Lipid meter detection data processing method and system

Abstract

The invention provides a fat meter detection data processing method and system, and relates to the technical field of data processing, and the method comprises the steps: carrying out the parameter extraction of form correction data, specifically, dividing a data sample into a plurality of continuous data layers along a detection time axis, taking each data layer as a cylinder section, and carrying out the parameter extraction; the equivalent volume of each data layer is calculated layer by layer by calculating the equivalent radius and height of each data layer, and the equivalent volumes of all the data layers are accumulated to generate volume data containing total equivalent volume parameters of the sample; and performing normalization and concentration conversion on the volume data containing the total equivalent volume parameter of the sample to output a final fat meter detection result. According to the method, signal form distortion can be more intelligently corrected, and characteristic parameters which can reflect sample characteristics more essentially and are higher in robustness are extracted, so that the overall performance of fat meter detection is improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a method and system for processing lipid analyzer data. Background Technology

[0002] In clinical testing and life science research, rapid and accurate detection of lipid indicators is crucial. Lipid analyzers, based on optical principles (such as turbidimetry and scattering methods), monitor changes in light signals (such as transmitted light intensity and scattered light intensity) over time during the reaction between the sample and reagents to reflect the concentration of specific components in the sample. This change typically manifests as a response curve with a specific shape; accurate data analysis and feature extraction of this curve are key to obtaining reliable detection results.

[0003] Existing lipid analyzer data processing methods typically include the following steps: acquiring raw photoelectric signals, performing simple filtering and baseline correction, identifying curve feature points (such as peaks and starting points), and finally calculating concentrations by calculating single or combined parameters such as the area under the curve (AUC), peak height, or signal values ​​at fixed time points. However, these traditional methods have some limitations in practice:

[0004] Existing methods typically extract parameters directly from the preprocessed raw curve morphology, neglecting distortions in the response curve morphology caused by individual sample differences, varying reaction kinetics, or non-ideal detection conditions (such as air bubbles or microparticle interference). For example, reaction delays may lead to changes in the slope of the initial segment of the curve, and incomplete reactions may result in abnormal morphology of the descending branch. These morphological variations directly lead to a poorer correlation between parameters (such as area and peak value) extracted based on fixed patterns (such as fixed time window integration or simple geometric features) and the true concentration of the sample, reducing the accuracy and repeatability of the detection. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a lipid analyzer data processing system that can more intelligently correct signal morphological distortion and extract more robust feature parameters that can more essentially reflect the characteristics of the sample, thereby improving the overall performance of lipid analyzer detection.

[0006] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows:

[0007] In a first aspect, a method for processing lipid analyzer data, the method comprising:

[0008] Step 1: Obtain the raw detection data collected by the lipid analyzer;

[0009] Step 2: Perform preprocessing on the raw detection data, including filtering and noise reduction and baseline calibration, to generate preprocessed data;

[0010] Step 3: Construct virtual feature triangles by identifying key data points in the preprocessed data, calculate the interior angle distribution of the virtual feature triangles, and perform angle compensation and geometric correction on the positions of key data points based on the interior angle distribution to generate morphological correction data.

[0011] Step 4: Extract parameters from the morphological correction data. Specifically, the data sample is divided into multiple consecutive data layers along the detection time axis. Each data layer is regarded as a cylindrical cross-section. The equivalent volume of each data layer is calculated by calculating the equivalent radius and height of each data layer. The equivalent volumes of all data layers are accumulated to generate volume data containing the total equivalent volume parameters of the sample.

[0012] Step 5: Normalize and convert the volume data, which includes the total equivalent volume parameter of the sample, to output the final lipid analyzer results.

[0013] Secondly, a lipid analyzer data processing system includes:

[0014] The acquisition module is used to acquire the raw test data collected by the lipid analyzer;

[0015] The preprocessing module is used to perform preprocessing on the raw detection data, including filtering and noise reduction and baseline calibration, to generate preprocessed data;

[0016] The calculation module is used to construct virtual feature triangles by identifying key data points in preprocessed data, calculate the distribution of interior angles of the virtual feature triangles, and perform angle compensation and geometric correction on the positions of key data points based on the distribution of interior angles to generate morphological correction data.

[0017] The extraction module is used to extract parameters from the morphology correction data. Specifically, it divides the data sample into multiple continuous data layers along the detection time axis, treats each data layer as a cylindrical cross-section, calculates the equivalent volume of each data layer by calculating the equivalent radius and height, and accumulates the equivalent volumes of all data layers to generate volume data containing the total equivalent volume parameters of the sample.

[0018] The processing module is used to normalize and convert the volume data, which includes the total equivalent volume parameter of the sample, to output the final lipid analyzer test results.

[0019] Thirdly, a computing device including a memory and a processor;

[0020] The memory stores one or more computer programs, the one or more computer programs including instructions; when the instructions are executed by the processor, the computing device performs the method as described in the first aspect.

[0021] Fourthly, a computer-readable storage medium for storing a computer program for performing the method as described in the first aspect.

[0022] The above-described solution of the present invention has at least the following beneficial effects:

[0023] By employing a two-stage noise reduction strategy that combines sliding window averaging with impulse noise identification and interpolation replacement based on slope changes, and a baseline calibration method based on linear fitting of stable baseline segments before and after the reaction, this invention can more thoroughly filter out random noise and sudden interference, and more accurately eliminate the influence of instrument baseline drift and background signals, providing a cleaner and more stable signal foundation for subsequent analysis.

[0024] By using virtual feature triangles and their interior angle distribution as geometric features to describe the overall shape of the signal, and by analyzing this interior angle distribution to perform targeted angle compensation and position correction on key feature points (starting point, peak point, and ending point), this invention can automatically identify and partially correct curve morphology variations caused by non-ideal response conditions or sample differences. This process is equivalent to aligning the response curves of different samples to a more standard geometric shape, so that the subsequently extracted parameters can better reflect the essential concentration information of the sample, rather than being interfered with by morphological distortion, thereby significantly improving the repeatability of the detection method and its adaptability to different samples.

[0025] This model reconstructs the signal sequence along the time axis into a series of continuous cylindrical cross-sections and characterizes the reaction process by calculating the accumulated equivalent total volume. This model organically combines signal intensity (equivalent radius) with reaction time (layer height), and the extracted total equivalent volume parameter of the sample more closely reflects the physical process of analyte accumulation in the reaction system, generating or consuming scattering / absorbing substances, thus possessing a more intuitive physical meaning. Compared to traditional area parameters, this volume parameter is more sensitive to both overall and local changes in the signal, enabling a more robust and accurate characterization of the total amount of analyte in the sample, thereby achieving higher quantitative detection accuracy and a wider linear range. Attached Figure Description

[0026] Figure 1 This is a flowchart illustrating a lipid analyzer data processing method according to the present invention.

