Environmental interference identification method and system for sample analysis and sample analyzer

Abstract

The invention relates to the technical field of photon detection, in particular to an environment interference identification method and system for sample analysis and a sample analyzer. According to the scheme, firstly, the photometric curve data of the target object to be measured are obtained, then, the photometric curve data are input into the preset feature model for feature parameter extraction, at least one feature parameter is extracted, then, the extracted feature parameters are analyzed to obtain the analysis result, and finally, the analysis result is obtained. And determining whether the photometric curve data of the target to-be-measured object is interfered by the environment according to the analysis result. By adopting the above scheme, the photometric curve data abnormity caused by environmental interference in any link of sample analysis can be detected, and the accuracy of the detection result can be ensured while the abnormal result can be efficiently detected by using the feature model, so that the accuracy of the detection result is improved. And the risks of false negative and false positive in clinical diagnosis are effectively reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of photon detection technology, and particularly relates to a sample analysis environmental interference identification method and system and a sample analyzer. BACKGROUND

[0002] Generally, a chemiluminescence immunoassay analyzer is used for sample analysis detection. The core principle of the detection is divided into two systems: an immune reaction system and a chemiluminescence analysis system. The immune reaction system is to combine a luminescent substance acrid ester-labeled antigen or antibody with a to-be-detected substance. The chemiluminescence analysis system is to use the oxidation of acrid ester under an alkaline environment by an oxidizing agent H2O2 to form an excited-state intermediate. When the excited-state intermediate returns to a stable ground state, light photons are emitted at the same time, and the number of light photons is measured by a measuring device.

[0003] The specific test process of the chemiluminescence immunoassay analyzer is: sample addition and transfer to a to-be-detected cup, addition of reagents, mixing, incubation, magnetic separation and cleaning, addition of a substrate liquid, and photon detection. In the actual test process, each stage may be interfered by environmental interference such as static electricity and electromagnetic interference, resulting in abnormal photon detection of the detection signal and abnormal concentration of the detection result. SUMMARY

[0004] The present application provides a sample analysis environmental interference identification method, system and sample analyzer, which effectively solves the problem that the environmental interference cannot be effectively identified in the existing sample analysis detection process, resulting in inaccurate test results.

[0005] According to a first aspect, a sample analysis environmental interference identification method is provided in an embodiment, comprising:

[0006] Obtaining measurement light curve data of a target to-be-detected substance;

[0007] Inputting the measurement light curve data into a preset feature model to obtain one or more feature parameters, analyzing the one or more feature parameters, and outputting an analysis result; wherein the one or more feature parameters include one or more of a high-frequency energy ratio, an instantaneous amplitude change rate, a low-frequency oscillation energy entropy, a wavelet entropy, a kurtosis coefficient, a marginal spectrum energy distribution, a dynamic threshold, a zero-crossing rate, a waveform symmetry index, and an autocorrelation function decay time.

[0008] Determining whether the measurement light curve data is interfered by the environment according to the analysis result.

[0009] In an implementable embodiment, the inputting of the measurement light curve data into the preset feature model to obtain one or more feature parameters comprises:

[0010] The preset feature model is a pre-trained artificial intelligence model; the photometric curve data is input into the artificial intelligence model; one-dimensional or multi-dimensional feature extraction is performed on the photometric curve data by the artificial intelligence model to obtain one or more feature parameters; or,

[0011] The preset feature model includes a plurality of preset feature functions; and the one or more feature parameters corresponding to the photometric curve data are calculated by using the plurality of preset feature functions.

[0012] In an implementable embodiment, the analysis of the one or more feature parameters and the output of the analysis result include:

[0013] The artificial intelligence model has a feature parameter abnormal range corresponding to the feature parameter, the artificial intelligence model matches the feature parameter with the corresponding feature parameter abnormal range respectively, obtains a matching result of whether the feature parameter matches the feature parameter abnormal range, and outputs the matching result; or,

[0014] The feature parameter is compared with a preset feature parameter normal range to obtain a comparison result of whether the feature parameter exceeds the preset feature parameter normal range and output the comparison result.

[0015] In an implementable embodiment, the determination of whether the photometric curve data is disturbed by the environment according to the analysis result includes:

[0016] If each of the feature parameters does not match the corresponding feature parameter abnormal range, it is determined that the photometric curve data is not disturbed by the environment; if at least one of the feature parameters matches the corresponding feature parameter abnormal range, it is determined that the photometric curve data is disturbed by the environment; or,

[0017] If the comparison result of each of the feature parameters is that the feature parameter does not exceed the corresponding preset feature parameter normal range, it is determined that the photometric curve data is not disturbed by the environment; if at least one of the feature parameters exceeds the corresponding preset feature parameter normal range, it is determined that the photometric curve data is disturbed by the environment.

[0018] In an implementable embodiment, the acquisition of the photometric curve data of the target to-be-measured object includes:

[0019] The original photometric curve data of the target to-be-measured object is acquired, and the original photometric curve data is processed to obtain the photometric curve data.

