Compound monitoring methods, systems, and media based on chromatographic column analysis

Abstract

The application discloses a compound monitoring method and system based on chromatographic column analysis and a medium, relates to the technical field of chromatographic detection, and comprises the following steps: taking a complex matrix sample, collecting a first chromatographic column response distribution of the complex matrix sample by using a diode array detector; performing optimization on an enhancement strategy according to adjustable action parameters of the diode array detector, and collecting a second chromatographic column response distribution of the complex matrix sample by using the diode array detector under the first enhancement strategy; identifying a difference chromatographic column response distribution of the first chromatographic column response distribution and the second chromatographic column response distribution, and constructing a high-dimensional feature fusion fingerprint; and performing similarity measurement based on the high-dimensional feature fusion fingerprint and a pre-constructed compound feature fingerprint library to obtain a compound detection result. The application solves the technical problem that the compound chromatographic response information is insufficient under the complex matrix condition in the prior art, and the detection accuracy is low, and achieves the technical effect of improving the compound detection accuracy.

Description

Technical Field

[0001] This invention relates to the field of chromatographic detection technology, specifically to compound monitoring methods, systems, and media based on chromatographic column analysis. Background Technology

[0002] In the chromatographic analysis of complex matrix samples, co-elution between different compounds is common, background interference is strong, and response signal differences are small. This is especially true for compounds with low content or weak absorption characteristics, whose chromatographic responses are often overwhelmed by matrix noise, making it difficult to fully characterize their features. Traditional detection methods typically rely on a single detection mode or limited wavelength information, resulting in a low dimensionality of the chromatographic response. This makes it difficult to comprehensively reflect the differences in compound characteristics across time and spectral dimensions, leading to insufficient utilization of feature information and limiting the accuracy of compound identification and quantitative analysis. Summary of the Invention

[0003] This application provides a compound monitoring method, system, and medium based on column analysis, which addresses the technical problem of insufficient chromatographic response information of compounds under complex matrix conditions in the prior art, resulting in low detection accuracy.

[0004] In view of the above problems, this application provides a compound monitoring method, system and medium based on column analysis.

[0005] The first aspect of this application provides a compound monitoring method based on column chromatography, the method comprising:

[0006] A complex matrix sample is taken, and the first column response distribution of the complex matrix sample is acquired using a diode array detector in a first detection mode. An enhancement strategy is optimized according to the adjustable operating parameters of the diode array detector, and the second column response distribution of the complex matrix sample is acquired under the first enhancement strategy. The differential column response distributions between the first and second column response distributions are identified. Based on the first, second, and differential column response distributions, a high-dimensional feature fusion fingerprint is constructed. Similarity measurement is performed based on the high-dimensional feature fusion fingerprint and a pre-constructed compound feature fingerprint library to obtain the compound detection results.

[0007] A second aspect of this application provides a compound monitoring system based on column chromatography analysis, the system comprising:

[0008] The first acquisition module is used to acquire complex matrix samples and acquire the first column response distribution of the complex matrix samples using a diode array detector according to a first detection mode; the second acquisition module is used to optimize the enhancement strategy according to the adjustable operating parameters of the diode array detector, and the diode array detector acquires the second column response distribution of the complex matrix samples under the first enhancement strategy; the identification module is used to identify the differential column response distributions between the first and second column response distributions, and construct a high-dimensional feature fusion fingerprint based on the first, second, and differential column response distributions; the similarity measurement module is used to perform similarity measurement based on the high-dimensional feature fusion fingerprint and a pre-constructed compound feature fingerprint library to obtain compound detection results.

[0009] A third aspect of the embodiments of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the compound monitoring method based on chromatographic column analysis provided in this application.

[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0011] This application takes a complex matrix sample and uses a diode array detector to acquire the first column response distribution of the complex matrix sample according to a first detection mode. An enhancement strategy is optimized according to the adjustable operating parameters of the diode array detector, and the diode array detector acquires the second column response distribution of the complex matrix sample under the first enhancement strategy. The difference in column response distribution between the first and second column response distributions is identified. Based on the first, second, and difference column response distributions, a high-dimensional feature fusion fingerprint is constructed. Similarity measurement is performed based on the high-dimensional feature fusion fingerprint and a pre-constructed compound feature fingerprint library to obtain the compound detection result. This invention solves the technical problem of insufficient chromatographic response information of compounds under complex matrix conditions in the prior art, leading to low detection accuracy. By optimizing the enhancement strategy of the diode array detector and fusing conventional chromatographic responses, enhanced chromatographic responses, and their difference responses to construct a high-dimensional feature fusion fingerprint, the technical effect of improving the accuracy of compound detection is achieved. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 A schematic flowchart of a compound monitoring method based on column chromatography provided in this application embodiment;

[0014] Figure 2 This is a schematic diagram of the compound monitoring system based on column analysis provided in an embodiment of this application.

[0015] Explanation of reference numerals in the attached figures: First acquisition module 11, Second acquisition module 12, Identification module 13, Similarity measurement module 14. Detailed Implementation

[0016] This application provides a compound monitoring method, system, and medium based on column analysis. It addresses the technical problem of insufficient chromatographic response information of compounds under complex matrix conditions, which leads to low detection accuracy in the prior art. By optimizing the diode array detector with enhancement strategies and fusing conventional chromatographic responses, enhanced chromatographic responses, and their differential responses to construct a high-dimensional feature fusion fingerprint, the technical effect of improving the accuracy of compound detection is achieved.

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0018] It should be noted that any variation of the terms "comprising" and "having" is intended to cover non-exclusive inclusion, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such processes, methods, products, or devices.

[0019] Example 1, as Figure 1 As shown, this application provides a compound monitoring method based on column chromatography analysis, the method comprising:

[0020] Step S100: Take a complex matrix sample and collect the first column response distribution of the complex matrix sample using a diode array detector in the first detection mode.

[0021] In this embodiment of the application, during compound monitoring, a complex matrix sample to be tested is first acquired, and then injected into the chromatographic column after being processed according to conventional liquid chromatography analysis requirements. The complex matrix sample passes through the chromatographic column under the influence of the mobile phase, and the compounds are separated sequentially over time due to their different retention capacities in the stationary phase, thus forming chromatographic effluents that vary with retention time.

