A method and system for identifying abnormal chromatographic peak morphology based on image processing

Abstract

This invention relates to the field of image processing technology, and discloses a method and system for identifying abnormal morphology of chromatographic peaks based on image processing. The method includes: segmenting the chromatographic peak regions of the chromatogram to be identified to obtain candidate peak images of the chromatogram; performing contour extraction analysis on the candidate peak images to obtain chromatographic peak contour curves of the candidate peak images; quantitatively characterizing the candidate peak images to obtain morphological feature parameters of the candidate peak images; training and learning the feature vector and peak morphology category labels of the chromatogram to be identified to obtain an abnormal morphology discrimination mapping table of the chromatogram to be identified; based on the abnormal morphology discrimination mapping table, performing anomaly evaluation on the morphological feature parameters to obtain a comprehensive evaluation value of the morphological feature parameters; and performing threshold comparison and identification on the comprehensive evaluation value to obtain the anomaly identification result of the chromatogram to be identified. This invention can improve the efficiency of identifying abnormal morphology of chromatographic peaks.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a method and system for identifying abnormal morphology of chromatographic peaks based on image processing. Background Technology

[0002] In the application of chromatographic peak morphology identification technology, existing identification methods mostly rely on manual judgment or simple numerical threshold analysis, making it difficult to comprehensively capture and analyze the overall morphological characteristics of chromatographic peaks. For identifying abnormal peak morphologies, they can only focus on deviations from single numerical indicators, failing to achieve a comprehensive assessment of multi-dimensional morphological characteristics such as peak shape, symmetry, and peak width consistency. This results in incomplete feature extraction in the identification of abnormal peak morphologies. Furthermore, existing technologies lack standardized image optimization processes in the chromatogram preprocessing stage. Noise interference and background specks in the chromatogram are not effectively eliminated, easily leading to insufficient accuracy in peak region segmentation, introducing more errors into subsequent morphology identification, and affecting the reliability of the identification results.

[0003] Existing methods for identifying abnormal morphology of chromatographic peaks lack systematic training based on historical samples and fail to establish a precise mapping relationship between feature vectors and peak morphology category labels. They rely solely on simple comparisons using fixed thresholds, failing to provide tiered assessments and accurate judgments based on the actual morphological characteristics of the peaks. This results in poor flexibility and adaptability. Furthermore, existing methods do not perform weighted allocation and comprehensive quantification of morphological feature parameters, making it difficult to scientifically measure the overall degree of abnormality of chromatographic peaks. This often leads to missed or false positives, and the identification efficiency and accuracy fall short of the application requirements for practical detection and analysis, failing to provide accurate and effective morphology determination criteria for the analysis of chromatographic detection results. Summary of the Invention

[0004] This invention provides a method and system for identifying abnormal morphology of chromatographic peaks based on image processing, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a method for identifying abnormal morphology of chromatographic peaks based on image processing, comprising:

[0006] S1. Segment the chromatogram to be identified by chromatographic peak regions to obtain chromatographic candidate peak images of the chromatogram to be identified;

[0007] S2. Perform contour extraction analysis on the chromatographic candidate peak image to obtain the chromatographic peak contour curve of the chromatographic candidate peak image;

[0008] S3. Based on the chromatographic peak profile curve, the chromatographic candidate peak image is quantitatively characterized to obtain the morphological feature parameters of the chromatographic candidate peak image;

[0009] S4. Based on the historical chromatogram sample set, train and learn the feature vector and peak morphology category label of the chromatogram to be identified to obtain the abnormal morphology discrimination mapping table of the chromatogram to be identified.

[0010] S5. Based on the abnormal morphology discrimination mapping table, perform anomaly evaluation on the morphological feature parameters to obtain a comprehensive evaluation value of the morphological feature parameters;

[0011] S6. The comprehensive evaluation value is compared and identified by a threshold to obtain the abnormal identification result of the chromatogram to be identified.

[0012] In a preferred embodiment, the step of segmenting the chromatogram to be identified into chromatographic peak regions to obtain a chromatographic candidate peak image of the chromatogram to be identified includes:

[0013] The chromatogram to be identified is converted to grayscale to obtain an enhanced chromatographic image of the chromatogram to be identified;

[0014] The enhanced chromatographic image is subjected to median filtering to obtain a denoised chromatographic image of the chromatogram to be identified;

[0015] The foreground pixels corresponding to the peak regions in the denoised chromatographic image are segmented into chromatographic peaks to obtain a binary chromatographic image of the chromatogram to be identified.

[0016] A morphological opening operation is performed on the binary chromatographic image to obtain a connected binary image of the chromatogram to be identified;

[0017] The connected binary image is analyzed and labeled with connected components to obtain the chromatographic candidate peak image of the chromatogram to be identified.

[0018] In a preferred embodiment, the step of performing contour extraction analysis on the chromatographic candidate peak image to obtain the chromatographic peak contour curve of the chromatographic candidate peak image includes:

[0019] The pixels in the peak region of the chromatographic candidate peak image are assigned a first gray value, and the background pixels are assigned a second gray value;

[0020] Based on the first gray value and the second gray value, contour detection is performed on the pixels at the boundary between the peak region and the background of the chromatographic candidate peak image to obtain the edge pixel set of the chromatographic candidate peak image;

[0021] Contour tracking is performed on the set of edge pixels, and adjacent edge pixels are connected sequentially according to the eight-neighbor connection relationship to obtain the initial closed contour line of the peak region.

[0022] The initial closed contour line is smoothed to obtain the chromatographic peak contour curve of the chromatographic candidate peak image.

[0023] In a preferred embodiment, the step of quantitatively characterizing the chromatographic candidate peak image based on the chromatographic peak profile curve to obtain the morphological feature parameters of the chromatographic candidate peak image includes:

[0024] Based on the chromatographic peak profile curve, identify the highest peak point, the left baseline start point, and the right baseline start point on the profile curve;

[0025] The horizontal distances between the highest point of the peak and the starting points of the left and right baselines are analyzed simultaneously to obtain the symmetry parameters of the chromatographic peak profile curve.

[0026] Based on the highest point of the peak, the trend of the change of the ordinate of the contour points in the chromatographic peak contour curve is statistically analyzed to obtain the tailing factor of the chromatographic peak contour curve.

[0027] Within a predetermined height range near the highest point of the peak, the width values ​​of the contour curve at different height positions are obtained. Based on the degree of difference between the width values, the consistency of the peak width in the chromatographic candidate peak image is measured to obtain the peak width variation coefficient of the chromatographic peak.

[0028] The symmetry parameter, the tailing factor, and the peak width variation coefficient are combined into morphological feature parameters of the chromatographic candidate peak image.

[0029] In a preferred embodiment, the step of measuring the consistency of the peak widths in the chromatographic candidate peak images based on the degree of difference between the width values, to obtain the peak width variation coefficient of the chromatographic peaks, includes:

[0030] Collect the peak width values ​​of the peak regions in the same chromatographic candidate peak image to obtain the peak width value sequence of the chromatographic candidate peak image;

[0031] The average peak width value of the peak width value sequence is obtained by averaging the peak width values ​​in the sequence.

[0032] The deviation between the peak width value and the average peak width value is evaluated to obtain the individual deviation value of the peak width value;

[0033] Based on the overall distribution of the individual deviation values, the consistency between peak width values ​​in the peak width value sequence is coupled and measured to obtain the peak width variation coefficient of the chromatographic peak.