[0020] In an implementable embodiment, the one-dimensional or multi-dimensional feature extraction of the photometric curve data by the artificial intelligence model to obtain one or more feature parameters includes:

[0021] the artificial intelligence model extracts high-frequency energy ratio and instantaneous amplitude change rate of the photometric curve data through time-frequency analysis method;

[0022] the artificial intelligence model extracts low-frequency oscillation energy entropy of the photometric curve data through frequency domain analysis method;

[0023] the artificial intelligence model extracts zero-crossing rate and waveform symmetry index of the photometric curve data through time domain analysis method;

[0024] the artificial intelligence model extracts autocorrelation function decay time of the photometric curve data through statistical feature method.

[0025] In an implementable embodiment, after determining whether the photometric curve data is disturbed by the environment according to the analysis result, the method further comprises:

[0026] if it is determined that the photometric curve data is disturbed by the environment according to the analysis result, outputting corresponding prompt information to prompt the user that the photometric curve data is disturbed by the environment; and / or,

[0027] if it is determined that the photometric curve data is disturbed by the environment according to the analysis result, determining the type of the environmental disturbance according to the analysis result and outputting the type of the environmental disturbance.

[0028] According to a second aspect, an embodiment provides an environmental disturbance identification system for sample analysis, comprising:

[0029] an acquisition module configured to acquire photometric curve data of a target sample;

[0030] an analysis processing module configured to:

[0031] input the photometric curve data into a preset feature model to obtain one or more feature parameters, analyze the one or more feature parameters, and output an analysis result; wherein the one or more feature parameters comprise one or more of high-frequency energy ratio, instantaneous amplitude change rate, low-frequency oscillation energy entropy, wavelet entropy, kurtosis coefficient, marginal spectrum energy distribution, dynamic threshold, zero-crossing rate, waveform symmetry index, and autocorrelation function decay time.

[0032] determine whether the photometric curve data is disturbed by the environment according to the analysis result.

[0033] According to a third aspect, an embodiment provides a sample analyzer, comprising:

[0034] a sample suction mechanism configured to suck a sample and transfer the sucked sample to a reaction container;

[0035] a reagent mechanism configured to add a reagent to the reaction container.

[0036] a mixing mechanism configured to mix the sample and the reagent in the reaction container;

[0037] a cleaning mechanism;

[0038] an optical detection device configured to perform optical detection on the reaction solution after the sample and the reagent are reacted, to obtain measurement curve data of a target analyte;

[0039] a processor configured to:

[0040] input the measurement curve data into a preset feature model, to obtain one or more feature parameters, analyze the one or more feature parameters, and output an analysis result; wherein the one or more feature parameters include one or more of a high-frequency energy ratio, an instantaneous amplitude change rate, a low-frequency oscillation energy entropy, a wavelet entropy, a kurtosis coefficient, a marginal spectrum energy distribution, a dynamic threshold, a zero-crossing rate, a waveform symmetry index, and an autocorrelation function decay time;

[0041] determine, according to the analysis result, whether the measurement curve data is disturbed by an environment.

[0042] According to a fourth aspect, a computer-readable storage medium is provided, and the medium stores a computer program executable by a processor to implement the method described above.

[0043] According to the above-mentioned sample analysis environment interference identification method / system, first, the measurement curve data of the target analyte is obtained, then the measurement curve data is input into a preset feature model to extract feature parameters, wherein at least one feature parameter is extracted, then the extracted feature parameters are analyzed to obtain an analysis result, and finally, according to the analysis result, it is determined whether the measurement curve data of the target analyte is disturbed by an environment. By using the above-mentioned scheme of the present application, the abnormal measurement curve data caused by environmental interference in any link of sample analysis can be detected, and the feature model can be used to efficiently detect abnormal results while ensuring the accuracy of the detection results, effectively reducing the risk of false negative and false positive in clinical diagnosis. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 a flowchart of a sample analysis environment interference identification method provided by the present embodiment;

[0045] Figure 2 normal measurement curve data diagrams provided by the present embodiment, wherein (a) and (b) are two normal measurement curve data diagrams, respectively;

[0046] Figure 3The abnormal photometric curve data schematic diagram provided for the embodiment is shown in (a), (b) and (c);

[0047] Figure 4 The structural block diagram of the sample analysis environment interference identification system provided for the embodiment is shown in (a), (b) and (c);

[0048] Figure 5 The structural block diagram of the sample analyzer provided for the embodiment is shown in (a), (b) and (c).

[0049] The drawing label: 100, acquisition module; 200, analysis processing module; 10, sample suction mechanism; 20, reagent mechanism; 30, mixing mechanism; 40, cleaning mechanism; 50, optical detection device; 60, processor. DETAILED DESCRIPTION

[0050] The application will be described in further detail below with specific embodiments and drawings. In different embodiments, similar elements are associated with similar element labels. In the following embodiments, many details are described in order to make the application better understood. However, those skilled in the art can easily recognize that some features can be omitted in different cases, or can be replaced by other elements, materials, methods. In some cases, some operations related to the application are not shown or described in the specification, in order to avoid the core part of the application being overwhelmed by too much description, and it is not necessary to describe these related operations in detail for those skilled in the art according to the description in the specification and general technical knowledge in the art.

