Chromatogram analysis method and device based on artificial intelligence and computer equipment

Abstract

The application provides a chromatogram analysis method and device based on artificial intelligence and computer equipment, and the method comprises the following steps: training a deep learning neural network to obtain a chromatogram analysis model; receiving digital chromatogram data images to be analyzed by the chromatogram analysis model, and segmenting the images into regional images; performing image content state examination and identification on each regional image by the chromatogram analysis model to obtain an examination prediction region and mark the image content state; evaluating and calculating the image content state of the examination prediction region and the possible probability degree of the image content state to obtain a first ratio, marking the image content state with a specific color when the first ratio is greater than a preset first threshold probability ratio; and drawing a chromatographic examination report and a chromatographic analysis report for reference by detection personnel according to the information of a chromatographic experiment being performed, information of a sample to be detected and analysis information of the chromatogram analysis model. The application has the beneficial effect of greatly improving the chromatographic analysis rate.

Description

Technical Field

[0001] This invention relates to the field of chromatographic analysis technology, and in particular to a chromatographic analysis method, apparatus, and computer equipment based on artificial intelligence. Background Technology

[0002] Chromatography is widely used in environmental, chemical analysis, and pharmaceutical fields. Practical chromatographic techniques include liquid chromatography, gas chromatography, and gel electrophoresis. These techniques require strict experimental conditions. Because chromatographic instruments are highly sensitive, the chromatograms obtained in actual experiments can be affected by environmental factors, the instrument's current condition, sample concentration, and experimental procedures, resulting in peak distortion, deformation, baseline drift, noise, and other abnormal changes that prevent accurate analysis. Therefore, all types of chromatographic detection require highly trained professionals, incurring significant learning and time costs. In practice, due to the influence of experimental instruments, experimental environment, and test samples, chromatograms from each experiment will exhibit varying degrees of image distortion. Because distorted chromatographic images affect subsequent quantitative analysis calculations, professional testing personnel must inspect the chromatographic instruments before each test and spend considerable time meticulously reviewing and analyzing each sample manually. Thus, chromatographic detection suffers from problems such as time-consuming and labor-intensive manual review, high subjectivity, and poor repeatability. Meanwhile, existing automated detection systems primarily rely on mathematical and statistical calculations integrated into the software to determine the state of chromatograms, which cannot automatically and accurately detect whether there is excessive distortion or deformation in the chromatograms. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a chromatogram analysis method, apparatus and computer equipment based on artificial intelligence, which aims to improve the chromatogram analysis rate.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is: a chromatogram analysis method based on artificial intelligence, comprising the following steps:

[0005] S10. Train the deep learning neural network to obtain the chromatogram analysis model;

[0006] S20. Receive the digital chromatographic data image to be reviewed and analyzed through the chromatographic analysis model, and segment the digital chromatographic data image into regional images;

[0007] S30. The image content status of each region image is reviewed and identified by the chromatogram analysis model to obtain the review prediction region, and the review prediction region is marked with the image content status. The image content status includes the highest peak, leading peak, tailing peak, broad peak, split peak, negative peak, peak with insufficient resolution, baseline noise and baseline drift of the chromatogram.

[0008] S40. Evaluate and calculate the image content state and the possible probability of the image content state in the predicted area to obtain a first ratio. If the first ratio is greater than the preset first threshold probability ratio, then filter it out and mark it with a specific color.

[0009] S50. Based on the ongoing chromatographic experiment information, the sample information to be tested, and the analytical information of the chromatographic analysis model, draw up a chromatographic review report and a chromatographic analysis report for reference by the testing personnel.

[0010] Furthermore, step S10, training the deep learning neural network to obtain the chromatogram analysis model, specifically includes:

[0011] S11. Obtain training sample images with pre-annotated image content states through a deep learning neural network, and segment the digital chromatographic data images of the training set into region images; the deep learning neural network adopts the DeepLabv3+ network; the backbone network of the DeepLabv3+ network is set to GhostNetv2.

[0012] S12. After performing row convolution and column convolution on the region image, the feature parameters are standardized by batch normalization and the linearity of the feature parameters is reduced by nonlinear activation to obtain a first feature image with compressed length and width of the region image.

[0013] S13. The first feature image is input into the seven channels of the ASPP encoder with dilation rates of 1, 6, 12, 18, 24, 28, and 32, respectively. The first feature image is convolved in these seven channels by convolution kernels that are dilated by 1, 6, 12, 18, 24, 28, and 32 times, respectively, to extract deep features. The extracted seven-channel deep features are then compressed through point convolution to obtain the second feature image.