[0022] When the chromatographic effluent enters the diode array detector, the detector is controlled to operate in the first detection mode. This first detection mode is a preset conventional detection mode used for continuous detection of the chromatographic effluent. In this mode, the diode array detector simultaneously acquires the absorption signals of the chromatographic effluent at multiple wavelengths and records the absorption signals at each time point as detection response data.

[0023] The collected detection response data are organized in order of retention time to form detection results containing information in terms of time dimension, response intensity dimension, and wavelength dimension, thereby obtaining the first column response distribution that characterizes the overall chromatographic response features of each compound in a complex matrix sample.

[0024] Step S200: Optimize the enhancement strategy according to the adjustable operating parameters of the diode array detector, and collect the second column response distribution of the complex matrix sample under the first enhancement strategy.

[0025] In this embodiment, when optimizing the enhancement strategy based on the adjustable operating parameters of the diode array detector, an enhancement strategy parameter space is first constructed based on the adjustable operating parameters of the diode array detector, focusing on the detection wavelength combination, wavelength weight distribution, and spectral sampling density adjustment parameters. Multiple candidate enhancement strategy combinations are then formed within this parameter space. Subsequently, the column response distribution under each candidate enhancement strategy combination is predicted to obtain the corresponding predicted differential column response distribution. The enhancement change rate is used as an evaluation index to select the candidate enhancement strategy combinations whose enhancement change rate meets the preset target conditions, which are then output as the first enhancement strategy.

[0026] The first enhancement strategy was then applied to the diode array detector, enabling it to detect complex matrix samples under this first enhancement strategy, thereby acquiring the second column response distribution that characterizes the compound response features under enhanced detection conditions.

[0027] Furthermore, the method provided in the application embodiment, which optimizes the enhancement strategy according to the adjustable operating parameters of the diode array detector, further includes:

[0028] An enhancement strategy parameter space is constructed based on the adjustable operating parameters of the diode array detector. The adjustable operating parameters include at least the detection wavelength combination, wavelength weight distribution, and spectral sampling density adjustment parameters. Multiple enhancement strategy candidate combinations are generated within this parameter space, each corresponding to a set of determined adjustable operating parameters. The column response distributions of these multiple enhancement strategy candidate combinations are predicted to obtain multiple predicted differential column response distributions. The enhancement strategy candidate combination whose enhancement change rate meets a preset target condition is selected as the first enhancement strategy output.

[0029] In this embodiment, when optimizing the enhancement strategy, the detection conditions are first gradually adjusted and organized based on the adjustable action parameters of the diode array detector. These adjustable action parameters include at least the detection wavelength combination, wavelength weight distribution, and spectral sampling density adjustment parameters. For the detection wavelength combination, different detection wavelengths are selected within the allowed detection wavelength range of the diode array detector, and multiple detection wavelength combinations are formed by adding, subtracting, or replacing the selected detection wavelengths. For the wavelength weight distribution, after determining the detection wavelength combination, different weights are set for the detection response corresponding to each detection wavelength, and multiple wavelength weight distributions are formed by adjusting the proportional relationship between the weights of each wavelength. For the spectral sampling density adjustment parameters, the spectral sampling density is adjusted by changing the wavelength acquisition interval, thereby forming different spectral sampling density adjustment parameters. The different value ranges of the above detection wavelength combinations, wavelength weight distributions, and spectral sampling density adjustment parameters are summarized and their combinable relationships are uniformly represented, thereby constructing the enhancement strategy parameter space.

[0030] Then, within the enhancement strategy parameter space, each adjustable action parameter is combined and configured to generate multiple sets of enhancement strategy candidate combination schemes. Each set of enhancement strategy candidate combination schemes corresponds to a set of determined adjustable action parameters, that is, each set of enhancement strategy candidate combination schemes includes a determined combination of detection wavelengths, a determined wavelength weight distribution, and a determined spectral sampling density adjustment parameter.

[0031] Next, the column response distribution of multiple enhancement strategy candidate combinations is predicted. In this process, based on standard absorption spectra of different compounds under different wavelength conditions and measured peak profiles obtained under full-wavelength scanning conditions, the corresponding absorption spectra of the compounds are read, where the absorption spectra include the molar absorptivity at each wavelength. Based on this, the peak profiles are functionalized by setting a peak shape function, and a distribution prediction model is constructed based on the peak shape function. Then, using the distribution prediction model, enhancement prediction is performed under the conditions of detection wavelength combinations, wavelength weight distributions, and spectral sampling density adjustment parameters corresponding to each enhancement strategy candidate combination, resulting in multiple enhanced chromatographic response distributions. Based on the enhanced chromatographic response distributions, multiple predicted differential column response distributions corresponding to the multiple enhancement strategy candidate combinations are calculated.

[0032] Finally, when screening candidate combinations of enhancement strategies whose enhancement rate meets the preset target conditions in multiple groups of predicted differential column response distributions, the corresponding enhancement rate is calculated for each predicted differential column response distribution, and the enhancement rate is compared with the preset target conditions. Based on the relationship between the magnitude of the enhancement rate, different candidate combinations of enhancement strategies are screened, and their influence on weakly responsive compounds and strongly responsive compounds is evaluated respectively. Thus, candidate combinations of enhancement strategies whose enhancement rate meets the preset target conditions are selected, and the screened candidate combinations of enhancement strategies are determined as the first enhancement strategy output.

[0033] Furthermore, in the method provided in the application embodiments, predicting the column response distribution of the multiple sets of enhancement strategy candidate combinations to obtain multiple sets of predicted differential column response distributions further includes:

[0034] A standard absorption spectrum sample library of different compounds at different wavelengths is obtained, along with the measured chromatographic peak profiles under full-wavelength scanning. The standard absorption spectrum sample library includes the molar absorptivity of samples at each wavelength. The standard absorption spectra corresponding to the compounds are read from the standard absorption spectrum sample library. A chromatographic peak shape function is set, and a distribution prediction model constructed using the chromatographic peak shape function is used to perform enhancement prediction based on the standard absorption spectra, resulting in multiple enhanced chromatographic response distributions under the multiple enhancement strategy candidate combinations. Multiple predicted differential column response distributions are calculated based on the multiple enhanced chromatographic response distributions.