[0034] In a preferred embodiment, the step of training and learning the feature vector and peak morphology category labels of the chromatogram to be identified based on a historical chromatogram sample set to obtain an abnormal morphology discrimination mapping table for the chromatogram to be identified includes:

[0035] Obtain a historical chromatogram sample set, which contains multiple historical chromatogram peak images, and the historical chromatogram peak images are pre-labeled with peak morphology category labels to characterize their peak shape status;

[0036] The historical chromatographic peak images are retrospectively analyzed to obtain the historical morphological feature parameters of the historical chromatographic peak images, and the historical morphological feature parameters are used as historical feature vectors.

[0037] A correlation analysis is performed on the historical feature vector and the peak shape category label to obtain the correspondence between the numerical distribution range of the historical feature vector and the peak shape category label;

[0038] Based on the aforementioned correspondence, an anomaly morphology discrimination mapping table is established, with the numerical range of the feature vector as the index and the peak morphology category label as the mapping value.

[0039] In a preferred embodiment, the step of performing anomaly assessment on the morphological feature parameters based on the anomaly morphology discrimination mapping table to obtain a comprehensive evaluation value for the morphological feature parameters includes:

[0040] The morphological characteristic parameters of the chromatogram to be identified are obtained, including symmetry parameters, tailing factor, and peak width variation coefficient.

[0041] Based on the abnormal morphology discrimination mapping table, the morphological feature parameters are matched item by item to obtain the similarity results of the morphological feature parameters;

[0042] Based on the peak morphology category labels corresponding to the historical feature vectors in the similarity results, the preliminary abnormal state category of the morphological feature parameters is determined.

[0043] Based on the preliminary abnormal state category, the contribution of the morphological feature parameters is evaluated to obtain the allocation weight of the morphological feature parameters;

[0044] Based on the weighted allocation and the abnormal morphology discrimination mapping table, the overall abnormality of the morphological feature parameters is graded and evaluated to obtain a comprehensive evaluation value for the morphological feature parameters.

[0045] In a preferred embodiment, the formula for calculating the comprehensive evaluation value is as follows:

[0046] ;

[0047] In the formula, The comprehensive evaluation value is... For the preliminary abnormal state category The The weighting of each morphological feature parameter. For the first The specific values ​​of each morphological feature parameter, when A value of 1 indicates a symmetry parameter; when... A value of 2 indicates a tailing factor. A value of 3 indicates the peak width variation coefficient. The first under the normal morphology category The baseline value of each morphological feature parameter The first under the normal morphology category The standard deviation of each morphological feature parameter The norm index is used to control the combination of different feature deviations.

[0048] In a preferred embodiment, the step of performing threshold comparison and identification on the comprehensive evaluation value to obtain the anomaly identification result of the chromatogram to be identified includes:

[0049] The comprehensive evaluation value is compared with the preset anomaly judgment threshold value one by one to obtain the anomaly judgment result of the comprehensive evaluation value.

[0050] Based on the anomaly determination results, the chromatographic candidate peak images are adapted and screened to obtain the anomaly peak images to be labeled in the chromatographic candidate peak images;

[0051] On the chromatogram to be identified, the chromatographic peak region where the abnormal peak image to be labeled is located is highlighted and marked;

[0052] Based on the position of the highlighted marker and the preliminary abnormality category of the abnormal peak image to be marked, the abnormality identification result of the chromatogram to be identified is obtained.

[0053] To address the above problems, the present invention also provides an image processing-based system for identifying abnormal morphology of chromatographic peaks, the system comprising:

[0054] The chromatographic peak region segmentation module is used to segment the chromatographic peak regions of the chromatogram to be identified, and obtain the chromatographic candidate peak image of the chromatogram to be identified;

[0055] The contour extraction and analysis module is used to perform contour extraction and analysis on the chromatographic candidate peak image to obtain the chromatographic peak contour curve of the chromatographic candidate peak image;

[0056] The quantization characterization module is used to quantify and characterize the chromatographic candidate peak image based on the chromatographic peak profile curve to obtain the morphological feature parameters of the chromatographic candidate peak image.

[0057] The discriminant mapping construction module is used to train and learn the feature vector and morphological category label of the chromatogram to be identified based on the historical chromatogram sample set, so as to obtain the abnormal morphology discriminant mapping table of the chromatogram to be identified;

[0058] An anomaly assessment module is used to assess the anomalies of the morphological feature parameters based on the anomaly morphology discrimination mapping table, and obtain a comprehensive evaluation value of the morphological feature parameters.

[0059] The identification output module is used to perform threshold comparison and identification on the comprehensive evaluation value to obtain the abnormal identification result of the chromatogram to be identified.

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] 1. This invention achieves accurate identification of abnormal chromatographic peak morphology through a standardized image processing workflow. First, the chromatogram undergoes a series of processes including grayscale conversion, filtering, and segmentation to accurately extract the chromatographic peak regions. Then, contour detection and tracking are used to obtain smooth chromatographic peak contour curves. Based on these curves, multi-dimensional morphological feature parameters such as symmetry parameters, tailing factors, and peak width variation coefficients are extracted, achieving a comprehensive quantitative characterization of chromatographic peak morphology. This significantly improves the completeness and accuracy of chromatographic peak morphology feature extraction, making the basic data for morphology identification more valuable. Simultaneously, an abnormal morphology discrimination mapping table is constructed based on a historical chromatogram sample set, achieving accurate association between feature vectors and peak morphology category labels. Combining weighted allocation and quantitative formulas to calculate a comprehensive evaluation value makes the abnormal morphology assessment more scientific and effectively improves the accuracy of chromatographic peak abnormality determination.

[0062] 2. This invention constructs a systematic method and system for identifying abnormal morphology of chromatographic peaks. Through modular division, it achieves efficient connection between each identification step, forming a standardized processing link from chromatographic peak region segmentation to abnormal identification result output. Automated abnormal morphology identification and marking can be completed without excessive manual intervention, significantly improving the overall efficiency of chromatographic peak abnormal morphology identification. Simultaneously, by comparing comprehensive evaluation value thresholds, abnormal peaks are screened and highlighted, accurately locating abnormal peak regions in the chromatogram and clarifying their abnormal state categories. This provides clear and accurate morphology judgment criteria for chromatographic detection and analysis, enhancing the practicality and intuitiveness of chromatographic peak abnormality identification results and providing reliable support for subsequent chromatographic detection data analysis. Attached Figure Description

[0063] Figure 1This is a flowchart illustrating a method for identifying abnormal morphology of chromatographic peaks based on image processing, provided in an embodiment of the present invention.

[0064] Figure 2 This is a functional block diagram of a chromatographic peak anomaly morphology recognition system based on image processing, provided in an embodiment of the present invention.

[0065] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0066] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0067] This application provides a method for identifying abnormal morphology of chromatographic peaks based on image processing. The executing entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cluster of cloud servers. The server can be an independent server or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0068] Reference Figure 1 The diagram shown is a flowchart illustrating a method for identifying abnormal chromatographic peak morphology based on image processing according to an embodiment of the present invention. In this embodiment, the method for identifying abnormal chromatographic peak morphology based on image processing includes:

[0069] S1. Segment the chromatogram to be identified by chromatographic peak regions to obtain chromatographic candidate peak images of the chromatogram to be identified;

[0070] In this embodiment of the invention, the step of segmenting the chromatogram to be identified into chromatographic peak regions to obtain chromatographic candidate peak images of the chromatogram to be identified includes:

[0071] The chromatogram to be identified is converted to grayscale to obtain an enhanced chromatographic image of the chromatogram to be identified;

[0072] The enhanced chromatographic image is subjected to median filtering to obtain a denoised chromatographic image of the chromatogram to be identified;

[0073] The foreground pixels corresponding to the peak regions in the denoised chromatographic image are segmented into chromatographic peaks to obtain a binary chromatographic image of the chromatogram to be identified.