[0051] In addition, the features, operations or characteristics described in the specification can be combined in any appropriate way to form various embodiments. At the same time, the steps or actions in the method description can also be sequentially adjusted or adjusted in a manner that is obvious to those skilled in the art. Therefore, the order in the specification and drawings is only for the purpose of clearly describing a certain embodiment, and does not mean that it is the necessary order, unless otherwise stated that a certain order must be followed.

[0052] The serial numbers of the components in this paper, such as "first", "second", etc., are only used to distinguish the described objects, and do not have any order or technical meaning. Unless otherwise specified, "connection" and "coupling" in this application include direct and indirect connection (coupling).

[0053] In the specific test process of the current sample analyzer (such as a chemiluminescence immunoassay analyzer, a biochemical analyzer, etc.), there may be electrostatic interference or electromagnetic interference or temperature and humidity influence in each process, thereby causing abnormal test results. The existing interference identification method cannot identify the interference abnormal characteristics in the photon detection stage at present. The traditional method is to use anti-static interference measures to avoid interference influence, but it cannot cope with instantaneous interference, resulting in low detection efficiency and low test result accuracy. In view of this, the present application provides a sample analysis environmental interference identification method, system and sample analyzer to solve this problem.

[0054] As shown in Figure 1 The sample analysis environmental interference identification method provided by the present embodiment comprises the following steps:

[0055] Step 100: acquiring the light measurement curve data of the target to-be-measured substance by the acquisition module.

[0056] Specifically, in the experiment, after the sample is sucked by the sample suction mechanism, the sucked sample is transferred to the reaction container, and then the reagent (i.e. substrate liquid) is added to the reaction container, so that the reagent and the sample are mixed and react sufficiently, and the photon signal is released after the reaction of the sample and the reagent. At this time, the optical detection device is used to detect the photon signal generated after the reaction to generate the light measurement curve data of the target to-be-measured substance. Then, the system acquires the light measurement curve data by the acquisition module.

[0057] In some cases, when the sample is transferred to the reaction container, the container may itself have static electricity. When the reagent is injected into the container, the mixing of the reagent and the sample will generate a large amount of photon signal, but due to the influence of static electricity, the light measurement curve data of the photon signal will be abnormal. There are also some cases, such as during the injection of the reagent, the influence of environmental factors (temperature, humidity, etc.) will also cause the light measurement curve data of the photon signal to be abnormal.

[0058] Step 200: the analysis processing module inputs the light measurement curve data into a preset feature model to obtain one or more feature parameters, analyzes the one or more feature parameters, and outputs an analysis result; wherein the one or more feature parameters include one or more of high-frequency energy ratio, instantaneous amplitude change rate, low-frequency oscillation energy entropy, wavelet entropy, kurtosis coefficient, marginal spectrum energy distribution, dynamic threshold, zero-crossing rate, waveform symmetry index and autocorrelation function decay time.

[0059] Specifically, the light measurement curve data of the target to-be-tested object obtained by the analysis processing module is input into a preset feature model, and the feature model is used to extract the feature model of the light measurement curve data to obtain at least one feature parameter. In actual detection, the extracted feature parameters include multiple parameters such as high-frequency energy ratio, instantaneous amplitude change rate, low-frequency oscillation energy entropy, wavelet entropy, kurtosis coefficient, marginal spectrum energy distribution, dynamic threshold, zero-crossing rate, waveform symmetry index, and autocorrelation function decay time. After the feature parameters are extracted, the feature parameters are analyzed by the preset feature model to obtain an analysis result.

[0060] Step 300: The analysis processing module determines whether the light measurement curve data is disturbed by the environment according to the analysis result.

[0061] After obtaining the analysis result, the analysis processing module can determine whether the light measurement curve data is disturbed by the environment according to the analysis result. By using the environmental disturbance identification method of the present application, the abnormal condition of the light measurement curve data caused by the environmental disturbance in any link of sample analysis can be detected. In addition, the preset feature model can efficiently detect abnormal results while ensuring the accuracy of the detection results, effectively reducing the risk of false negatives and false positives in clinical diagnosis.

[0062] In some embodiments, the light measurement curve data is input into a preset feature model to obtain one or more feature parameters, which can be achieved by two ways:

[0063] One (the first way) is to use an artificial intelligence model for feature extraction. Specifically, the preset feature model is a pre-trained artificial intelligence model. The light measurement curve data is input into the artificial intelligence model. The artificial intelligence model is used to extract one-dimensional or multi-dimensional features of the light measurement curve data to obtain one or more feature parameters.

[0064] In this embodiment, the artificial intelligence model is pre-trained by using the normal light measurement curve data set (reference Figure 2 ) and the abnormal light measurement curve data set (reference Figure 3 ) as feature data sets, so that the artificial intelligence model can extract one-dimensional or multi-dimensional features of the input light measurement curve data of the target to-be-tested object and classify the normal light measurement curve data and the abnormal light measurement curve data. Specifically, in actual application, the pre-trained artificial intelligence model can include multiple sub-models, some of which are used to extract feature parameters and perform feature fusion, some of which are used to perform feature matching, and some of which are used to determine whether the environment is disturbed and give the specific disturbance type.

