Dual-wavelength dry chemical detection method and system

Abstract

The invention provides a dual-wavelength dry chemical detection method and system. The method comprises the following steps: acquiring a first reaction stationary point signal of a first to-be-detected substance reaction area of a to-be-detected sample under a first wavelength and a second reaction stationary point signal of the first to-be-detected substance reaction area under a second wavelength; inputting the first reaction stationary point signal into a first concentration quantification model corresponding to the first to-be-detected substance to obtain a first concentration value; comparing the first concentration value with a preset concentration switching threshold value; if the first concentration value is smaller than the concentration switching threshold value, determining a concentration detection result of a first to-be-detected substance in the to-be-detected sample based on the first concentration value; and if the first concentration value is greater than or equal to the concentration switching threshold value, inputting the second reaction stationary point signal into a second concentration quantification model corresponding to the first to-be-detected substance to obtain a second concentration value, and determining a concentration detection result of the first to-be-detected substance in the to-be-detected sample based on the second concentration value. According to the invention, the detection result accuracy of dry chemical detection can be improved.

Description

Technical Field

[0001] This application relates to the field of dry chemical detection technology, specifically to a dual-wavelength dry chemical detection method and system. Background Technology

[0002] Dry chemistry testing technology is widely used in clinical and point-of-care testing (POCT) fields due to its advantages such as solid-phase reagents, no need for liquid flow paths, and ease of operation, especially for daily blood glucose and hemoglobin monitoring in patients with diabetes and its complications. The basic principle of this technology is to use a photoelectric detection system to measure the color depth change after a blood sample reacts chemically with a solid-phase reagent on a test strip. By detecting the intensity of the reflected light signal at a specific wavelength, the concentration of the analyte in the blood can be calculated.

[0003] However, in high-concentration sample detection scenarios, the color depth of some reaction products increases significantly, causing a physical shift in their maximum absorption peak. Simultaneously, the darker reactants cause the photoelectric sensor to enter the nonlinear response region or even approach saturation. This change in optical characteristics means that conventional electrical signal gain amplification cannot physically reproduce the true light absorption, resulting in a significant reduction in signal resolution in the high-concentration range and severely impacting the accuracy of the detection results. Summary of the Invention

[0004] The embodiments of this application provide a dual-wavelength dry chemical detection method and system, which aims to improve the accuracy of detection results in dry chemical detection.

[0005] In a first aspect, embodiments of this application provide a dual-wavelength dry chemical detection method, the dual-wavelength dry chemical detection method comprising:

[0006] Acquire the first reaction station point signal of the first analyte reaction region of the sample under test at the first wavelength, and the second reaction station point signal at the second wavelength;

[0007] The first reaction station point signal is input into the first concentration quantification model corresponding to the first analyte to obtain the first concentration value;

[0008] The first concentration value is compared with the preset concentration switching threshold.

[0009] If the first concentration value is less than the concentration switching threshold, the concentration detection result of the first analyte in the sample to be tested is determined based on the first concentration value.

[0010] If the first concentration value is greater than or equal to the concentration switching threshold, the second reaction stationary point signal is input into the second concentration quantification model corresponding to the first analyte to obtain the second concentration value. Based on the second concentration value, the concentration detection result of the first analyte in the sample to be tested is determined.

[0011] In the above embodiments, by comparing the first concentration value obtained based on the first reaction station point signal with the concentration switching threshold, the first concentration value is directly used when the first concentration value is less than the concentration switching threshold to determine that the sample to be tested is of low concentration. When the first concentration value is greater than or equal to the concentration switching threshold, the second concentration value obtained based on the second reaction station point signal and the second concentration quantization model is used. In this way, the optimal wavelength signal and quantization model are automatically matched for different concentration ranges of the sample to be tested, eliminating the error caused by the nonlinear response of a single wavelength signal at high concentrations, thereby improving the accuracy of the detection results of dry chemical detection.

[0012] In one embodiment, acquiring the first reaction stationary point signal of the first analyte reaction region of the sample under a first wavelength includes:

[0013] The continuous light signal of the first analyte reaction zone of the sample under test is acquired at a first wavelength, and the continuous light signal is converted into a discrete time series signal.

[0014] Within a preset sliding time window, the discrete time series signal is subjected to an augmented Dickey-Fuller test to obtain the t-statistic value;

[0015] The probability value corresponding to the t-statistic is obtained by looking up a table.

