Multimodal resolved time-resolved raman spectroscopy thermographic interactive analysis method and system

Abstract

The application discloses a multi-peak decomposition time-resolved Raman spectrum thermal map interactive analysis method and system, and belongs to the field of spectrum analysis. The method comprises the following steps: continuously or time-divisionally collecting a to-be-measured sample to obtain Raman spectrum data corresponding to different time points; preprocessing the original Raman spectrum data; extracting spectral characteristic peaks by using an adaptive peak value recognition algorithm, and performing characteristic peak integral calculation; obtaining time contribution coefficients of different chemical components or characteristic signals in the time dimension by using a multi-peak decomposition algorithm; generating a Raman thermal map in the time dimension according to the characteristic peak intensity and / or the time contribution coefficients; establishing an index mapping relationship between the position of the Raman thermal map and the corresponding time Raman spectrum data; and responding to the interactive operation of the user on the thermal map and displaying the Raman spectrum curve at the corresponding time point in real time. The application can realize visual display, dynamic monitoring and interactive analysis of time sequence data of Raman spectrum, and improves the efficiency and accuracy of Raman signal analysis of dynamic samples.

Description

Technical Field

[0001] This invention relates to the field of spectral analysis technology, and in particular to an interactive analysis method and system for time-resolved Raman spectral thermograms with multi-peak decomposition. Background Technology

[0002] Raman spectroscopy can characterize molecular vibrational information in samples and has wide applications in biological detection, chemical analysis, medical diagnostics, and materials characterization. For samples such as sweat, blood, residual bodily fluids, and other samples that change over time, the Raman signal is often not only related to the sample composition but also affected by dynamic processes such as evaporation, concentration, reaction, deposition, and structural evolution. Therefore, continuously or time-series acquiring samples and studying the changes in Raman spectra over time has significant application value.

[0003] In existing technologies, Raman data analysis mostly focuses on peak position determination or static intensity comparison of single-acquisition spectra. For large amounts of time-series spectral data obtained through continuous acquisition, analysis is typically performed using only tables, single-curve overlays, or single-peak intensity trend plots. These methods have the following problems: 1. It is not conducive to the intuitive observation of the overall variation pattern of Raman signals over a long time scale; 2. It is difficult to simultaneously analyze the evolution of multiple characteristic peaks or multiple components over time; 3. Users find it difficult to quickly locate the complete Raman spectrum corresponding to a specific moment from heatmaps or trend charts; 4. When the amount of time-series data is large, the efficiency of querying and linked display is low; 5. For complex samples such as sweat and blood, single peak analysis is difficult to accurately characterize the dynamic changes of multiple components.

[0004] Therefore, there is a need for an interactive method and system that can construct heatmaps from time-resolved Raman spectral data, enable the linked display of heatmaps and spectra, and support multi-peak decomposition analysis. Summary of the Invention

[0005] To address the shortcomings of existing technologies in visualizing time-resolved Raman data, the inconvenience of analyzing dynamic changes, and the difficulty in linking heatmaps and spectra, this invention provides an interactive analysis method and system for time-resolved Raman spectral heatmaps with multi-peak decomposition.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: The present invention provides an interactive analysis method for time-resolved Raman spectroscopy thermograms with multi-peak decomposition, comprising the following steps: S1: Continuously or time-segmentedly acquire Raman spectral data corresponding to different time points of the sample to be tested; S2: Baseline correction, noise smoothing and normalization are performed on the Raman spectral data to obtain preprocessed Raman spectral data; S3: Perform peak identification on the preprocessed Raman spectral data to extract characteristic peak information; S4: Perform integral calculations based on the characteristic peak information to obtain the characteristic peak intensities at different time points; S5: Perform multi-peak decomposition on the preprocessed Raman spectra at different time points to solve for the time contribution coefficients of different chemical components or characteristic signals; S6: Generate a Raman heatmap based on the characteristic peak intensity and / or the time contribution coefficient; S7: Establish an index mapping relationship between the Raman heatmap location and the corresponding Raman spectral data at the time point; S8: Respond to the user's interactive operation on the Raman heatmap, locate the corresponding time point and retrieve the corresponding Raman spectrum; S9: Display the Raman spectrum curve corresponding to the time point in real time in the visualization interface.

[0007] In S2, the baseline correction uses a polynomial fitting method to construct a background baseline function, and the corrected Raman spectrum is obtained by subtracting the background baseline function from the original spectrum; the noise smoothing uses one of the moving average method, Savitzky-Golay smoothing method, or Gaussian smoothing method; the normalization method uses one of the maximum value normalization, total area normalization, or reference peak normalization.

