Raman spectroscopy fluorescence baseline correction method and apparatus

Abstract

The application discloses a Raman spectrum fluorescence baseline correction method and device, the method comprises the following steps: collecting a time domain signal of a sample to be measured, and obtaining an IRF function of a sample collection system of the sample to be measured; pre-processing the time domain signal, and performing deconvolution on the pre-processed time domain signal and the IRF function to obtain a deconvolution result; fitting the latter half data of the deconvolution result by using a support vector regression model to obtain a fitting curve; wherein, a radial basis function is used as a kernel function, and the hyperparameters of the support vector regression model are optimized by using a grid search and cross-validation method; performing difference operation on the deconvolution result and the fitting curve to obtain a Raman decay curve; and performing windowing and averaging on the Raman decay curve to obtain a Raman spectrum. The application does not need to preset a fluorescence decay model and parameters thereof, utilizes the characteristics of the time domain signal data itself, processes the IRF deconvolution result by using the support vector regression model, predicts a time domain baseline, can greatly improve the signal-to-noise ratio of extracted Raman signals, achieves good effects, and is suitable for practical application.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Raman signal processing, and particularly relates to a Raman spectrum fluorescence baseline correction method and device, a terminal equipment and a computer readable storage medium. BACKGROUND

[0002] With the continuous progress of spectral analysis technology, Raman spectrum analysis method as a non-destructive, non-contact analysis method, has been widely used in the fields of material science, biomedical, chemistry and other fields. Raman spectrum can detect the molecular vibration mode of the sample by detecting the non-elastic scattering effect between the sample photons, obtain the molecular vibration information of the sample, and provide in-depth information about the chemical composition and structure of the material, so as to realize the composition analysis and structure identification of the sample. Compared with the traditional infrared spectrum, Raman spectrum is not sensitive to water, and can detect asymmetric vibration mode, so it has significant advantages in the analysis of liquid samples and biological tissues. In recent years, with the continuous development of laser, spectrometer and detector technology, the application range of Raman spectrum is continuously expanding, especially in the fields of drug detection, food safety and environmental monitoring, which shows great potential. However, Raman signal is usually much weaker than other scattering signals, and is easily disturbed by background noise, so high-sensitivity detectors and signal enhancement techniques are needed. At the same time, the fluorescence interference of the sample and the environmental noise will also affect the accuracy of the signal, which limits the application of Raman spectrum technology.

[0003] A major challenge for Raman spectroscopy application is the removal of the interference of fluorescence background. Since the Raman scattering signal is usually several orders of magnitude weaker than the excitation light and fluorescence signal, the fluorescence background will significantly interfere with the analysis of Raman signal and affect its accuracy. Therefore, effective removal of the fluorescence background is a key task in Raman data processing. For the baseline correction of conventional Raman, there are many data processing methods. The Partial Least Squares (PLS) method was proposed in the literature (Wold S, Martens H, Wold H, The multivariate calibration problem in chemistry solved by the PLS method, Lecture Notes in Mathematics. Springer, 1983); the Adaptive Iterative Re-weighted Penalized Least Squares (airPLS) method was improved and proposed in the literature (Zhang ZM, Chen S, Liang YZ, Baseline correction using adaptive iteratively reweighted penalized least squares, Analyst. 2010 May; 135(5): 1138-46); the Asymmetric Least Squares (ALS) method was proposed in the literature (Eilers, Paul and Boelens, Hans, Baseline Correction with Asymmetric Least Squares Smoothing. Unpubl. Manuscr [J]. 2005); and there are also methods for baseline correction of conventional Raman data using deep learning. The above methods for fluorescence baseline correction of conventional Raman data cannot effectively suppress strong fluorescence background noise, and the effect is very limited. Based on the difference in time between Raman scattering and fluorescence response, time-resolved technology is increasingly used to suppress fluorescence interference. Time-resolved Raman spectroscopy (TRRS) technology is usually combined with pulsed laser and high time resolution detector (such as photomultiplier tube PMT, high-speed optical shutter based on Kerr cell, intensified charge-coupled device ICCD, single-photon avalanche diode (SPAD)) to achieve. For example, the combination of picosecond pulsed laser and ICCD detector can be used for detecting explosives through polymer or plastic packaging, and identifying minerals at different depths; the TCSPC technology based on SPAD can realize the fast acquisition of TRFS.Among the time-resolved techniques, time-correlated single photon counting (TCSPC) is an ideal choice for suppressing the fluorescence background due to its low noise, high detection efficiency and compact volume. A number of studies have shown that the time-domain Raman signals obtained by TCSPC technology contain a wealth of Raman information. If the fluorescence baseline can be corrected for the time-domain Raman signals, the sensitivity and accuracy of spectral detection will be significantly improved. However, the intensity of Raman signals is usually very low (about one percent) compared to that of fluorescence signals, and due to system noise, instrument response function (IRF) and nonlinear effects, it is not a small challenge to design a suitable fluorescence baseline correction algorithm.

