiment is a first-order polynomial.

[0129] Optionally, the calculation formula for peak normalization processing in the high-concentration gas detection system based on direct absorption spectroscopy in this embodiment is expressed as:

[0130]

[0131] Among them, MAX represents the peak normalized value, X i represents the original peak value of the i-th absorption spectrum signal, cofe i Represents the coefficient after the normalized peak of the i-th absorption spectrum signal.

[0132] Optionally, the process of acquiring the training data set required for the deep learning concentration detection model in the high-concentration gas detection system based on direct absorption spectroscopy in this embodiment is as follows:

[0133] The direct absorption and transmission spectrum signals of the sample gas at different concentrations are obtained through simulation;

[0134] A low-order polynomial is used to perform baseline fitting processing on the direct absorption and transmission spectrum signals of the sample gas at different concentrations to obtain the absorption spectrum signals of the sample gas at different concentrations;

[0135] Perform peak normalization processing on the absorption spectrum signals of sample gas at different concentrations;

[0136] Selecting a preset number of concentration points at equal intervals from the nonlinear region of the absorption spectrum signal of the normalized sample gas at different concentrations, and determining the corresponding absorption spectrum signal according to the selected concentration points;

[0137] A preset number of selected concentration points and the corresponding absorption spectrum signals are constructed into a training data set to train the deep learning concentration detection model.

[0138] Optionally, the process of selecting a preset number of concentration points in the high-concentration gas detection system based on direct absorption spectroscopy in this embodiment is as follows:

[0139] Determine a contour line based on the peak normalized value;

[0140] Determine the initial position and the end position where the contour line intersects the nonlinear region of the absorption spectrum signal at different concentrations of the sample gas;

[0141] Determine the concentration interval based on the concentration point corresponding to the initial position and the concentration point corresponding to the end position;

[0142] A preset number of concentration points are selected at equal intervals from the concentration range.

[0143] Optionally, in the high-concentration gas detection system based on direct absorption spectroscopy of the present embodiment, simulation is performed based on the parameters of the sample gas at the corresponding set wavenumbers in the spectral database and the preset optical path under standard atmospheric pressure and normal temperature conditions to obtain direct absorption transmission spectral signals of the sample gas at different concentrations.

[0144] Embodiment three:

[0145] The present invention discloses an electronic device, which includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of a high-concentration gas detection method based on direct absorption spectroscopy in any one of the first embodiments disclosed in the present invention are implemented.

[0146] Figure 7 is a structural diagram of an electronic device according to an embodiment of the present invention, such as Figure 7 As shown, the electronic device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. Among them, the processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the electronic device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, near field communication (NFC) or other technologies. The display screen of the electronic device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the electronic device can be a touch layer covered on the display screen, or a button, a trackball or a touch pad set on the housing of the electronic device, or an external keyboard, touch pad or mouse, etc.

[0147] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a structural diagram of the part related to the technical solution of the present disclosure, and does not constitute a limitation on the electronic device to which the technical solution of the present application is applied. The specific electronic device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0148] Embodiment 4:

[0149] The embodiment of the present invention discloses a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of any one of the high-concentration gas detection methods based on direct absorption spectroscopy in the first embodiment of the present invention are implemented.

[0150] Please note that the technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification. The above embodiments only express several implementation methods of the present application, and their descriptions are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that for ordinary technicians in this field, without departing from the concept of the present application, several variations and improvements can be made, which all belong to the scope of protection of the present application. Therefore, the scope of protection of the patent in this application shall be based on the attached claims.

[0151] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

[0152] Although some specific embodiments of the present invention have been described in detail by way of example, it will be appreciated by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It will be appreciated by those skilled in the art that the above embodiments may be modified without departing from the scope and spirit of the present invention. The scope of the present invention is defined by the appended claims.

Claims

1. A high concentration gas detection method based on direct absorption spectroscopy, characterized in that: The said party The law includes: Step S1: obtaining direct absorption and transmission spectrum signals of the gas to be tested; Step S2: using a low-order polynomial to perform baseline fitting processing on the direct absorption and transmission spectrum signal of the gas to be measured to obtain the absorption spectrum signal of the gas to be measured; Step S3: performing peak normalization processing on the absorption spectrum signal of the gas to be tested; Step S4: input the normalized absorption spectrum signal of the gas to be tested into a pre-trained deep learning concentration detection model for processing to obtain a predicted concentration value of the gas to be tested.