[0035] In this embodiment, a standard absorption spectrum sample library of different compound samples under different wavelength conditions and the measured chromatographic peak profile under full wavelength scanning conditions are first obtained. The standard absorption spectrum sample library includes the molar absorptivity of each wavelength sample, which is used to characterize the absorption strength of the compound at the corresponding wavelength. The measured chromatographic peak profile is obtained by chromatographic separation of the compound sample and continuous acquisition under full wavelength scanning conditions, which is used to characterize the peak height, peak width and peak shape characteristics of the chromatographic peak as retention time changes.

[0036] Next, the standard absorption spectrum corresponding to the compound is read from the absorption spectrum sample library. That is, the standard absorption spectrum corresponding to the compound is retrieved in the absorption spectrum sample library according to the compound identifier, and the molar absorptivity of the compound at each wavelength sample is extracted as the spectral input data for predicting the column response distribution.

[0037] Subsequently, chromatographic peak shape functions were used to functionalize the measured chromatographic peak profiles. These functions, with retention time as the independent variable and chromatographic response intensity as the dependent variable, describe the peak shape changes over time. Chromatographic peak shape functions include, but are not limited to, Gaussian peak shape functions, exponentially modified Gaussian peak shape functions, or other functional forms characterizing peak width and shape. Taking the Gaussian peak shape function as an example, its expression is as follows: Where t represents the retention time, This indicates the retention time at the peak of the chromatographic peak, and A represents the peak height parameter. This represents the peak width parameter of the chromatographic peak. The measured peak profile is mapped to a continuous time-dimensional response function using a peak shape function, and the fundamental time-dimensional response is calculated at a preset retention time position. As the time dimension basis input for the distribution prediction model.

[0038] After setting the chromatographic peak shape function, a distribution prediction model is constructed based on the chromatographic peak shape function. The distribution prediction model uses the time-dimensional fundamental response of the chromatographic peak shape function output as its basis. The provided peak structure serves as the basic response framework, upon which molar absorptivity corresponding to different detection wavelengths in the standard absorption spectrum is introduced. This ensures that the predicted response intensity at the same retention time position is jointly determined by the time-dimensional basic response and the molar absorptivity corresponding to the wavelength. A predicted chromatographic response distribution is formed by weighted superposition of the response results corresponding to different detection wavelengths. The overall calculation relationship is expressed as follows: .in, This represents the chromatographic response intensity output by the distribution prediction model at retention time t, where n represents the number of detection wavelengths involved in the prediction calculation. This represents the value of the i-th detection wavelength involved in the prediction calculation. This indicates the target compound at the detection wavelength. The molar absorptivity at that wavelength is used to characterize the absorption capacity at that wavelength. Indicates the detection wavelength The corresponding wavelength weights are used to characterize the proportion of the response corresponding to that wavelength to the final chromatographic response. This represents the time-dimensional basic response calculated by the chromatographic peak shape function and is used to provide a uniform peak shape structure. The wavelength weights satisfy a preset normalization relationship to ensure that the amplitude of the output chromatographic response is consistent under different conditions. This allows the output of the distribution prediction model to inherit the peak shape characteristics of the chromatographic peak in the time dimension and reflect the absorption differences of the compounds in the wavelength dimension.

[0039] After the distribution prediction model is constructed, when performing enhanced prediction based on the standard absorption spectrum, the detection wavelength values ​​involved in the prediction calculation, the weight parameters corresponding to different detection wavelengths, and the density of wavelength sampling positions are changed so that the distribution prediction model outputs the corresponding enhanced chromatographic response distribution under different detection parameter configurations. The enhanced chromatographic response distribution represents the predicted chromatographic response intensity output by the distribution prediction model at each retention time position under the corresponding detection parameter configuration.

[0040] Finally, the enhanced chromatographic response distributions obtained under different detection parameter configurations are compared with those obtained under the baseline detection conditions. The difference between the enhanced and baseline chromatographic responses is calculated at the same retention time position to obtain the predicted differential column response distribution. The calculation relationship is expressed as follows: ,in This represents the chromatographic response intensity output by the distribution prediction model at retention time t under enhanced detection conditions. This represents the chromatographic response intensity obtained at retention time t under baseline detection conditions. This represents the response difference introduced by the enhanced detection conditions relative to the baseline detection conditions at the corresponding retention time position. The response difference is output as the predicted differential column response distribution.

[0041] Furthermore, the method provided in the application embodiments also includes:

[0042] The first enhancement strategy output is selected from the multiple sets of predicted differential column response distributions where the enhancement rate of change meets a preset target condition. The preset target condition includes a first enhancement rate threshold. The peak height, peak area, and spectral energy of each chromatographic peak in the first column response distribution are extracted. Chromatographic peak intervals where the peak height, peak area, and spectral energy are all less than the corresponding preset threshold are marked as weakly responding compounds. The weakly responding compound-predicted differential column response distribution is predicted, and the weakly responding compound-enhancement rate is calculated based on this distribution. Finally, the first enhancement strategy output is selected from the weakly responding compound-enhancement rate where the enhancement rate of change is greater than the first enhancement rate threshold.

[0043] In this embodiment, when screening multiple candidate combinations of enhancement strategies, the enhancement rate of change for each candidate combination is first calculated based on multiple predicted differential column response distributions. A quantitative comparison is then made between the first column response distribution and the predicted column response distribution corresponding to the candidate combination. Within the same peak range, the response amount corresponding to the first column response distribution and the response amount corresponding to the predicted column response distribution under the candidate combination are obtained, with peak area selected as the uniform comparison object. The ratio of the peak area under the candidate combination to the peak area in the first column response distribution is calculated to obtain the change ratio of the column response under the candidate combination relative to the first column response distribution. This change ratio is used as the enhancement rate of change. Simultaneously, preset target conditions are set, including at least a first enhancement rate of change threshold.

[0044] Next, based on the response distribution of the first chromatographic column, feature extraction is performed on each chromatographic peak. Specifically, the peak height, peak area, and spectral energy of each chromatographic peak are extracted from the response distribution of the first chromatographic column. The peak height is obtained by reading the response intensity at the peak apex, the peak area is obtained by integrating the response intensity within the retention time interval of the chromatographic peak, and the spectral energy is obtained by summing the response intensities within the corresponding wavelength range of the chromatographic peak.