[0004] Currently, the signal processing methods commonly used for correcting fluorescence curves to obtain Raman spectra from time-domain data are mainly direct method and inversion method based on exponential decay model fitting. The direct method is to deconvolve the time-domain data with the IRF, and then directly take the window of the deconvolution result to obtain the Raman spectrum. This method is simple and easy to operate, but it idealistically assumes that the Raman signal intensity is obvious, ignoring the fact that the Raman intensity usually accounts for only one percent of the fluorescence intensity or even lower. The Raman signal is easily submerged in the fluorescence background noise, and this method often has difficulty in obtaining high signal-to-noise ratio Raman spectrum. The inversion method based on exponential decay model fitting is to fit the fluorescence curve with an exponential decay model, and then remove the fluorescence signal. It is the mainstream method for obtaining Raman spectrum at present. However, there are still many problems. For example, although the traditional multi-exponential fitting method can fit the fluorescence decay curve, it is easy to lead to inaccurate fitting results when facing high noise environment or unknown composition leading to uncertain exponent. The setting of initial parameters will also bring unpredictable results. For example, the literature (J. Kekkonen, T. Talala and I. Nissinen, Time-and spectrally-resolved mesoscopic Raman and fluorescence imaging of carious enamel by a CMOS SPAD-based spectrometer, 2023 IEEE International Instrumentation and Measurement Technology Conference (I2MTC), Kuala Lumpur, Malaysia, 2023) uses a single exponential decay model to fit the fluorescence baseline of tooth sample data. The literature states that the fluorescence decay curve of the tooth biological sample is assumed to be a single exponential decay, and then the data after deconvolution is fitted with an exponential function to correct the fluorescence baseline and extract the Raman signal of the tooth biological sample. This approach is an idealized simplification of the fluorescence decay pattern and lacks rigorous theoretical support. Similarly, the patent number CN202110290121.4 uses the least squares method to fit the fluorescence lifetime curve to establish the relationship between multi-photon and single-photon measurement. This method also presets the fluorescence decay model and its parameters.Document (N. Krstajić et al., A256x8 SPAD line sensor for time resolved fluorescence and raman sensing, ESSCIRC 2014-40th European Solid State Circuits Conference (ESSCIRC), Venice Lido, Italy, 2014, pp143-146) uses a nonlinear least squares method to extract the Raman signal, also using an exponential decay model. Similarly, patent CN201810940171.0 also uses a fluorescence decay model to process time-domain signals to obtain Raman spectra. It can be seen that the existing extraction method using an exponential decay model to fit the fluorescence curve (whether it is a single exponential model or a multi-exponential model) often needs to rely on a large amount of prior knowledge and pre-set a large number of parameters, and the processing complexity is high, which cannot guarantee the accuracy while realizing efficient processing and high generalization. It is pointed out in document (Joseph R, Lakowicz, Principles of Fluorescence Spectroscop, Springer, 2006) that the fitting results of different exponential models can be very close, and the same exponential model can be completely different because of the different pre-set parameters. Therefore, even if the exponential decay model is used to fit the fluorescence lifetime curve, the obtained parameters may be far from the actual fluorescence lifetime. The composition of the fluorophore in reality is often complex, making it more difficult and unpredictable to remove the fluorescence baseline using the fluorescence decay model. SUMMARY

[0005] In order to solve the above problems of the prior art, the present application provides a Raman spectrum fluorescence baseline correction method, device, terminal equipment and computer readable storage medium, which does not pre-set the fluorescence decay model and its parameters, uses the characteristics of the time-domain signal data itself, processes the IRF deconvolution result using support vector regression (SVR), and predicts the time-domain baseline, which can greatly improve the signal-to-noise ratio of the extracted Raman signal, achieve good results, and have practical significance.