[0016] The discrete time series signal at the moment when the probability value is less than the preset significance level threshold is taken as the first reaction stationary point signal.

[0017] In the above embodiments, by acquiring the continuous light signal of the reaction zone of the first analyte at the first wavelength and converting it into a discrete time series signal, the augmented Dickey-Fuller test is performed within a preset sliding time window to obtain the t-statistic value and the corresponding probability value. The signal stationarity is determined by comparing the probability value with the significance level threshold. Thus, the time series statistical characteristics are used to replace the fixed time method to identify the chemical reaction endpoint, ensuring that the discrete time series signal, which is the first reaction stationary point signal, corresponds to the thermodynamic equilibrium state of the chemical reaction, and eliminating the detection error caused by the reaction time fluctuation.

[0018] In one embodiment, the step of performing an augmented Dickey-Fuller test on the discrete time series signal within a preset sliding time window to obtain the t-statistic value includes:

[0019] Obtain the regression equation that includes displacement, time trend, and random error terms;

[0020] Regression analysis is performed on the regression equation using the discrete time series signal within the sliding time window;

[0021] Determine the estimated value of the coefficient of existence of the unit root in the regression equation, and the standard deviation of the estimated value;

[0022] The t-statistic value is determined based on the ratio of the estimated value to the standard deviation.

[0023] In the above embodiments, regression analysis is performed on discrete time series signals within a sliding time window based on a regression equation containing displacement, time trend, and random error terms. The t-statistic value is determined by the ratio of the estimated value of the unit root existence coefficient to its standard deviation. This quantifies the deterministic trend and random fluctuation components in the discrete time series signal, providing a standardized statistical basis for determining whether a unit root exists in the signal. This ensures the accuracy of judging the stationary state of the response using the t-statistic value in the presence of noise interference or signal drift.

[0024] In one embodiment, the first concentration quantification model is generated through the following steps:

[0025] Multiple first standard solutions of the first analyte are obtained, and the concentration of the first analyte in the different first standard solutions is different;

[0026] Using a pre-set first standard detection device, determine the first standard concentration value of each first standard solution;

[0027] Acquire the first calibration reaction stationary point signal of each first standard solution at the first wavelength;

[0028] Based on the first calibrated reaction stationary point signal and the corresponding first standard concentration value, the first concentration quantification model is generated.

[0029] In the above embodiments, by acquiring the first calibration reaction stationary point signals of multiple first standard solutions with different concentrations at the first wavelength, and associating them with the first standard concentration value determined by a preset first standard detection device, a mathematical mapping relationship between the photoelectric response value at the first wavelength and the true concentration of the first analyte is established. The generated first concentration quantification model can accurately characterize the concentration characteristics within the linear response range of the first wavelength, providing a calculation benchmark based on standard value calibration for the accurate quantification of low-concentration samples.

[0030] In one embodiment, after determining the first standard concentration value of each first standard solution using a preset first standard detection device, the method further includes:

[0031] Acquire the second calibration reaction stationary point signal of each of the first standard solutions at the second wavelength;

[0032] Based on the second calibrated reaction stationary point signal and the corresponding first standard concentration value, the second concentration quantification model is generated.

[0033] In the above embodiments, by acquiring the second calibration reaction station point signal of the first standard solution at the second wavelength, and generating a second concentration quantification model based on the signal and the first standard concentration value, a high concentration calculation model adapted to the optical characteristics of the second wavelength is constructed using the same set of standard concentration benchmarks. This ensures that when the concentration of the sample to be tested exceeds the linear range of the first wavelength, the system can call the second concentration quantification model based on the response characteristics of the second wavelength for accurate calculation, thus achieving traceability of measurement results across the entire range.

[0034] In one embodiment, while performing the steps of acquiring the first reaction stationary point signal of the first analyte reaction region of the test sample at a first wavelength and the second reaction stationary point signal at a second wavelength, the method further includes:

[0035] Obtain the signal of the third reaction station point of the second analyte reaction region of the sample under the second wavelength;

[0036] The dual-wavelength dry chemical detection method also includes:

[0037] The third reaction station point signal is input into the third concentration quantification model corresponding to the second analyte to obtain the third concentration value;

[0038] Based on the third concentration value, the concentration detection result of the second analyte in the sample to be tested is determined.

[0039] In the above embodiments, by acquiring the third reaction station point signal of the reaction region of the second analyte at the second wavelength while performing the first analyte detection step, and inputting the signal into the third concentration quantification model to obtain the third concentration value, the concentration determination of the second analyte is completed in parallel using the same detection process and partially shared optical wavelength resources. This achieves the joint detection of two different components in the sample in a single operation, improving detection efficiency and the richness of sample information acquisition.