[0045] Then, the peak height, peak area, and spectral energy are compared with the corresponding preset thresholds. The chromatographic peak intervals where the peak height, peak area, and spectral energy are all less than the corresponding preset thresholds are marked, and the compounds corresponding to the chromatographic peak intervals are identified as weakly responsive compounds.

[0046] For the weakly responsive compound, its column response is predicted under multiple enhancement strategy candidate combinations. Specifically, based on the aforementioned constructed distribution prediction model, the standard absorption spectrum corresponding to the weakly responsive compound, as well as the detection wavelength combination, wavelength weight distribution, and spectral sampling density adjustment parameters corresponding to each enhancement strategy candidate combination, are input into the distribution prediction model. This allows the distribution prediction model to output the predicted column response distribution of the weakly responsive compound under each enhancement strategy candidate combination within the same chromatographic peak interval. Then, the predicted column response distribution is compared with the corresponding chromatographic peak interval of the weakly responsive compound in the first column response distribution. Specifically, within the same chromatographic peak interval, the difference between the predicted column response distribution and the first column response distribution is calculated, thereby obtaining the weakly responsive compound-predicted differential column response distribution, which reflects the change in the weakly responsive compound's response under the enhancement strategy candidate combination.

[0047] Subsequently, the enhancement rate of the weak response compound was calculated based on the predicted differential column response distribution. In this process, within the chromatographic peak range corresponding to the weak response compound, the peak area corresponding to the predicted column response distribution under the candidate enhancement strategy combinations and the corresponding peak area in the first column response distribution were obtained, and the enhancement rate of the weak response compound was determined by the ratio of these two values.

[0048] Finally, the weak response compound-enhanced change rate is compared with the first enhanced change rate threshold. Candidate combinations of enhancement strategies that are greater than the first enhanced change rate threshold are selected from the weak response compound-enhanced change rates, and the candidate combinations of enhancement strategies that meet the selection criteria are determined as the first enhancement strategy output.

[0049] Furthermore, the method provided in the application embodiments also includes:

[0050] The preset target conditions also include a second enhancement rate threshold, which is less than the first enhancement rate threshold; chromatographic peak intervals with peak height, peak area, and spectral energy all greater than the corresponding preset thresholds are marked as strong-response compounds; the strong-response compound-predicted differential column response distribution is predicted, and the strong-response compound-enhancement rate is calculated based on the strong-response compound-predicted differential column response distribution; and enhancement strategy candidate combinations with values ​​less than the second enhancement rate threshold are selected from the strong-response compound-enhancement rate as the first enhancement strategy output.

[0051] In this embodiment of the application, a second enhanced rate of change threshold is first set by a technical expert in the preset target conditions. The second enhanced rate of change threshold is less than the first enhanced rate of change threshold. The second enhanced rate of change threshold is used to limit the allowable range of change of the candidate combination scheme of the enhancement strategy for the strong response compound, so as to avoid generating excessive response perturbation of the strong response compound while improving the response of the weak response compound.

[0052] Next, based on the peak height, peak area, and spectral energy extracted from the response distribution of the first chromatographic column, threshold judgments are performed on each chromatographic peak interval. Chromatographic peak intervals in which the peak height, peak area, and spectral energy are all greater than the corresponding preset thresholds are marked, and compounds corresponding to the chromatographic peak intervals are identified as strongly responsive compounds. Among them, peak height is used to characterize the maximum response intensity at the peak apex of the chromatographic peak, peak area is used to characterize the cumulative response of the chromatographic peak within the retention time interval, and spectral energy is used to characterize the overall response level of the chromatographic peak in the wavelength dimension.

[0053] Subsequently, for strongly responsive compounds, the column response is predicted under multiple enhancement strategy candidate combinations. Specifically, based on the aforementioned distribution prediction model, the standard absorption spectrum corresponding to the strongly responsive compound, as well as the detection wavelength combination, wavelength weight distribution, and spectral sampling density adjustment parameters corresponding to each enhancement strategy candidate combination are input into the distribution prediction model. This allows the distribution prediction model to output the predicted column response distribution of the strongly responsive compound under each enhancement strategy candidate combination within the same chromatographic peak range.

[0054] Next, the predicted column response distribution is compared with the corresponding peak intervals of the strongly responsive compounds in the first column response distribution. Specifically, the difference between the predicted and first column response distributions is calculated within the same peak interval to obtain the strongly responsive compound-predicted difference column response distribution, reflecting the change in the response of the strongly responsive compounds under the enhancement strategy candidate combination scheme. Subsequently, the strongly responsive compound-enhancement change rate is calculated based on the strongly responsive compound-predicted difference column response distribution. This involves obtaining the peak area corresponding to the predicted column response distribution under the enhancement strategy candidate combination scheme and the corresponding peak area in the first column response distribution within the peak interval corresponding to the strongly responsive compound, and determining the strongly responsive compound-enhancement change rate through the ratio of these two ratios.

[0055] Finally, the enhanced change rate of the strong response compound is compared with the second enhanced change rate threshold. Candidate combinations of enhancement strategies that are less than the second enhanced change rate threshold are selected from the enhanced change rates of the strong response compounds. The candidate combinations of enhancement strategies that meet the selection criteria are determined as the first enhancement strategy output.

[0056] Step S300: Identify the differential column response distribution between the first column response distribution and the second column response distribution, and construct a high-dimensional feature fusion fingerprint based on the first column response distribution, the second column response distribution, and the differential column response distribution.

[0057] In this embodiment, to identify the differential column response distributions of the first and second chromatographic columns, the time and wavelength axes of the first and second column response distributions are first aligned. Based on the aligned first and second column response distributions, the differential column response distribution is extracted. The differential column response distribution characterizes the response changes introduced by enhanced detection conditions relative to conventional detection conditions. This includes a time-response difference distribution reflecting the difference in response over retention time, a time-response rate of change distribution reflecting the magnitude of response change, and a wavelength-dimensional response distribution reflecting the characteristics of response changes over wavelength.