[0006] The first object of the present application is to provide a Raman spectrum fluorescence baseline correction method.

[0007] The second object of the present application is to provide a Raman spectrum fluorescence baseline correction device.

[0008] The third object of the present application is to provide a terminal equipment.

[0009] The fourth object of the present application is to provide a computer readable storage medium.

[0010] The first object of the application can be achieved by adopting the following technical solutions:

[0011] A Raman spectrum fluorescence baseline correction method, the method comprising:

[0012] Collecting a time domain signal of a sample to be measured, and obtaining an IRF function of a sample collection system of the sample to be measured;

[0013] Pretreating the time domain signal, and deconvolving the pretreated time domain signal and the IRF function to obtain a deconvolution result;

[0014] Fitting the latter half data of the deconvolution result by using a support vector regression model to obtain a fitting curve; wherein a radial basis function is used as a kernel function, and the hyperparameters of the support vector regression model are optimized by a grid search and cross-validation method;

[0015] Subtracting the deconvolution result from the fitting curve to obtain a Raman decay curve;

[0016] Windowing and averaging the Raman decay curve to obtain a Raman spectrum.

[0017] Further, the support vector regression model finds an optimal regression function on the input data set, so that most data points are within the tolerance of the function; at the same time, a relaxation variable is introduced to handle inevitable errors.

[0018] Further, the introduced relaxation variable is and The objective function is:

[0019]

[0020] And the constraint condition is satisfied:

[0021]

[0022] In the formula, ω and b are parameters of the model; C is a regularization parameter, which controls the trade-off between the complexity of the model and the error; y i is the actual output corresponding to the i th data sample point x i , n is the total number of data sample points, and ϵ is the tolerance.

[0023] Further, the regression function of the support vector regression model is:

[0024]

[0025] The Gaussian radial basis function is:

[0026]

[0027] wherein m is the number of support vectors, x i is the i th data sample point; and is the Lagrange multiplier, representing the weight of x i ; is a kernel function, used to calculate the similarity between input data sample points x and x i , the kernel function determines the way of mapping to the high-dimensional feature space; b is a parameter of the model; γ is a parameter of the kernel function, used to control the width of the Gaussian function; x j is the j th data sample point.

[0028] Further, the preprocessing includes noise reduction and outlier rejection.

[0029] Further, the sample collection system is a time-correlated single photon counting system, which is composed of a pulsed laser, an optical path module, a photodetector, a single photon avalanche diode detector and a TCSPC system, the pulsed laser is divided into a first light beam and a second light beam through the optical path module; the first light beam transmits the obtained electrical signal to the TCSPC system as a reference synchronization signal to record the collected time domain signal after passing through the photodetector; the second light beam is coupled into the microscope objective through the optical path module and focused on the sample to be measured to obtain a time domain signal; the time domain signal is coupled into the optical path module after collimation through the microscope objective, coupled into the single photon avalanche diode detector through the spectrometer for signal amplification and transmitted to the TCSPC system for data recording.

[0030] Further, the second light beam is coupled into a 25um optical fiber for transmission, and white light is added as an illumination light to assist the laser to focus on the surface of the sample to be measured to obtain the best Raman signal.

[0031] The time-correlated single photon counting system is a coupling system of the microscope as the excitation light and the signal light.

[0032] The second object of the application can be achieved by adopting the following technical scheme:

[0033] A Raman spectrum fluorescence baseline correction device, the device comprises:

[0034] A signal acquisition unit is configured to acquire a time domain signal of a sample to be measured and obtain an IRF function of a sample collection system;

[0035] A deconvolution unit is configured to preprocess the time domain signal, deconvolve the preprocessed time domain signal and the IRF function, and obtain a deconvolution result;

[0036] A data fitting unit is configured to fit the latter half data of the deconvolution result by using a support vector regression (SVR) model to obtain a fitting curve; wherein a radial basis function is used as a kernel function, and the hyperparameters of the SVR model are optimized by using a grid search and cross-validation method.