[0040] In one embodiment, the third concentration quantification model is generated through the following steps:

[0041] Multiple second standard solutions of the second analyte are obtained, and the concentration of the second analyte in the different second standard solutions is different;

[0042] The second standard concentration value of each second standard solution is determined using a pre-set second standard detection device.

[0043] Acquire the third calibration reaction stationary point signal of each second standard solution at the second wavelength;

[0044] Based on the second standard concentration value and the corresponding third calibration reaction stationary point signal, the third concentration quantification model is generated.

[0045] In the above embodiments, by acquiring the third calibration reaction stationary point signals of multiple second standard solutions with different concentrations at the second wavelength, and associating them with the second standard concentration values ​​determined by the preset second standard detection device, a third concentration quantification model specifically for the second analyte is constructed, eliminating the optical response differences between different analytes, and ensuring the accuracy of the second analyte concentration detection results determined based on the second wavelength and the third reaction stationary point signals through independent calibration.

[0046] In one embodiment, the first test substance is blood glucose, and the second test substance is hemoglobin.

[0047] In the above embodiments, by limiting the first analyte to blood glucose and the second analyte to hemoglobin, the dual-wavelength dry chemical detection and multi-index joint detection scheme is specifically applied to the screening scenario of diabetes and its complications. The dual-wavelength segmented calculation is used to solve the problem of inaccurate measurement of high-concentration blood glucose, while the parallel detection process is used to obtain the hemoglobin index, providing a comprehensive concentration detection result for the metabolic level and anemia status of diabetic patients in a single test.

[0048] Secondly, embodiments of this application provide a dual-wavelength dry biochemical measurement system, the dual-wavelength dry biochemical measurement system comprising a first photoelectric detection module and a data analysis module, the first photoelectric detection module comprising a first light source corresponding to a first wavelength, a second light source corresponding to a second wavelength, and a light signal receiving unit, wherein the data analysis module is used for:

[0049] Based on the optical signal at the first wavelength and the optical signal at the second wavelength collected by the optical signal receiving unit, the dual-wavelength dry chemical detection method described in any one of the above statements is executed.

[0050] In the above embodiments, by configuring a first light source corresponding to the first wavelength and a second light source corresponding to the second wavelength in the first photoelectric detection module, the optical signal receiving unit collects dual-wavelength signals, and the data analysis module executes dual-wavelength dry chemical detection logic, thereby providing the necessary dual-channel optical hardware foundation and data processing platform for segmented threshold determination and adaptive stationary point identification, ensuring that the system can physically realize the function of generating two different wavelength light sources and calculation models.

[0051] In one embodiment, the dual-wavelength dry biochemical measurement system further includes a second photoelectric detection module and a sample holding space.

[0052] The sample accommodating space is used to accommodate the sample to be tested. A first detection hole corresponding to the first analyte reaction zone of the sample to be tested and a second detection hole corresponding to the second analyte reaction zone of the sample to be tested are provided on the side wall of the shell forming the sample accommodating space.

[0053] The photoelectric detection direction of the first photoelectric detection module is towards the first detection hole, and the photoelectric detection direction of the second photoelectric detection module is towards the second detection hole.

[0054] In the above embodiments, by setting a first detection hole and a second detection hole on the side wall of the shell forming the sample accommodating space, and making the first photoelectric detection module and the second photoelectric detection module face the corresponding detection hole respectively, an independent optical detection path is constructed in physical space, preventing optical signal crosstalk between the first analyte reaction area and the second analyte reaction area, and ensuring that the dual-wavelength dry biochemical measurement system can simultaneously and independently collect and process photoelectric signals from two different reaction areas. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 This is a schematic flowchart of an embodiment of the dual-wavelength dry chemical detection method provided in this application;

[0057] Figure 2 This is a flowchart illustrating the overall technical route of the dual-wavelength dry chemical detection method provided in the embodiments of this application;

[0058] Figure 3 This is a flowchart of the dual-wavelength segmented detection logic for a first analyte (such as blood glucose) provided in an embodiment of this application;

[0059] Figure 4 This is a schematic diagram of the hardware structure layout of the first photoelectric detection module and the second photoelectric detection module provided in the embodiments of this application;

[0060] Figure 5 This is a graph showing the correspondence between the light signal change curve of the first analyte during the reaction process and the real-time ADF test probability value (p-value) provided in the embodiments of this application;

[0061] Figure 6 This is a scatter plot showing the relationship between the discrete time series signal of the reaction region of the first analyte at a first wavelength (e.g., 640 nm) and the standard concentration, provided by an embodiment of this application.