[0065] The characteristic parameters of the present application specifically include high-frequency energy ratio, instantaneous amplitude change rate, low-frequency oscillation energy entropy, wavelet entropy, kurtosis coefficient, marginal spectrum energy distribution, dynamic threshold, zero-crossing rate, waveform symmetry index, and autocorrelation function decay time. When training the artificial intelligence model, the functions for extracting the above characteristic parameters are written in the model, and the artificial intelligence model can extract the corresponding characteristic parameters according to the functions. The judgment criteria and extraction methods for the above characteristic parameters are as follows:

[0066] For the high-frequency energy ratio, the strength proportion of the high-frequency interference can be reflected by calculating the high-frequency energy ratio. The specific extraction method is as follows:

[0067] High frequency definition: signal component with frequency ≥ 1 kHz (corresponding to D1-D3 layers of wavelet decomposition).

[0068] The calculation formula is:

[0069] E total =∑ i |s denoised (i)| 2 ;

[0070] E high =∑ j∈{D1,D2,D3} ∑ k |C j (k)| 2 ;

[0071]

[0072] In the above formula, E total represents the total energy, E high represents the high-frequency energy, R high represents the high-frequency energy ratio, C j (k) represents the j-th layer wavelet coefficient, j represents the number of wavelet decomposition layers, D1, D2, and D3 represent D1, D2, and D3 layers, and s denoised (i) represents the value of the denoised signal at the i-th sampling point.

[0073] When R high ≥ 0.15, it is determined to be significant high-frequency interference.

[0074] For the instantaneous amplitude change rate , the signal mutation rate can be characterized by calculating the instantaneous amplitude change rate, and when , an alarm is triggered.

[0075] where, for the time window Δt:

[0076] Basic window: Δt base = 1 ms (corresponding to a sampling frequency of 10 kHz);

[0077] Dynamic adjustment: when continuous mutations are detected, the window is shortened to 0.2ms.

[0078] For amplitude difference ΔA:

[0079] Based on the amplitude calculation formula of the denoised signal:

[0080] ΔA = |S denoised (t2) - S denoised (t1) |;

[0081] In the above formula, S denoised (t) represents the denoised signal.

[0082] For low-frequency oscillation energy entropy, the complexity of low-frequency interference can be quantified by calculating the low-frequency oscillation energy entropy.

[0083] First, the denoised signal is decomposed by wavelet packet (number of layers = 4); Then extract the energy distribution of the lowest frequency subband (such as node 0-0), and then calculate the entropy value:

[0084]

[0085] In the above formula, H low represents the energy entropy, p i represents the proportion of i-layer energy, E i represents the energy of each layer coefficient.

[0086] Where H low ≥ 40% and more than 2 times the second highest peak, it is determined as interference and triggers an alarm.

[0087] For wavelet entropy:

[0088] Wavelet entropy calculates the complexity of the signal through the energy distribution after wavelet decomposition, reflecting the energy confusion degree of the signal in different frequency bands.

[0089] First, wavelet decomposition is performed: the denoised signal S denoised is decomposed by N-layer wavelet (preferably db4 wavelet basis); Then the decomposition obtains each layer coefficient: the approximate coefficient A N , the detail coefficient D1, D2, …D N .

[0090] Then energy calculation is performed:

[0091] E i = ∑ k |D i (k) | 2 (i = 1, 2, …, N);

[0092]

[0093] In the formula, E i represents the energy of each layer coefficient, D i (k) represents the wavelet coefficient at scale i and time position k, and the size (module) of the coefficient reflects the energy intensity of the signal at the corresponding scale (frequency range) and position (time point).

[0094] Then the energy proportion and entropy value are calculated:

[0095]

[0096] In the formula, H wavelet represents the wavelet entropy.

[0097] where H wavelet ≤1.2bit is determined as interference, triggering an alarm.

[0098] For kurtosis coefficient, the sharpness of signal distribution is measured by kurtosis coefficient to reflect the probability of extreme value.

[0099] The calculation formula of kurtosis coefficient is as follows:

[0100]

[0101] In the formula, μ4 is the fourth order central moment, μ2 represents the second order central moment (i.e. variance), and μ is the mean of the population. When K>0, it is determined as high-frequency burr interference, triggering an alarm.

[0102] For marginal spectrum energy distribution, by calculating the marginal spectrum energy distribution, combined with the global energy distribution of time-frequency analysis, the energy accumulation characteristics of the signal in each frequency band are reflected.

[0103] The specific extraction steps are: through short-time Fourier transform, the window length L=256 points, the overlap rate is 75%, and then the spectrum S(t, f) is calculated; then the spectrum is integrated according to frequency:

[0104] M(f) = ∑ t |S(t,f)| 2 ; In the formula, M(f) represents the marginal spectrum;

[0105] Finally, the frequency bands are divided (such as 1kHz step), and the energy proportion of each frequency band is counted. When the main frequency band energy proportion is greater than or equal to 40%, and the secondary peak energy proportion is less than 50% of the main peak energy, it is determined as interference, triggering an alarm.

[0106] For dynamic threshold, whether RLU mutation occurs is determined by calculating the threshold.