[0027] Figure 2 This is a schematic diagram of a lipid analyzer data processing system according to the present invention.

[0028] Figure 3 A schematic diagram of a computing device Detailed Implementation

[0029] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0030] like Figure 1 As shown, an embodiment of the present invention proposes a method for processing lipid analyzer detection data, including:

[0031] Step 1: Obtain the raw detection data collected by the lipid analyzer;

[0032] Step 2: Perform preprocessing on the raw detection data, including filtering and noise reduction and baseline calibration, to generate preprocessed data;

[0033] Step 3: Construct virtual feature triangles by identifying key data points in the preprocessed data, calculate the interior angle distribution of the virtual feature triangles, and perform angle compensation and geometric correction on the positions of key data points based on the interior angle distribution to generate morphological correction data.

[0034] Step 4: Extract parameters from the morphological correction data. Specifically, the data sample is divided into multiple consecutive data layers along the detection time axis. Each data layer is regarded as a cylindrical cross-section. The equivalent volume of each data layer is calculated by calculating the equivalent radius and height of each data layer. The equivalent volumes of all data layers are accumulated to generate volume data containing the total equivalent volume parameters of the sample.

[0035] Step 5: Normalize and convert the volume data, which includes the total equivalent volume parameter of the sample, to output the final lipid analyzer results.

[0036] In this embodiment, the filtering and noise reduction and baseline calibration in the preprocessing stage can effectively remove invalid information such as environmental interference, instrument noise, and baseline drift in the original data, significantly improving data purity and laying a high-quality data foundation for subsequent processing, reducing error propagation in subsequent steps. The virtual feature triangle construction and angle compensation and geometric correction mechanism can accurately correct the positional deviation of key data points, restore the true shape of the detection data, avoid feature recognition errors caused by data shape distortion, and improve the accuracy and reliability of data shape. Based on the volume calculation method of data layer-cylindrical cross section, the complex sample volume measurement is transformed into a simplified calculation of layer-by-layer accumulation. This ensures the refinement of volume parameter extraction (layer-by-layer processing avoids the coarseness of overall estimation) and achieves accurate acquisition of total equivalent volume parameters through the quantitative calculation of equivalent radius and height, solving the problem of difficulty in accurately measuring the volume of irregular samples. The final normalization and concentration conversion steps can eliminate the systematic deviation caused by different detection conditions (such as the initial state of the sample and differences in the detection environment), realize the standardized output of detection results, improve the comparability of detection results between different batches and different samples, and meet the requirements of detection accuracy and consistency in practical applications.

[0037] In a preferred embodiment of the present invention, step 1, acquiring the raw detection data collected by the lipid analyzer, includes:

[0038] The lipid analyzer's detection process is initiated based on preset detection parameters (adapted to the optical detection modes of clinical lipid detection, such as turbidimetry or scattering). This triggers the lipid analyzer's photoelectric detection module to begin operation: the lipid analyzer emits detection light of a specific wavelength into the sample-reagent mixture in the reaction vessel, simultaneously starting the time axis timing. The photoelectric sensor captures the intensity of transmitted light (or scattered light) through the mixture in real time and converts the light signal into a continuous electrical signal value. The data acquisition module, according to a preset sampling frequency (adapted to the kinetics of lipid reactions to ensure the capture of signal changes throughout the reaction process), binds the electrical signal values ​​with corresponding timestamps, forming a raw data sequence of time-signal values. The acquisition process covers the entire time period from detection initiation (no sample baseline stage), the lipid reaction between the sample and reagent (signal dynamic change stage), to the complete termination of the reaction (final baseline stabilization stage). After acquisition, the raw detection data is temporarily stored in the lipid analyzer's data storage unit for subsequent preprocessing.

[0039] In a preferred embodiment of the present invention, step 2 involves preprocessing the original detection data, including filtering and noise reduction and baseline calibration, to generate preprocessed data, including:

[0040] Step 2.1 involves performing a sliding window averaging filter on the original test data sequence to obtain initial smoothed data. This includes: determining the length of the sliding window (the number of consecutive data points covered by the window, balancing smoothing effect with preserving the true shape of the response curve) based on the sampling frequency of the lipid analyzer and common noise characteristics in clinical lipid testing (such as high-frequency small-amplitude noise); then, using each data point in the original test data sequence as the center, defining a sliding window that includes that data point and several adjacent data points before and after it (for edge data points at the beginning and end of the sequence, the window only covers one side of the usable data points); sliding the window point by point, calculating the average value of the signal values ​​of all data points within each window, and using this average value as the smoothed value of the data point corresponding to the center of the window; and integrating the smoothed values ​​of all data points in chronological order to form an initial smoothed data sequence after removing high-frequency small-amplitude noise, which preserves the overall trend of the response curve while initially reducing the interference of random noise.

[0041] Step 2.2: Calculate the slope change between adjacent data points in the initial smoothed data, and identify local data segments with slope changes exceeding a preset threshold as noise pulses. Replace these noise pulses with interpolated values ​​from their preceding and following data points to obtain noise-filtered data. This includes: traversing the initial smoothed data sequence, sequentially calculating the signal value change amplitude between two adjacent data points (combining the time interval to reflect the slope change trend), and clarifying the slope change characteristics of each pair of adjacent data points; using a preset slope change threshold (set based on the normal slope fluctuation range of the clinical lipid response curve, excluding abnormal mutations caused by bubbles and microparticles) as the judgment criterion, filtering out local data segments with slope changes exceeding the threshold (if 1-3 consecutive pairs of adjacent data points all show slopes exceeding the threshold, they are judged as noise pulse segments); for each identified noise pulse segment, extract its immediately preceding and following normal data points (the last normal point before the start of the pulse segment and the first normal point after the end of the pulse segment), and use linear interpolation to complete the signal values ​​of each data point within the pulse segment, replacing the original abnormal values; after replacing all noise pulse segments, integrate them to obtain the noise-filtered data sequence that eliminates sudden noise pulses.

[0042] Step 2.3: On the detection time axis corresponding to the noise-filtered data, identify and locate the initial stable segment and the final stable segment corresponding to the absence of a sample baseline. This includes: dividing the detection time axis corresponding to the noise-filtered data into three core time periods: the initial period after detection starts when the sample and reagent do not react, the dynamic period when the sample and reagent undergo a lipid reaction, and the final period after the reaction has completely terminated; for all data points in the initial period, calculate the coefficient of variation (reflecting the degree of data dispersion), and determine the continuous data segments with a coefficient of variation lower than the preset stability threshold as the initial stable segment (this segment is the stage without a sample baseline, and the light signal does not fluctuate significantly); similarly, for all data points in the final period, calculate the coefficient of variation of the signal values, and select the continuous data segments with a coefficient of variation lower than the stability threshold as the final stable segment (this segment is the stage where the reaction has completely terminated and the light signal has returned to stability); accurately locate the start and end times of the two stable segments on the time axis, as well as the positions of all corresponding data points.