[0037] An operation unit is configured to perform a difference operation on the deconvolution result and the fitting curve to obtain a Raman attenuation curve.

[0038] A calculation unit is configured to perform a windowing and averaging operation on the Raman attenuation curve to obtain a Raman spectrum.

[0039] The third object of the present application can be achieved by adopting the following technical solution:

[0040] A terminal device includes a processor and a memory for storing a program executable by the processor, and the processor implements the Raman spectrum fluorescence baseline correction method described above when executing the program stored in the memory.

[0041] The fourth object of the present application can be achieved by adopting the following technical solution:

[0042] A computer-readable storage medium stores a program, and the program is executed by a processor to implement the Raman spectrum fluorescence baseline correction method described above.

[0043] The present application has the following beneficial effects relative to the prior art:

[0044] By introducing the post-processing algorithm based on the SVR model, the present application can still maintain high fitting accuracy under high noise conditions, and obtain the Raman signal of the fluorescence sample without requiring a large amount of prior knowledge; at the same time, the present application has high computing efficiency and is suitable for real-time processing of large-scale data. BRIEF DESCRIPTION OF DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of the drawings shown.

[0046] Figure 1 The flowchart of the Raman spectrum fluorescence baseline correction method of the present application embodiment 1.

[0047] Figure 2 The principle diagram of the Raman spectrum fluorescence baseline correction method of the present application embodiment 1.

[0048] Figure 3A structural schematic diagram of a time-correlated single photon counting system of embodiment 1 of the present application.

[0049] Figure 4 A comparison diagram of a fitted fluorescence lifetime curve and an original fluorescence lifetime of embodiment 1 of the present application.

[0050] Figure 5 An effect comparison diagram of embodiment 1 of the present application, wherein (a) is a Raman signal obtained by the method provided by the present application, and (b) is a Raman signal obtained by a multi-exponential decay method.

[0051] Figure 6 A structural block diagram of a Raman spectrum fluorescence baseline correction device of embodiment 2 of the present application.

[0052] Figure 7 A structural block diagram of a terminal device of embodiment 3 of the present application. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application. It should be understood that the described specific embodiments are only used to explain the present application, and are not used to limit the present application.

[0054] Embodiment 1:

[0055] As shown in Figure 1 , 2 , the present embodiment provides a Raman spectrum fluorescence baseline correction method, comprising the following steps:

[0056] S101, collecting a time-domain signal of a sample to be measured, and obtaining an IRF function of a sample collection system.

[0057] An IRF function of a time-correlated single photon counting system is obtained, and a time-domain signal of a sample is collected by using the time-correlated single photon counting system.

[0058] The time-domain signal collection of the present embodiment mainly relies on a time-correlated single photon counting system, as shown in Figure 3 , the system is composed of a laser source, an optical path module, a spectrometer, a single photon avalanche diode (SPAD) detector and a data acquisition and processing system (TCSPC system). The optical path module includes a beam splitter 1, an optical fiber 2, an optical fiber 3, a narrow-band filter 4, a dichroic mirror 6, a band-pass filter 7, a collection optical fiber 8 and a spectrometer.

[0059] The selection of a suitable laser source is critical, and a pulsed laser is usually used. The system uses a picosecond laser with a repetition frequency of 20 MHz to ensure high temporal resolution of the light source. In addition, in order to ensure the stability of the experimental conditions, the laser source needs to have stable output power and good beam quality.

[0060] After the pulsed laser passes through the beam splitter 1, a first beam and a second beam are obtained; the first beam passes through the photodetector, and the obtained electrical signal is transmitted to the time-correlated single photon counter as a reference synchronization signal to record the collected time-domain signal; the second beam passes through optical fibers 2 and 3, is collimated by a fiber collimator, is filtered by a narrow-band filter 4 to remove wavelengths other than 532 nm excitation light, passes through multiple mirrors, is coupled into the microscope objective by the dichroic mirror 6, and is focused onto the sample to obtain the time-domain signal; the time-domain signal is coupled into a 50um optical fiber 8 after collimation by the microscope objective, is coupled to a single photon avalanche diode detector (SPAD) via a beam splitter to amplify the signal and is transmitted to a time-correlated single photon counter (TCSPC) for data recording. In order to reduce the loss of laser energy in the optical path, the reflective mirror is coated with a high reflectivity coating corresponding to the wavelength of the laser. In order to ensure that the laser is accurately irradiated onto the sample surface, a white light is added as an illumination light in the microscopic coaxial system to assist the laser in focusing onto the sample surface to obtain the best Raman signal collection. The split beams are due to the need for the system to record accurate time-correlated data.