[0107] According to the real-time characteristics of the signal (such as noise level, baseline fluctuation), the determination threshold is automatically adjusted, and the specific extraction steps are:

[0108] First, noise baseline estimation is performed, and standard deviation σ is calculated by acquiring a signal silent section (when no light emission trigger) noise ; Then, the initial threshold value T base = k · σ is calculated noise ; After that, dynamic adjustment is performed: T dynamic = T base + a · |ΔA|, where a represents a sensitivity coefficient (usually 0.1-0.3), ΔA represents the average of the recent amplitude change rate, T base represents the initial threshold value, and T dynamic represents the dynamic threshold value; Finally, real-time updating is performed, and σ noise and T dynamic are recalculated every 10 seconds.

[0109] Among them, when the instantaneous change of the continuous 3 points exceeds the dynamic threshold value, an alarm is triggered, for example, assuming that the dynamic threshold value is calculated as 0.12, and the instantaneous changes of the continuous three points are (0.2, 0.3, 0.4), that is, it is judged as interference.

[0110] In addition, for the characteristic parameters: zero-crossing rate, waveform symmetry index, and autocorrelation function decay time, they can be extracted as auxiliary characteristic parameters. Among them, the zero-crossing rate is used to distinguish periodic interference from random noise, the waveform symmetry index is used to identify signal distortion caused by electrostatic interference, and the autocorrelation function decay time is used to judge the persistence of interference (transient / steady).

[0111] Among them, the photometric curve data can be photometric values with time labels, or graphs of photometric values and time relationships, etc. Generally, normal photometric curve data has the following characteristics:

[0112] (1) Stable baseline (low noise): Low signal distribution is uniform and has no obvious mutation, non-signal peak area is close to the baseline, and the fluctuation range is small.

[0113] (2) Single peak shape (signal area): a single steep peak appears, the peak value meets the expected range of immune response, the peak shape is symmetrical or close to a bell curve, the rising edge and the falling edge are smoothly transitioned, and there is no high-frequency oscillation.

[0114] Abnormal photometric curve data has the following characteristics:

[0115] (1) High-frequency burr (dense longitudinal scatter group / high-frequency oscillation baseline): dense longitudinal scatter groups appear in some areas of the abnormal graph, showing multiple frequencies and small amplitude RLU abnormal fluctuations, and high-frequency oscillation in the baseline area.

[0116] (2) Instantaneous amplitude mutation (discrete high amplitude point / unexpected spike): Discrete high amplitude point in the abnormal graph in the 100-200 horizontal axis and 100-180 vertical axis interval, which forms a significant difference with the surrounding data points, and an amplitude spike appears in an unexpected position.

[0117] (3) Baseline drift (overall data point shift / high baseline persistent rise): In the abnormal graph, the data points are shifted to the high vertical axis area as a whole, and the baseline is persistently deviated from zero.

[0118] As a way to extract feature parameters, the multi-dimensional feature extraction of the photometric curve data by the artificial intelligence model obtains a plurality of feature parameters, specifically including:

[0119] The artificial intelligence model extracts the high-frequency energy ratio and instantaneous amplitude change rate of the photometric curve data by time-frequency analysis method;

[0120] The artificial intelligence model extracts the low-frequency oscillation energy entropy of the photometric curve data by frequency domain analysis method;

[0121] The artificial intelligence model extracts the zero-crossing rate and waveform symmetry index of the photometric curve data by time domain analysis method;

[0122] The artificial intelligence model extracts the autocorrelation function decay time of the photometric curve data by statistical feature method.

[0123] In this embodiment, the artificial intelligence model extracts the feature parameters of the photometric curve data, such as high-frequency energy ratio, instantaneous amplitude change rate, low-frequency oscillation energy entropy, zero-crossing rate, waveform symmetry index and autocorrelation function decay time, and the specific extraction method has been described in the above embodiment, which will not be repeated here. Then analyze the feature parameters by the artificial intelligence module to determine whether the photometric curve data is disturbed by the environment.

[0124] When this method is used, one or more feature parameters are analyzed, and the analysis result is output, specifically including: the artificial intelligence model has a feature parameter abnormal range corresponding to the feature parameter, the artificial intelligence model matches the feature parameter with the corresponding feature parameter abnormal range, obtains the matching result of whether the feature parameter matches the feature parameter abnormal range, and outputs.

[0125] Since the artificial intelligence model is pre-trained, the artificial intelligence model has a feature parameter normal range and a feature parameter abnormal range corresponding to the feature parameters. Therefore, when analyzing the feature parameters extracted from the light measurement curve data of the target to-be-measured object, the artificial intelligence model can directly match the one or more feature parameters with the corresponding feature parameter abnormal range, and obtain the matching result of whether the feature parameters match the feature parameter abnormal range. When each feature parameter does not match the corresponding feature parameter abnormal range, it is determined that the light measurement curve data is not disturbed by the environment; if at least one feature parameter matches the corresponding feature parameter abnormal range, it is determined that the light measurement curve data is disturbed by the environment.

[0126] In addition, when it is determined by the method that the light measurement curve data is disturbed by the environment, a prompt information corresponding to the feature parameter abnormal type is outputted to prompt the user that the light measurement curve data is disturbed by the environment. For example, when the high-frequency energy ratio abnormality is detected, the system will issue a warning prompt information to remind the user that the test result is disturbed by the abnormality, and the user can take corresponding measures to retest according to the abnormality to check and confirm the source of the disturbance and perform corresponding processing.