[0043] Step 2.4: Based on the mean values ​​of data points within the initial and final stable segments, construct a baseline fitting line. Subtract the corresponding time point values ​​on the baseline fitting line from the values ​​of all data points in the noise-filtered data to generate preprocessed data after baseline calibration. This includes: calculating the mean signal values ​​of all data points within the initial stable segment and the mean signal values ​​of all data points within the final stable segment; constructing a baseline fitting line connecting the mean of the initial stable segment minus the corresponding time midpoint with the mean of the final stable segment minus the corresponding time midpoint, using detection time as the horizontal axis and signal value as the vertical axis (this line...). During the baseline measurement process using a line-adaptive lipid analyzer, a slight linear drift of the baseline over time is observed (instead of simply using a fixed baseline). Subsequently, each data point in the noise-filtered data is traversed, and the corresponding signal value on the baseline fitting line at that time point is determined based on its corresponding timestamp. The original signal value of each data point is subtracted from the baseline fitting line signal value at its corresponding time point to eliminate the systematic error caused by baseline drift. Finally, all calibrated data points are integrated in chronological order to generate preprocessed data that completes baseline calibration. This data can truly reflect the signal changes in the lipid reaction process of the sample and eliminate baseline interference.

[0044] In a preferred embodiment of the present invention, step 3, which involves identifying key data points in the preprocessed data to construct a virtual feature triangle and calculating the interior angle distribution of the virtual feature triangle, includes:

[0045] Step 3.1: On the response curve formed by the preprocessed data, identify and locate the signal start point, peak point, and signal end point as three key data points. The signal start point is the inflection point where the curve first continuously deviates from the baseline, the peak point is the global maximum value of the curve, and the signal end point is the inflection point where the curve finally falls back and stabilizes at the baseline. This includes: Based on the preprocessed data generated in Step 2, construct a response curve with detection time as the horizontal axis and signal amplitude as the vertical axis, and clarify the overall trend of the curve (from the baseline, rising to the peak, and then falling back to the baseline); For the identification of the signal start point: Starting from the initial stage of the response curve, traverse the data points along the time axis in a positive direction, and set a reasonable deviation threshold based on the signal fluctuation range after baseline calibration (this threshold is higher than the normal fluctuation range of the baseline to avoid misjudging random noise as the signal start). When the signal amplitude of three or more consecutive data points exceeds the deviation threshold, and subsequent data points continue to exceed the threshold, the first data point of the consecutive data segment is determined as the first continuous deviation. The inflection point of the baseline is the signal starting point. For peak point identification: traverse all data points of the entire response curve, record the signal amplitude of each data point, and filter out the data point with the largest signal amplitude. Further verify the trend of data changes before and after this data point (signal amplitude of preceding data points continuously increases, signal amplitude of subsequent data points continuously decreases), eliminate false peaks caused by local fluctuations, and confirm that this data point is the global maximum value point of the curve, i.e., the peak point. For signal endpoint identification: starting from the later stage of the response curve, traverse the data points backward along the time axis, set a regression threshold with reference to the stable fluctuation range of the baseline. When the signal amplitude of three or more consecutive data points falls back to within this regression threshold, and subsequent (forward time axis) data points remain stable within this range, the first data point of this consecutive data segment is determined as the inflection point where the signal finally falls back and stabilizes at the baseline, i.e., the signal endpoint. Finally, record the timestamps and signal amplitudes corresponding to the signal starting point, peak point, and signal endpoint to complete the precise location of the three key data points.

[0046] By using a recognition logic that combines continuous data point verification with trend judgment, we can effectively avoid misjudgment of key data points caused by curve shape distortion (such as start delay, local fluctuations, and incomplete fallback). This ensures that the three core feature points can accurately reflect the start, peak, and end of the response curve, providing reliable basic data for subsequent shape analysis. It also solves the problem of feature point positioning deviation caused by traditional methods that rely solely on a single threshold or simple extreme value judgment.

[0047] Step 3.2: Using the signal start point, peak point, and signal end point as vertices, construct a virtual feature triangle on a two-dimensional plane formed by the detection time axis and the signal amplitude axis. This includes: using the detection time as the horizontal axis (X-axis) and the signal amplitude as the vertical axis (Y-axis), convert the three key data points located in Step 3.1 into two-dimensional coordinate values. The signal start point corresponds to coordinates (T1, S1), where T1 is the start point timestamp and S1 is the start point signal amplitude; the peak point corresponds to coordinates (T2, S2), where T2 is the peak point timestamp and S2 is the peak point signal amplitude; and the signal end point corresponds to coordinates (T3, S3), where T3 is the end point timestamp and S3 is the signal amplitude. The endpoint signal amplitude is then determined. Subsequently, within this two-dimensional plane, three key data points are sequentially connected by line segments: the signal start point (T1, S1) and the peak point (T2, S2) are connected by a line segment to form the first side of a triangle; the peak point (T2, S2) and the signal endpoint (T3, S3) are connected by a line segment to form the second side of the triangle; and the signal endpoint (T3, S3) and the signal start point (T1, S1) are connected by a line segment to form the third side of the triangle. The three line segments are mutually closed, forming a unique virtual feature triangle in the two-dimensional plane. The shape of this triangle is determined by the relative positional relationship of the three key data points, intuitively reflecting the core morphological characteristics of the response curve.

[0048] By transforming the abstract response curve shape into a concrete geometric figure (virtual feature triangle), the curve shape is simplified and quantified. The overall shape differences of the curve (such as the combined changes of the initial slope, peak height, and fall speed) that are difficult to describe by traditional methods are transformed into the side length and angle relationship of the triangle. This provides an intuitive geometric carrier for the subsequent quantitative analysis of shape distortion and solves the limitation of traditional methods in being unable to fully capture the overall shape characteristics of the curve.