[0008] In step S3, peak identification is performed by calculating the first derivative of the Raman spectrum and determining candidate peak positions based on the zero-crossing positions where the derivative changes from positive to negative. Then, characteristic peaks are selected based on a peak intensity threshold. This peak intensity threshold is used to distinguish effective characteristic peaks from noise signals.

[0009] In S3, the peak intensity threshold is adaptively determined based on the current average intensity of the spectrum and the current standard deviation of the spectral intensity, satisfying T = μ + kσ, where T is the peak intensity threshold, μ is the current average intensity of the spectrum, σ is the current standard deviation of the spectral intensity, and k is an empirical coefficient.

[0010] In S4, the integral calculation of the characteristic peak information involves setting an integration interval near the center wavenumber of the characteristic peak and accumulating or integrating the spectral intensity within the interval to obtain the characteristic peak intensity at the corresponding time point.

[0011] In S5, multi-peak decomposition is performed on the preprocessed Raman spectra at different time points. This involves representing the Raman spectrum to be measured as a linear combination of multiple reference spectra and solving the contribution coefficient of each reference spectrum by least squares to characterize the contribution changes of the corresponding chemical components or characteristic signals at different time points.

[0012] In S6, the Raman heatmap is divided into a time-dimensional Raman heatmap or a component contribution heatmap according to the data source. The time-dimensional Raman heatmap is generated by mapping the characteristic peak intensity value to a color value; the component contribution heatmap is generated by mapping the time contribution coefficient to a color value.

[0013] In S7, the index mapping relationship is established by establishing a one-to-one correspondence between the heat map position and the corresponding time point spectral array address, so as to realize fast access to the Raman spectrum corresponding to any time point.

[0014] In S8, the visualization interface includes a heatmap display area and a spectrum display area. When the user clicks on any time position in the heatmap display area, the spectrum display area is simultaneously refreshed and displays the Raman spectrum curve of the corresponding time point.

[0015] This invention discloses an interactive analysis system for time-resolved Raman spectroscopy heatmaps with multi-peak decomposition, comprising a data acquisition module, a spectral preprocessing module, a peak identification module, a characteristic peak integration module, a multi-peak decomposition module, a heatmap generation module, an index mapping module, an interactive control module, and a visualization display module. The data acquisition module is connected to the spectral preprocessing module, which in turn is connected to the peak identification module, the peak identification module to the characteristic peak integration module, the characteristic peak integration module to the multi-peak decomposition module, the multi-peak decomposition module to the heatmap generation module, the heatmap generation module to the index mapping module, the index mapping module to the interactive control module, and the interactive control module to the visualization display module.

[0016] The data acquisition module is used to acquire Raman spectral data corresponding to different time points; The aforementioned spectral preprocessing module is used to perform baseline correction, noise smoothing, and normalization on the Raman spectrum; The peak identification module is used to identify characteristic peaks in Raman spectra; The characteristic peak integration module is used to perform integration calculations on the target peak in the spectrum corresponding to each time point based on the identified characteristic peak position and the preset integration interval to obtain the characteristic peak intensity. The multi-peak decomposition module is used to perform multi-peak decomposition on Raman spectra at different time points and solve for the time contribution coefficients of each component or characteristic signal. The aforementioned heatmap generation module is used to generate a time-dimensional Raman heatmap or a component contribution heatmap based on the characteristic peak intensity and / or time contribution coefficient. The index mapping module is used to establish an index mapping relationship between heat map locations and corresponding time point spectral data; The interactive control module is used to respond to the user's click operation on the heatmap; The visualization module is used to display the heat map and the Raman spectral curves at the corresponding time points.

[0017] The aforementioned interactive analysis system for time-resolved Raman spectroscopy heatmaps with multi-peak decomposition further includes a data storage module, a parameter setting module, a result export module, a reference spectral library management module, and a sample recording module. The data storage module saves the original spectra, time labels, and relevant acquisition parameters. The parameter setting module allows users to set the smoothing window, baseline fitting order, peak threshold, integration interval, and heatmap display range. The result export module exports heatmap images, peak tables, integration results, decomposition results, and original spectra. The reference spectral library management module saves standard Raman spectra of different substances for use during multi-peak decomposition. The sample recording module saves metadata such as sample number, sample type, acquisition time, and experimental environment.