[0058] Next, the response distributions of the first column, the second column, and the differential column are mapped to the same time axis and wavelength axis coordinate system, so that the response data of the three at the same retention time position and the same detection wavelength position are established to establish a correspondence, thereby obtaining the first column response distribution, the second column response distribution, and the differential column response distribution that are completely corresponding on the time axis and wavelength axis.

[0059] Subsequently, within the same chromatographic peak interval, the response intensity sequences at each sampling time point and each detection wavelength were read from the response distribution of the first chromatographic column to obtain the first response feature reflecting the compound's response characteristics under conventional detection conditions. Under the same chromatographic peak interval and the same time axis and wavelength axis index, the corresponding response intensity sequences were read from the response distribution of the second chromatographic column to obtain the second response feature reflecting the compound's response characteristics under the first enhancement strategy. At the same time, within the same chromatographic peak interval, the time-response difference distribution feature, the time-response change rate distribution feature, and the wavelength dimension response distribution feature were read from the differential chromatographic column response distribution to obtain the differential response feature reflecting the response changes between the two detection conditions.

[0060] Next, the response values ​​in each feature are uniformly scaled so that the first response feature, the second response feature, and the differential response feature are all represented by a consistent numerical range, and the relative relationship between the response and retention time and detection wavelength within each chromatographic peak interval remains unchanged, thus obtaining the first response feature, the second response feature, and the differential response feature with consistent scale.

[0061] Finally, following a preset feature arrangement order, the response features of the first column, the second column, and the differential column are sequentially spliced ​​together. This combines the conventional detection response information, the detection response information under the first enhancement strategy, and the response change information within the same chromatographic peak range into a high-dimensional feature vector. Corresponding high-dimensional feature vectors are then generated for each chromatographic peak range in the complex matrix sample and combined according to the peak arrangement order on the retention time axis. This yields a high-dimensional feature fusion fingerprint characterizing the comprehensive response properties of compounds in the complex matrix sample.

[0062] Furthermore, in the method provided in the application embodiments, identifying the difference in chromatographic column response distribution between the first chromatographic column response distribution and the second chromatographic column response distribution further includes:

[0063] The time axis and wavelength axis of the response distribution of the first column and the response distribution of the second column are aligned. Based on the aligned response distribution of the first column and the response distribution of the second column, the differential column response distribution is extracted. The differential column response distribution includes time-response difference distribution features, time-response change rate distribution features, and wavelength dimension response distribution features.

[0064] In this embodiment, when extracting the difference in column response distribution between the first and second column response distributions, the time axis of the first and second column response distributions is first aligned. This time axis alignment is achieved by uniformly processing the retention time information in the two column response distributions. Specifically, the retention time sampling points in the first column response distribution are used as a reference, and the retention time sampling points in the second column response distribution are matched accordingly. This ensures that the second column response distribution has corresponding response data at each retention time position of the first column response distribution, thereby establishing a one-to-one correspondence between the two column response distributions in the retention time dimension and completing the time axis alignment.

[0065] After completing the time axis alignment, the response distributions of the first and second chromatographic columns are aligned along the wavelength axis. Wavelength axis alignment is achieved by uniformly processing the detection wavelength information in the response distributions of the two columns. That is, the detection wavelength sampling points in the response distribution of the first column are used as a reference, and the detection wavelength sampling points in the response distribution of the second column are matched accordingly. This ensures that the response distribution of the second column has corresponding response data at each detection wavelength position of the response distribution of the first column, thus forming a one-to-one correspondence between the two column response distributions in the wavelength dimension, and completing the wavelength axis alignment.

[0066] After aligning the time axis and wavelength axis, the response intensity data of the first column response distribution and the second column response distribution are read at the same retention time position and the same detection wavelength position, respectively. By calculating the difference in response intensity at the corresponding positions and calculating the response change ratio, the differential column response distribution characterizing the response change of the second column response distribution relative to the first column response distribution is obtained.

[0067] In the process of constructing the differential chromatographic column response distribution, the response differences corresponding to each retention time position are arranged in chronological order along the retention time dimension to obtain the time-response difference distribution characteristics that reflect the absolute change of response over time. At the same time, the response change ratios corresponding to each retention time position are arranged in chronological order along the retention time dimension to obtain the time-response change rate distribution characteristics that reflect the magnitude of response change over time. Finally, the response differences and response change ratios corresponding to different detection wavelength positions are arranged in chronological order along the wavelength dimension to obtain the wavelength dimension response distribution characteristics that reflect the response change in the spectral dimension, thus forming the differential chromatographic column response distribution.

[0068] Step S400: Perform similarity measurement based on the high-dimensional feature fusion fingerprint and the pre-constructed compound feature fingerprint library to obtain the compound detection result.

[0069] In this embodiment, similarity measurement is performed based on a high-dimensional feature fusion fingerprint and a pre-constructed compound feature fingerprint library. The high-dimensional feature fusion fingerprint is formed by fusing the baseline feature vector corresponding to the response distribution of the first chromatographic column, the enhancement feature vector corresponding to the response distribution of the second chromatographic column, and the change feature vector corresponding to the response distribution of the differential chromatographic column. According to the pre-constructed compound feature fingerprint library, the first high-dimensional feature fusion fingerprint library corresponding to the first enhancement strategy is read, and the high-dimensional feature fusion fingerprint is similar to the compound feature fingerprints in the first high-dimensional feature fusion fingerprint library one by one to obtain the similarity calculation result.

[0070] Finally, compound matching is completed based on the similarity calculation results, and the matching results are output as compound detection results.

[0071] Furthermore, in the method provided in the application embodiments, the similarity measurement based on the high-dimensional feature fusion fingerprint and the pre-constructed compound feature fingerprint library to obtain the compound detection result further includes:

[0072] The high-dimensional feature fusion fingerprint includes extracting the baseline feature vector extracted from the response distribution of the first chromatographic column, the enhanced feature vector extracted from the response distribution of the second chromatographic column, and the variation feature vector extracted from the differential chromatographic column response distribution. The baseline feature vector, enhanced feature vector, and variation feature vector are then normalized and weighted and concatenated. Based on the pre-constructed compound feature fingerprint library, a first high-dimensional feature fusion fingerprint library under the first enhancement strategy is read. The high-dimensional feature fusion fingerprint and the first high-dimensional feature fusion fingerprint library are subjected to cosine similarity calculation, and the similarity calculation result is output. Compound matching is performed according to the similarity calculation result to obtain the compound detection result.