[0049] Step 3.3: Calculate the angle values ​​of the three interior angles of the virtual feature triangle, obtaining the angle distribution of the first, second, and third interior angles. The first interior angle has the peak point as its vertex, the second interior angle has the signal start point as its vertex, and the third interior angle has the signal end point as its vertex. This includes: defining the three interior angles and their corresponding vertices: the first interior angle has the peak point (T2, S2) as its vertex and is formed by the angle between the first side (start point - peak point) and the second side (peak point - end point); the second interior angle has the signal start point (T1, S1) as its vertex and is formed by the angle between the third side (end point - start point) and the first side (start point - peak point); the third interior angle has the signal end point (T3, S3) as its vertex and is formed by the angle between the second side (peak point - end point) and the third side (end point - start point); subsequently, based on the relative positional relationship of the three points connected in the two-dimensional plane and the line... The analysis focuses on the direction of line segments and the tilt trend and length ratio of each side. By determining the upward / downward direction of the line segments and the turning relationship between adjacent line segments (such as the upward slope of the first side from the starting point to the peak point, and the downward slope of the second side from the peak point to the end point), and combining the relative differences (time difference and amplitude difference) of the coordinates of the two endpoints of the line segments, the angle relationship between each side is determined. According to the geometric rule of larger side to larger angle and smaller side to smaller angle, and combined with the length ratio of the line segments (indirectly determined through the relative difference of coordinates), the angle values ​​of the first, second, and third interior angles are determined in sequence to ensure that the sum of the three angle values ​​conforms to the basic geometric rule of the sum of the interior angles of a triangle. Finally, the angle values ​​of the three interior angles are arranged in the order of first interior angle - second interior angle - third interior angle to form a complete interior angle distribution. This distribution directly quantifies the shape of the virtual feature triangle, thereby reflecting the morphological distortion characteristics of the response curve.

[0050] By quantifying the shape of the virtual feature triangle through the distribution of interior angles, the shape distortion of the response curve (such as the increase of the second interior angle due to the initial delay and the decrease of the third interior angle due to incomplete fallback) is transformed into a quantifiable angle index, providing a clear quantitative basis for subsequent angle compensation and geometric correction. This solves the problem that traditional methods cannot quantify the degree of curve shape distortion, making shape correction more targeted and accurate.

[0051] In a preferred embodiment of the present invention, the positions of key data points are compensated for and geometrically corrected based on the distribution of interior angles to generate morphological correction data, including:

[0052] Step 3.4: Based on the interior angle distribution, determine the required angle compensation amount for the amplitude coordinates of the signal start point, peak point, and signal end point according to the preset angle-compensation mapping rule. This includes: constructing the preset angle-compensation mapping rule: This rule is established based on the detection data of a large number of standard lipid samples (without morphological distortion and ideal reaction conditions). First, the standard angle range of the three interior angles of the virtual feature triangle corresponding to the ideal response curve is statistically determined through experiments. Then, for each interior angle, the difference interval of the angle deviating from the standard range is divided, and the corresponding amplitude compensation amount is matched for each interval (the larger the deviation difference, the larger the compensation amount, to ensure that the compensation strength is adapted to the degree of morphological distortion). Subsequently, the actual interior angle distribution obtained in step 3.3 is compared with the preset standard angle range one by one, and the deviation difference between the actual angle and the standard angle for the first interior angle, the second interior angle, and the third interior angle is calculated. Then, according to the above mapping rules, the corresponding angle compensation amount is determined according to the interval to which the deviation difference of each interior angle belongs. Among them, the deviation difference of the second interior angle corresponds to the amplitude compensation amount at the starting point of the signal, the deviation difference of the first interior angle corresponds to the amplitude compensation amount at the peak point, and the deviation difference of the third interior angle corresponds to the amplitude compensation amount at the ending point of the signal, thus completing the accurate matching of the amplitude compensation amount of each key data point.

[0053] Beneficial effects: By using preset mapping rules, the quantitative index of morphological distortion caused by the deviation of the inner angle is transformed into an operable amplitude compensation amount, avoiding the blindness of compensation operation; at the same time, the compensation amount is strongly correlated with the degree of angle deviation, ensuring that the compensation strength matches the actual distortion situation.

[0054] Step 3.5: Based on the compensation amounts for each angle, compensate the amplitude coordinates of the signal start point, peak point, and signal end point respectively to obtain three key data points after compensation, including:

[0055] Step 3.51: Determine the compensation direction of the angle compensation amounts of the first, second, and third interior angles for the corresponding vertex amplitude coordinates. Specifically, the angle compensation amount of the first interior angle determines the upward or downward compensation direction of the peak point amplitude coordinates; the angle compensation amount of the second interior angle determines the upward or downward compensation direction of the signal starting point amplitude coordinates; and the angle compensation amount of the third interior angle determines the upward or downward compensation direction of the signal ending point amplitude coordinates. This includes: First, based on the ideal morphological characteristics of the standard sample, establish the correlation logic between the interior angle deviation and the compensation direction: For the first interior angle (peak point as vertex), if the actual angle is greater than the standard angle, it indicates that the peak point amplitude may be too low (the curve peak shape is flat), and the compensation direction is set upward; if the actual angle is less than the standard angle, it indicates that the peak point amplitude may be too high (the curve peak is abnormally convex), and the compensation direction is set downward. For the second interior angle (with the signal starting point as the vertex), if the actual angle is greater than the standard angle, it indicates a signal start delay (the initial slope is too small), and the starting point amplitude is too low; the compensation direction is set upwards. If the actual angle is less than the standard angle, it indicates an abnormal signal jump in the initial segment, and the starting point amplitude is too high; the compensation direction is set downwards. For the third interior angle (with the signal ending point as the vertex), if the actual angle is greater than the standard angle, it indicates incomplete signal fallback (the ending point has not fully returned to the baseline), and the ending point amplitude is too high; the compensation direction is set downwards. If the actual angle is less than the standard angle, it indicates that the signal fallback is too fast, and the ending point amplitude is too low; the compensation direction is set upwards. Through the above logic, the vertex amplitude compensation direction corresponding to the compensation amount of each interior angle is clearly defined.

[0056] Step 3.52: Based on the compensation amount at each angle and its corresponding compensation direction, increase or decrease the amplitude coordinate values ​​of the signal start point, peak point, and signal end point respectively to obtain three key data points whose amplitude coordinates have been adjusted, while keeping their time coordinates unchanged. This includes: while keeping the time coordinates of the signal start point, peak point, and signal end point unchanged (ensuring the integrity of the detection timing and avoiding timing chaos), adjusting the amplitude coordinates of the three key data points according to the compensation direction and compensation amount: if the compensation direction is upward, increase the corresponding compensation amount based on the original amplitude coordinate values; if the compensation direction is downward, decrease the corresponding compensation amount based on the original amplitude coordinate values, thus completing the accurate correction of the amplitude coordinates of each key data point.

[0057] Step 3.53: The three key data points whose amplitude coordinates have been adjusted but whose time coordinates remain unchanged are respectively designated as the compensated signal start point, compensated peak point, and compensated signal end point, and are used as morphological feature control points. This includes: naming the three key data points with adjusted amplitudes and unchanged time coordinates as the compensated signal start point, compensated peak point, and compensated signal end point, respectively. Since these three points have eliminated the positional deviation caused by morphological distortion and can reflect the core morphological features of the ideal response curve, they are used as morphological feature control points for subsequent curve reconstruction, providing a benchmark for morphological repair of the full time series signal.

