Chromatographic state intelligent identification method based on atlas

Abstract

The invention relates to the field of image data processing, in particular to a chromatogram state intelligent identification method based on an atlas. The method comprises the steps that multiple matching templates of target chromatographic peaks are established, and one matching template corresponds to the peak shape of one type of target chromatographic peaks; extracting noise peaks removed during denoising of the to-be-identified chromatogram; determining the similarity between the noise peak and each matching template to obtain a similarity set; determining the comprehensive similarity of the noise peak and the target chromatographic peak according to the similarity set; and identifying whether the noise peak is a target chromatographic peak or not according to a relationship between the comprehensive similarity and a set threshold value. According to the method, the chromatographic peak which is misjudged as noise and is removed during denoising can be identified, and the accuracy of chromatographic peak identification is improved, so that the accuracy of chromatographic state identification is improved.

Description

Technical Field

[0001] The present invention relates to the field of image data processing, and more specifically, to a spectrum-based intelligent identification method for chromatographic states. Background Art

[0002] Chromatographic analysis is one of the important instrumental analysis methods. It has fast analysis speed, high separation efficiency and accurate qualitative and quantitative analysis of complex multi-component mixtures. It has been applied to various industries such as petroleum, chemical industry, medicine, health, metallurgy, geology, agriculture, environmental protection, etc. Chromatographic state recognition is the recognition of chromatographic peaks in real-time monitoring and analysis of chromatograms in chromatographic analysis. The recognition of chromatographic peaks is the basis and key to qualitative analysis and quantitative calculation in chromatographic analysis, and affects the determination of the name and concentration of the final component chromatographic state recognition. Therefore, there is a high requirement for the accuracy of chromatographic peak recognition.

[0003] For the identification of chromatographic peaks, the Fourier deconvolution method, the derivative method, etc. are currently used more frequently, but the identification accuracy of these methods cannot meet the requirements. In this regard, a chromatographic peak identification method based on a convolutional neural network is disclosed in a Chinese patent application document with publication number CN11767790A. This method identifies chromatographic peaks by using a convolutional neural network. Although this method does effectively improve the accuracy of chromatographic peak identification, this method, like other methods, pays more attention to the identification of the positions of chromatographic peaks with strong signal intensity, and does not take into account that the signal intensity of chromatographic peaks with less content is weak and the frequency characteristics are similar to noise. Therefore, these chromatographic peaks may be misjudged as noise and removed in the denoising process, resulting in low accuracy in chromatographic peak identification.

[0004] Therefore, how to identify these chromatographic peaks with weak signal intensity and frequency characteristics similar to noise is an urgent problem to be solved. Summary of the invention

[0005] In order to solve the technical problem that the chromatographic peaks with less content are mistakenly judged as noise and removed, resulting in low recognition accuracy, the present invention provides solutions in the following aspects.

[0006] The present invention provides a spectrum-based chromatographic state intelligent recognition method, comprising: establishing multiple matching templates of target chromatographic peaks, wherein one matching template corresponds to the peak shape of a type of target chromatographic peak; extracting noise peaks removed during denoising of a chromatogram to be identified; determining the similarity between the noise peak and each matching template to obtain a similarity set, wherein the calculation expression of the similarity is: ; In the formula, is the first The similarity between the noise peak and the matching template; is the first The noise peak and The number of regions where the matching templates overlap; is the first The noise peak and The total number of overlapping points in the overlapping area l of the matching templates; is the first The total number of points of the noise peak; The comprehensive similarity between the noise peak and the target chromatographic peak is determined according to the similarity set; and whether the noise peak is the target chromatographic peak is identified according to the relationship between the comprehensive similarity and a set threshold.

[0007] By calculating the similarity between the noise peak removed during denoising of the chromatogram to be identified and the matching template of the established target chromatographic peak (i.e., the chromatographic peak with weak signal intensity and frequency characteristics similar to the noise), it is determined whether the noise peak is the target chromatographic peak, avoiding the problem of removing the target chromatographic peak that is mistakenly identified as noise, thereby improving the accuracy of chromatographic peak identification and further improving the accuracy of chromatographic state identification.

[0008] Furthermore, multiple matching templates for the target chromatographic peak are established, including: obtaining the target chromatographic peak in multiple historical denoised chromatograms; determining the peak shape feature vector of the target chromatographic peak, wherein the peak shape feature vector includes curvature, skewness and peak width ratio; clustering the peak shape feature vectors, and for any obtained cluster, fitting the target chromatographic peak in the cluster using multiple fitting functions, wherein the independent variable of the fitting function is time and the dependent variable is the signal intensity of the chromatographic peak; and using the peak shape generated by the fitting function with the largest fitting degree as the matching template of the cluster.