[0014] S14. The first feature image and the second feature image are used as input features of the decoder, and the review prediction features are obtained by fusing redefinition size, point convolution and bilinear interpolation.

[0015] S15. Minimize the loss function by examining the predicted features and their corresponding labels. When the loss value meets the convergence requirement, the trained chromatogram analysis model is obtained.

[0016] Furthermore, in step S12, the formulas used for performing row convolution and column convolution on the region image are as follows:

[0017]

[0018] Among them, y i The values ​​represent the output features, x represents the input features, w represents the convolution kernel, and r represents the stride.

[0019] Furthermore, in step S12, the feature parameters are standardized through batch standardization, and the formula is as follows:

[0020]

[0021] Here, Θ represents the parameters of the training network, and gradient estimation is performed on a training dataset of N total numbers.

[0022] Furthermore, in step S12, the nonlinear activation reduces the linearity of the feature parameters by using the ReLU function and the softmax function as the minimization loss function; the minimization loss function is:

[0023]

[0024] FocalLoss = -log p t (1-p t ) 2

[0025]

[0026] Among them, l n Influenced by the set weights, p represents the average predicted probability of the training dataset; t P represents the model's predicted probability for the target class, indicating an adjustment to the model's focus on positive samples in the training dataset; i and G i It is the value of the prediction result at the i-th pixel, representing the similarity between the prediction result of the trained model and the true label.

[0027] Furthermore, the digital chromatographic data image is a two-dimensional gas / liquid chromatographic data image, with the x-axis unit being time and the y-axis unit being signal value.

[0028] Furthermore, step S50, based on the ongoing chromatographic experiment information, the sample information to be tested, and the analytical information of the chromatographic analysis model, generates a chromatographic review report and a chromatographic analysis report for reference by the testing personnel. Specifically, this includes...

[0029] S51. Compile the ongoing chromatographic experiment information, the sample information to be tested, and the analytical information of the chromatographic analysis model into a chromatographic review report for reference by the testing personnel.

[0030] S52. Extract the signal values ​​and time parameters from qualified chromatograms without abnormal peaks, and calculate the chromatographic analysis parameters, including retention time t. R Peak height h, peak area A, half peak height W 1 / 2 The peak width W and standard deviation σ are calculated, and a chromatographic analysis report is prepared for reference by the testing personnel based on the chromatographic experimental information, sample information, and calculated analytical parameters.

[0031] The present invention also provides an artificial intelligence-based chromatogram analysis device, comprising:

[0032] The model training module is used to train the deep learning neural network to obtain the chromatogram analysis model;

[0033] The image segmentation module is used to receive digital chromatographic data images to be reviewed and analyzed through a chromatographic analysis model, and to segment the digital chromatographic data images into regional images.

[0034] The image content status recognition module is used to review and identify the image content status of each region image through the chromatogram analysis model to obtain the review prediction region, and to mark the review prediction region with the image content status. The image content status includes the highest peak, leading peak, tailing peak, broad peak, split peak, negative peak, peak with insufficient resolution, baseline noise and baseline drift of the chromatogram.

[0035] The image content status evaluation module is used to evaluate and calculate the image content status of the review prediction area and the possible probability of the image content status to obtain a first ratio. If the first ratio is greater than the preset first threshold probability ratio, it is filtered out and marked with a specific color.

[0036] The report generation module is used to generate chromatographic review reports and chromatographic analysis reports for reference by testing personnel based on the information of the ongoing chromatographic experiment, the information of the sample to be tested, and the analytical information of the chromatographic analysis model.

[0037] The present invention also provides a computer device, the computer device including a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the artificial intelligence-based chromatogram analysis method as described in any of the preceding claims.

[0038] The present invention also provides a storage medium storing a computer program that, when executed by a processor, can implement the artificial intelligence-based chromatogram analysis method as described in any of the preceding claims.

[0039] The beneficial effects of this invention are as follows: This solution provides a method for automatically checking for distortion and deformation of chromatograms and analyzing qualified chromatograms. It utilizes artificial intelligence technology for automatic review and analysis of chromatograms. The automated chromatogram review and analysis technology can greatly improve the chromatographic analysis rate. Its automatic review and analysis function can reduce the subjectivity of manual review and shorten the detection and testing time of chromatographic detection experiments in different fields, effectively helping to speed up chromatographic detection experiments. Attached Figure Description

[0040] The specific structure of the present invention will now be described in detail with reference to the accompanying drawings.