[0073] In this embodiment, the high-dimensional feature fusion fingerprint consists of a baseline feature vector, an enhanced feature vector, and a variation feature vector. The baseline feature vector comprises response values ​​extracted in a predetermined order within each chromatographic peak interval from the response distribution of the first chromatographic column. The enhanced feature vector comprises response values ​​extracted in the same order within the corresponding chromatographic peak interval from the response distribution of the second chromatographic column. The variation feature vector comprises the response difference and response change rate extracted within the corresponding chromatographic peak interval from the differential chromatographic column response distribution. When normalizing the baseline, enhanced, and variation feature vectors, the maximum and minimum response values ​​of each feature vector are obtained. Each response value in the feature vector is then subtracted from its minimum response value and divided by the difference between the maximum and minimum response values, thus mapping all response values ​​in the baseline, enhanced, and variation feature vectors to a uniform numerical range. After normalization, the normalized values ​​in the baseline feature vector, enhanced feature vector, and changed feature vector are numerically scaled according to a preset weight ratio. The three types of feature vectors are then concatenated end to end according to a fixed feature arrangement order to obtain a high-dimensional feature fusion fingerprint.

[0074] Next, based on the pre-constructed compound feature fingerprint library, the first high-dimensional feature fusion fingerprint library corresponding to the first enhancement strategy is read. The first high-dimensional feature fusion fingerprint library stores high-dimensional feature fusion fingerprints of multiple known compounds constructed under the first enhancement strategy. The first high-dimensional feature fusion fingerprint is consistent with the high-dimensional feature fusion fingerprint in terms of feature dimension composition, feature arrangement order, and numerical processing method.

[0075] When performing similarity measurement, the cosine similarity is calculated sequentially between the high-dimensional feature fusion fingerprint and each high-dimensional feature fusion fingerprint in the first high-dimensional feature fusion fingerprint database. During the calculation, the values ​​of the high-dimensional feature fusion fingerprint and the target high-dimensional feature fusion fingerprint at the same dimensional position are multiplied and summed to obtain the inner product value. At the same time, the values ​​of each dimension of the high-dimensional feature fusion fingerprint and the target high-dimensional feature fusion fingerprint are squared, summed, and squared to obtain their respective modulus lengths. Then, the inner product value is divided by the product of the two modulus lengths to obtain the corresponding cosine similarity value, and the cosine similarity value is output as the similarity calculation result.

[0076] Finally, the similarity calculation results are compared and processed. The similarity values ​​are sorted from high to low, and the high-dimensional feature fusion fingerprint with the highest similarity value or that meets the preset matching conditions is selected. The compound corresponding to the high-dimensional feature fusion fingerprint in the first high-dimensional feature fusion fingerprint library is determined as the compound detection result.

[0077] Furthermore, in the method provided in the application embodiments, constructing the compound feature fingerprint library further includes:

[0078] The system sets up known compound detection samples, initial column response distribution, and enhanced column response distribution collected based on enhancement strategy parameter space samples. The diode array detector performs multiple detections on the known compound detection samples, and constructs the compound feature fingerprint library based on the mean features of multiple detections of each compound under each enhancement strategy by fusing fingerprint samples.

[0079] In this embodiment of the application, when constructing the compound feature fingerprint library, a known compound detection sample is first set up. The known compound detection sample is a compound sample with known composition and category. Under uniform chromatographic conditions, the known compound detection sample is detected by a diode array detector to obtain the corresponding initial chromatographic column response distribution. The initial chromatographic column response distribution is used to record the chromatographic response information of the known compound under the current detection conditions.

[0080] After obtaining the initial column response distribution, different combinations of enhancement strategy parameters are selected according to the enhancement strategy parameter space. Under each combination of enhancement strategy parameters, the known compound sample is detected by a diode array detector, and the corresponding enhanced column response distribution is collected. The enhanced column response distribution is used to record the chromatographic response information of the known compound under different enhancement strategy parameters.

[0081] During the acquisition of initial and enhanced column response distributions, for each known compound sample and each enhancement strategy parameter combination, the diode array detector performs multiple detections of the known compound sample under the same detection conditions, and records the column response distribution obtained from each detection. After completing multiple detections, for the same known compound sample under the same enhancement strategy parameter combination, corresponding high-dimensional feature fusion fingerprints are generated according to the aforementioned high-dimensional feature fusion fingerprint construction method. The mean of the high-dimensional feature fusion fingerprints generated from multiple detections is calculated in each feature dimension to obtain the corresponding mean feature fusion fingerprint sample.

[0082] After obtaining mean feature fusion fingerprint samples of different known compounds under different combinations of enhancement strategy parameters, the mean feature fusion fingerprint samples are organized and stored according to compound identifier and enhancement strategy parameter identifier to construct a compound feature fingerprint library.

[0083] In summary, the embodiments of this application have at least the following technical effects:

[0084] This application takes a complex matrix sample and uses a diode array detector to acquire the first column response distribution of the complex matrix sample according to a first detection mode. An enhancement strategy is optimized according to the adjustable operating parameters of the diode array detector, and the diode array detector acquires the second column response distribution of the complex matrix sample under the first enhancement strategy. The difference in column response distribution between the first and second column response distributions is identified. Based on the first, second, and difference column response distributions, a high-dimensional feature fusion fingerprint is constructed. Similarity measurement is performed based on the high-dimensional feature fusion fingerprint and a pre-constructed compound feature fingerprint library to obtain the compound detection result. This invention solves the technical problem of insufficient chromatographic response information of compounds under complex matrix conditions in the prior art, leading to low detection accuracy. By optimizing the enhancement strategy of the diode array detector and fusing conventional chromatographic responses, enhanced chromatographic responses, and their difference responses to construct a high-dimensional feature fusion fingerprint, the technical effect of improving the accuracy of compound detection is achieved.