[0127] Another (second) way is to use a preset feature function for feature extraction. Specifically, the preset feature model includes a plurality of preset feature functions; and the one or more feature parameters corresponding to the light measurement curve data are calculated by using the plurality of preset feature functions.

[0128] Specifically, the preset feature model includes a plurality of preset feature functions, and each feature parameter corresponds to a preset feature function. When performing feature extraction, the feature parameters to be extracted are determined according to the actual situation. In actual application, all feature parameters are generally extracted. Specifically, the light measurement curve data is inputted into the preset feature model, and the preset feature functions in the model directly perform feature calculation on the light measurement curve data to obtain a plurality of feature parameters. The plurality of preset feature functions are the same as the feature parameter calculation functions described in the above embodiments, and will not be described in detail in this embodiment.

[0129] Then, after obtaining the feature parameters, the feature parameters are directly compared with the preset feature parameter normal range in the preset feature model to obtain a comparison result of whether the feature parameters exceed the preset feature parameter normal range. Each calculated feature parameter has a comparison result. When the comparison result of each feature parameter is that the feature parameter does not exceed the corresponding preset feature parameter normal range, it is determined that the light measurement curve data is not disturbed by the environment; if at least one feature parameter has a comparison result that the feature parameter exceeds the corresponding preset feature parameter normal range, it is determined that the light measurement curve data is disturbed by the environment.

[0130] In some embodiments, the light measurement curve data of the target object to be tested is acquired, specifically including: acquiring original light measurement curve data of the target object to be tested, and performing noise reduction processing on the original light measurement curve data to obtain the light measurement curve data.

[0131] Specifically, in the process of acquiring the light measurement curve data of the target object to be tested, the original light measurement curve data of the target object to be tested is first acquired. Since the original light measurement curve data contains various noises, it will also interfere with the test results. Therefore, noise reduction processing needs to be performed on the original light measurement curve data. Specifically, wavelet noise reduction processing is performed: db4 wavelet basis is selected for 5-layer decomposition, low-frequency approximation coefficients (A5) are retained, and high-frequency noise coefficients (D1-D3) are filtered out. Then, convolution filtering, length interception and alignment, and signal recovery are performed. The reconstruction signal formula is:

[0132]

[0133] wherein x recon [n] is the reconstructed signal after noise reduction, A5[n] is the reconstructed component of the 5th layer low-frequency approximation coefficient, D5[n] is the reconstructed component of the 5th layer detail coefficient, D4[n] is the reconstructed component of the 4th layer detail coefficient, g rec is a low-pass reconstruction filter (db4 wavelet basis), h rec is a high-pass reconstruction filter (db4 wavelet basis), represents convolution operation, and IDWT() represents inverse discrete wavelet transform.

[0134] The wavelet noise reduction is performed on the original light measurement curve data in the above manner to obtain the light measurement curve data, so that the signal-to-noise ratio (SNR) of the light measurement curve data after noise reduction can be improved by at least 15 dB.

[0135] After the light measurement curve data is determined to be disturbed by the environment, the preset feature model of the present scheme can calculate each feature parameter in the light measurement curve data of the target object to be tested. Therefore, when the light measurement curve data is determined to be disturbed by the environment, the specific abnormal reason or abnormal type can also be determined according to how much each feature parameter exceeds the corresponding preset feature parameter normal range. In this way, the staff can pay attention to the abnormal reason in the next experiment, or improve the experimental process according to the abnormal type, so as to avoid the same abnormal disturbance, thereby improving the detection efficiency.

[0136] Reference Figure 4 The sample analysis environmental disturbance identification system provided in the present embodiment includes an acquisition module 100 and an analysis processing module 200. The acquisition module 100 is used to acquire the light measurement curve data of the target object to be tested.

[0137] The analysis processing module 200 is configured to: input the photometric curve data into a preset feature model to obtain one or more feature parameters, analyze the one or more feature parameters, and output an analysis result; wherein the one or more feature parameters include one or more of a high-frequency energy ratio, an instantaneous amplitude change rate, a low-frequency oscillation energy entropy, a wavelet entropy, a kurtosis coefficient, a marginal spectrum energy distribution, a dynamic threshold, a zero-crossing rate, a waveform symmetry index, and an autocorrelation function decay time; and determine whether the photometric curve data is disturbed by the environment according to the analysis result.

[0138] The environment interference identification system for sample analysis provided in the embodiment first acquires the photometric curve data of the target sample to be measured by the acquisition module 100, then inputs the photometric curve data into a preset feature model to extract feature parameters by the analysis processing module 200, wherein at least one feature parameter is extracted, then analyzes the extracted feature parameters to obtain an analysis result, and finally determines whether the photometric curve data of the target sample to be measured is disturbed by the environment according to the analysis result. By using the environment interference identification system, the abnormal photometric curve data caused by the environment interference in any link of sample analysis can be detected, and the accuracy of the detection result can be ensured while the abnormal result is efficiently detected by using the feature model, thereby effectively reducing the risk of false negative and false positive in clinical diagnosis. In addition, since the functions and effects of the acquisition module and the analysis processing module have been described in detail in the above embodiment of the environment interference identification method for sample analysis, no further description is given herein.