[0074] A morphological opening operation is performed on the binary chromatographic image to obtain a connected binary image of the chromatogram to be identified;

[0075] The connected binary image is analyzed and labeled with connected components to obtain the chromatographic candidate peak image of the chromatogram to be identified.

[0076] Each pixel in the chromatogram to be identified is numerically converted according to a preset grayscale conversion relationship. The pixel information of the three channels (red, green, and blue) in the original image is uniformly converted into single-channel grayscale pixel information. During the conversion process, the grayscale difference features between the peak area and the baseline area in the chromatogram are preserved. The grayscale conversion process is completed to obtain the enhanced chromatographic image of the chromatogram to be identified.

[0077] Using each pixel in the enhanced chromatographic image as the center processing pixel, all neighboring pixels around the pixel are selected according to a fixed neighborhood range. The gray values ​​of these neighboring pixels and the center processing pixel are extracted. All extracted gray values ​​are sorted in ascending order. The gray value in the middle of the sorted sequence is selected as the new gray value of the center processing pixel. The above replacement operation is performed on all pixels in the enhanced chromatographic image in turn to eliminate randomly appearing isolated noise pixels in the image and obtain the denoised chromatographic image of the chromatogram to be identified.

[0078] The denoised chromatographic image is distinguished from the background based on a preset grayscale threshold. Pixels with grayscale values ​​greater than or equal to the grayscale threshold are identified as foreground pixels of the corresponding chromatographic peak region, while pixels with grayscale values ​​less than the grayscale threshold are identified as background pixels. Foreground and background pixels are assigned values ​​and labels respectively. Only the region formed by the foreground pixels is retained, and the region corresponding to the background pixels is removed. This completes the chromatographic peak segmentation and yields a binary chromatographic image of the chromatogram to be identified.

[0079] First, an erosion operation is performed on the binary chromatogram image. Each foreground pixel in the image is traversed to determine whether there are continuous foreground pixel connections around the pixel. Small foreground pixel blocks that do not meet the continuous connection condition and scattered pixels at the edges are removed. Then, a dilation operation is performed on the eroded image to expand the edges of the remaining foreground region outward, restore the overall shape and area of ​​the foreground region, and connect the originally broken foreground regions to each other. This completes the morphological opening operation and obtains the connected binary image of the chromatogram to be identified.

[0080] Starting from the top left corner of the connected binary image, traverse all pixels row by row and column by column. When an unassigned foreground pixel is detected, start from that pixel and search for all adjacent foreground pixels in the horizontal and vertical directions. Divide these adjacent foreground pixels into the same connected region and assign a unique identifier to each independent connected region. Complete the division and labeling of all connected regions to obtain the chromatographic candidate peak image of the chromatogram to be identified.

[0081] The beneficial effects are that by sequentially performing grayscale conversion, median filtering, chromatographic peak segmentation, morphological opening operation, and connected domain analysis and identification on the chromatogram to be identified, the grayscale difference between the peak region and the baseline region in the chromatogram is effectively preserved, isolated noise in the image is eliminated, the foreground pixels and background pixels of the chromatographic peak region are accurately distinguished, small and scattered pixels are removed and the original outline of the foreground region is restored, the broken foreground region is connected, and the independent connected regions of the chromatographic peak are clearly divided and identified, and finally, an accurate and effective chromatographic candidate peak image of the chromatogram to be identified is obtained.

[0082] S2. Perform contour extraction analysis on the chromatographic candidate peak image to obtain the chromatographic peak contour curve of the chromatographic candidate peak image;

[0083] In this embodiment of the invention, the step of performing contour extraction analysis on the chromatographic candidate peak image to obtain the chromatographic peak contour curve of the chromatographic candidate peak image includes:

[0084] The pixels in the peak region of the chromatographic candidate peak image are assigned a first gray value, and the background pixels are assigned a second gray value;

[0085] Based on the first gray value and the second gray value, contour detection is performed on the pixels at the boundary between the peak region and the background of the chromatographic candidate peak image to obtain the edge pixel set of the chromatographic candidate peak image;

[0086] Contour tracking is performed on the set of edge pixels, and adjacent edge pixels are connected sequentially according to the eight-neighbor connection relationship to obtain the initial closed contour line of the peak region.

[0087] The initial closed contour line is smoothed to obtain the chromatographic peak contour curve of the chromatographic candidate peak image.

[0088] The process iterates through every pixel in the chromatographic candidate peak image, identifying pixels belonging to the chromatographic peak region and pixels belonging to the background region. All pixels belonging to the chromatographic peak region are uniformly set to a predetermined first grayscale value, and all pixels belonging to the background region are uniformly set to a predetermined second grayscale value, clearly distinguishing the peak region from the background region in terms of grayscale values. Using the first and second grayscale values ​​as the distinguishing criteria, each pixel in the chromatographic candidate peak image is detected row by row and column by column. It is determined whether the grayscale value of the current pixel and the grayscale values ​​of its adjacent pixels belong to the first and second grayscale values, respectively. Pixels that meet this boundary judgment condition are identified as contour pixels. All detected contour pixels are collected to form the edge pixel set of the chromatographic candidate peak image. A starting edge pixel is selected from the edge pixel set. Following the eight-neighborhood connection relationship, the next adjacent edge pixel is sequentially searched for. The found adjacent edge pixels are connected in the detection order, and this connection operation is continued until all edge pixels form a continuous line connecting end to end, resulting in the initial closed contour line of the peak region. Traverse each contour point on the initial closed contour line, select multiple consecutive contour points around the contour point as references, and adjust the position of the current contour point according to the position distribution of the reference contour points to eliminate sharp turns and abrupt fluctuations on the initial closed contour line, making the contour lines smoother and completing the smoothing process of the initial closed contour line, thus obtaining the chromatographic peak contour curve of the chromatographic candidate peak image.

[0089] The beneficial effect is that by assigning different gray values ​​to the peak region pixels and background pixels in the chromatographic candidate peak image, the peak region and background region are clearly distinguished. Then, based on the two gray values, the edge pixels at the junction of the peak region and the background are accurately detected and the edge pixel set is formed. The edge pixels are contour tracked through the eight-neighbor connection relationship to form an initial closed contour line. Finally, the initial closed contour line is smoothed to eliminate sharp turns and abrupt fluctuations, resulting in a smooth and gentle chromatographic candidate peak image with a chromatographic peak contour curve, ensuring that the contour curve can accurately reflect the actual shape of the chromatographic peak.

[0090] S3. Based on the chromatographic peak profile curve, the chromatographic candidate peak image is quantitatively characterized to obtain the morphological feature parameters of the chromatographic candidate peak image;

[0091] In this embodiment of the invention, the step of quantitatively characterizing the chromatographic candidate peak image based on the chromatographic peak profile curve to obtain the morphological feature parameters of the chromatographic candidate peak image includes:

[0092] Based on the chromatographic peak profile curve, identify the highest peak point, the left baseline start point, and the right baseline start point on the profile curve;

[0093] The horizontal distances between the highest point of the peak and the starting points of the left and right baselines are analyzed simultaneously to obtain the symmetry parameters of the chromatographic peak profile curve.

[0094] Based on the highest point of the peak, the trend of the change of the ordinate of the contour points in the chromatographic peak contour curve is statistically analyzed to obtain the tailing factor of the chromatographic peak contour curve.

[0095] Within a predetermined height range near the highest point of the peak, the width values ​​of the contour curve at different height positions are obtained. Based on the degree of difference between the width values, the consistency of the peak width in the chromatographic candidate peak image is measured to obtain the peak width variation coefficient of the chromatographic peak.