[0009] The efficiency of chromatographic peak recognition can be improved by clustering the target chromatographic peaks and then matching them. In addition, the accuracy of the generated matching templates is improved by generating matching templates corresponding to each cluster using the fitting function with the largest fitting degree, thereby improving the accuracy of the subsequent determination of the similarity between the noise peak and the matching template. Furthermore, the peak shape generated by the fitting function with the greatest degree of fit is used as the matching template of the cluster cluster, including: according to the average peak shape feature vector of the cluster cluster, the fitting function with the greatest degree of fit is corrected so that the peak shape feature vector corresponding to the peak shape generated by the fitting function is close to the average peak shape feature vector of the cluster cluster; and the peak shape generated by the corrected fitting function is used as the matching template corresponding to the cluster cluster.

[0010] By modifying the fitting function, the generated matching template can better characterize the peak shape characteristics of all target chromatographic peaks in the cluster, thereby improving the accuracy of chromatographic peak recognition.

[0011] Furthermore, the calculation expression of the maximum fit is: ; In the formula, For the The maximum fitness among all fitness of clusters; is the maximum value of the horizontal coordinate of the fitting function; For the Clusters within The mean signal intensity of the chromatographic peak at time; For the The clusters are used The fitting function is The fitted value of the signal intensity of the chromatographic peak at time.

[0012] Furthermore, the calculation expression of the comprehensive similarity is: ; In the formula, is the first The comprehensive similarity between the noise peak and the target chromatographic peak; is the first The noise peak and Similarity of matching templates; is the first The mean of the similarities between the noise peaks and each matching template; is the total number of types of matching templates; For the The maximum fitness among all fitness of the clusters.

[0013] because describes the deviation of each similarity from the overall similarity, and the maximum fit describes the accuracy of each matching template. As well as the maximum fitting degree, it is possible to improve the comprehensive similarity between the determined noise peak and the target chromatographic peak, thereby improving the accuracy of chromatographic peak identification.

[0014] Furthermore, the calculation expression of the comprehensive similarity is: ; In the formula, is the first The comprehensive similarity between the noise peak and the target chromatographic peak; is the first The variance of the similarity between the noise peaks and each matching template; is the total number of types of matching templates; For the The maximum fitness among all fitness of clusters; is the normalization function; For An exponential function with base .

[0015] Since the variance of the similarity describes the fluctuation degree of the similarity between the noise peak and each matching template, and the maximum fit describes the accuracy of each matching template, therefore, by comprehensively considering the variance of the similarity and the maximum fit, the comprehensive similarity between the noise peak and the target chromatographic peak can be improved, thereby improving the accuracy of chromatographic peak identification.

[0016] Furthermore, the expression of the peak shape feature vector is: ; In the formula, For the The historical chromatogram The peak shape feature vector of the target chromatographic peak; For the The historical chromatogram Target chromatographic peak value Peak width The ratio of For the The historical chromatogram The average curvature of the target chromatographic peaks; For the The historical chromatogram The skewness of the target chromatographic peak.

[0017] Further, determining the similarity between the noise peak and each matching template includes: matching the noise peak with each matching template according to a template matching algorithm, and determining the similarity between the noise peak and each matching template according to the overlap during matching.

[0018] Furthermore, extracting the noise peaks removed from the chromatogram to be identified during denoising includes: acquiring the chromatogram to be identified, denoising the chromatogram to be identified according to wavelet transform; comparing the chromatograms before and after denoising, and extracting the noise peaks by moving the window.

[0019] Furthermore, whether the noise peak is a chromatographic peak is identified based on the relationship between the comprehensive similarity and the set threshold, including: if the comprehensive similarity is greater than or equal to the set threshold, the noise peak is determined as a target chromatographic peak and restored to the chromatogram to be identified after denoising; if the comprehensive similarity is less than the set threshold, the noise peak is determined as a real noise peak.

[0020] By restoring the target chromatographic peak misjudged as noise to the chromatogram, the accuracy and reliability of the obtained chromatogram are improved, thereby improving the accuracy of chromatographic state identification.