[0061] In order to ensure the size of the system and facilitate coupling with the signal collection system, the laser source is coupled into 25um optical fibers 2 and 3. The time-correlated single photon counting system uses a microscope as the coupling system for excitation light and signal light.

[0062] In the optical path design, the system also considers the suppression of stray light and spatial light, including the use of suppression materials to absorb stray light and the use of structural closure to suppress spatial light. In addition, the use of low-noise fiber coupling systems can effectively improve the transmission efficiency of the Raman signal and reduce the loss during the fiber connection process.

[0063] (1) Collection of IRF (Instrument Response Function).

[0064] The accurate measurement of the instrument response function (IRF) is an indispensable part of the TCSPC technique, as the width and shape of the IRF directly affect the measurement accuracy of the fluorescence lifetime.

[0065] The IRF is usually acquired by measuring the fluorescence signal of a standard sample with a very short lifetime (e.g. picoseconds). In this embodiment, a fluorescent dye with an extremely short fluorescence lifetime is selected as the standard sample, and its fluorescence decay curve is acquired using the same laser excitation and the same optical path configuration as described above. The shape of this curve is the IRF of the system, and by fitting this curve, the time resolution of the IRF can be accurately obtained.

[0066] (2) Acquisition of the time-domain signal.

[0067] After the optical path design and IRF acquisition are completed, the device signal of the sample is acquired. In actual operation, the sample should be selected to avoid materials with strong absorption or strong scattering as much as possible to reduce the influence of background noise. After the sample is excited by the laser, the time-domain signal generated is received by the SPAD detector and transmitted to the data acquisition system. In order to obtain high-quality fluorescence lifetime data, long-time signal acquisition is required to ensure sufficient photon counting rate. The acquired time-domain signal usually contains system noise, dark noise and background noise.

[0068] S102, deconvolving the pre-processed time-domain signal with the IRF function to obtain a deconvolution result.

[0069] The time-domain signal is mainly pre-processed to remove abnormal points.

[0070] The pre-processing of this embodiment includes noise reduction and rejection of abnormal values.

[0071] The pre-processed time-domain signal will be deconvolved with the acquired IRF function to obtain the deconvolved time-domain signal as the deconvolution result.

[0072] S103, fitting the deconvolution result using a support vector regression model to obtain a fitting curve.

[0073] The deconvolved time-domain signal is input into a support vector regression (SVR) model for fitting. The selection and parameter optimization of the SVR model are the core of the entire data post-processing. In this embodiment, a radial basis function (RBF) is used as the kernel function, and the hyperparameters (including the penalty coefficient C and the kernel parameter γ) of the SVR model are optimized by the grid search (Grid Search) and cross-validation (Cross-Validation) methods to ensure that the model has good fitting performance. The fitting effect is as follows: Figure 4 .

[0074] The SVR model fits a smooth and physically plausible fluorescence lifetime curve by minimizing the balance between prediction error and model complexity. Compared with the traditional multi-exponential fitting method, the SVR model can maintain high fitting accuracy under high noise conditions and does not need to rely on prior exponential model assumptions. The fitted fluorescence lifetime curve can be further used to extract Raman signals. Specifically, the characteristic peaks of the Raman signal are found by derivative analysis of the fitted curve, and the intensity thereof is calculated.

[0075] The SVR model finds an optimal regression function on the input data set, so that most data points are within a certain distance (i.e., tolerance) of the function. At the same time, the SVR also allows some data points to deviate from the function by introducing slack variables, but tries to minimize these deviations as much as possible.

[0076] The goal of the SVR model is to find a regression function:

[0077]

[0078] So that:

[0079]

[0080] Where y i represents the actual output corresponding to the i-th data sample point x i , and e is the tolerance, and w and b are parameters of the model.

[0081] In order to deal with inevitable errors, slack variables and are introduced, and the objective function becomes:

[0082]

[0083] Where n represents the total number of data sample points. Usually, the value of i ranges from 1 to n.