[0096] The symmetry parameter, the tailing factor, and the peak width variation coefficient are combined into morphological feature parameters of the chromatographic candidate peak image.

[0097] The step of measuring the consistency of the peak widths in the chromatographic candidate peak images based on the degree of difference between the width values, and obtaining the peak width variation coefficient of the chromatographic peaks, includes:

[0098] Collect the peak width values ​​of the peak regions in the same chromatographic candidate peak image to obtain the peak width value sequence of the chromatographic candidate peak image;

[0099] The average peak width value of the peak width value sequence is obtained by averaging the peak width values ​​in the sequence.

[0100] The deviation between the peak width value and the average peak width value is evaluated to obtain the individual deviation value of the peak width value;

[0101] Based on the overall distribution of the individual deviation values, the consistency between peak width values ​​in the peak width value sequence is coupled and measured to obtain the peak width variation coefficient of the chromatographic peak.

[0102] Traverse all contour points in the chromatographic peak profile curve, record the coordinate information of each contour point, and determine the highest peak point by comparing the ordinate values ​​of all contour points. At the same time, check each point along the chromatographic peak profile curve from left to top, and determine the left baseline starting point when the ordinate of the contour point first stabilizes at the preset baseline gray value and no longer changes significantly. Check each point along the chromatographic peak profile curve from right to top, and determine the right baseline starting point when the ordinate of the contour point first stabilizes at the preset baseline gray value and no longer changes significantly.

[0103] Using a preset horizontal distance calculation standard, the horizontal distance between the highest point of the peak and the starting point of the left baseline, and the horizontal distance between the highest point of the peak and the starting point of the right baseline are measured respectively. The two horizontal distances are compared and analyzed simultaneously. The relative relationship between the two reflects the degree of symmetry of the chromatographic peak profile curve, and thus the symmetry parameter of the chromatographic peak profile curve is obtained.

[0104] Using the highest point of the peak as the reference point, the ordinate values ​​of each contour point in the chromatographic peak profile curve are recorded point by point. Following the order from the starting point of the left baseline through the highest point of the peak to the starting point of the right baseline, the changing trend of the ordinate of all contour points is statistically analyzed. The difference between the rate of decrease of the ordinate of the contour point in the direction of the starting point of the right baseline and the rate of increase of the ordinate of the contour point in the direction of the starting point of the left baseline is analyzed. The tailing factor of the chromatographic peak profile curve is obtained based on the degree of difference.

[0105] A predetermined height range near the highest point of the peak is set in advance. This range is based on the vertical coordinate of the highest point of the peak and a fixed height interval is set downwards. Multiple different height positions are uniformly selected within this height interval, and the width value of the chromatographic peak profile curve at each height position is measured. All the measured width values ​​are compared with each other, and the degree of deviation between each width value is calculated. The consistency of the peak width in the chromatographic candidate peak image is measured based on the degree of deviation, and the peak width variation coefficient of the chromatographic peak is obtained.

[0106] The symmetry parameters, tailing factor, and peak width variation coefficient obtained in the previous steps are collected and organized to ensure that the three parameters completely and accurately correspond to the characteristics of the chromatographic peak profile curve. These three parameters are then combined and integrated to form a set that can comprehensively reflect the morphological characteristics of the chromatographic candidate peak image, thus obtaining the morphological characteristic parameters of the chromatographic candidate peak image.

[0107] In the same chromatographic candidate peak image, the horizontal span values ​​corresponding to the chromatographic peak profile curves obtained from multiple different height positions selected within a predetermined height range near the highest point of the peak are all collected centrally. During the collection process, it is ensured that no peak width value corresponding to any selected height position is missed, while abnormal invalid values ​​appearing during the collection process are discarded. Then, all the collected valid peak width values ​​are arranged in an orderly manner according to the order of each height position from low to high, forming a complete, continuous and complete set of values, thus obtaining the peak width value sequence of the chromatographic candidate peak image.

[0108] The effective peak width values ​​in the peak width value sequence are accumulated one by one to obtain a total accumulated value. Then, the specific number of effective peak width values ​​in the peak width value sequence is counted to ensure that the count is completely consistent with the accumulated peak width values, without over-counting or under-counting. Subsequently, the total accumulated value and the counted number of effective peak widths are converted accordingly. The total accumulated value is divided by the number of effective peak widths. Through this overall mean conversion method, a value that can represent the concentration level of all peak width values ​​is determined, and the average peak width value of the peak width value sequence is obtained.

[0109] Each valid peak width value in the peak width value sequence is compared position by position with the calculated average peak width value. For each peak width value, the difference between it and the average peak width value is calculated. The sign of the difference is used to represent the direction of the shift of the peak width value relative to the average peak width value, and the magnitude of the difference is used to represent the degree of shift of the peak width value relative to the average peak width value. This degree of shift corresponding to each peak width value is recorded as an independent deviation characterization value to ensure that each peak width value has a unique corresponding deviation characterization value, thus obtaining the individual deviation value of the peak width value.

[0110] All individual deviation values ​​are aggregated to form a complete set. The magnitude, positive / negative distribution, and correlation between different individual deviation values ​​are analyzed one by one to determine whether the individual deviation values ​​are concentrated within a certain fixed range. Based on the uniformity of the distribution of individual deviation values, the overall stability of all peak width values ​​in the peak width value sequence is quantitatively determined. The more uniform the distribution, the better the peak width consistency; the more dispersed the distribution, the worse the peak width consistency. Finally, by comprehensively calculating the overall distribution of individual deviation values ​​and the overall stability of peak width, a quantitative index that can accurately reflect the consistency of chromatographic peak width, namely the peak width variation coefficient, is obtained.

[0111] The beneficial effects are that the highest point of the peak, the starting point of the left baseline, and the starting point of the right baseline can be accurately identified through the chromatographic peak profile curve. The symmetry parameters of the chromatographic peak profile curve are obtained by combining these feature points. The tailing factor is obtained by statistically analyzing the change trend of the ordinate of the profile points based on the highest point of the peak. The peak width consistency is measured by the difference in the width values ​​of the profile curve at different height positions within a predetermined height range near the highest point of the peak, and the peak width variation coefficient is obtained. The three types of parameters are combined to form the morphological feature parameters of the chromatographic candidate peak image, which comprehensively and accurately characterizes the shape features of the chromatographic candidate peak, and provides reliable parameter support for subsequent chromatographic peak identification and analysis.

[0112] S4. Based on the historical chromatogram sample set, train and learn the feature vector and peak morphology category label of the chromatogram to be identified to obtain the abnormal morphology discrimination mapping table of the chromatogram to be identified.

[0113] In this embodiment of the invention, the step of training and learning the feature vector and peak morphology category label of the chromatogram to be identified based on a historical chromatogram sample set to obtain an abnormal morphology discrimination mapping table for the chromatogram to be identified includes:

[0114] Obtain a historical chromatogram sample set, which contains multiple historical chromatogram peak images, and the historical chromatogram peak images are pre-labeled with peak morphology category labels to characterize their peak shape status;

[0115] The historical chromatographic peak images are retrospectively analyzed to obtain the historical morphological feature parameters of the historical chromatographic peak images, and the historical morphological feature parameters are used as historical feature vectors.

[0116] A correlation analysis is performed on the historical feature vector and the peak shape category label to obtain the correspondence between the numerical distribution range of the historical feature vector and the peak shape category label;

[0117] Based on the aforementioned correspondence, an anomaly morphology discrimination mapping table is established, with the numerical range of the feature vector as the index and the peak morphology category label as the mapping value.