[0021] In summary, the present invention improves the accuracy of chromatographic peak recognition by restoring the target chromatographic peak misjudged as noise to the chromatogram, thereby avoiding the removal of the chromatographic peak misjudged as noise, and further improves the accuracy of chromatographic state recognition. In addition, by generating matching templates corresponding to each clustering cluster using the fitting function with the largest fitting degree, the accuracy of the generated matching template is improved, thereby improving the accuracy of the subsequent determination of the similarity between the chromatographic peak and the matching template. BRIEF DESCRIPTION OF THE DRAWINGS By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein: Figure 1 is a flow chart schematically illustrating a method for intelligently identifying chromatographic states based on a spectrum according to an embodiment of the present invention; Figure 2 FIG. 4 is a flow chart schematically illustrating the process of establishing a matching template for a target chromatographic peak according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0023] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0024] Figure 1 It is a flow chart schematically showing a method for intelligently identifying chromatographic states based on a spectrum according to an embodiment of the present invention.

[0025] The present invention provides a method for intelligently identifying chromatographic states based on a spectrum, such as Figure 1As shown, the method of the present invention comprises: S1. Establish multiple matching templates for target chromatographic peaks.

[0026] In this embodiment, the target chromatographic peak is a low-frequency chromatographic peak (ie, a chromatographic peak with a relatively low content, which is easily misjudged as noise and removed), and one matching template corresponds to the peak shape of one type of target chromatographic peak.

[0027] Specifically, Figure 2 As shown in the figure, multiple matching templates for establishing target chromatographic peaks include: S101, obtaining target chromatographic peaks from multiple historical denoised chromatograms.

[0028] Specifically, multiple historical denoised chromatograms can be obtained from the database, and each chromatogram is annotated with the peak name, retention time, corresponding components, etc. Further, the target chromatographic peak in the historical chromatogram can be obtained by wavelet transform.

[0029] S102, determining a peak shape feature vector of the target chromatographic peak; Specifically, the starting point, apex and end point of the target chromatographic peak on the historical chromatogram can be identified by the derivative method (combination of the first-order derivative and the second-order derivative), thereby obtaining the peak shape feature vector of the target chromatographic peak. In this embodiment, the peak shape feature vector includes curvature, skewness and peak width ratio.

[0030] In one embodiment, the expression of the peak shape feature vector is: ; In the formula, For the The historical chromatogram The peak shape feature vector of the target chromatographic peak; For the The historical chromatogram Target chromatographic peak value Peak width The ratio of For the The historical chromatogram The average curvature of the target chromatographic peaks; For the The historical chromatogram The skewness of the target chromatographic peak.

[0031] Understandably, Characterizes the sharpness of the chromatographic peak. The larger it is, the sharper the peak shape of the chromatographic peak is. Characterizes the degree of change in the shape of the chromatographic peak. The larger the value is, the greater the change in the peak shape of the chromatographic peak is. Characterize the symmetry of the peak shape of the chromatographic peak. If the skewness of the chromatographic peak is 0, the chromatographic peak is a symmetrical peak; if the skewness of the chromatographic peak is greater than 0, the chromatographic peak is right-skewed; if the skewness of the chromatographic peak is less than 0, the chromatographic peak is left-skewed. By selecting three parameters that can more accurately describe the peak shape of the chromatographic peak, the accuracy of subsequent chromatographic peak clustering can be improved, and the obtained clustering clusters can better represent the peak shape of a class of chromatographic peaks. In other optional embodiments, those skilled in the art can make a selection according to actual needs.

[0032] S103, clustering the peak shape feature vectors, and for any obtained cluster, fitting the target chromatographic peaks in the cluster using a plurality of fitting functions.

[0033] Specifically, the peak shape feature vectors can be clustered according to the K-means clustering algorithm. Specifically, first, the optimal K value is determined according to the silhouette coefficient, the K-means clustering algorithm is run to initialize the cluster center, and iterative calculation is performed until the clustering result is stable, so that multiple clusters are obtained. It can be understood that each cluster corresponds to a type of peak shape feature vector, and a type of peak shape feature vector corresponds to a type of target chromatographic peak.

[0034] By clustering the target chromatographic peaks, the noise peaks only need to be matched with the chromatographic peaks that can represent a type of target chromatographic peaks, avoiding the problem of matching the noise peaks with each target chromatographic peak, thereby shortening the time for chromatographic peak identification and improving the efficiency of subsequent matching of noise peaks with target chromatographic peaks.