[0085] Example 2, based on the same inventive concept as the compound monitoring method based on chromatographic column analysis in the foregoing examples, such as... Figure 2 As shown, this application provides a compound monitoring system based on column chromatography analysis. The system and method embodiments in this application are based on the same inventive concept. The system includes:

[0086] The first acquisition module 11 is used to acquire complex matrix samples and acquire the first column response distribution of the complex matrix samples using a diode array detector in a first detection mode; the second acquisition module 12 is used to optimize the enhancement strategy according to the adjustable operating parameters of the diode array detector, and the diode array detector acquires the second column response distribution of the complex matrix samples under the first enhancement strategy; the identification module 13 is used to identify the differential column response distribution between the first column response distribution and the second column response distribution, and construct a high-dimensional feature fusion fingerprint based on the first column response distribution, the second column response distribution, and the differential column response distribution; the similarity measurement module 14 is used to perform similarity measurement based on the high-dimensional feature fusion fingerprint and a pre-constructed compound feature fingerprint library to obtain compound detection results.

[0087] Furthermore, the system is also used to implement the following functions:

[0088] An enhancement strategy parameter space is constructed based on the adjustable operating parameters of the diode array detector. The adjustable operating parameters include at least the detection wavelength combination, wavelength weight distribution, and spectral sampling density adjustment parameters. Multiple enhancement strategy candidate combinations are generated within this parameter space, each corresponding to a set of determined adjustable operating parameters. The column response distributions of these multiple enhancement strategy candidate combinations are predicted to obtain multiple predicted differential column response distributions. The enhancement strategy candidate combination whose enhancement change rate meets a preset target condition is selected as the first enhancement strategy output.

[0089] Furthermore, the system is also used to implement the following functions:

[0090] The time axis and wavelength axis of the response distribution of the first column and the response distribution of the second column are aligned. Based on the aligned response distribution of the first column and the response distribution of the second column, the differential column response distribution is extracted. The differential column response distribution includes time-response difference distribution features, time-response change rate distribution features, and wavelength dimension response distribution features.

[0091] Furthermore, the system is also used to implement the following functions:

[0092] A standard absorption spectrum sample library of different compounds at different wavelengths is obtained, along with the measured chromatographic peak profiles under full-wavelength scanning. The standard absorption spectrum sample library includes the molar absorptivity of samples at each wavelength. The standard absorption spectra corresponding to the compounds are read from the standard absorption spectrum sample library. A chromatographic peak shape function is set, and a distribution prediction model constructed using the chromatographic peak shape function is used to perform enhancement prediction based on the standard absorption spectra, resulting in multiple enhanced chromatographic response distributions under the multiple enhancement strategy candidate combinations. Multiple predicted differential column response distributions are calculated based on the multiple enhanced chromatographic response distributions.

[0093] Furthermore, the system is also used to implement the following functions:

[0094] The first enhancement strategy output is selected from the multiple sets of predicted differential column response distributions where the enhancement rate of change meets a preset target condition. The preset target condition includes a first enhancement rate threshold. The peak height, peak area, and spectral energy of each chromatographic peak in the first column response distribution are extracted. Chromatographic peak intervals where the peak height, peak area, and spectral energy are all less than the corresponding preset threshold are marked as weakly responding compounds. The weakly responding compound-predicted differential column response distribution is predicted, and the weakly responding compound-enhancement rate is calculated based on this distribution. Finally, the first enhancement strategy output is selected from the weakly responding compound-enhancement rate where the enhancement rate of change is greater than the first enhancement rate threshold.

[0095] Furthermore, the system is also used to implement the following functions:

[0096] The preset target conditions also include a second enhancement rate threshold, which is less than the first enhancement rate threshold; chromatographic peak intervals with peak height, peak area, and spectral energy all greater than the corresponding preset thresholds are marked as strong-response compounds; the strong-response compound-predicted differential column response distribution is predicted, and the strong-response compound-enhancement rate is calculated based on the strong-response compound-predicted differential column response distribution; and enhancement strategy candidate combinations with values ​​less than the second enhancement rate threshold are selected from the strong-response compound-enhancement rate as the first enhancement strategy output.

[0097] Furthermore, the system is also used to implement the following functions:

[0098] The high-dimensional feature fusion fingerprint includes extracting the baseline feature vector extracted from the response distribution of the first chromatographic column, the enhanced feature vector extracted from the response distribution of the second chromatographic column, and the variation feature vector extracted from the differential chromatographic column response distribution. The baseline feature vector, enhanced feature vector, and variation feature vector are then normalized and weighted and concatenated. Based on the pre-constructed compound feature fingerprint library, a first high-dimensional feature fusion fingerprint library under the first enhancement strategy is read. The high-dimensional feature fusion fingerprint and the first high-dimensional feature fusion fingerprint library are subjected to cosine similarity calculation, and the similarity calculation result is output. Compound matching is performed according to the similarity calculation result to obtain the compound detection result.

[0099] Furthermore, the system is also used to implement the following functions:

[0100] The system sets up known compound detection samples, initial column response distribution, and enhanced column response distribution collected based on enhancement strategy parameter space samples. The diode array detector performs multiple detections on the known compound detection samples, and constructs the compound feature fingerprint library based on the mean features of multiple detections of each compound under each enhancement strategy by fusing fingerprint samples.

[0101] Example 3: Based on the compound monitoring method based on chromatographic column analysis in the foregoing examples, and with the same inventive concept, this application also provides a computer-readable storage medium storing a computer program, which, when executed, implements the steps of any of the methods described in Example 1 above.