[0118] Historical chromatograms under various operating conditions and for different detection objects were collected. These historical chromatograms were then categorized and organized, and invalid historical chromatograms with blurred images, incomplete peak regions, or severe baseline drift were removed. Valid historical chromatograms with clear images, complete peak regions, and stable baselines were retained. All valid historical chromatograms together constitute a historical chromatogram sample set. Each valid historical chromatogram corresponds to a historical chromatographic peak image, and each historical chromatographic peak image is pre-labeled with a unique peak morphology category label by manual annotation. This peak morphology category label is used to clearly characterize the peak shape state of the corresponding historical chromatographic peak image. The label types cover all preset peak shape categories, such as normal peak, tailing peak, leading peak, and split peak, ensuring that the peak shape state of each historical chromatographic peak image can be accurately reflected by the corresponding peak morphology category label.

[0119] Each historical chromatographic peak image in the historical chromatographic sample set is selected, and all operations S2 and S3 are performed sequentially on each historical chromatographic peak image. That is, firstly, the contour extraction analysis of the historical chromatographic peak image is performed to obtain the chromatographic peak contour curve of the historical chromatographic peak image. Then, based on the chromatographic peak contour curve, the historical chromatographic peak image is quantitatively characterized to obtain the symmetry parameter, tailing factor and peak width variation coefficient of the historical chromatographic peak contour curve. These three types of parameters are integrated in a fixed order to form a parameter set that can comprehensively characterize the peak shape features of the historical chromatographic peak image. This parameter set serves as a historical feature vector, ensuring that each historical chromatographic peak image has a unique corresponding historical feature vector, and that the historical feature vector completely matches the peak shape state of the historical chromatographic peak image.

[0120] All historical feature vectors are associated one-to-one with their corresponding peak morphology category labels to ensure that each historical feature vector can accurately match the peak morphology category label of its corresponding historical chromatographic peak image. Then, a comprehensive analysis is performed on the associated historical feature vectors and peak morphology category labels. The value ranges of symmetry parameters, tailing factors, and peak width variation coefficients are statistically analyzed for each historical feature vector corresponding to each peak morphology category label. The numerical distribution pattern of historical feature vectors corresponding to different peak morphology category labels is clarified. It is determined that historical feature vectors within a certain numerical distribution interval all correspond to the same peak morphology category label, forming a clear correspondence between the numerical distribution interval of historical feature vectors and peak morphology category labels. This ensures that the correspondence pattern can accurately reflect the association between historical feature vectors and peak morphology category labels.

[0121] Based on the corresponding patterns obtained in the steps, a basic framework for the mapping table is constructed. The numerical distribution intervals of the symmetry parameter, tailing factor, and peak width variation coefficient in the historical feature vector are used as indices of the mapping table. Each index corresponds to a unique peak morphology category label as a mapping value. The mapping relationship between each numerical interval index and the corresponding peak morphology category label is clearly defined to ensure that the correspondence between the index and the mapping value fully conforms to the correspondence pattern between the numerical distribution interval of the historical feature vector and the peak morphology category label. At the same time, the mapping table is verified to remove content with incorrect mapping relationships and duplicate indexes, and to supplement missing numerical interval indices and corresponding mapping values. Finally, a complete and accurate abnormal morphology discrimination mapping table is formed. This mapping table can quickly map the corresponding peak morphology category label to the numerical interval to which the feature vector of the chromatogram to be identified belongs.

[0122] The beneficial effects are as follows: by collecting and organizing historical chromatograms under different operating conditions and different detection objects, invalid images are removed and valid images are retained to form a historical chromatogram sample set. Each historical chromatogram peak image is labeled with a unique peak morphology category label. Then, contour extraction and quantitative characterization operations are performed on each historical chromatogram peak image to obtain historical feature vectors. The correlation analysis between historical feature vectors and peak morphology category labels is used to obtain the correspondence between the two. Based on the correspondence, a mapping table framework is built and verified, and finally a complete and accurate abnormal morphology discrimination mapping table is formed. This provides a reliable basis for the peak morphology discrimination of the chromatogram to be identified, ensuring that the discrimination process is accurate, efficient and reproducible.

[0123] S5. Based on the abnormal morphology discrimination mapping table, perform anomaly evaluation on the morphological feature parameters to obtain a comprehensive evaluation value of the morphological feature parameters;

[0124] In this embodiment of the invention, the step of performing anomaly assessment on the morphological feature parameters based on the anomaly morphology discrimination mapping table to obtain a comprehensive evaluation value for the morphological feature parameters includes:

[0125] The morphological characteristic parameters of the chromatogram to be identified are obtained, including symmetry parameters, tailing factor, and peak width variation coefficient.

[0126] Based on the abnormal morphology discrimination mapping table, the morphological feature parameters are matched item by item to obtain the similarity results of the morphological feature parameters;

[0127] Based on the peak morphology category labels corresponding to the historical feature vectors in the similarity results, the preliminary abnormal state category of the morphological feature parameters is determined.

[0128] Based on the preliminary abnormal state category, the contribution of the morphological feature parameters is evaluated to obtain the allocation weight of the morphological feature parameters;

[0129] Based on the weighted allocation and the abnormal morphology discrimination mapping table, the overall abnormality of the morphological feature parameters is graded and evaluated to obtain a comprehensive evaluation value for the morphological feature parameters.

[0130] The formula for calculating the comprehensive evaluation value is as follows:

[0131] ;

[0132] In the formula, The comprehensive evaluation value is... For the preliminary abnormal state category The The weighting of each morphological feature parameter. For the first The specific values ​​of each morphological feature parameter, when A value of 1 indicates a symmetry parameter; when... A value of 2 indicates a tailing factor. A value of 3 indicates the peak width variation coefficient. The first under the normal morphology category The baseline value of each morphological feature parameter The first under the normal morphology category The standard deviation of each morphological feature parameter The norm index is used to control the combination of different feature deviations.

[0133] The comprehensive evaluation value is derived from the quantitative value obtained by classifying and assessing the overall abnormality of the morphological feature parameters. This value is determined by combining the assigned weights and the abnormal morphology discrimination mapping table, and is used to quantitatively characterize the overall abnormality of the morphological feature parameters.

[0134] Preliminary Abnormal Status Category The The weighting of each morphological feature parameter is derived from the contribution assessment process based on the initial abnormal state category. By analyzing the influence of each type of parameter on the initial abnormal state category, the degree of deviation of the values ​​of each type of parameter from the normal range and the degree of fit with the initial abnormal state category are determined. The higher the degree of fit, the higher the contribution. Then, corresponding weights are assigned according to the contribution, following the principle that the higher the contribution, the greater the weight, and the lower the contribution, the smaller the weight. The sum of the weights of the three types of parameters is a fixed value, and the final weighting is obtained.

[0135] No. The specific values ​​of each morphological feature parameter are derived from the morphological feature parameters extracted after quantitative characterization of the chromatogram to be identified. The first parameter is the symmetry parameter, the second is the tailing factor, and the third is the peak width variation coefficient. After extraction, each parameter is checked to ensure that it is complete, valid, without missing values, and without abnormal values, so as to accurately reflect the shape characteristics of the chromatographic peaks in the chromatogram to be identified.

[0136] Normal morphology category The baseline value of each morphological feature parameter is derived from the historical feature vector corresponding to the normal peak morphology category label in the abnormal morphology discrimination mapping table. By statistically analyzing the values ​​of each morphological feature parameter in the historical feature vectors corresponding to all normal peak morphology categories, the representative value of all values ​​is taken as the baseline value to ensure that the baseline value can accurately reflect the standard level of the corresponding parameter under normal morphology.