[0058] By clearly defining the correspondence between the compensation direction and the key vertices, targeted corrections for different types of morphological distortions (start delay, peak anomaly, and incomplete fallback) are achieved. Keeping the time coordinate unchanged avoids the loss of temporal information, and marking them as feature control points provides a reliable benchmark for subsequent overall curve reconstruction, thus solving the problem that traditional methods cannot accurately correct the morphological deviations of key feature points.

[0059] Step 3.6: Using the three compensated key data points as morphological feature control points, an interpolation algorithm is used on the detection time axis to reconstruct the signal amplitude of the entire detection time series and generate morphological correction data. This includes: clearly defining the reconstruction range to cover the entire lipid analyzer detection time series (from detection start to reaction termination) to ensure the integrity of the reconstructed data; subsequently, using the three compensated key data points (morphological feature control points) obtained in step 3.53 as the core, and combining the original detection time axis scale in the preprocessed data (the time coordinate of each time point remains unchanged), an interpolation algorithm adapted to the continuity characteristics of the lipid response curve is selected (ensuring that the reconstructed curve has no abrupt changes and a smooth morphology). Within the time interval between two adjacent morphological feature control points, based on the time interval between the two control points and the compensated amplitude difference, all missing signal amplitude data within this interval are smoothly filled in. For example, between the compensated signal start point and the compensated signal peak point, the signal amplitude is filled in point by point in chronological order to ensure that the signal change trend after filling is consistent with the rising segment characteristics of the ideal response; similarly, the signal amplitude is filled in between the compensated peak point and the compensated signal end point to restore the falling segment characteristics of the ideal response. Finally, all signal amplitudes (including compensated feature control points and interpolated intermediate points) within the entire detection time series are integrated in chronological order to generate morphological correction data that eliminates morphological distortion. This data can truly reflect the ideal signal change process of the sample's lipid response.

[0060] By reconstructing the entire time series signal using the corrected feature control points as a benchmark through interpolation algorithms, a complete closed loop from key feature point correction to overall curve morphology restoration was achieved, completely eliminating curve morphology distortion caused by individual sample differences, abnormal reaction dynamics, or detection interference. The reconstructed morphology-corrected data retains the true reaction signal characteristics, providing a high-quality data foundation for the accurate extraction of volume parameters in subsequent steps. This fundamentally solves the core problem of inaccurate detection results and poor repeatability caused by morphology distortion in traditional methods.

[0061] In a preferred embodiment of the present invention, step 4, parameter extraction of the morphological correction data, specifically involves dividing the data sample into multiple consecutive data layers along the detection time axis, treating each data layer as a cylindrical cross-section, calculating the equivalent volume layer by layer by calculating the equivalent radius and height of each data layer, and accumulating the equivalent volumes of all data layers to generate volume data containing the total equivalent volume parameters of the sample, including:

[0062] Step 4.1: Along the detection time axis corresponding to the morphology correction data, divide the entire data sample into several continuous data layers with equal time intervals. This includes: extracting the complete detection time axis information corresponding to the morphology correction data, clarifying the start time, end time, and total detection duration, ensuring that the division covers the entire cycle of the lipid reaction (from reaction initiation to complete termination, with no time segments omitted). Subsequently, a reasonable equal time interval duration is set based on the kinetic characteristics of the lipid reaction: the interval duration needs to consider two core requirements: firstly, it must be small enough to retain the detailed changes in the signal during the reaction process (avoiding the loss of key reaction stage features due to excessively large intervals); secondly, it must avoid excessively small intervals leading to redundant computation (balancing accuracy and efficiency). Based on the set interval duration, starting from the detection start time, several continuous time intervals are sequentially divided along the positive time axis, with each time interval being a data layer. Simultaneously, ensure seamless connection and no overlap between adjacent data layers, and that the end time of the last data layer is completely consistent with the detection end time. Finally, clarify the time range corresponding to each data layer and extract all morphology correction data points contained within each layer, completing the data layer division.

[0063] Beneficial effects: By dividing the data layer into equal time intervals, the continuous time series signal is discretized into standardized processing units, which not only ensures the integrity of the signal reflecting the whole process, but also avoids the fragmentation problem when extracting parameters in the traditional fixed time window; the reasonable interval length setting achieves a balance between accuracy and efficiency.

[0064] Step 4.2: Each data layer is considered as the cross-section of a cylinder. The equivalent radius of a continuous data layer is proportional to the average amplitude of all data points within that layer. The height of the continuous data layer is the duration of the equal time interval. This includes clarifying the core significance of morphological correction data—its signal amplitude has eliminated morphological distortion and truly reflects the dynamic changes in the effective reaction amount during the lipid reaction process. Therefore, the magnitude of the signal amplitude can be directly related to the activity level of the reaction. Based on this, a geometric mapping logic between the data layer and the cylinder cross-section is established: each data layer with equal time intervals is considered as the cross-section of a thin cylinder. The equivalent radius of the continuous data layer is proportionally related to the average signal amplitude of all data points within that layer (the larger the average amplitude, the more intense the reaction at that time stage, and the larger the corresponding equivalent radius). The height of the continuous data layer directly follows the equal time interval duration set in Step 4.1 (the span of each data layer in the time dimension is fixed, corresponding to a fixed cylinder height). Meanwhile, through a large number of standard lipid samples in the early stage of detection experiments, the rationality of the mapping logic was verified, ensuring that the geometric model can accurately represent the correspondence between signal amplitude and lipid response amount.

[0065] By transforming abstract signal data into a concrete cylindrical geometric model, the problem of traditional methods being unable to effectively correlate signal characteristics with total response is solved. By setting the amplitude mean to be proportional to the equivalent radius, the quantitative transformation of signal characteristics into geometric parameters is realized, providing an intuitive and operable carrier for subsequent volume calculations.

[0066] Step 4.3: For each data layer obtained from the division, calculate its equivalent radius and height according to the duration of the equal time intervals. This includes: Calculating the height of each data layer: Directly extract the duration of the equal time intervals determined in Step 4.1 as the height of the equivalent cylinder of the corresponding data layer. Since all data layers are divided by equal time intervals, the height of all data layers is consistent and does not need to be calculated repeatedly; only the accuracy of the interval duration needs to be confirmed. Then, calculate the equivalent radius of each data layer: For each data layer, first traverse all the morphological correction data points it contains, extract the signal amplitude value of each data point, and then calculate the average value of these amplitude values ​​(to eliminate small fluctuations in individual data points within the layer and ensure representativeness); then call the amplitude mean - equivalent radius proportionality coefficient obtained from the previous calibration (this coefficient is calibrated through the correspondence between the signal amplitude of the standard concentration sample and the known reaction amount), and according to the rule that the equivalent radius is proportional to the amplitude mean, convert the calculated amplitude mean within the layer into the equivalent radius of the data layer, completing the calculation of the core geometric parameters of each data layer.