[0035] Furthermore, for any cluster, a plurality of fitting functions (such as Gaussian function, Lorentzian function, etc.) are used to fit the target chromatographic peak in the cluster. The independent variable of the fitting function is the retention time, and the dependent variable is the signal intensity of the chromatographic peak. It can be understood that each fitting function can generate a peak shape.

[0036] S104, using the peak shape generated by the fitting function with the largest fitting degree as the matching template of the cluster.

[0037] It is understandable that different fitting functions are used, and the fitting effect is also different. Therefore, it is necessary to determine the fitting degree of each fitting function. The larger the fitting degree, the better the fitting effect of the cluster, that is, the higher the accuracy of the generated matching template, indicating that the peak shape generated by the fitting function can better characterize the peak shape corresponding to the cluster.

[0038] In one embodiment, the calculation expression of the maximum fit is: ; In the formula, For the The maximum fitness among all fitness of clusters; is the maximum value of the abscissa (i.e., retention time) of the fitting function; For the Clusters within The mean signal intensity of the chromatographic peak at time; For the th cluster, use the th The fitting function is The fitted value of the signal intensity of the chromatographic peak at time.

[0039] According to the above calculation expression of maximum fitting degree, For each chromatographic peak The average signal strength at time The chromatographic peaks generated by the fitting function are The absolute error of the signal strength at the moment, therefore, The smaller the value, the more The smaller the error between the chromatographic peak generated by the fitting function and all the chromatographic peaks in the cluster, the higher the fitting degree.

[0040] It can be understood that the fitting function with the maximum fitting degree can be determined through the calculation expression of the maximum fitting degree. Furthermore, the peak shape generated by the fitting function with the maximum fitting degree is used as the matching template of the cluster. Specifically, the fitting function with the maximum fitting degree can be corrected (the coefficients in the fitting function are adjusted) according to the average peak shape feature vector of the cluster, so that the peak shape feature vector corresponding to the peak shape generated by the fitting function is close to the average peak shape feature vector of the cluster; the peak shape generated by the corrected fitting function is used as the matching template corresponding to the cluster. Repeat the above process to generate matching templates for all clusters.

[0041] By correcting the fitting function, the matching template generated after correction can better characterize the peak shapes of all target chromatographic peaks in the cluster, thereby improving the accuracy of chromatographic peak recognition.

[0042] In other optional embodiments, the fitting function with the largest degree of fit may not be corrected, that is, the peak shape generated by the fitting function with the largest degree of fit may be directly used as the matching template of the cluster.

[0043] S2. Extract the noise peaks removed from the chromatogram to be identified during denoising.

[0044] Specifically, a chromatogram detected by a chromatograph (i.e., a chromatogram to be identified) is obtained, wavelet transform is performed on the chromatogram for denoising, the denoised chromatogram is compared with the original chromatogram, and the noise peaks removed from the original chromatogram during the denoising process are extracted by moving a window.

[0045] S3. Determine the similarity between the noise peak and each matching template to obtain a similarity set.

[0046] Specifically, for any noise peak, the noise peak is matched with each matching template in turn according to the template matching algorithm, and the similarity between the noise peak and each matching template is determined according to the overlap during matching, thereby obtaining a similarity set.

[0047] In one embodiment, the similarity is calculated as: ; In the formula, is the first The noise peak and Similarity of matching templates; is the first The noise peak and The number of regions where the matching templates overlap; is the first The noise peak and The total number of overlapping points in the overlapping area l of the matching templates; is the first The total number of points of the noise peak.

[0048] S4. Determine the comprehensive similarity between the noise peak and the target chromatographic peak according to the similarity set.

[0049] In one embodiment, the calculation expression of the comprehensive similarity is: ; In the formula, is the first The comprehensive similarity between the noise peak and the target chromatographic peak; is the first The noise peak and Similarity of matching templates; is the first The mean of the similarities between the noise peaks and each matching template; is the total number of types of matching templates; For the The maximum fitness among all fitness of the clusters.

[0050] In the above calculation expression of comprehensive similarity, Characterizes the outlier degree of the mean similarity between the noise peak and all matching templates. The smaller the value, the smaller the comprehensive similarity between the noise peak and the chromatographic peak with less content (i.e., the target chromatographic peak), and the more likely the noise peak is to be a true noise peak. It can be understood that since the similarity between the real noise and each matching template is relatively low (noise is not a chromatographic peak, so the similarity with the chromatographic peak is low), if the noise peak is real noise, the corresponding outlier degree will be relatively low, that is, the comprehensive similarity between the noise peak and the chromatographic peak with less content is low. Conversely, the larger the value, the greater the comprehensive similarity between the noise peak and the chromatographic peak with less content. This is because the larger the value, the greater the similarity between the noise peak and a certain matching template, and the lower the similarity with other matching templates. At this time, the noise peak and the matching template have outliers, that is, the comprehensive similarity between the noise peak and the chromatographic peak with less content is large. This is because a chromatographic peak generally has a high similarity with only one matching template, and a low similarity with other matching templates.