[0137] Normal morphology category The standard deviation of each morphological feature parameter is derived from all historical feature vectors corresponding to the normal peak morphology category label in the abnormal morphology discrimination mapping table. By statistically analyzing all values ​​of the corresponding morphological feature parameter under this category, the deviation of each value from the benchmark value is calculated. Then, all deviations are comprehensively calculated to obtain the standard deviation that reflects the dispersion of the corresponding parameter values ​​under normal morphology.

[0138] The norm index is used to control the combination of different feature deviations. Its value is determined based on the analysis results of historical chromatogram sample sets. By analyzing the deviation of various parameters from the normal value range in the historical feature vectors, and combining the impact of different deviation combinations on the accuracy of the comprehensive evaluation value, a fixed norm index is determined to ensure that the deviation of various parameters can be reasonably integrated so that the comprehensive evaluation value can accurately reflect the overall degree of abnormality.

[0139] The significance of this formula lies in integrating and calculating the deviation degree of each type of morphological feature parameter with its corresponding assigned weight. By calculating the deviation degree of each type of parameter relative to the normal benchmark value, and then assigning influence weights corresponding to different deviation degrees according to the weight size, the formula obtains a comprehensive evaluation value that can quantitatively characterize the overall abnormality degree of the morphological feature parameters through the norm index control of the combination of deviation degrees. This achieves accurate quantification of the abnormality degree of the morphological feature parameters of the chromatogram to be identified, providing a scientific and accurate quantitative basis for the subsequent anomaly identification of the chromatogram to be identified, ensuring the reliability and reproducibility of the anomaly identification results, and closely connecting with the morphological feature parameter anomaly assessment process described above. It fully utilizes the assigned weights, morphological feature parameters, and other products obtained above, completing the transformation from parameters to quantitative evaluation values.

[0140] The morphological feature parameters of the chromatogram to be identified were extracted and quantitatively characterized. It was determined that the morphological feature parameters consist of three types of parameters: symmetry parameter, tailing factor, and peak width variation coefficient. The extracted parameters were checked one by one to confirm that each type of parameter was complete, valid, without missing values ​​or abnormal values. This ensured that the extracted morphological feature parameters could comprehensively and accurately reflect the shape characteristics of the chromatographic peaks in the chromatogram to be identified, providing reliable basic data for subsequent anomaly assessment. Finally, it was determined that the morphological feature parameters of the obtained chromatogram to be identified included only the symmetry parameter, tailing factor, and peak width variation coefficient.

[0141] A pre-established abnormal morphology discrimination mapping table is retrieved. This mapping table uses the numerical range of the feature vector as the index and the peak morphology category label as the mapping value. Each type of morphological feature parameter in the chromatogram to be identified is compared and matched item by item with the corresponding numerical range of the parameter in the mapping table. Specifically, the symmetry parameter is compared with each numerical range of the symmetry parameter in the mapping table, the tailing factor is compared with each numerical range of the tailing factor in the mapping table, and the peak width variation coefficient is compared with each numerical range of the peak width variation coefficient in the mapping table. The corresponding numerical range matched for each type of parameter and the historical feature vector associated with that numerical range are recorded. All comparison results are integrated to obtain the similarity results of the morphological feature parameters, ensuring that the similarity results can accurately reflect the matching situation between the morphological feature parameters to be identified and the historical feature vectors in the mapping table.

[0142] From the similarity results of morphological feature parameters, the historical feature vector corresponding to each matching result is extracted. Then, based on the correspondence between the historical feature vector and the peak morphology category label in the abnormal morphology discrimination mapping table, the peak morphology category label corresponding to each historical feature vector is determined. All extracted peak morphology category labels are summarized and analyzed, and the frequency of occurrence of each category label is counted. The peak morphology category label with the highest frequency is used as the core reference. Combined with the degree of matching between the morphological feature parameter to be identified and the corresponding historical feature vector, the preliminary abnormal state category corresponding to the morphological feature parameter is determined. The preliminary abnormal state category covers four preset categories: normal, slight abnormality, moderate abnormality, and severe abnormality, to ensure that the preliminary abnormal state category can accurately reflect the basic abnormality of the morphological feature parameter.

[0143] Based on the established preliminary anomaly category, the influence of each type of morphological feature parameter on this category is analyzed. Specifically, it is determined which parameter among the symmetry parameter, tailing factor, and peak width variation coefficient has a value that deviates more from the normal range and better matches the preliminary anomaly category. The parameter with a higher degree of deviation and better match with the preliminary anomaly category has a higher contribution to the anomaly. Each type of parameter is assigned a corresponding weight according to its contribution, following the principle that the higher the contribution, the greater the weight, and the lower the contribution, the smaller the weight. The total weight of the three types of parameters is a fixed value. This completes the contribution assessment of the morphological feature parameters and obtains the assigned weights for the morphological feature parameters, ensuring that the assigned weights accurately reflect the influence of each type of parameter on the anomaly.

[0144] Combining the obtained weighted distribution and abnormal morphology discrimination mapping table, the values ​​of each type of parameter in the morphological feature parameters are compared again with the normal value range and the value range corresponding to different abnormal levels in the mapping table. According to the weight of each type of parameter, the degree of abnormality of each type of parameter is weighted and calculated. In the weighted calculation process, the parameter with a larger weight has a greater impact on the overall degree of abnormality. Based on the weighted calculation results, the overall degree of abnormality of the morphological feature parameters is graded and evaluated according to the preset abnormality level classification standard. The grading and evaluation standard corresponds to the four preset categories of the preliminary abnormal state category. Finally, a value that can quantitatively characterize the overall degree of abnormality of the morphological feature parameters is obtained. This value is the comprehensive evaluation value of the morphological feature parameters, ensuring that the comprehensive evaluation value can accurately and comprehensively reflect the abnormality of the morphological feature parameters of the chromatogram to be identified.

[0145] The beneficial effects are as follows: by extracting morphological feature parameters from the chromatogram to be identified through quantitative characterization and verifying their completeness and validity, reliable basic data is provided for anomaly assessment. Then, the anomaly morphology discrimination mapping table is retrieved, and each morphological feature parameter is compared and matched with the corresponding parameter value range in the mapping table to obtain similar results. Peak morphology category labels corresponding to historical feature vectors are extracted from the similar results. The initial anomaly state category is determined by combining the frequency of label occurrence and the degree of matching. Based on the initial anomaly state category, the contribution of various parameters to the anomaly state is analyzed and corresponding weights are assigned. Finally, the degree of anomaly of the morphological feature parameters is weighted and graded by combining the assigned weights with the anomaly morphology discrimination mapping table, resulting in a comprehensive evaluation value that can accurately and comprehensively characterize the overall degree of anomaly of the morphological feature parameters. This provides scientific and accurate quantitative support for the anomaly discrimination of the chromatogram to be identified and ensures the reliability of the anomaly discrimination results.

[0146] S6. The comprehensive evaluation value is compared and identified by a threshold to obtain the abnormal identification result of the chromatogram to be identified.

[0147] In this embodiment of the invention, the step of performing threshold comparison and identification on the comprehensive evaluation value to obtain the anomaly identification result of the chromatogram to be identified includes:

[0148] The comprehensive evaluation value is compared with the preset anomaly judgment threshold value one by one to obtain the anomaly judgment result of the comprehensive evaluation value.

[0149] Based on the anomaly determination results, the chromatographic candidate peak images are adapted and screened to obtain the anomaly peak images to be labeled in the chromatographic candidate peak images;

[0150] On the chromatogram to be identified, the chromatographic peak region where the abnormal peak image to be labeled is located is highlighted and marked;

[0151] Based on the position of the highlighted marker and the preliminary abnormality category of the abnormal peak image to be marked, the abnormality identification result of the chromatogram to be identified is obtained.