[0018] This invention acquires Raman spectral data at multiple time points by continuously or time-segmented Raman acquisition of the sample under test, and then preprocesses, identifies characteristic peaks, and decomposes the spectral data. Furthermore, a time-dimensional Raman heatmap is constructed based on the intensity of characteristic peaks or the time contribution coefficients obtained from the decomposition, and an index mapping relationship is established between the heatmap position and the corresponding Raman spectrum. This allows users to view the Raman spectrum at any time point in real time by clicking on any time position in the heatmap, thereby achieving linked analysis of the time and spectral dimensions.

[0019] The present invention provides an interactive analysis method and system for time-resolved Raman spectroscopy thermograms with multi-peak decomposition. Compared with the prior art, its advantages are as follows: 1. It can convert continuously or time-division acquired Raman spectral data into time-dimensional heatmaps, enhancing the visualization of time-series changes; 2. Improve the stability of dynamic Raman signal extraction through adaptive peak identification and peak region integration; 3. Improve the accuracy of dynamic analysis of multiple components in complex samples by using multi-peak decomposition algorithms; 4. Through the linkage mechanism between heatmaps and spectra, users can quickly locate the complete spectrum corresponding to any point in time; 5. Applicable to long-term monitoring and Raman time-series analysis of sweat, blood and other dynamically changing samples. Attached Figure Description

[0020] Figure 1 This is a schematic flowchart of an interactive analysis method for time-resolved Raman spectroscopy thermograms with multi-peak decomposition, as described in an embodiment of the present invention.

[0021] Figure 2 This is a schematic diagram of the structure of an interactive analysis system for time-resolved Raman spectroscopy thermograms with multi-peak decomposition, as described in an embodiment of the present invention.

[0022] Figure 3 This is a schematic diagram of the interactive visualization analysis interface for time-resolved Raman spectroscopy thermograms in an embodiment of the present invention. Detailed Implementation

[0023] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the following embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. Without departing from the basic concept of the present invention, those skilled in the art can make equivalent substitutions or conventional adjustments to the acquisition parameters, preprocessing methods, peak identification conditions, multi-peak decomposition models, heatmap construction methods, interface interaction forms, and data organization methods, all of which should fall within the scope of protection of the present invention.

[0024] This invention provides an interactive analysis method and system for time-resolved Raman spectroscopy thermograms with multi-peak decomposition. It is applicable to samples such as sweat, blood, serum, residual bodily fluids, evaporating droplets, reaction process samples, and other analytes whose Raman signals change over time during detection. Unlike traditional analysis methods that organize data using spatial coordinates, this invention primarily uses the time dimension for organization. It processes a large number of Raman spectra obtained during continuous or time-segmented acquisition of the sample and constructs a two-dimensional thermogram based on the time axis. This allows users to intuitively observe the signal change patterns of the sample during long-term measurement and interactively view the complete Raman spectrum at any given time point through the thermogram.

[0025] The overall structure of an interactive analysis system for time-resolved Raman spectroscopy thermograms with multi-peak decomposition is as follows: Figure 2 As shown, Figure 2 The diagram illustrates the connection relationships between the Raman spectroscopy data acquisition module, spectral preprocessing module, peak identification module, characteristic peak integration module, multi-peak decomposition module, heatmap generation module, index mapping module, interactive control module, and visualization display module. The system of this invention includes at least these modules: data acquisition module, spectral preprocessing module, peak identification module, characteristic peak integration module, multi-peak decomposition module, heatmap generation module, index mapping module, interactive control module, and visualization display module. These modules can be integrated into the same computing device or implemented collaboratively by the Raman spectroscopy acquisition device and the host computer analysis software.

[0026] The data acquisition module is a Raman spectroscopy acquisition device used to continuously or time-divisionally acquire Raman data of the sample to be tested, so as to obtain Raman spectral data arranged in chronological order; the spectral preprocessing module is used to perform baseline correction, smoothing and denoising, and normalization on the raw Raman spectral data; the peak identification module is used to automatically extract spectral characteristic peaks at different time points; the characteristic peak integration module is used to calculate the integrated intensity of the characteristic peak at different time points to obtain the characteristic peak intensity; the multi-peak decomposition module is used to decompose complex mixed spectra to obtain the time contribution coefficients of different chemical components or characteristic signals as a function of time; the heatmap generation module is used to generate Raman heatmaps or component contribution heatmaps in the time dimension based on the characteristic peak intensity or time contribution coefficients; the index mapping module is used to establish the index mapping relationship between the heatmap position and the spectrum at the corresponding time point; the interactive control module is used to respond to user clicks, drags, selections, or switching of heatmaps; and the visualization display module is used to display the Raman heatmaps in the time dimension and the component contribution heatmaps, as well as the Raman spectral curves at the corresponding time points.