[0041] Figure 1 This is a flowchart of the chromatogram analysis method based on artificial intelligence according to an embodiment of the present invention;

[0042] Figure 2 This is a flowchart of the deep learning neural network training process according to an embodiment of the present invention;

[0043] Figure 3 This is a block diagram of an artificial intelligence-based chromatogram analysis device according to an embodiment of the present invention;

[0044] Figure 4 This is a block diagram of the model training module in an embodiment of the present invention;

[0045] Figure 5 This is a schematic block diagram of a computer device according to an embodiment of the present invention. Detailed Implementation

[0046] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0047] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0048] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0049] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0050] like Figure 1 As shown, the first embodiment of the present invention is: a chromatogram analysis method based on artificial intelligence, comprising the following steps:

[0051] S10. Train the deep learning neural network to obtain the chromatogram analysis model;

[0052] S20. Receive the digital chromatographic data image to be reviewed and analyzed through the chromatographic analysis model, and segment the digital chromatographic data image into regional images;

[0053] In this step, digital chromatographic data refers to the data file in chromatographic data format processed by a professional chromatographic detector, as well as the digital chromatographic data image depicted by the chromatographic data format data file after data processing.

[0054] S30. The image content status of each region image is reviewed and identified by the chromatogram analysis model to obtain the review prediction region, and the review prediction region is marked with the image content status. The image content status includes the highest peak, leading peak, tailing peak, broad peak, split peak, negative peak, peak with insufficient resolution, baseline noise and baseline drift of the chromatogram.

[0055] S40. Evaluate and calculate the image content state and the possible probability of the image content state in the predicted area to obtain a first ratio. If the first ratio is greater than the preset first threshold probability ratio, then filter it out and mark it with a specific color.

[0056] In this step, the highest peak is marked with color code #CD0000, the leading peak with color code #ECE513, the tailing peak with color code #2W8B57, the broad peak with color code #0000CD, the split peak with color code #CD853F, the negative peak with color code #9400D3, the peak with insufficient separation with color code #FF1493, and the baseline problem with color code #008000.

[0057] S50. Based on the ongoing chromatographic experiment information, the sample information to be tested, and the analytical information of the chromatographic analysis model, draw up a chromatographic review report and a chromatographic analysis report for reference by the testing personnel.

[0058] In this step, the chromatographic review report and chromatographic analysis report contain complete public experimental information and their respective attribute data, thus providing analytical guidance for testing personnel and a standardized data transmission format for subsequent system upgrades. This improves the testing efficiency of personnel and provides a standardized data input format for system expansion and upgrades.

[0059] This solution provides a method for automatically identifying distortions in chromatograms and analyzing qualified chromatograms. It utilizes artificial intelligence technology for automatic review and analysis of chromatograms. This automated chromatogram review and analysis technology can greatly improve the speed of chromatographic analysis. Its automatic review and analysis function can reduce the subjectivity of manual review and shorten the detection and testing time of chromatographic detection experiments in different fields, effectively helping to speed up chromatographic detection experiments.

[0060] like Figure 2 As shown, in a specific embodiment, step S10, training the deep learning neural network to obtain the chromatogram analysis model, specifically includes:

[0061] S11. Obtain training sample images with pre-annotated image content states through a deep learning neural network, and segment the digital chromatographic data images of the training set into region images; the deep learning neural network adopts the DeepLabv3+ network; the backbone network of the DeepLabv3+ network is set to GhostNetv2.

[0062] S12. After performing row convolution and column convolution on the region image, the feature parameters are standardized by batch normalization and the linearity of the feature parameters is reduced by nonlinear activation to obtain a first feature image with compressed length and width of the region image.

[0063] S13. The first feature image is input into the seven channels of the ASPP encoder with dilation rates of 1, 6, 12, 18, 24, 28, and 32, respectively. The first feature image is convolved in these seven channels by convolution kernels that are dilated by 1, 6, 12, 18, 24, 28, and 32 times, respectively, to extract deep features. The extracted seven-channel deep features are then compressed through point convolution to obtain the second feature image.

[0064] S14. The first feature image and the second feature image are used as input features of the decoder, and the review prediction features are obtained by fusing redefinition size, point convolution and bilinear interpolation.