[0051] It is understandable that due to It describes the deviation of each similarity from the overall similarity, and the maximum fit describes the accuracy of each matching template. Therefore, by using the maximum fit to correct and calibrate the deviation, the accuracy of determining the comprehensive similarity between the noise peak and the target chromatographic peak can be improved, thereby improving the accuracy of chromatographic peak identification.

[0052] In another embodiment, the calculation expression of the comprehensive similarity can also be: ; In the formula, is the first The comprehensive similarity between the noise peak and the target chromatographic peak; is the first The variance of the similarity between the noise peaks and each matching template; is the total number of types of matching templates; For the The maximum fitness among all fitness of clusters; is the normalization function; For An exponential function with base .

[0053] From the above calculation expression of comprehensive similarity, it can be known that the smaller the variance of similarity, the smaller the comprehensive similarity between the noise peak and the target chromatographic peak, and the more likely the noise peak is to be a real noise peak. This is because the smaller the variance of similarity, the smaller the difference between each similarity in the similarity set of the noise peak, and because the similarity between the real noise peak and each matching template is relatively low (that is, the difference between each similarity is small), therefore, the smaller the variance of similarity, the smaller the comprehensive similarity between the noise peak and the target chromatographic peak. Conversely, the larger the variance of similarity, the higher the comprehensive similarity between the noise peak and the target chromatographic peak. This is because the larger the variance, the larger the difference between each similarity in the similarity set of the noise peak, that is, the similarity between the noise peak and a certain matching template is large, and the similarity with other matching templates is low (generally speaking, a chromatographic peak has a high similarity with only one matching template, and a low similarity with other matching templates), therefore, the larger the variance of similarity, the higher the comprehensive similarity between the noise peak and the target chromatographic peak.

[0054] It can be understood that since the variance of the similarity describes the fluctuation degree of the similarity between the noise peak and each matching template, and the maximum fit describes the accuracy of each matching template, by using the maximum fit to correct and calibrate the variance, the comprehensive similarity between the noise peak and the target chromatographic peak can be improved, thereby improving the accuracy of chromatographic peak identification. By determining whether the noise peak is a chromatographic peak with a small content based on the comprehensive similarity, the accuracy of chromatographic peak identification can be further ensured.

[0055] S5. Identify whether the noise peak is a target chromatographic peak based on the relationship between the comprehensive similarity and a set threshold.

[0056] If the comprehensive similarity is greater than or equal to the set threshold, the noise peak is determined as the target chromatographic peak and restored to the chromatogram to be identified after denoising; if the comprehensive similarity is less than the set threshold, the noise peak is determined as the real noise peak.

[0057] In one embodiment, the threshold value is set to 0.5. In other optional embodiments, the threshold value may also be set to 0.6. Those skilled in the art may make a selection based on actual needs.

[0058] It should be noted that the noise peaks extracted above are low-frequency noise peaks, while for high-frequency noise peaks, a threshold can be set based on historical experience to filter out the high-frequency noise in the wavelet transform, thereby obtaining a chromatogram with noise removed (at this time, the chromatogram contains restored chromatographic peaks).

[0059] Furthermore, the chromatogram is input into the trained CNN model, the input is the chromatogram with noise removed as mentioned above, and the output is the recognition probability that each chromatographic peak on the chromatogram is a certain abnormal peak. At this point, the recognition of the chromatographic peak is completed.

[0060] By restoring the chromatographic peaks misjudged as noise to the chromatogram, omission of chromatographic peaks is avoided, thereby improving the accuracy and reliability of the obtained chromatogram, thereby improving the accuracy of chromatographic peak identification and the accuracy of chromatographic state identification.