[0027] In a preferred embodiment, the system may further include a data storage module, a parameter setting module, a result export module, a reference spectral library management module, and a sample recording module. The data storage module stores the original spectra, time stamps, and related acquisition parameters; the parameter setting module allows users to set the smoothing window, baseline fitting order, peak threshold, integration interval, and heatmap display range; the result export module exports heatmap images, peak tables, integration results, decomposition results, and original spectra; the reference spectral library management module stores standard Raman spectra of different substances for use during multi-peak decomposition; and the sample recording module stores metadata such as sample number, sample type, acquisition time, and experimental environment.

[0028] An interactive analysis method for time-resolved Raman spectroscopy thermograms with multi-peak decomposition is described in the flowchart below. Figure 1 This document illustrates the processing flow for Raman spectral data acquisition, spectral preprocessing, peak identification, characteristic peak integration calculation, multi-peak spectral decomposition, heatmap generation, index mapping, and interactive display. Specifically, it includes S1: Time-resolved Raman data acquisition: In this embodiment, the Raman spectral acquisition device continuously acquires data from the sample under test or acquires data at preset time intervals. The sample under test can be a droplet sample, a body fluid sample during the drying process, a blood sample in long-term measurement, or other dynamic samples that change over time. During the acquisition process, the sample position remains basically fixed, and the system sequentially acquires Raman spectra at multiple time points along a time-based framework.

[0029] Let the acquisition time points be t1, t2, t3, ..., tn, and each time point corresponds to a complete Raman spectrum. Then all the data can be organized into a two-dimensional matrix: S(t, λ) Where t represents the time dimension, λ represents the Raman shift, and S represents the Raman intensity value at the corresponding time point and wavenumber position. If the total acquisition time is T and the sampling time interval is Δt, then the total number of sampling points n is approximately T / Δt. The acquisition time interval can be set to the second, ten-second, minute, or longer intervals according to actual needs.

[0030] In one specific embodiment, Raman spectroscopy can be performed on sweat samples for 30 minutes, recording a complete Raman spectrum every 10 seconds. If each spectrum covers -221 cm⁻¹... -1 Up to 3304cm -1 By measuring the Raman shift range and including 2048 sampling points, 180 spectra can be obtained, forming a time-resolved Raman data matrix with dimensions of 180×2048. A similar method can be used for continuous acquisition of blood samples to observe changes in Raman signals caused by agglutination, evaporation, or component migration.

[0031] In addition to the spectrum itself, the system preferably stores the following information: acquisition start time, sampling interval, laser wavelength, integration time, laser power, objective lens magnification, ambient temperature, ambient humidity, and sample number. This information can serve as auxiliary parameters for subsequent analysis or experimental reproduction.

[0032] Data organization and index establishment: To ensure the efficiency of subsequent processing and interactive display, this invention preferably standardizes the organization of time-series spectral data. Specifically, the following data structures can be established: a wavenumber array Wavenumber[λ], used to record the Raman shift corresponding to each sampling point; a time array Time[t], used to record the acquisition time of each spectrum; an intensity matrix Spectrumt, used to store the intensity values ​​of the complete spectrum at each time point; and a metadata structure MetaInfo, used to store instrument parameters and experimental conditions.

[0033] In a preferred embodiment, the system immediately establishes a time index mapping relationship after importing or acquiring data, meaning that for any time number i, the corresponding Spectrum can be directly located. After the heatmap is generated, the system further establishes an index mapping relationship between the heatmap display position and the original spectrum. For example, when the horizontal axis of the heatmap is time and the vertical axis is Raman shift, clicking on a time column will directly locate the entire spectrum corresponding to that time point; when the horizontal axis of the heatmap is time and the vertical axis is characteristic peak number or component number, the system can also look up the original spectrum or local peak region data of the corresponding time point based on the clicked position.

[0034] This index structure ensures that even with a large amount of time-series spectral data, users can still quickly query spectral information at any given time, thereby improving the efficiency of interactive analysis.

[0035] S2: Spectral Preprocessing: During long-term continuous acquisition, Raman spectra are often affected by factors such as fluorescence background drift, random noise, instrument stability fluctuations, and overall intensity changes. Therefore, the original spectrum needs to be preprocessed before performing subsequent peak identification and multi-peak decomposition. The preprocessing preferably includes baseline correction, smoothing and denoising, and normalization.