[0065] S15. Minimize the loss function by examining the predicted features and their corresponding labels. When the loss value meets the convergence requirement, the trained chromatogram analysis model is obtained.

[0066] In one specific embodiment, in step S12, the formulas used for performing row convolution and column convolution on the region image are as follows:

[0067]

[0068] Among them, y i The values ​​represent the output features, x represents the input features, w represents the convolution kernel, and r represents the stride.

[0069] In one specific embodiment, in step S12, the feature parameters are standardized by batch standardization, and the formula is:

[0070]

[0071] Here, Θ represents the parameters of the training network, and gradient estimation is performed on a training dataset of N total numbers.

[0072] In one specific embodiment, in step S12, the loss function used to reduce the linearity of the feature parameters through nonlinear activation is the ReLU function and the softmax function; the loss function is:

[0073]

[0074] FocalLoss = -log p t (1-p t ) 2

[0075]

[0076] Among them, l n Influenced by the set weights, p represents the average predicted probability of the training dataset; t P represents the model's predicted probability for the target class, indicating an adjustment to the model's focus on positive samples in the training dataset; i and G i It is the value of the prediction result at the i-th pixel, representing the similarity between the prediction result of the trained model and the true label.

[0077] In this embodiment, the loss functions are categorized by name: CE loss function, Focal loss function, and Dice loss function. Model training and evaluation are performed by calculating the loss value using the loss function. The chromatogram analysis model is obtained through iterative training. Each training iteration calculates the loss function by minimizing the predicted features and their corresponding labels. Model training is complete when the loss value meets the convergence requirement; if the loss value does not meet the convergence requirement, iterative training is required until the loss value converges.

[0078] In one specific embodiment, the digital chromatographic data image is a two-dimensional gas / liquid chromatographic data image, where the x-axis is in units of time and the y-axis is in units of signal value.

[0079] In one specific embodiment, step S50, drawing a chromatographic review report and a chromatographic analysis report for reference by testing personnel based on the ongoing chromatographic experiment information, the sample information to be tested, and the analytical information of the chromatographic analysis model, specifically includes:

[0080] S51. Compile the ongoing chromatographic experiment information, the sample information to be tested, and the analytical information of the chromatographic analysis model into a chromatographic review report for reference by the testing personnel.

[0081] S52. Extract the signal values ​​and time parameters from qualified chromatograms without abnormal peaks, and calculate the chromatographic analysis parameters, including retention time t. R Peak height h, peak area A, half peak height W 1 / 2 The peak width W and standard deviation σ are calculated, and a chromatographic analysis report is prepared for reference by the testing personnel based on the chromatographic experimental information, sample information, and calculated analytical parameters.

[0082] like Figure 3 As shown, another embodiment of the present invention is a chromatogram analysis device based on artificial intelligence, comprising:

[0083] The model training module is used to train the deep learning neural network to obtain the chromatogram analysis model;

[0084] The image segmentation module is used to receive digital chromatographic data images to be reviewed and analyzed through a chromatographic analysis model, and to segment the digital chromatographic data images into regional images.

[0085] The image content status recognition module is used to review and identify the image content status of each region image through the chromatogram analysis model to obtain the review prediction region, and to mark the review prediction region with the image content status. The image content status includes the highest peak, leading peak, tailing peak, broad peak, split peak, negative peak, peak with insufficient resolution, baseline noise and baseline drift of the chromatogram.

[0086] The image content status evaluation module is used to evaluate and calculate the image content status of the review prediction area and the possible probability of the image content status to obtain a first ratio. If the first ratio is greater than the preset first threshold probability ratio, it is filtered out and marked with a specific color.

[0087] The report generation module is used to generate chromatographic review reports and chromatographic analysis reports for reference by testing personnel based on the information of the ongoing chromatographic experiment, the information of the sample to be tested, and the analytical information of the chromatographic analysis model.

[0088] like Figure 4 As shown, the model training module 10 specifically includes,

[0089] Image segmentation unit 11 is used to acquire training sample images with pre-annotated image content states through a deep learning neural network, and to segment the digital color data images of the training set into region images; the deep learning neural network adopts the DeepLabv3+ network; the backbone network of the DeepLabv3+ network is set to GhostNetv2.

[0090] The first feature extraction unit 12 is used to perform row convolution and column convolution on the region image, and then standardize the feature parameters through batch standardization and reduce the linearity of the feature parameters through nonlinear activation to obtain a first feature image with compressed length and width of the region image.