[0139] Reference Figure 5 The sample analyzer provided in the embodiment includes a sample suction mechanism 10, a reagent mechanism 20, a mixing mechanism 30, a cleaning mechanism 40, an optical detection device 50, and a processor 60. The sample suction mechanism 10 is configured to suck a sample and transfer the sucked sample to a reaction container. The reagent mechanism 10 is configured to add a reagent to the reaction container. The mixing mechanism 30 is configured to mix the sample and the reagent in the reaction container. The cleaning mechanism 40. The optical detection device 50 is configured to perform optical detection on a reaction solution after the sample and the reagent react to obtain photometric curve data of a target sample to be measured. The processor 60 is configured to: input the photometric curve data into a preset feature model to obtain one or more feature parameters, analyze the one or more feature parameters, and output an analysis result; wherein the one or more feature parameters include one or more of a high-frequency energy ratio, an instantaneous amplitude change rate, a low-frequency oscillation energy entropy, a wavelet entropy, a kurtosis coefficient, a marginal spectrum energy distribution, a dynamic threshold, a zero-crossing rate, a waveform symmetry index, and an autocorrelation function decay time; and determine whether the photometric curve data is disturbed by the environment according to the analysis result.

[0140] The sample analyzer provided in the embodiment is used for transferring the sample to the reaction container through the sample suction mechanism 10, then adding the reagent (i.e. the substrate solution) to the reaction container through the reagent mechanism 20 (a measuring cup or other instrument can be used), mixing the sample and the reagent through the mixing mechanism 30, so that the sample and the reagent can fully react, in addition, the reaction solution can be incubated to keep the reaction environment at a certain temperature and humidity to generate the photon signal in a more suitable environment, at this time, the photon signal is collected through the optical detection device 50, and after processing, the photometric curve data is generated to obtain the photometric curve data of the target analyte. However, in the process of the experiment, some environmental factors will inevitably interfere, and before the sample is added to the reaction container, or before the substrate solution is added to the reaction container through the reagent mechanism 20, static electricity will exist in the container itself, and after the sample or the substrate solution is added, the reaction process will be disturbed by the static electricity, resulting in abnormal photon signal, and then the photometric curve data generated by the optical detection device 50 is abnormal, which seriously affects the detection result. In addition, during the incubation process, the temperature and humidity are not controlled well, which will also affect the reaction, resulting in abnormal photon signal, and then the photometric curve data generated by the optical detection device 50 is abnormal, which seriously affects the detection result. Therefore, the sample analyzer of the embodiment detects and identifies the photometric curve data generated by the optical detection device 50 through the processor 60 to exclude the influence of environmental interference on the experiment, and the specific environmental interference identification method is described in detail in the above embodiment, which will not be described in detail in the embodiment. If it is detected that the photometric curve data is not disturbed by the environment, the photometric curve data is analyzed and processed in the next step to obtain the concentration of the target analyte, and then the concentration is displayed. In addition, the concentration can be judged, specifically by judging whether the concentration exceeds the preset threshold value, and if it exceeds, an alarm prompt is given to prompt the user to detect and analyze the photometric curve data again. Finally, after the experiment is completed, the experimental containers are cleaned through the cleaning mechanism 40.

[0141] The computer readable storage medium provided in the embodiment stores a computer program, and the computer program can be executed by the processor to realize the method described above. Since the environmental interference identification method for sample analysis has been described in detail in the above embodiment, the embodiment will not be described in detail here.

[0142] Those skilled in the art can understand that all or part of the functions of various methods in the above embodiments can be realized by hardware or by a computer program. When all or part of the functions in the above embodiments are realized by a computer program, the program can be stored in a computer readable storage medium, which can include a read-only memory, a random access memory, a magnetic disk, an optical disk, a hard disk, and the like. The above functions are realized by executing the program by a computer. For example, the program is stored in a memory of a device, and the above functions are realized by executing the program in the memory by a processor. In addition, when all or part of the functions in the above embodiments are realized by a computer program, the program can also be stored in a storage medium such as a server, another computer, a disk, an optical disk, a flash disk, or a mobile hard disk, and is saved in a memory of a local device by downloading or copying, or the system of the local device is updated, and the above functions are realized by executing the program in the memory by a processor.

[0143] The above application of specific examples to the present application is described, which is only used to help understand the present application and does not limit the present application. For those skilled in the art, according to the idea of the present application, a number of simple deductions, deformations or substitutions can be made.

Claims

1. A method for identifying environmental interference in sample analysis, characterized in that: include: Obtain photometric curve data of the target object to be measured; Inputting the photometric curve data into a preset characteristic model to obtain one or more characteristic parameters, analyzing the one or more characteristic parameters, and outputting an analysis result; wherein the one or more characteristic parameters include one or more of high-frequency energy ratio, instantaneous amplitude change rate, low-frequency oscillation energy entropy, wavelet entropy, kurtosis coefficient, marginal spectral energy distribution, dynamic threshold, zero-crossing rate, waveform symmetry index, and autocorrelation function decay time; Determine whether the photometric curve data is interfered with by the environment according to the analysis result.