[0152] The system retrieves pre-set anomaly judgment thresholds, which are determined based on the comprehensive evaluation values ​​of normal and abnormal chromatographic peaks in a historical chromatographic sample set. There are three fixed thresholds, corresponding to the judgment criteria for slight, moderate, and severe anomalies, respectively. The first threshold corresponds to the distinction between slight and normal, the second to moderate and slight, and the third to severe and moderate. The comprehensive evaluation value of the obtained morphological characteristic parameters is compared with these three anomaly judgment thresholds one by one. If the comprehensive evaluation value is lower than the first threshold, it is judged as normal; if the comprehensive evaluation value is between the first and second thresholds, it is judged as slight anomaly; if the comprehensive evaluation value is between the second and third thresholds, it is judged as moderate anomaly; and if the comprehensive evaluation value is higher than the third threshold, it is judged as severe anomaly. The final anomaly judgment result is obtained based on the comprehensive evaluation value.

[0153] The anomaly determination results of the comprehensive evaluation values ​​obtained above are combined with the chromatographic candidate peak images of the corresponding chromatograms to be identified. The anomaly determination results are then matched and filtered according to their categories. The filtering rule is to retain only the chromatographic candidate peak images corresponding to the anomaly determination results of slight, moderate, and severe anomalies, and to remove the chromatographic candidate peak images corresponding to the anomaly determination results of normal. The filtered chromatographic candidate peak images are checked one by one to confirm that each image corresponds to a valid anomaly determination result and that the image is complete and the peak region is clear, ensuring that the filtered images are all chromatographic peak images with anomalies. Finally, the anomaly peak images to be labeled are obtained from the chromatographic candidate peak images.

[0154] Retrieve the original chromatogram to be identified, clearly identify the chromatographic peak region corresponding to the abnormal peak image to be labeled in the chromatogram, and mark the region using a preset highlighting marking method. The marking method is to draw continuous exclusive marking lines along the outline edge of the chromatographic peak region. The color of the lines is clearly distinguishable from the baseline and peak region color of the chromatogram to be identified. At the same time, mark the corresponding abnormality level below the marked region. The marking content is consistent with the abnormality judgment result, namely slight abnormality, moderate abnormality, and severe abnormality. During the marking process, ensure that the lines are continuous and the markings are clear, without obscuring the key features of the chromatographic peak. After marking is completed, confirm that the highlighted marking can accurately point to the chromatographic peak region where the abnormal peak image to be labeled is located, and that the marking content completely matches the abnormality situation.

[0155] The system records the exact locations of all highlighted markers on the chromatogram to be identified. These locations are clearly defined by the coordinate range of the chromatogram to ensure accurate positioning of the chromatographic peak region where each abnormal peak image is located. Simultaneously, it extracts the preliminary abnormality category corresponding to each abnormal peak image and associates the coordinate locations of the highlighted markers with the corresponding preliminary abnormality categories. All associated information is integrated to form a complete information set containing the abnormal peak locations and abnormality levels. This information set can clearly reflect the specific location and degree of abnormality of the abnormal chromatographic peaks in the chromatogram to be identified, ultimately yielding the abnormality identification results of the chromatogram to be identified.

[0156] The beneficial effects are as follows: by retrieving the anomaly judgment threshold determined based on the analysis of historical chromatogram sample sets, the comprehensive evaluation value is compared with the threshold value one by one to obtain the anomaly judgment result. Combined with the anomaly judgment result, the images of abnormal peaks to be marked are screened and their validity is verified. Then, the abnormal peak areas are clearly and distinctly highlighted on the chromatogram to be identified, the marking position is recorded and associated with the corresponding preliminary abnormality status category, and all related information is integrated to form a complete information set. Finally, the anomaly identification result of the chromatogram to be identified can be obtained, which can clearly reflect the specific location and degree of abnormality of the abnormal chromatogram peak. This provides a clear and definite basis for the anomaly judgment of the chromatogram to be identified, ensuring the reliability and practicality of anomaly identification.

[0157] like Figure 2 The diagram shown is a functional block diagram of a chromatographic peak anomaly morphology recognition system based on image processing provided in an embodiment of the present invention.

[0158] The image processing-based chromatographic peak anomaly morphology recognition system described in this invention can be installed in an electronic device. Depending on the functions implemented, the image processing-based chromatographic peak anomaly morphology recognition system may include a chromatographic peak region segmentation module M1, a contour extraction and analysis module M2, a quantification and characterization module M3, a discriminant mapping construction module M4, an anomaly assessment module M5, and a recognition output module M6. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by the processor of an electronic device and perform a fixed function, stored in the memory of the electronic device.

[0159] In this embodiment, the functions of each module / unit are as follows:

[0160] The chromatographic peak region segmentation module M1 is used to segment the chromatographic peak region of the chromatogram to be identified, and obtain the chromatographic candidate peak image of the chromatogram to be identified.

[0161] The contour extraction and analysis module M2 is used to perform contour extraction and analysis on the chromatographic candidate peak image to obtain the chromatographic peak contour curve of the chromatographic candidate peak image;

[0162] The quantization characterization module M3 is used to quantify the chromatographic candidate peak image based on the chromatographic peak profile curve to obtain the morphological feature parameters of the chromatographic candidate peak image.

[0163] The discrimination mapping construction module M4 is used to train and learn the feature vector and morphological category label of the chromatogram to be identified based on the historical chromatogram sample set, so as to obtain the abnormal morphology discrimination mapping table of the chromatogram to be identified.

[0164] The anomaly assessment module M5 is used to assess the anomalies of the morphological feature parameters based on the anomaly morphology discrimination mapping table, and obtain a comprehensive evaluation value of the morphological feature parameters.

[0165] The identification output module M6 is used to perform threshold comparison and identification on the comprehensive evaluation value to obtain the abnormal identification result of the chromatogram to be identified.

[0166] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0167] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0168] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0169] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0170] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0171] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for identifying abnormal morphology of chromatographic peaks based on image processing, characterized in that, The method includes: S1. Segment the chromatogram to be identified by chromatographic peak regions to obtain chromatographic candidate peak images of the chromatogram to be identified; S2. Perform contour extraction analysis on the chromatographic candidate peak image to obtain the chromatographic peak contour curve of the chromatographic candidate peak image; S3. Based on the chromatographic peak profile curve, the chromatographic candidate peak image is quantitatively characterized to obtain the morphological feature parameters of the chromatographic candidate peak image; S4. Based on the historical chromatogram sample set, train and learn the feature vector and peak morphology category label of the chromatogram to be identified to obtain the abnormal morphology discrimination mapping table of the chromatogram to be identified. S5. Based on the abnormal morphology discrimination mapping table, perform anomaly evaluation on the morphological feature parameters to obtain a comprehensive evaluation value of the morphological feature parameters; S6. The comprehensive evaluation value is compared and identified by a threshold to obtain the abnormal identification result of the chromatogram to be identified.

2. The method for identifying abnormal morphology of chromatographic peaks based on image processing as described in claim 1, characterized in that, The step of segmenting the chromatogram to be identified into chromatographic peak regions to obtain chromatographic candidate peak images of the chromatogram to be identified includes: The chromatogram to be identified is converted to grayscale to obtain an enhanced chromatographic image of the chromatogram to be identified; The enhanced chromatographic image is subjected to median filtering to obtain a denoised chromatographic image of the chromatogram to be identified; The foreground pixels corresponding to the peak regions in the denoised chromatographic image are segmented into chromatographic peaks to obtain a binary chromatographic image of the chromatogram to be identified. A morphological opening operation is performed on the binary chromatographic image to obtain a connected binary image of the chromatogram to be identified; The connected binary image is analyzed and labeled with connected components to obtain the chromatographic candidate peak image of the chromatogram to be identified.