[0036] Baseline correction: For any original spectrum S_raw(t,λ) corresponding to a time point t, a background baseline function B(λ) is first constructed. In a preferred embodiment, the baseline function is expressed as an nth-order polynomial: B(λ) = a0 + a1λ + a2λ 2 +...+anλ n After solving for the coefficients using least-squares fitting, the background function is subtracted from the original spectrum to obtain the baseline-corrected spectrum. Scorrected(t,λ)=Sraw(t,λ)-B(λ) The polynomial order n can be set to 3rd, 4th, 5th, or other suitable values ​​based on the actual background morphology. When the fluorescence background varies significantly, piecewise fitting, adaptive iterative fitting, or other equivalent baseline subtraction methods can also be used.

[0037] Smoothing and Denoising: To reduce the impact of high-frequency noise on peak identification and integration results, the system performs smoothing processing on the corrected spectrum. Smoothing methods can include moving average, Savitzky-Golay smoothing, or Gaussian smoothing. In one specific embodiment, the smoothing window width can be set to 5, 7, or 9 points to reduce random noise while preserving peak shape information as much as possible.

[0038] Normalization: Since sample states, illumination conditions, and overall signal intensity may fluctuate at different time points, this invention preferably performs normalization processing on the spectra to improve the comparability between spectra at different times. Normalization methods may include maximum value normalization, total area normalization, or reference peak normalization. For example, each spectrum can be divided by its maximum intensity value to obtain a normalized spectrum; alternatively, a relatively stable reference peak can be selected, and its intensity can be used as the normalization benchmark.

[0039] After the above processing, a preprocessed spectrum S_pre(t,λ) suitable for subsequent analysis can be obtained.

[0040] S3: Adaptive Peak Identification: After preprocessing, the system performs peak identification on the Raman spectra corresponding to each time point to automatically extract the main characteristic peaks. Peak identification optimization includes the following steps: First, calculate the first derivative D(λ)=dS(λ) / dλ on the preprocessed spectrum; then find the zero crossover point where the derivative changes from positive to negative as the candidate peak position; further calculate the peak height, peak width and local signal-to-noise ratio of the candidate peaks, and screen out the true characteristic peaks according to the threshold conditions.

[0041] In a preferred embodiment, the peak threshold T is defined adaptively as follows: T=μ+kσ Where μ is the current average spectral intensity, σ is the current standard deviation of spectral intensity, and k is an empirical coefficient. This method can adapt to changes in overall signal intensity at different time points, avoiding missed or false detections caused by fixed thresholds.

[0042] In a further implementation, minimum peak spacing, minimum peak width, and local signal-to-noise ratio thresholds can be set to avoid mistaking noise spikes for valid peaks. For cases with wide peaks or shoulder peaks, the peak center position can be determined by combining the second derivative or local fitting methods. Finally, the system can generate a set of characteristic peaks P(t)={p1,p2,p3,...} for each time point and record parameters such as peak position, peak height, peak width, and integration boundary for each peak.

[0043] S4: Characteristic Peak Intensity Calculation: In this invention, to enhance the stability of dynamic signal extraction, it is preferable not to directly use the instantaneous intensity of a single wavenumber point, but to integrate the intensity within a certain waveband near the target peak. Let a target peak be x. p Then, define the integration interval [x] in its vicinity. p -Δ,x p Integrating or summing the spectral intensities within this interval, we obtain the characteristic peak intensity I(t) of the peak at time point t: I(t) == Where I(t) represents the characteristic peak intensity at time t; t represents the Raman spectroscopy acquisition time; S(t,x) represents the spectral intensity at time t and wavenumber position x; x represents the wavenumber position of the Raman spectrum; x p The value represents the center wavenumber position corresponding to the target peak; Δ represents the integration half-width, used to determine the upper and lower limits of the integration interval. The parameter Δ can be set to 5cm based on the peak width, spectral resolution, and sample signal characteristics. -1 10cm -1 15cm -1Or other suitable values. If a characteristic peak drifts slightly during long-term acquisition, the system can also adaptively adjust the integration center based on the location of the local maximum value, thereby ensuring the stability of the extraction results.

[0044] S5: Multi-peak spectral decomposition: For a single target peak, a characteristic peak intensity curve varying with time can be obtained; for multiple target peaks, multiple time-series characteristic curves can be obtained. Furthermore, when the system processes the entire Raman shift range or multiple target peaks simultaneously, a two-dimensional intensity matrix can be formed for heatmap construction.