[0084] And satisfy the constraint condition:

[0085]

[0086] This embodiment uses the kernel function Gaussian radial basis function RBF for nonlinear regression, and the regression function of the SVR model can be expressed as:

[0087]

[0088] Where m is the number of support vectors; and are Lagrange multipliers, used to represent the weight of x i ; and is a kernel function used to calculate the similarity between input data sample points x and x i The choice of kernel function determines the way of mapping to the high-dimensional feature space.

[0089] where the Gaussian radial basis function is represented as:

[0090]

[0091] where γ is the parameter of the kernel function, used to control the width of the Gaussian function; x j represents the jth data sample point.

[0092] Since the duration of Raman is extremely short (usually hundreds of picoseconds), while the fluorescence decay is usually several nanoseconds to tens of nanoseconds, the latter half of the photon decay curve can be considered as the fluorescence decay curve. In this embodiment, the last 50% of the data is taken as the sample to fit the trajectory of the entire fluorescence background curve.

[0093] S104, difference operation between the deconvolution result and the fitting curve to obtain the Raman decay curve.

[0094] The deconvolution result obtained by step S102 is subtracted by the fitting curve obtained by step S103 to obtain the Raman decay curve.

[0095] S105, windowing and averaging the Raman decay curve to obtain the Raman spectrum.

[0096] At the initial position of excitation, the Raman decay curve is windowed and averaged to finally obtain the Raman spectrum data. By analyzing the change rate and peak position of the Raman spectrum data, the signal peak corresponding to Raman scattering is identified. Due to the good fitting performance of the SVR model, the extracted Raman signal has high signal-to-noise ratio and can accurately reflect the molecular vibration information of the sample.

[0097] In the last stage of data processing, the extracted Raman signal will be compared with the known standard Raman spectrum to determine the composition and structure information of the sample. If the Raman signal of the sample is weak, the signal resolution and accuracy can be further improved by accumulating signals of multiple acquisition periods. The final comparison effect is as follows Figure 5 .

[0098] This embodiment employs a post-processing algorithm based on a Support Vector Regression (SVR) model to process the photon attenuation curves acquired by a time-domain device to obtain Raman spectra. In the processing, firstly, significant noise and outliers are removed through data preprocessing. Then, the preprocessed data is deconvolved with the measured IRF function, and the last 50% of the resulting curve is input into the SVR model. Utilizing the model's good generalization ability, an accurate fluorescence background curve is fitted. The fitted background curve is subtracted from the deconvolution result to obtain a high signal-to-noise ratio (SNR) Raman attenuation curve. Finally, windowing is applied to the initial excitation position to obtain the high SNR Raman spectrum.

[0099] Those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware, and the corresponding program can be stored in a computer-readable storage medium.

[0100] It should be noted that although the method operations of the above embodiments are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the order of execution of the described steps may be changed. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0101] Example 2:

[0102] like Figure 6 As shown, this embodiment provides a Raman spectroscopy fluorescence baseline correction device, which includes a signal acquisition unit 601, a deconvolution unit 602, a data fitting unit 603, a calculation unit 604, and a computation unit 605, wherein:

[0103] The signal acquisition unit 601 is used to acquire the time-domain signal of the sample under test and obtain the IRF function of the sample acquisition system.

[0104] The deconvolution unit 602 is used to preprocess the time-domain signal, deconvolve the preprocessed time-domain signal with the IRF function, and obtain the deconvolution result.

[0105] The data fitting unit 603 is used to fit the latter half of the deconvolution result data using a support vector regression model to obtain a fitted curve. Specifically, a radial basis function is used as the kernel function, and the hyperparameters of the support vector regression model are optimized through grid search and cross-validation.

[0106] The operation unit 604 is used to perform a subtraction operation between the deconvolution result and the fitted curve to obtain the Raman attenuation curve.

[0107] The computing unit 605 is configured to window and average the Raman attenuation curve to obtain a Raman spectrum.

[0108] The specific implementation of each unit in this embodiment can refer to the above-mentioned embodiment 1, which will not be repeated here. It should be noted that the device provided in this embodiment is only exemplified by the division of the above-mentioned functional units. In actual application, the above-mentioned functions can be completed by different functional units according to the needs, that is, the internal structure is divided into different functional units to complete all or part of the functions described above.