2. The environmental interference identification method according to claim 1, wherein: The step of inputting the photometric curve data into a preset characteristic model to obtain one or more characteristic parameters includes: The preset feature model is a pre-trained artificial intelligence model; the photometric curve data is input into the artificial intelligence model; one-dimensional or multi-dimensional feature extraction is performed on the photometric curve data by the artificial intelligence model to obtain one or more feature parameters; or The preset characteristic model includes a plurality of preset characteristic functions; the plurality of preset characteristic functions are used to respectively calculate one or more characteristic parameters corresponding to the photometric curve data.

3. The environmental interference identification method according to claim 2, wherein: The analyzing of one or more characteristic parameters and outputting analysis results include: The artificial intelligence model has a characteristic parameter abnormal range corresponding to the characteristic parameter, and the artificial intelligence model matches the characteristic parameter with the corresponding characteristic parameter abnormal range respectively, obtains a matching result of whether the characteristic parameter matches the characteristic parameter abnormal range and outputs the result; or, The characteristic parameter is compared with a preset characteristic parameter normal range, and a comparison result of whether the characteristic parameter exceeds the preset characteristic parameter normal range is obtained and outputted.

4. The environmental interference identification method according to claim 3, wherein: Determining whether the photometric curve data is interfered with by the environment according to the analysis result includes: If each of the characteristic parameters does not match the corresponding characteristic parameter abnormal range, it is determined that the photometric curve data is not interfered with by the environment; if at least one of the characteristic parameters matches the corresponding characteristic parameter abnormal range, it is determined that the photometric curve data is interfered with by the environment; or, If the comparison result of each of the characteristic parameters is that the characteristic parameter does not exceed the corresponding preset characteristic parameter normal range, it is determined that the photometric curve data is not affected by environmental interference; if the comparison result of at least one of the characteristic parameters is that the characteristic parameter exceeds the corresponding preset characteristic parameter normal range, it is determined that the photometric curve data is affected by environmental interference.

5. The environmental interference identification method according to claim 1, wherein: The step of obtaining the photometric curve data of the target object to be measured includes: The original photometric curve data of the target object to be measured is obtained, and the original photometric curve data is subjected to noise reduction processing to obtain photometric curve data.

6. The environmental interference identification method according to claim 2, wherein: The one-dimensional or multi-dimensional feature extraction of the photometric curve data by the artificial intelligence model to obtain one or more feature parameters includes: The artificial intelligence model extracts the high-frequency energy ratio and instantaneous amplitude change rate of the photometric curve data through a time-frequency analysis method; The artificial intelligence model extracts the low-frequency oscillation energy entropy of the photometric curve data through a frequency domain analysis method; The artificial intelligence model extracts the zero-crossing rate and waveform symmetry index of the photometric curve data through a time domain analysis method; The artificial intelligence model extracts the decay time of the autocorrelation function of the photometric curve data through a statistical feature method.

7. The environmental interference identification method according to claim 1, wherein: After determining whether the photometric curve data is interfered with by the environment according to the analysis result, the method further includes: If it is determined according to the analysis result that the photometric curve data is interfered with by the environment, outputting corresponding prompt information to prompt the user that the photometric curve data is interfered with by the environment; and / or, If it is determined according to the analysis result that the photometric curve data is affected by environmental interference, the type of environmental interference is determined according to the analysis result and the type of environmental interference is output.

8. A system for identifying environmental interference in sample analysis, characterized in that: include: An acquisition module is used to obtain the photometric curve data of the target object to be measured; Analysis and processing module for: Inputting the photometric curve data into a preset characteristic model to obtain one or more characteristic parameters, analyzing the one or more characteristic parameters, and outputting an analysis result; wherein the one or more characteristic parameters include one or more of high-frequency energy ratio, instantaneous amplitude change rate, low-frequency oscillation energy entropy, wavelet entropy, kurtosis coefficient, marginal spectral energy distribution, dynamic threshold, zero-crossing rate, waveform symmetry index, and autocorrelation function decay time; Determine whether the photometric curve data is interfered with by the environment according to the analysis result.

9. A sample analyzer, characterized in that: include: A sample suction mechanism, used for sucking a sample and transferring the sucked sample to a reaction container; a reagent mechanism, for adding reagents into the reaction vessel; A mixing mechanism, used to mix the sample and reagent in the reaction container; cleaning mechanism; An optical detection device is used to perform optical detection on the reaction liquid after the sample reacts with the reagent to obtain the photometric curve data of the target object to be tested; Processor for: Inputting the photometric curve data into a preset characteristic model to obtain one or more characteristic parameters, analyzing the one or more characteristic parameters, and outputting an analysis result; wherein the one or more characteristic parameters include one or more of high-frequency energy ratio, instantaneous amplitude change rate, low-frequency oscillation energy entropy, wavelet entropy, kurtosis coefficient, marginal spectral energy distribution, dynamic threshold, zero-crossing rate, waveform symmetry index, and autocorrelation function decay time; Determine whether the photometric curve data is interfered with by the environment according to the analysis result.

10. A computer-readable storage medium, characterized in that A computer program is stored on the medium, and the computer program can be executed by a processor to implement the method according to any one of claims 1 to 7.

Images