[0045] S6: Time-Dimensional Heatmap Generation: A key step in this invention is to construct a heatmap based on time-series Raman data, thereby visually representing the variation of the Raman signal over time. The heatmap has at least one time axis, and the other axis can be set as a Raman shift axis, target peak number axis, characteristic band number axis, or component number axis, depending on the analysis objective. Colors in the heatmap represent the signal strength, integral value, or contribution coefficient obtained from decomposition at the corresponding location.

[0046] In one implementation, using Raman shift on the vertical axis of the heatmap creates a "time-wavenumber Raman spectrum." In this case, each column in the heatmap corresponds to a time point, and each row corresponds to a wavenumber position, with color indicating the Raman intensity at that time point and wavenumber. This method is suitable for visually observing the changes in the entire spectrum over time.

[0047] In another implementation, if the vertical axis of the heatmap uses characteristic peak numbers or characteristic peak bands, a "time-characteristic peak intensity heatmap" can be formed. This method is suitable for long-term monitoring of a small number of key peaks and facilitates the analysis of the enhancement, weakening, shifting, or disappearance of multiple characteristic peaks over time.

[0048] In a further embodiment, to generate the component contribution heatmap, the vertical axis of the heatmap uses the component numbers obtained through multi-peak decomposition, thus forming a "time-component contribution heatmap". This component contribution heatmap can intuitively display the relative change patterns of different chemical components throughout the entire collection process.

[0049] To improve display quality, the system can first normalize the heatmap data and then generate a color heatmap using a preset color mapping table. The color mapping table can use blue-green-yellow-red gradients, grayscale gradients, or other color schemes suitable for scientific visualization. The system also supports automatic scaling of the display area, manual setting of color scale upper and lower limits, and magnification of local time intervals for observation.

[0050] Multi-peak decomposition and dynamic component analysis: For complex biological samples such as sweat and blood, a single peak intensity is often insufficient to accurately characterize changes in the sample's state, as multiple components may have overlapping peaks within the same wavelength band. Therefore, this invention further introduces a multi-peak decomposition step to enhance the dynamic resolution capability of complex mixed spectra.

[0051] For any time point t, the corresponding Raman spectrum can be expressed as a linear combination of multiple reference spectra: S(t,λ)=Σai(t)Ri(λ)+ε Where Ri(λ) represents the standard or template spectrum of the i-th reference component, ai(t) represents the contribution coefficient of the component at time point t, and ε represents noise or fitting residual. The reference spectrum can be derived from a pre-established standard library, measured spectra of pure substances, or extracted from representative spectra in the current dataset.

[0052] In a preferred embodiment, the system uses the least squares method or non-negative least squares method to solve for each ai(t). When it is required that the contributions of all components are non-negative, a constraint condition of ai(t) ≥ 0 can be added. Finally, time contribution curves of multiple components can be obtained, and a time-component contribution heatmap can be further constructed. For samples with evaporation, concentration, crystallization, or reaction processes, this step can more realistically reflect the dynamic changes of multiple components within the sample during the detection process.

[0053] S7: Time-Spectral Index Construction - Index Mapping of Heatmap Locations to Spectra: To enable linked analysis between heatmaps and spectra, the system prioritizes establishing a one-to-one correspondence between the heatmap display position and the original spectrum. When the horizontal axis of the heatmap represents time and the vertical axis represents Raman shift, clicking on a time column allows the system to look up the corresponding time point spectrum in the original spectral matrix based on that time position. If the user clicks on a specific cell, the system not only displays the complete spectrum for that time point but also additionally marks the wavenumber position corresponding to that cell.

[0054] When the vertical axis of the heatmap is a characteristic peak number or a component number, the system can also resolve the corresponding time point, peak number or component number by clicking on the location, and display the complete spectrum, local magnified curve of the target peak or component decomposition results at that time point.

[0055] In a preferred embodiment, the system directly maps the horizontal position of the heatmap to the time array Time[i], thereby establishing the following mapping relationship: Heatmap(i, j) → Spectrum(Time[i],λ) With the help of this index mapping relationship, the system can still achieve a relatively fast interactive response even with a large amount of time-series data.