[0102] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0103] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for monitoring a compound based on analysis of a chromatographic column, characterized in that, The method comprises: taking a complex matrix sample, collecting a first column chromatographic response distribution of the complex matrix sample by a diode array detector according to a first detection mode; performing optimization of an enhancement strategy according to adjustable action parameters of the diode array detector, and collecting a second column chromatographic response distribution of the complex matrix sample under a first enhancement strategy; identifying a differential column chromatographic response distribution of the first column chromatographic response distribution and the second column chromatographic response distribution, and constructing a high-dimensional feature fusion fingerprint based on the first column chromatographic response distribution, the second column chromatographic response distribution and the differential column chromatographic response distribution; performing similarity measurement based on the high-dimensional feature fusion fingerprint and a pre-constructed compound feature fingerprint library to obtain a compound detection result; performing optimization of an enhancement strategy according to adjustable action parameters of the diode array detector, the method comprising: constructing an enhancement strategy parameter space according to the adjustable action parameters of the diode array detector, wherein the adjustable action parameters at least include a detection wavelength combination, a wavelength weight distribution and a spectral sampling density adjustment parameter; generating a plurality of groups of enhancement strategy candidate combination schemes in the enhancement strategy parameter space, wherein each group of enhancement strategy candidate combination schemes corresponds to a group of determined adjustable action parameters; performing column chromatographic response distribution prediction on the plurality of groups of enhancement strategy candidate combination schemes to obtain a plurality of groups of predicted differential column chromatographic response distributions; screening an enhancement strategy candidate combination scheme whose enhancement change rate meets a preset target condition from the plurality of groups of predicted differential column chromatographic response distributions as a first enhancement strategy output.

2. The method of claim 1, wherein, identifying a differential column chromatographic response distribution of the first column chromatographic response distribution and the second column chromatographic response distribution, the method comprising: aligning the first column chromatographic response distribution and the second column chromatographic response distribution in time axis and wavelength axis, and extracting a differential column chromatographic response distribution according to the aligned first column chromatographic response distribution and the second column chromatographic response distribution; wherein the differential column chromatographic response distribution comprises a time sequence-response difference distribution feature, a time sequence-response change rate distribution feature and a wavelength dimension response distribution feature.

3. The method of claim 1, wherein, performing column chromatographic response distribution prediction on the plurality of groups of enhancement strategy candidate combination schemes to obtain a plurality of groups of predicted differential column chromatographic response distributions, the method comprising: obtaining a standard absorption spectrum sample library of different compound samples at different wavelengths, and a measured chromatographic peak profile under full wavelength scanning, wherein the standard absorption spectrum sample library includes a molar absorption coefficient at each wavelength sample; reading a standard absorption spectrum corresponding to a compound according to the standard absorption spectrum sample library; setting a chromatographic peak shape function, and performing enhancement prediction based on the standard absorption spectrum by using a distribution prediction model constructed by the chromatographic peak shape function to obtain a plurality of enhanced chromatographic response distributions under the plurality of groups of enhancement strategy candidate combination schemes; calculating a plurality of groups of predicted differential column chromatographic response distributions according to the plurality of enhanced chromatographic response distributions.

4. The method of claim 1, wherein, Screening the multiple groups of prediction difference chromatographic column response distribution from the enhanced change rate meeting the preset target condition as the first enhanced strategy output, the preset target condition includes the first enhanced change rate threshold; Extracting the peak height, peak area and spectral energy of each chromatographic peak in the first chromatographic column response distribution; Marking the chromatographic peak interval with peak height, peak area and spectral energy less than the corresponding preset threshold as a weak response compound; Predicting the weak response compound-predicted difference chromatographic column response distribution of the weak response compound, and calculating the weak response compound-enhanced change rate according to the weak response compound-predicted difference chromatographic column response distribution; Screening the enhanced strategy candidate combination scheme greater than the first enhanced change rate threshold from the weak response compound-enhanced change rate as the first enhanced strategy output.

5. The method of claim 4, wherein, The preset target condition also includes a second enhanced change rate threshold, which is less than the first enhanced change rate threshold; Marking the chromatographic peak interval with peak height, peak area and spectral energy greater than the corresponding preset threshold as a strong response compound; Predicting the strong response compound-predicted difference chromatographic column response distribution of the strong response compound, and calculating the strong response compound-enhanced change rate according to the strong response compound-predicted difference chromatographic column response distribution; Screening the enhanced strategy candidate combination scheme less than the second enhanced change rate threshold from the strong response compound-enhanced change rate as the first enhanced strategy output.

6. The method of claim 1, wherein, Based on the high-dimensional feature fusion fingerprint and the pre-constructed compound feature fingerprint library, similarity measurement is performed to obtain a compound detection result, and the method comprises: The high-dimensional feature fusion fingerprint comprises a reference feature vector extracted from the first chromatographic column response distribution, an enhanced feature vector extracted from the second chromatographic column response distribution, and a change feature vector extracted from the difference chromatographic column response distribution, and the reference feature vector, the enhanced feature vector and the change feature vector are normalized and weighted to obtain a splicing result; According to the pre-constructed compound feature fingerprint library, a first high-dimensional feature fusion fingerprint library under the first enhanced strategy is read; The high-dimensional feature fusion fingerprint and the first high-dimensional feature fusion fingerprint library are subjected to cosine similarity calculation, and a similarity calculation result is outputted; According to the similarity calculation result, compound matching is performed to obtain a compound detection result.

7. The method of claim 6, wherein, The method for constructing the compound feature fingerprint library comprises: Setting known compound detection samples, initial chromatographic column response distribution and enhanced chromatographic column response distribution acquired based on enhanced strategy parameter space samples; The diode array detector detects the known compound detection samples multiple times, and the compound feature fingerprint library is constructed according to the mean feature fusion fingerprint sample of each compound under multiple detections of each enhanced strategy.

8. A compound monitoring system based on chromatographic column analysis, characterized by, The system is used to execute the compound monitoring method based on chromatographic column analysis according to any one of claims 1-7, and the system comprises: A first acquisition module is configured to take a complex matrix sample, and a diode array detector is configured to acquire a first chromatographic column response distribution of the complex matrix sample in a first detection mode; a second acquisition module configured to perform an enhanced strategy optimization according to an adjustable action parameter of the diode array detector, and the diode array detector is configured to acquire a second column response distribution of the complex matrix sample under a first enhanced strategy; an identification module configured to identify a difference column response distribution of the first column response distribution and the second column response distribution, and construct a high-dimensional feature fusion fingerprint based on the first column response distribution, the second column response distribution, and the difference column response distribution; a similarity measurement module configured to perform similarity measurement based on the high-dimensional feature fusion fingerprint and a pre-constructed compound feature fingerprint library, and obtain a compound detection result.

9. A computer readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the compound monitoring method based on the column analysis according to any one of claims 1-7. The program is executed by the processor to implement the compound monitoring method based on the column analysis according to any one of claims 1-7.

Images