[0067] The height is directly adopted from the interval duration, which simplifies the calculation process while ensuring the stability of the time dimension parameters; the equivalent radius is calculated based on the mean amplitude within the layer, which effectively avoids the interference of fluctuations in a single data point. Combined with the calibrated scaling factor, it ensures the accurate correspondence between the geometric parameters and the actual response.

[0068] Step 4.4: Based on the equivalent radius and height of each data layer, calculate its equivalent volume layer by layer using the cylinder volume formula. This includes ensuring that the cylindrical geometric model corresponding to each data layer has complete core parameters (the equivalent radius and height calculated in Step 4.3). Then, based on the basic calculation logic of cylinder volume, using the equivalent radius of each data layer as a foundation, first determine its corresponding equivalent base area (reflecting the lateral distribution of the reaction quantity at that time stage). Then, combine the equivalent base area with the height of the data layer (reflecting the time span) to calculate the equivalent volume corresponding to a single data layer. During the calculation process, calculations are performed independently for each data layer to avoid parameter interference between different data layers. Simultaneously, the equivalent volume results for each data layer are recorded, and its corresponding time interval is labeled, forming a correlation table of data layer-time interval-equivalent volume for easy subsequent traceability and accumulation operations.

[0069] By calculating layer by layer using cylindrical volume logic, the reaction amount at each time stage is transformed into a quantifiable volume parameter. Compared with the overall estimation of the area under the curve (AUC) in traditional methods, it can more accurately capture the difference in reaction intensity at different reaction stages (such as the rapid rise, peak, and slow fall). The volume of each data layer is calculated and recorded independently, providing clear basic data for subsequent accumulation and avoiding the loss of details caused by overall calculation.

[0070] Step 4.5 involves summing the equivalent volumes of all data layers to generate volume data containing the total equivalent volume parameter of the sample. This includes: organizing the equivalent volume results of all data layers obtained in Step 4.4 and sorting them according to the time sequence corresponding to the data layers (from the first data layer corresponding to the start time of detection to the last data layer corresponding to the end time of detection), ensuring that the summation order is consistent with the time progress of the lipid reaction. Subsequently, the equivalent volume values ​​of each sorted data layer are extracted sequentially and summed. During the summation process, the integrity of the data is verified in real time to avoid omitting or repeatedly summing the volume of a certain data layer (consistency can be ensured by comparing the number of summed data layers with the total number of data layers divided in Step 4.1). After the summation is completed, the obtained sum is the total equivalent volume parameter of the sample corresponding to the entire lipid reaction process. Finally, this total equivalent volume parameter is integrated with the basic information of the sample detection (such as sample number, detection time, lipid analyzer model, etc.) to generate complete volume data, providing core parameter support for subsequent normalization and concentration conversion.

[0071] By accumulating the equivalent volumes of all data layers in chronological order, the signal characteristics of the entire lipid reaction process are fully covered. Compared with traditional methods that extract single parameters such as peak values ​​or fixed time point signal values, the total equivalent volume parameter can more comprehensively reflect the true total amount of lipid components in the sample. The volume data generated by combining the basic information of the sample provides a complete data carrier for subsequent standardization processing, effectively improving the accuracy and reliability of subsequent concentration conversion.

[0072] In a preferred embodiment of the present invention, step 5, which involves normalizing and converting the volume data including the total equivalent volume parameter of the sample to output the final lipid analyzer detection result, includes:

[0073] Step 5.1: Obtain the pre-established total equivalent volume parameter of the standard through standard sample testing. Calculate the ratio between the total equivalent volume parameter of the sample and the total equivalent volume parameter of the standard to obtain the normalized volume ratio. This includes: retrieving the pre-established total equivalent volume parameter library: This library is a set of total equivalent volume parameters of the standard samples after a complete testing process (including preprocessing, morphology correction, and volume extraction in steps 1-4) of a series of known concentrations of standard samples (such as clinically commonly used lipid calibrators covering low, medium, and high concentration gradients) using the same lipid analyzer and the same batch of reagents. Each standard parameter is labeled with the corresponding testing conditions (ensuring consistency with the testing conditions of the sample to be tested, such as testing temperature, reaction time, and reagent volume). Subsequently, based on the type of sample to be tested (such as cholesterol or triglycerides) and the testing item, match the corresponding total equivalent volume parameter of the standard from the parameter library (if there are multiple concentration standard samples, the parameter of the baseline concentration standard can be selected as the normalization baseline, or the baseline point parameter of the concentration-volume fitting curve can be used). Finally, the total equivalent volume parameter of the sample generated in step 4.5 is extracted, and its ratio is calculated with the matched standard total equivalent volume parameter. The result is the normalized volume ratio, thereby eliminating the systematic bias caused by different test batches and minor fluctuations in the instrument.

[0074] By using standard parameters as a benchmark for normalization, systematic interferences such as differences in lipid analyzer equipment, reagent batch fluctuations, and changes in the detection environment are effectively offset. This solves the problems of lack of unified reference standards for sample parameters and incomparability of test results from different batches in traditional methods. The matching design of concentration gradient standards provides a reliable ratio basis for subsequent accurate concentration conversion.

[0075] Step 5.2: Based on the pre-calibrated concentration-volume conversion coefficients corresponding to the sample type, the normalized volume ratio is converted into a preliminary concentration value of the sample. This includes: retrieving a pre-calibrated concentration-volume conversion coefficient library. This library is established for different types of samples (such as serum total cholesterol, serum triglycerides, low-density lipoprotein, etc.). The calibration process involves selecting multiple sets of standards covering common clinical concentration ranges, processing them sequentially to obtain the total equivalent volume parameter of each standard, calculating its normalized volume ratio with the benchmark standard, and then establishing a correspondence by using the known concentration of the standard as the vertical axis and the normalized volume ratio as the horizontal axis through linear or nonlinear fitting. From this, the concentration-volume conversion coefficient specific to each sample type is extracted (ensuring that the coefficient strictly matches the sample type and avoiding cross-use). Subsequently, the specific type of the current sample to be tested is confirmed, and the corresponding conversion coefficient is accurately matched from the coefficient library. Finally, the normalized volume ratio obtained in step 5.1 is combined with this conversion coefficient to complete the conversion from volume ratio to concentration value, obtaining the preliminary concentration value of the sample to be tested.