[0109] Embodiment 3

[0110] The computer device provided in this embodiment can be a computer, such as Figure 7 As shown in the figure, the processor 702, the memory, the input device 703, the display 704 and the network interface 705 are connected through the system bus 701. The processor is used to provide computing and control capabilities. The memory includes a non-volatile storage medium 706 and an internal memory 707. The non-volatile storage medium 706 stores an operating system, a computer program and a database. The internal memory 707 provides an environment for the running of the operating system and the computer program in the non-volatile storage medium. When the processor 702 executes the computer program stored in the memory, the Raman spectrum fluorescence baseline correction method of the above-mentioned embodiment 1 is realized, as follows:

[0111] The time-domain signal of the to-be-tested sample is collected, and an IRF function of a sample collection system of the to-be-tested sample is obtained.

[0112] The time-domain signal is preprocessed, and the preprocessed time-domain signal is deconvoluted with the IRF function to obtain a deconvolution result.

[0113] The support vector regression model is used to fit the latter half data of the deconvolution result to obtain a fitting curve. The radial basis function is used as a kernel function, and the hyperparameters of the support vector regression model are optimized by the grid search and cross-validation methods.

[0114] The deconvolution result is subtracted from the fitting curve to obtain a Raman attenuation curve.

[0115] The Raman attenuation curve is windowed and averaged to obtain a Raman spectrum.

[0116] Embodiment 4

[0117] The storage medium provided in this embodiment is a computer readable storage medium, which stores a computer program. When the computer program is executed by the processor, the Raman spectrum fluorescence baseline correction method of the above-mentioned embodiment 1 is realized, as follows:

[0118] Collecting a time domain signal of a to-be-tested sample, and obtaining an IRF function of a to-be-tested sample collection system.

[0119] Preprocessing the time domain signal, and deconvolving the preprocessed time domain signal with the IRF function to obtain a deconvolution result.

[0120] Fitting the latter half data of the deconvolution result by using a support vector regression model to obtain a fitting curve, wherein a radial basis function is used as a kernel function, and the hyperparameters of the support vector regression model are optimized by using a grid search and cross-validation method.

[0121] Subtracting the deconvolution result from the fitting curve to obtain a Raman decay curve.

[0122] Windowing and averaging the Raman decay curve to obtain a Raman spectrum.

[0123] It should be noted that the computer readable storage medium of the embodiment can be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium may, for example, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination of the above. More specific examples of computer readable storage media can include, but are not limited to, an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0124] In summary, the present application first proposes a data post-processing technology based on the SVR model to process the baseline correction problem of the time domain signal in the Raman spectrum device, improves the signal-to-noise ratio of the data processing of the obtained Raman spectrum, and solves the problems of noise sensitivity and high processing complexity in the traditional method. By introducing the powerful support vector regression (SVR) model to fit the fluorescence background curve of the time domain signal in the Raman spectrum, instead of using the traditional exponential decay curve fitting, the dependence on prior knowledge is reduced by avoiding fitting the fluorescence lifetime curve based on the fluorescence lifetime model.

[0125] The above is only a preferred embodiment of the present application, but the protection scope of the present application is not limited thereto. Any skilled person in the art can make equivalent replacements or changes to the technical solutions and inventive concepts of the present application within the scope disclosed by the present application, and such replacements or changes are also within the protection scope of the present application.