[0056] S8: User interaction clicks and real-time spectrum display Interactive display and linkage analysis: like Figure 3 As shown, the visualization interface of this invention preferably adopts a multi-region linked layout, and more preferably a four-region layout. Figure 3 The layout of the original time-series heatmap display area, characteristic peak heatmap display area, component contribution heatmap display area, and Raman spectrum curve display area is shown. The interface can simultaneously display the original time-series heatmap, characteristic peak intensity heatmap, component contribution heatmap, and the corresponding Raman spectrum curve. Specifically, the original time-series heatmap shows the overall distribution of the complete Raman spectrum as a function of Raman shift at different time points; the characteristic peak intensity heatmap shows the integral intensity change of the selected characteristic peak over time; the component contribution heatmap shows the contribution changes of each component at different time points after multi-peak decomposition; and the Raman spectrum curve displays the complete Raman spectrum corresponding to the currently selected time point.

[0057] In a preferred embodiment, the interface can adopt a two-dimensional grid arrangement, where the upper left area is the "original spectral heatmap" display area, the upper right area is the "characteristic peak heatmap" display area, the lower left area is the "component contribution heatmap" display area, and the lower right area is the "Raman spectrum curve" display area. A system title can be set at the top of the interface, and parameter adjustment areas, peak selection areas, time range selection areas, heatmap display control areas, and result saving areas can be set around or below the interface.

[0058] When a user clicks, drags, or selects a region in any heatmap area, the system performs the following steps: First, it acquires the user's interaction location; then, it analyzes the corresponding time point or time interval; next, it locates the corresponding original spectral data, characteristic peak integral data, or component contribution data based on the index mapping relationship; finally, it plots the complete Raman spectral curve for the corresponding time point in real time in the Raman spectral curve display area. Preferably, the system can also simultaneously mark the currently selected time position in multiple heatmaps to achieve linked display between the original spectral heatmap, characteristic peak heatmap, component contribution heatmap, and single-time Raman spectrum.

[0059] For operations involving selecting a specific time interval, the system can also average all spectra within that time period and display the corresponding average spectral curve to reduce the impact of fluctuations in a single sampling. Furthermore, when the user switches between different characteristic peaks or different component channels, the characteristic peak heatmap and component contribution heatmap can be refreshed synchronously, and the Raman spectral curve in the lower right area can also be updated accordingly.

[0060] In a further embodiment, the system can also simultaneously overlay and display the original spectrum, preprocessed spectrum, peak position markers, fitted spectrum, and residual curve in the Raman spectral curve display area, thereby helping users to more comprehensively analyze the signal characteristics at a specific time point. Furthermore, the system can display the current time point, selected characteristic peaks, peak integration results, component contribution values, and other auxiliary information in the status bar or information display area to improve interactive analysis efficiency and result interpretability.

[0061] Technical advantages: Compared with existing methods that mainly rely on single-spectrum judgment or static trend chart analysis, this invention has at least the following advantages: First, it can centrally display a large amount of Raman spectral data collected continuously over a long period of time in the form of a heatmap, which facilitates the observation of the overall temporal change pattern; Second, it can improve the stability of feature signal extraction through adaptive peak identification and peak region integration; Third, it can improve the accuracy of dynamic analysis of multiple components in complex samples through multi-peak decomposition; Fourth, it can establish a mapping relationship between the heatmap position and the spectrum at the corresponding time, realizing time-spectrum linkage display; Fifth, it is suitable for long-term monitoring and interactive analysis of dynamically changing samples such as sweat and blood.

[0062] In summary, this invention can effectively solve the problems of unintuitive visualization, difficulty in comparing dynamic changes, inconvenience in querying spectra at any time, and difficulty in analyzing multi-component complex samples in time-resolved Raman data analysis. It has good practical value and prospects for promotion.

Claims

1. An interactive analysis method for time-resolved Raman spectroscopy thermograms with multi-peak decomposition, characterized in that, Includes the following steps: S1. Continuously or time-segmentedly acquire Raman spectral data corresponding to different time points of the sample to be tested; S2. The acquired Raman spectral data is subjected to baseline correction, noise smoothing and normalization to obtain preprocessed Raman spectral data; S3. Perform peak identification on the preprocessed spectral data to extract characteristic peak information; S4. Perform integral calculations based on the characteristic peak information to obtain the characteristic peak intensities at different time points; S5. Perform multi-peak decomposition on the preprocessed Raman spectra at different time points to solve for the time contribution coefficients of different chemical components or characteristic signals; S6. Generate a Raman heatmap based on the characteristic peak intensity and / or the time contribution coefficient; S7. Establish an index mapping relationship between the Raman heatmap location and the corresponding time point Raman spectral data; S8. Respond to the user's interactive operation on the Raman heat map, locate the corresponding time point and retrieve the corresponding Raman spectrum; S9. The Raman spectral curve corresponding to the time point is displayed in real time in the visualization interface.