[0091] The second feature extraction unit 13 is used to input the first feature image into the seven channels of the ASPP encoder with dilation rates of 1, 6, 12, 18, 24, 28, and 32, respectively. The first feature image is convolved in these seven channels by convolution kernels that are dilated by 1, 6, 12, 18, 24, 28, and 32 times, respectively, to extract deep features. The extracted seven-channel deep features are compressed through point convolution to obtain the second feature image.

[0092] The feature fusion unit 14 is used to take the first feature image and the second feature image as input features of the decoder, and obtain the review prediction features by fusing redefinition size, point convolution and bilinear interpolation.

[0093] The loss function calculation unit 15 is used to calculate the loss function by minimizing the review prediction features and their corresponding labels. When the loss value meets the convergence requirement, the trained chromatogram analysis model is obtained.

[0094] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned artificial intelligence-based chromatogram analysis device can be referred to the corresponding description in the foregoing method embodiments. For the sake of convenience and brevity, it will not be repeated here.

[0095] The aforementioned AI-based chromatogram analysis device can be implemented as a computer program, which can perform operations such as... Figure 5 It runs on the computer device shown.

[0096] Please see Figure 5 , Figure 5This is a schematic block diagram of a computer device provided in an embodiment of this application. The computer device 500 can be a terminal or a server. The terminal can be an electronic device with communication functions, such as a smartphone, tablet, laptop, desktop computer, personal digital assistant, or wearable device. The server can be a standalone server or a server cluster composed of multiple servers.

[0097] See Figure 5 The computer device 500 includes a processor 502, a memory, and a network interface 505 connected via a system bus 501. The memory may include a non-volatile storage medium 503 and internal memory 504.

[0098] The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 includes program instructions that, when executed, cause the processor 502 to perform an artificial intelligence-based chromatogram analysis method.

[0099] The processor 502 provides computing and control capabilities to support the operation of the entire computer device 500.

[0100] The internal memory 504 provides an environment for the operation of the computer program 5032 in the non-volatile storage medium 503. When the computer program 5032 is executed by the processor 502, the processor 502 can execute an artificial intelligence-based chromatogram analysis method.

[0101] This network interface 505 is used for network communication with other devices. Those skilled in the art will understand that... Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device 500 to which the present application is applied. The specific computer device 500 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0102] The processor 502 is used to run a computer program 5032 stored in the memory to implement the artificial intelligence-based chromatogram analysis method described above.

[0103] It should be understood that, in the embodiments of this application, the processor 502 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0104] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.

[0105] Therefore, the present invention also provides a storage medium. This storage medium can be a computer-readable storage medium. The storage medium stores a computer program, wherein the computer program includes program instructions. When executed by a processor, the program instructions cause the processor to perform the artificial intelligence-based chromatogram analysis method as described above.

[0106] The storage medium can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), magnetic disk, or optical disk.

[0107] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0108] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For example, the division of each unit is merely a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0109] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the device of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this 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.

[0110] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.

[0111] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A chromatogram analysis method based on artificial intelligence, characterized in that, Including the following steps: S10. Train the deep learning neural network to obtain the chromatogram analysis model; S20. Receive the digital chromatographic data image to be reviewed and analyzed through the chromatographic analysis model, and segment the digital chromatographic data image into regional images; S30. The image content status of each region image is reviewed and identified by the chromatogram analysis model to obtain the review prediction region, and the review prediction region is marked with the image content status. The image content status includes the highest peak, leading peak, tailing peak, broad peak, split peak, negative peak, peak with insufficient resolution, baseline noise and baseline drift of the chromatogram. S40. Evaluate and calculate the image content state and the possible probability of the image content state in the predicted area to obtain a first ratio. If the first ratio is greater than the preset first threshold probability ratio, then filter it out and mark it with a specific color. S50. Based on the ongoing chromatographic experiment information, the sample information to be tested, and the analytical information of the chromatographic analysis model, draw up a chromatographic review report and a chromatographic analysis report for reference by the testing personnel.