Claims

1. A method of Raman spectrofluorimetric baseline correction, characterized in that, The method comprises: collecting a time domain signal of a to-be-tested sample, and obtaining an IRF function of a to-be-tested sample collection system; preprocessing the time domain signal, and performing deconvolution on the preprocessed time domain signal and the IRF function to obtain a deconvolution result; fitting the latter half data of the deconvolution result by using a support vector regression model to obtain a fitting curve; the support vector regression model finds an optimal regression function on the input data set, so that most data points are within the tolerance of the function; meanwhile, a slack variable is introduced to process inevitable errors; wherein, a radial basis function is used as a kernel function, and the hyperparameters of the support vector regression model are optimized by using a grid search and cross-validation method; performing difference operation on the deconvolution result and the fitting curve to obtain a Raman attenuation curve; windowing and averaging the Raman attenuation curve to obtain a Raman spectrum; where the introduced slack variables are and The objective function is and the constraint condition is met: where ω and b are parameters of the model; C is a regularization parameter that controls the trade-off between model complexity and error; y i is the corresponding actual output, and n is the total number of data sample points, and ε is the tolerance. i is the corresponding actual output, and n is the total number of data sample points, and ε is the tolerance. the regression function of the support vector regression model is: wherein the Gaussian radial basis function is: where m is the number of support vectors, x i is the i-th data sample point; and is the Lagrange multiplier, representing the weight of x i ; is the kernel function, used to calculate the similarity between input data sample points x and x i , the kernel function determines the way of mapping to the high-dimensional feature space; b is the parameter of the model; γ is the parameter of the kernel function, used to control the width of the Gaussian function; x j is the j-th data sample point.

2. The Raman spectral fluorescence baseline correction method according to claim 1, wherein, the preprocessing comprises noise reduction and abnormal value rejection.

3. The Raman spectral fluorescence baseline correction method according to claim 1, wherein, The to-be-tested sample collection system is a time-dependent single photon counting system, which is composed of a pulsed laser, an optical path module, a photodetector, a single photon avalanche diode detector and a TCSPC system. The pulsed laser is divided into a first light beam and a second light beam by the optical path module. After the first light beam passes through the photodetector, the obtained electrical signal is transmitted to the TCSPC system as a reference synchronization signal to record the collected time domain signal. The second light beam passes through the optical path module, and the laser is coupled into the microscope objective and focused on the to-be-tested sample to obtain the time domain signal. The time domain signal is collimated by the microscope objective and coupled into the optical path module, and then coupled into the single photon avalanche diode detector through the spectrometer for signal amplification and transmitted to the TCSPC system for data recording.

4. The Raman spectral fluorescence baseline correction method according to claim 3, wherein, The second light beam is coupled into a 25um optical fiber for transmission, and white light is added as an illumination light to assist the laser to focus on the surface of the to-be-tested sample to obtain the best Raman signal. The time-dependent single photon counting system is a coupling system of the microscope as the excitation light and the signal light.

5. A Raman spectrofluorometric baseline correction device, characterized by, The device comprises: a signal collection unit configured to collect a time domain signal of a to-be-tested sample, and obtain an IRF function of a to-be-tested sample collection system; a deconvolution unit configured to preprocess the time domain signal, and perform deconvolution on the preprocessed time domain signal and the IRF function to obtain a deconvolution result; a data fitting unit configured to fit the latter half data of the deconvolution result by using a support vector regression model to obtain a fitting curve; the support vector regression model finds an optimal regression function on the input data set, so that most data points are within the tolerance of the function; meanwhile, a slack variable is introduced to process inevitable errors; wherein, a radial basis function is used as a kernel function, and the hyperparameters of the support vector regression model are optimized by using a grid search and cross-validation method; an operation unit configured to perform difference operation on the deconvolution result and the fitting curve to obtain a Raman attenuation curve; a calculation unit configured to window and average the Raman attenuation curve to obtain a Raman spectrum; where the introduced slack variables are and the objective function is and the constraint condition is met: where ω and b are parameters of the model; C is a regularization parameter that controls the trade-off between model complexity and error; y i is the corresponding actual output, and n is the total number of data sample points, and ε is the tolerance. i is the corresponding actual output, and n is the total number of data sample points, and ε is the tolerance. the regression function of the support vector regression model is: wherein the Gaussian radial basis function is: where m is the number of support vectors, x i is the ith data sample point; and is the Lagrange multiplier, representing the weight of x i ; is the kernel function, used to calculate the similarity between input data sample points x and x i , the kernel function determines the way of mapping to the high-dimensional feature space; b is the parameter of the model; γ is the parameter of the kernel function, used to control the width of the Gaussian function; x j is the jth data sample point.

6. A computer device comprising a processor and a memory for storing a processor executable program, characterized in that, The processor implements the Raman spectrum fluorescence baseline correction method in any one of claims 1-4 when executing the program stored in the memory.

7. A storage medium storing a program, characterized by comprising: The processor implements the Raman spectrum fluorescence baseline correction method in any one of claims 1-4 when executing the program stored in the memory.

Images