3. The method for identifying abnormal morphology of chromatographic peaks based on image processing as described in claim 1, characterized in that, The step of performing contour extraction analysis on the chromatographic candidate peak image to obtain the chromatographic peak contour curve of the chromatographic candidate peak image includes: The pixels in the peak region of the chromatographic candidate peak image are assigned a first gray value, and the background pixels are assigned a second gray value; Based on the first gray value and the second gray value, contour detection is performed on the pixels at the boundary between the peak region and the background of the chromatographic candidate peak image to obtain the edge pixel set of the chromatographic candidate peak image; Contour tracking is performed on the set of edge pixels, and adjacent edge pixels are connected sequentially according to the eight-neighbor connection relationship to obtain the initial closed contour line of the peak region. The initial closed contour line is smoothed to obtain the chromatographic peak contour curve of the chromatographic candidate peak image.

4. The method for identifying abnormal morphology of chromatographic peaks based on image processing as described in claim 1, characterized in that, The step of quantitatively characterizing the chromatographic candidate peak image based on the chromatographic peak profile curve to obtain the morphological feature parameters of the chromatographic candidate peak image includes: Based on the chromatographic peak profile curve, identify the highest peak point, the left baseline start point, and the right baseline start point on the profile curve; The horizontal distances between the highest point of the peak and the starting points of the left and right baselines are analyzed simultaneously to obtain the symmetry parameters of the chromatographic peak profile curve. Based on the highest point of the peak, the trend of the change of the ordinate of the contour points in the chromatographic peak contour curve is statistically analyzed to obtain the tailing factor of the chromatographic peak contour curve. Within a predetermined height range near the highest point of the peak, the width values ​​of the contour curve at different height positions are obtained. Based on the degree of difference between the width values, the consistency of the peak width in the chromatographic candidate peak image is measured to obtain the peak width variation coefficient of the chromatographic peak. The symmetry parameter, the tailing factor, and the peak width variation coefficient are combined into morphological feature parameters of the chromatographic candidate peak image.

5. The method for identifying abnormal morphology of chromatographic peaks based on image processing as described in claim 4, characterized in that, The step of measuring the consistency of the peak widths in the chromatographic candidate peak images based on the degree of difference between the width values, and obtaining the peak width variation coefficient of the chromatographic peaks, includes: Collect the peak width values ​​of the peak regions in the same chromatographic candidate peak image to obtain the peak width value sequence of the chromatographic candidate peak image; The average peak width value of the peak width value sequence is obtained by averaging the peak width values ​​in the sequence. The deviation between the peak width value and the average peak width value is evaluated to obtain the individual deviation value of the peak width value; Based on the overall distribution of the individual deviation values, the consistency between peak width values ​​in the peak width value sequence is coupled and measured to obtain the peak width variation coefficient of the chromatographic peak.

6. The method for identifying abnormal morphology of chromatographic peaks based on image processing as described in claim 1, characterized in that, The step of training and learning the feature vector and peak morphology category labels of the chromatogram to be identified based on a historical chromatogram sample set to obtain an abnormal morphology discrimination mapping table for the chromatogram to be identified includes: Obtain a historical chromatogram sample set, which contains multiple historical chromatogram peak images, and the historical chromatogram peak images are pre-labeled with peak morphology category labels to characterize their peak shape status; The historical chromatographic peak images are retrospectively analyzed to obtain the historical morphological feature parameters of the historical chromatographic peak images, and the historical morphological feature parameters are used as historical feature vectors. A correlation analysis is performed on the historical feature vector and the peak shape category label to obtain the correspondence between the numerical distribution range of the historical feature vector and the peak shape category label; Based on the aforementioned correspondence, an anomaly morphology discrimination mapping table is established, with the numerical range of the feature vector as the index and the peak morphology category label as the mapping value.

7. The method for identifying abnormal morphology of chromatographic peaks based on image processing as described in claim 1, characterized in that, The step of evaluating the morphological feature parameters based on the abnormal morphology discrimination mapping table to obtain a comprehensive evaluation value for the morphological feature parameters includes: The morphological characteristic parameters of the chromatogram to be identified are obtained, including symmetry parameters, tailing factor, and peak width variation coefficient. Based on the abnormal morphology discrimination mapping table, the morphological feature parameters are matched item by item to obtain the similarity results of the morphological feature parameters; Based on the peak morphology category labels corresponding to the historical feature vectors in the similarity results, the preliminary abnormal state category of the morphological feature parameters is determined. Based on the preliminary abnormal state category, the contribution of the morphological feature parameters is evaluated to obtain the allocation weight of the morphological feature parameters; Based on the weighted allocation and the abnormal morphology discrimination mapping table, the overall abnormality of the morphological feature parameters is graded and evaluated to obtain a comprehensive evaluation value for the morphological feature parameters.

8. The method for identifying abnormal morphology of chromatographic peaks based on image processing as described in claim 7, characterized in that, The formula for calculating the comprehensive evaluation value is as follows: ； In the formula, The comprehensive evaluation value is... For the preliminary abnormal state category The The weighting of each morphological feature parameter. For the first The specific values ​​of each morphological feature parameter, when A value of 1 indicates a symmetry parameter; when... A value of 2 indicates a tailing factor. A value of 3 indicates the peak width variation coefficient. The first under the normal morphology category The baseline value of each morphological feature parameter The first under the normal morphology category The standard deviation of each morphological feature parameter The norm index is used to control the combination of different feature deviations.

9. The method for identifying abnormal morphology of chromatographic peaks based on image processing as described in claim 1, characterized in that, The step of performing threshold comparison and identification on the comprehensive evaluation value to obtain the anomaly identification result of the chromatogram to be identified includes: The comprehensive evaluation value is compared with the preset anomaly judgment threshold value one by one to obtain the anomaly judgment result of the comprehensive evaluation value. Based on the anomaly determination results, the chromatographic candidate peak images are adapted and screened to obtain the anomaly peak images to be labeled in the chromatographic candidate peak images; On the chromatogram to be identified, the chromatographic peak region where the abnormal peak image to be labeled is located is highlighted and marked; Based on the position of the highlighted marker and the preliminary abnormality category of the abnormal peak image to be marked, the abnormality identification result of the chromatogram to be identified is obtained.

10. A system for identifying abnormal morphology of chromatographic peaks based on image processing, characterized in that, The system is used to implement any one of the image processing-based methods for identifying abnormal chromatographic peak morphologies as described in claims 1-9, the system comprising: The chromatographic peak region segmentation module is used to segment the chromatographic peak regions of the chromatogram to be identified, and obtain the chromatographic candidate peak image of the chromatogram to be identified; The contour extraction and analysis module is used to perform contour extraction and analysis on the chromatographic candidate peak image to obtain the chromatographic peak contour curve of the chromatographic candidate peak image; The quantization characterization module is used to quantify and characterize the chromatographic candidate peak image based on the chromatographic peak profile curve to obtain the morphological feature parameters of the chromatographic candidate peak image. The discriminant mapping construction module is used to train and learn the feature vector and morphological category label of the chromatogram to be identified based on the historical chromatogram sample set, so as to obtain the abnormal morphology discriminant mapping table of the chromatogram to be identified; An anomaly assessment module is used to assess the anomalies of the morphological feature parameters based on the anomaly morphology discrimination mapping table, and obtain a comprehensive evaluation value of the morphological feature parameters. The identification output module is used to perform threshold comparison and identification on the comprehensive evaluation value to obtain the abnormal identification result of the chromatogram to be identified.

Images