2. The chromatogram analysis method based on artificial intelligence as described in claim 1, characterized in that: Step S10: Training the deep learning neural network to obtain the chromatogram analysis model specifically includes, S11. Obtain training sample images with pre-annotated image content states through a deep learning neural network, and segment the digital chromatographic data images of the training set into region images; the deep learning neural network adopts the DeepLabv3+ network; the backbone network of the DeepLabv3+ network is set to GhostNetv2. S12. After performing row convolution and column convolution on the region image, the feature parameters are standardized by batch normalization and the linearity of the feature parameters is reduced by nonlinear activation to obtain a first feature image with compressed length and width of the region image. S13. The first feature image is input into the seven channels of the ASPP encoder with dilation rates of 1, 6, 12, 18, 24, 28, and 32, respectively. The first feature image is convolved in these seven channels by convolution kernels that are dilated by 1, 6, 12, 18, 24, 28, and 32 times, respectively, to extract deep features. The extracted seven-channel deep features are then compressed through point convolution to obtain the second feature image. S14. The first feature image and the second feature image are used as input features of the decoder, and the review prediction features are obtained by fusing redefinition size, point convolution and bilinear interpolation. S15. Minimize the loss function by examining the predicted features and their corresponding labels. When the loss value meets the convergence requirement, the trained chromatogram analysis model is obtained.

3. The chromatogram analysis method based on artificial intelligence as described in claim 2, characterized in that: In step S12, the formulas used for row convolution and column convolution of the region image are as follows: in, This represents the numerical values ​​of the output features. Represents the features of the input. Represents the convolution kernel. Indicates the step size.

4. The chromatogram analysis method based on artificial intelligence as described in claim 3, characterized in that: In step S12, the feature parameters are standardized through batch standardization, and the formula is as follows: in, To train the network parameters, gradient estimation is performed on a training dataset of N elements.

5. The chromatogram analysis method based on artificial intelligence as described in claim 4, characterized in that: In step S12, the nonlinear activation reduces the linearity of the feature parameters by using the ReLU function and the softmax function as the minimization loss functions; the minimization loss function is: in, Influenced by the set weights, it represents the average predicted probability of the training dataset; It is the model's predicted probability of the target class, representing the adjustment of the model's attention to positive samples in the training dataset; and The prediction result is in the first... The value at each pixel Let i be the model's predicted value at the i-th pixel. Let be the actual label value at the i-th pixel.

6. The chromatogram analysis method based on artificial intelligence as described in claim 1, characterized in that: The digital chromatographic data image is a two-dimensional gas / liquid chromatographic data image, with the x-axis in units of time and the y-axis in units of signal value.

7. The chromatogram analysis method based on artificial intelligence as described in claim 1, characterized in that: Step S50: Based on the ongoing chromatographic experiment information, the sample information to be tested, and the analytical information of the chromatographic analysis model, a chromatographic review report and a chromatographic analysis report are prepared for reference by the testing personnel. Specifically, these include... S51. Compile the ongoing chromatographic experiment information, the sample information to be tested, and the analytical information of the chromatographic analysis model into a chromatographic review report for reference by the testing personnel. S52. Extract the signal values ​​and time parameters from qualified chromatograms without abnormal peaks, and calculate the chromatographic analysis parameters, including retention time. Peak height Peak area Half peak height Peak width Standard deviation Based on the chromatographic experimental information, sample information, and calculated analytical parameters, a chromatographic analysis report is generated for reference by testing personnel.

8. A chromatogram analysis device based on artificial intelligence, characterized in that, include: The model training module is used to train the deep learning neural network to obtain the chromatogram analysis model; The image segmentation module is used to receive digital chromatographic data images to be reviewed and analyzed through a chromatographic analysis model, and to segment the digital chromatographic data images into regional images. The image content status recognition module is used to review and identify the image content status of each region image through the chromatogram analysis model to obtain the review prediction region, and to mark the review prediction region with the image content status. The image content status includes the highest peak, leading peak, tailing peak, broad peak, split peak, negative peak, peak with insufficient resolution, baseline noise and baseline drift of the chromatogram. The image content status evaluation module is used to evaluate and calculate the image content status of the review prediction area and the possible probability of the image content status to obtain a first ratio. If the first ratio is greater than the preset first threshold probability ratio, it is filtered out and marked with a specific color. The report generation module is used to generate chromatographic review reports and chromatographic analysis reports for reference by testing personnel based on the information of the ongoing chromatographic experiment, the information of the sample to be tested, and the analytical information of the chromatographic analysis model.

9. A computer device, characterized in that: The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the artificial intelligence-based chromatogram analysis method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that: The storage medium stores a computer program that, when executed by a processor, can implement the artificial intelligence-based chromatogram analysis method as described in any one of claims 1 to 7.

Images