[0076] The conversion factor design adapts to the different reaction characteristics of different lipid samples, avoiding conversion deviations caused by universal factors. Based on the coefficients fitted and calibrated by multiple concentration standards, it ensures the linear or nonlinear accuracy of concentration conversion, transforming abstract volume ratios into intuitive concentration values, and solving the core problems of poor correlation between parameters and concentration and low conversion accuracy in traditional methods.

[0077] Step 5.3 involves standardizing units and rounding off significant figures for the initial concentration values ​​to form the concentration detection result, which is then used as the result of the lipid analyzer. This includes converting the initial concentration values ​​obtained in Step 5.2 to the standard units of the corresponding test items according to industry standards for clinical testing or life science research (such as the International Federation for Clinical Chemistry (IFCC) standards and domestic clinical testing operating procedures). (For example, cholesterol and triglyceride tests commonly use mmol / L or mg / dL, which needs to be standardized according to the needs of the testing scenario. For example, mmol / L is preferred for clinical testing, while mg / dL can be used in some research scenarios.) This ensures that the units conform to industry-standard specifications, facilitating subsequent interpretation and application of the results. Then, significant figures are rounded off: considering the detection accuracy of the lipid analyzer (such as the limit of detection and repeatability error) and the accuracy level of the standard calibration, the significant figures of the concentration values ​​are determined (for example, when the lipid analyzer detection accuracy is 0.01 mmol / L, the significant figures are retained to two decimal places). The initial concentration values ​​are rounded off according to numerical rounding rules (such as rounding to the nearest whole number), eliminating meaningless redundant figures and avoiding false accuracy. Finally, the rounded concentration test results are integrated with the basic sample information (such as sample number, test time, lipid analyzer model, reagent batch, and test item name) to generate structured lipid analyzer test results, which can be output, printed, or transmitted to the Laboratory Information Management System (LIMS) for archiving through the lipid analyzer's display interface.

[0078] The uniformity of the testing institutions ensures the universality and industry compatibility of the test results, avoiding clinical misjudgments or misuse of research data caused by institutional confusion; the rounding of significant digits ensures the rigor of the results, conforming to the actual detection accuracy of the lipid analyzer and eliminating the misleading effects of false accuracy; the structured output and data transmission adapt to the actual application needs of clinical and research, improving the practicality and traceability of the test results.

[0079] like Figure 2 As shown, a lipid analyzer data processing system includes:

[0080] The acquisition module is used to acquire the raw test data collected by the lipid analyzer;

[0081] The preprocessing module is used to perform preprocessing on the raw detection data, including filtering and noise reduction and baseline calibration, to generate preprocessed data;

[0082] The calculation module is used to construct virtual feature triangles by identifying key data points in preprocessed data, calculate the distribution of interior angles of the virtual feature triangles, and perform angle compensation and geometric correction on the positions of key data points based on the distribution of interior angles to generate morphological correction data.

[0083] The extraction module is used to extract parameters from the morphology correction data. Specifically, it divides the data sample into multiple continuous data layers along the detection time axis, treats each data layer as a cylindrical cross-section, calculates the equivalent volume of each data layer by calculating the equivalent radius and height, and accumulates the equivalent volumes of all data layers to generate volume data containing the total equivalent volume parameters of the sample.

[0084] The processing module is used to normalize and convert the volume data, which includes the total equivalent volume parameter of the sample, to output the final lipid analyzer test results.

[0085] The prediction system according to embodiments of the present invention can correspond to performing the methods described in the embodiments of the present invention, and the above and other operations and / or functions of each module of the prediction system are respectively for implementing Figure 1 The corresponding process of the method in the illustrated embodiment will not be described in detail here for the sake of brevity.

[0086] This application also provides a computing device. This computing device can utilize a server.

[0087] like Figure 3 As shown in the figure, this is a schematic diagram of a computing device provided in an embodiment of this application. The computing device 700 includes a bus 701, a processor 702, a communication interface 703, and a memory 704. The processor 702, the memory 704, and the communication interface 703 communicate with each other via the bus 701.

[0088] The 701 bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 3 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0089] The processor 702 can be any one or more of the following processors: central processing unit (CPU), graphics processing unit (GPU), microprocessor (MP), or digital signal processor (DSP).

[0090] Communication interface 703 is used for external communication. Memory 704 may include volatile memory, such as random access memory (RAM). Memory 704 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard disk drive (HDD), or solid-state drive (SSD). Memory 704 stores executable code, which processor 702 executes to perform the aforementioned industrial equipment remaining life prediction method.

[0091] Specifically, in implementing the industrial equipment remaining life prediction system shown in the above embodiments, and where each module or unit of the industrial equipment remaining life prediction system described in the above embodiments is implemented by software, the software or program code required to execute the functions of each module / unit in the industrial equipment remaining life prediction system described in the above embodiments can be partially or entirely stored in the memory 704. The processor 702 executes the program code corresponding to each unit stored in the memory 704 to execute the aforementioned industrial equipment remaining life prediction method.

[0092] This application also provides a computer-readable storage medium. The computer-readable storage medium can be any available medium that a computing device can store, or a data storage device such as a data center containing one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive). The computer-readable storage medium includes instructions that instruct a computing device to execute the aforementioned industrial equipment remaining life prediction method.

[0093] This application also provides a computer program product comprising one or more computer instructions. When the computer instructions are loaded and executed on a computing device, all or part of the processes or functions described in this application are generated.

[0094] The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, or data center to another website, computer, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means.

[0095] When the computer program product is executed by a computer, the computer performs any of the aforementioned methods for predicting the remaining useful life of industrial equipment. The computer program product can be a software installation package; when any of the aforementioned methods for predicting the remaining useful life of industrial equipment is required, the computer program product can be downloaded and executed on the computer.

[0096] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for processing lipid analyzer data, characterized in that, The method includes: Obtain the raw test data collected by the lipid analyzer; The raw detection data is preprocessed, including filtering and noise reduction and baseline calibration, to generate preprocessed data; By identifying key data points in the preprocessed data to construct virtual feature triangles, calculating the interior angle distribution of the virtual feature triangles, and performing angle compensation and geometric correction on the positions of key data points based on the interior angle distribution, morphological correction data is generated. The morphological correction data is extracted by dividing the data sample into multiple consecutive data layers along the detection time axis. Each data layer is regarded as a cylindrical cross-section. The equivalent volume of each data layer is calculated by calculating the equivalent radius and height of each data layer. The equivalent volumes of all data layers are accumulated to generate volume data containing the total equivalent volume parameters of the sample. The volume data, including the total equivalent volume parameter of the sample, is normalized and the concentration is converted to output the final lipid analyzer results.