2. The high-concentration gas detection method based on direct absorption spectroscopy according to claim 1, characterized in that: The step S2 specifically includes: Step S21: using a low-order polynomial to perform baseline fitting processing on the non-absorption data in the direct absorption transmission spectrum signal of the gas to be measured, so as to obtain a transmission spectrum signal of the gas to be measured after fitting; Step S22: performing baseline removal processing on the fitted transmission spectrum signal of the gas to be measured to obtain the transmission spectrum signal of the gas to be measured with the baseline removed; Step S23: performing an inversion operation on the transmission spectrum signal of the gas to be measured with the baseline removed to obtain the absorption spectrum signal of the gas to be measured.

3. The high-concentration gas detection method based on direct absorption spectroscopy according to claim 2 is characterized in that: The low-order polynomial is a first-order polynomial.

4. The high-concentration gas detection method based on direct absorption spectroscopy according to claim 1, characterized in that: In step S3, the calculation formula for the peak normalization process is expressed as: Among them, MAX represents the peak normalized value, X i represents the original peak value of the i-th absorption spectrum signal, cofe i Represents the coefficient after the normalized peak of the i-th absorption spectrum signal.

5. The high-concentration gas detection method based on direct absorption spectroscopy according to claim 1, characterized in that: In step S4, the process of acquiring the training data set required for the deep learning concentration detection model is as follows: Step 41: obtaining direct absorption and transmission spectrum signals of sample gas at different concentrations through simulation; Step 42: performing baseline fitting processing on the direct absorption and transmission spectrum signals of the sample gas at different concentrations using a low-order polynomial to obtain absorption spectrum signals of the sample gas at different concentrations; Step 43: performing peak normalization processing on the absorption spectrum signals at different concentrations of the sample gas; Step 44: selecting a preset number of concentration points at equal intervals from the nonlinear region of the normalized absorption spectrum signal of the sample gas at different concentrations, and determining corresponding absorption spectrum signals according to the selected concentration points; Step 45: construct a training data set with the selected preset number of concentration points and the corresponding absorption spectrum signals to train the deep learning concentration detection model.

6. The high-concentration gas detection method based on direct absorption spectroscopy according to claim 5, characterized in that: In step S44, the process of selecting the preset number of concentration points is as follows: Determine a contour line based on the peak normalized value; Determine the initial position and the end position where the contour line intersects with the nonlinear region of the absorption spectrum signal at different concentrations of the sample gas; Determining a concentration interval based on a concentration point corresponding to the initial position and a concentration point corresponding to the end position; A preset number of concentration points are selected from the concentration range at equal intervals.

7. The high-concentration gas detection method based on direct absorption spectroscopy according to claim 5, characterized in that: The step S41 specifically includes: under standard atmospheric pressure and room temperature conditions, simulation is performed based on the parameters of the sample gas at the corresponding set wavenumber in the spectral database and the preset optical path to obtain direct absorption and transmission spectral signals of the sample gas at different concentrations.

8. A high concentration gas detection system based on direct absorption spectroscopy, characterized in that: The system comprises: A first processing module is configured to obtain a direct absorption and transmission spectrum signal of the gas to be tested; The second processing module is configured to perform baseline fitting processing on the direct absorption and transmission spectrum signal of the gas to be measured using a low-order polynomial to obtain the absorption spectrum signal of the gas to be measured; A third processing module is configured to perform peak normalization processing on the absorption spectrum signal of the gas to be measured; The fourth processing module is configured to input the normalized absorption spectrum signal of the gas to be tested into a pre-trained deep learning concentration detection model for processing to obtain a predicted concentration value of the gas to be tested.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the steps in a high-concentration gas detection method based on direct absorption spectroscopy described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of a high-concentration gas detection method based on direct absorption spectroscopy according to any one of claims 1 to 7 are implemented.

Images