[0061] In the description of this specification, “plurality” means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0062] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A method for intelligently identifying chromatographic states based on a spectrum, characterized in that: include: Establishing multiple matching templates for target chromatographic peaks, wherein one matching template corresponds to a type of peak shape of target chromatographic peaks; Extract the noise peaks removed during denoising of the chromatogram to be identified; Determine the similarity between the noise peak and each matching template to obtain a similarity set, and the calculation expression of the similarity is: ； In the formula, is the first The noise peak and Similarity of matching templates; is the first The noise peak and The number of regions where the matching templates overlap; is the first The noise peak and The total number of overlapping points in the overlapping area l of the matching templates; is the first The total number of points of the noise peak; Determining the comprehensive similarity between the noise peak and the target chromatographic peak according to the similarity set; It is determined whether the noise peak is a target chromatographic peak according to the relationship between the comprehensive similarity and a set threshold.

2. The method for intelligently identifying chromatographic states based on a spectrum according to claim 1, characterized in that: Create multiple matching templates for target chromatographic peaks, including: Obtain the target chromatographic peak in multiple historical denoised chromatograms; Determining a peak shape characteristic vector of the target chromatographic peak, wherein the peak shape characteristic vector includes curvature, skewness, and peak width ratio; Clustering the peak shape feature vectors, and for any obtained cluster, fitting the target chromatographic peaks in the cluster using a plurality of fitting functions, wherein the independent variable of the fitting function is time, and the dependent variable is the signal intensity of the chromatographic peak; The peak shape generated by the fitting function with the largest fitting degree is used as the matching template of the cluster.

3. The method for intelligently identifying chromatographic states based on a spectrum according to claim 2, characterized in that: The peak shape generated by the fitting function with the largest fitting degree is used as the matching template of the cluster, including: According to the average peak shape feature vector of the cluster, the fitting function with the largest fitting degree is modified so that the peak shape feature vector corresponding to the peak shape generated by the fitting function is close to the average peak shape feature vector of the cluster; The peak shape generated by the modified fitting function is used as the matching template corresponding to the cluster.

4. The method for intelligently identifying chromatographic states based on a spectrum according to claim 2, characterized in that: The calculation expression of the maximum fit is: ； In the formula, For the The maximum fitness among all fitness of clusters; is the maximum value of the horizontal coordinate of the fitting function; For the Clusters within The mean signal intensity of the chromatographic peak at time; For the The clusters are used The fitting function is The fitted value of the signal intensity of the chromatographic peak at time.

5. The method for intelligently identifying chromatographic states based on a spectrum according to claim 2, characterized in that: The calculation expression of comprehensive similarity is: ； In the formula, is the first The comprehensive similarity between the noise peak and the target chromatographic peak; is the first The noise peak and Similarity of matching templates; is the first The mean of the similarities between the noise peaks and each matching template; is the total number of types of matching templates; For the The maximum fitness among all fitness of the clusters.

6. The method for intelligently identifying chromatographic states based on a spectrum according to claim 2, characterized in that: The calculation expression of comprehensive similarity is: ； In the formula, is the first The comprehensive similarity between the noise peak and the target chromatographic peak; is the first The variance of the similarity between the noise peaks and each matching template; is the total number of types of matching templates; For the The maximum fitness among all fitness of clusters; is the normalization function; For An exponential function with base .

7. The method for intelligently identifying chromatographic states based on a spectrum according to claim 2, characterized in that: The expression of the peak shape characteristic vector is: ； In the formula, For the The historical chromatogram The peak shape feature vector of the target chromatographic peak; For the The historical chromatogram Target chromatographic peak value Peak width The ratio of For the The historical chromatogram The average curvature of the target chromatographic peaks; For the The historical chromatogram The skewness of the target chromatographic peak.

8. The method for intelligently identifying chromatographic states based on a spectrum according to claim 1, characterized in that: Determining the similarity between the noise peak and each matching template includes: According to the template matching algorithm, the noise peak is matched with each matching template, and the similarity between the noise peak and each matching template is determined according to the overlap during matching.

9. The method for intelligently identifying chromatographic states based on a spectrum according to claim 1, characterized in that: Extract the noise peaks removed from the chromatogram to be identified during denoising, including: Acquire the chromatogram to be identified, and denoise the chromatogram to be identified according to wavelet transform; The chromatograms before and after denoising are compared, and the noise peaks are extracted by moving the window.

10. The method for intelligently identifying chromatographic states based on a spectrum according to claim 1, characterized in that: Identifying whether the noise peak is a target chromatographic peak according to the relationship between the comprehensive similarity and the set threshold value includes: If the comprehensive similarity is greater than or equal to the set threshold, the noise peak is determined as the target chromatographic peak and restored to the chromatogram to be identified after denoising; If the comprehensive similarity is less than the set threshold, the noise peak is determined to be a true noise peak.

Images