2. The interactive analysis method for time-resolved Raman spectroscopy thermograms with multi-peak decomposition according to claim 1, characterized in that, In S2, the baseline correction uses a polynomial fitting method to construct a background baseline function, and the corrected Raman spectrum is obtained by subtracting the background baseline function from the original spectrum.

3. The interactive analysis method for time-resolved Raman spectroscopy thermograms with multi-peak decomposition according to claim 1, characterized in that, In S3, peak identification is performed by calculating the first derivative of the Raman spectrum and determining the candidate peak position by combining the zero crossover position where the derivative changes from positive to negative. Then, characteristic peaks are selected based on the peak intensity threshold.

4. The interactive analysis method for time-resolved Raman spectroscopy thermograms with multi-peak decomposition according to claim 3, characterized in that, The peak intensity threshold is adaptively determined based on the current average intensity and the current standard deviation of the spectral intensity, satisfying T = μ + kσ, where T is the peak intensity threshold, μ is the current average intensity, σ is the current standard deviation of the spectral intensity, and k is an empirical coefficient.

5. The interactive analysis method for time-resolved Raman spectroscopy thermograms with multi-peak decomposition according to claim 1, characterized in that, In S4, the characteristic peak integral calculation involves setting an integration interval at the center wavenumber of the characteristic peak and accumulating or integrating the spectral intensity within the interval to obtain the characteristic peak intensity at the corresponding time point.

6. The interactive analysis method for time-resolved Raman spectroscopy thermograms with multi-peak decomposition according to claim 1, characterized in that, In S5, multi-peak decomposition represents the Raman spectrum to be measured as a linear combination of multiple reference spectra, and solves the contribution coefficient of each reference spectrum by least squares to characterize the contribution change of the corresponding chemical component at different time points.

7. The interactive analysis method for time-resolved Raman spectroscopy thermograms with multi-peak decomposition according to claim 1, characterized in that, In step S6, the Raman heatmap is divided into a time-dimensional Raman heatmap or a component contribution heatmap based on the data source. The time-dimensional Raman heatmap or the component contribution heatmap is generated by mapping the characteristic peak intensity value or contribution coefficient to color value.

8. The interactive analysis method for time-resolved Raman spectroscopy thermograms with multi-peak decomposition according to claim 1, characterized in that, In step S7, the index mapping relationship is established by establishing a one-to-one correspondence between the heat map position and the corresponding time point spectral array address, so as to realize fast access to the Raman spectrum corresponding to any time point.

9. The interactive analysis method for time-resolved Raman spectroscopy thermograms with multi-peak decomposition according to claim 1, characterized in that, In S8, the visualization interface includes a heat map display area and a spectrum display area. After the user clicks on any time position in the heat map display area, the spectrum display area is refreshed synchronously and displays the Raman spectrum curve of the corresponding time point.

10. An interactive analysis system for time-resolved Raman spectroscopy thermograms with multi-peak decomposition, characterized in that, include: It includes a data acquisition module, a spectral preprocessing module, a peak identification module, a characteristic peak integration module, a multi-peak decomposition module, a heatmap generation module, an index mapping module, an interactive control module, and a visualization display module. The data acquisition module is connected to the spectral preprocessing module, the spectral preprocessing module is connected to the peak identification module, the peak identification module is connected to the characteristic peak integration module, the characteristic peak integration module is connected to the multi-peak decomposition module, the multi-peak decomposition module is connected to the heatmap generation module, the heatmap generation module is connected to the index mapping module, the index mapping module is connected to the interactive control module, and the interactive control module is connected to the visualization display module. The data acquisition module is used to acquire Raman spectral data at different time points; The spectral preprocessing module is used to perform baseline correction, noise smoothing, and normalization on Raman spectra. Peak identification module, used to identify characteristic peaks in Raman spectra; The characteristic peak integration module is used to perform integration calculations on the target peak in the spectrum corresponding to each time point based on the identified characteristic peak position and the preset integration interval to obtain the characteristic peak intensity. The multi-peak decomposition module is used to perform multi-peak decomposition on Raman spectra at different time points and solve for the time contribution coefficients of each component or characteristic signal. A heatmap generation module is used to generate a time-dimensional Raman heatmap or a component contribution heatmap based on the characteristic peak intensity and / or time contribution coefficient. The index mapping module is used to establish the index mapping relationship between heat map locations and corresponding time point spectral data; The interactive control module is used to respond to user clicks on the heatmap; The visualization module is used to display heatmaps and Raman spectral curves at corresponding time points.

Images