2. The method for processing lipid analyzer data according to claim 1, characterized in that, The raw detection data undergoes preprocessing, including filtering and noise reduction, and baseline calibration, to generate preprocessed data, including: The original detection data sequence is subjected to sliding window averaging filtering to obtain initial smoothed data; Calculate the slope change between adjacent data points in the initial smoothed data, and identify local data segments whose slope changes exceed a preset threshold as noise pulses. Replace them with the interpolation of the adjacent data points before and after them to obtain noise-filtered data. On the detection time axis corresponding to the noise-filtered data, identify and locate the initial stable segment and the final stable segment corresponding to the absence of sample baseline; Based on the mean values ​​of data points within the initial and final stable segments, a baseline fitting line is constructed. The values ​​of all data points in the noise-filtered data are then subtracted from the values ​​of the corresponding time points on the baseline fitting line to generate preprocessed data after baseline calibration.

3. The method for processing lipid analyzer data according to claim 2, characterized in that, By identifying key data points in the preprocessed data to construct virtual feature triangles, the distribution of the interior angles of the virtual feature triangles is calculated, including: On the response curve formed by the preprocessed data, the signal start point, peak point and signal end point are identified and located as three key data points. The signal start point is the inflection point where the curve first deviates continuously from the baseline, the peak point is the global maximum value point of the curve, and the signal end point is the inflection point where the curve finally falls back and stabilizes at the baseline. A virtual feature triangle is constructed on a two-dimensional plane formed by the detection time axis and the signal amplitude axis, with the signal start point, peak point and signal end point as vertices; Calculate the angle values ​​of the three interior angles of the virtual feature triangle to obtain the interior angle distribution composed of the first interior angle, the second interior angle, and the third interior angle, where the first interior angle is based on the peak point, the second interior angle is based on the signal start point, and the third interior angle is based on the signal end point.

4. The method for processing lipid analyzer data according to claim 3, characterized in that, Based on the distribution of interior angles, the positions of key data points are compensated for by angles and geometrically corrected to generate morphological correction data, including: Based on the interior angle distribution, the angle compensation amount to be applied to the amplitude coordinates of the signal start point, peak point and signal end point is determined according to the preset angle-compensation mapping rule. Based on the compensation amount at each angle, the amplitude coordinates of the signal start point, peak point and signal end point are compensated respectively to obtain three key data points after compensation; Using the three compensated key data points as morphological feature control points, the signal amplitude of the entire detection time series is reconstructed using an interpolation algorithm on the detection time axis to generate morphological correction data.

5. The method for processing lipid analyzer data according to claim 4, characterized in that, Parameter extraction is performed on the morphological correction data. Specifically, the data samples are divided into multiple continuous data layers along the detection time axis. Each data layer is considered as a cylindrical cross-section. The equivalent volume of each data layer is calculated by calculating its equivalent radius and height, and the equivalent volumes of all data layers are accumulated to generate volume data containing the total equivalent volume parameters of the samples, including: Along the detection time axis corresponding to the morphological correction data, the entire data sample is divided into several continuous data layers with equal time intervals. Each data layer is considered as the cross-section of a cylinder, where the equivalent radius of a continuous data layer is proportional to the average amplitude of all data points within that data layer, and the height of a continuous data layer is the duration of the equal time interval. For each data layer obtained from the division, its equivalent radius and height are calculated according to the duration of equal time intervals. Based on the equivalent radius and height of each data layer, the equivalent volume of each layer is calculated layer by layer using the cylinder volume formula. The equivalent volumes of all data layers are summed to generate volume data containing the total equivalent volume parameter of the samples.

6. The method for processing lipid analyzer data according to claim 5, characterized in that, Based on the compensation amounts at each angle, the amplitude coordinates of the signal's starting point, peak point, and ending point are compensated respectively, resulting in three key data points after compensation: Determine the compensation direction of the angle compensation amount of the first interior angle, the second interior angle, and the third interior angle for the corresponding vertex amplitude coordinates. The angle compensation amount of the first interior angle determines the upward or downward compensation direction of the peak point amplitude coordinates, the angle compensation amount of the second interior angle determines the upward or downward compensation direction of the signal start point amplitude coordinates, and the angle compensation amount of the third interior angle determines the upward or downward compensation direction of the signal end point amplitude coordinates. Based on the compensation amount at each angle and its corresponding compensation direction, the amplitude coordinate values ​​of the signal start point, peak point and signal end point are increased or decreased respectively to obtain three key data points whose amplitude coordinates have been adjusted, while their time coordinates remain unchanged. The three key data points, whose amplitude coordinates have been adjusted but whose time coordinates remain unchanged, are respectively denoted as the signal start point, the signal peak point, and the signal end point after compensation, and are used as morphological feature control points.

7. The method for processing lipid analyzer data according to claim 6, characterized in that, The volume data, including the total equivalent volume parameter of the sample, is normalized and concentration converted to output the final lipid analyzer results, including: Obtain the pre-established total equivalent volume parameter of the standard through the testing of standard products, and calculate the ratio between the total equivalent volume parameter of the sample and the total equivalent volume parameter of the standard to obtain the normalized volume ratio. Based on the pre-calibrated concentration-volume conversion factor corresponding to the type of sample to be tested, the normalized volume ratio is converted into the preliminary concentration value of the sample. The initial concentration values ​​are standardized in units and rounded to obtain concentration test results, which are then used as the results of the lipid analyzer.

8. A lipid analyzer data processing system, wherein the system implements the method as described in any one of claims 1 to 7, characterized in that, include: The acquisition module is used to acquire the raw test data collected by the lipid analyzer; The preprocessing module is used to perform preprocessing on the raw detection data, including filtering and noise reduction and baseline calibration, to generate preprocessed data; The calculation module is used to construct virtual feature triangles by identifying key data points in preprocessed data, calculate the distribution of interior angles of the virtual feature triangles, and perform angle compensation and geometric correction on the positions of key data points based on the distribution of interior angles to generate morphological correction data. The extraction module is used to extract parameters from the morphology correction data. Specifically, it divides the data sample into multiple continuous data layers along the detection time axis, treats each data layer as a cylindrical cross-section, calculates the equivalent volume of each data layer by calculating the equivalent radius and height, and accumulates the equivalent volumes of all data layers to generate volume data containing the total equivalent volume parameters of the sample. The processing module is used to normalize and convert the volume data, which includes the total equivalent volume parameter of the sample, to output the final lipid analyzer test results.

9. A computing device, characterized in that, Including memory and processor; The memory stores one or more computer programs, the one or more computer programs including instructions; when the instructions are executed by the processor, the computing device performs the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store a computer program for performing the method as described in any one of claims 1 to 7.

Images