[0062] Figure 7 This is a scatter plot showing the relationship between the discrete time series signal of the reaction region of the first analyte at a second wavelength (e.g., 520 nm) and the standard concentration, provided in an embodiment of this application.

[0063] Figure 8 This is a linear regression analysis graph showing the relationship between the measurement results of the first analyte obtained by the single-wavelength detection method in related technologies and the standard values ​​of clinical biochemical equipment.

[0064] Figure 9 This is a linear regression analysis graph showing the relationship between the measurement results of the first analyte obtained by the dual-wavelength segmented detection method in an embodiment of this application and the standard value of a clinical biochemical device.

[0065] Figure 10 This is a comparison chart of the relative deviations between the dual-wavelength dry chemical detection method and the single-wavelength method provided in the embodiments of this application in the detection results of high-concentration samples (≥10 mmol / L).

[0066] in, Figure 4 The image is in color to reflect the actual hardware structure of the first and second photoelectric detection modules using a variety of different colors. Detailed Implementation

[0067] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. In addition, in the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0068] In a first aspect, embodiments of this application provide a dual-wavelength dry chemical detection method, applied to a dual-wavelength dry biochemical measurement system (hereinafter referred to as the "system"), with the system being the executing entity.

[0069] Specifically, refer to Figure 1 Dual-wavelength dry chemical detection methods may include:

[0070] S101. Obtain the first reaction station point signal of the first analyte reaction zone of the sample under the first wavelength, and the second reaction station point signal under the second wavelength.

[0071] In the embodiments of this application, the sample to be tested typically refers to a biological fluid sample such as whole blood, serum, or plasma. The first analyte reaction zone is a specific area set on a dry chemical test strip or test card, which is pre-filled with a chemical reagent capable of undergoing a specific biochemical reaction with the first analyte and producing a color change.

[0072] In some embodiments of this application, the first analyte is blood glucose. The first wavelength is typically selected to correspond to the maximum absorption peak of the chromogenic substance at low concentrations, such as red light in the range of 600 nanometers (nm) to 680 nanometers, preferably 640 nm. The second wavelength is typically selected to correspond to high concentrations or to avoid the effects of spectral drift at high concentrations, such as green light in the range of 500 nm to 550 nm, preferably 520 nm.

[0073] The first and second reaction stationary point signals refer to the electrical signal values ​​collected during the biochemical reaction kinetics process when the reaction rate approaches zero and the optical reflectance or absorbance no longer changes significantly with time.

[0074] In some embodiments of this application, the dual-wavelength dry chemistry detection method is executed using a dual-wavelength dry biochemical measurement system. In this system, a first light source emits light of a first wavelength to irradiate the reaction zone of a first analyte, and a second light source emits light of a second wavelength to irradiate the same reaction zone. A light signal receiving unit receives the light intensity signal reflected from the reaction zone. The system converts the analog light signal into a digital signal using an analog-to-digital converter. During the reaction, the system does not simply read values ​​at fixed times, but continuously monitors the reaction curve, identifies the moment when the reaction reaches its endpoint (i.e., the plateau moment), and reads the signal of the first wavelength corresponding to that moment as the first plateau point signal, and reads the signal of the second wavelength corresponding to that moment as the second plateau point signal. This method eliminates errors caused by fluctuations in reaction time due to differences in ambient temperature or sample viscosity. Detailed data processing procedures are described in subsequent embodiments.

[0075] S102. Input the first reaction station point signal into the first concentration quantification model corresponding to the first analyte to obtain the first concentration value.

[0076] In this embodiment, the first concentration quantization model is a pre-constructed mathematical formula or lookup table used to map the optical signal value at a first wavelength to a concentration value. Since the first wavelength has high sensitivity and resolution for low-concentration samples, this model is mainly used to accurately retrieve values ​​in the low-concentration range. The first concentration value is a preliminary quantization result calculated based on the optical response characteristics of the first wavelength.

[0077] S103. Compare the first concentration value with the preset concentration switching threshold.

[0078] In this embodiment, the concentration switching threshold is a pre-set critical concentration value used to distinguish the linear response range of the photoelectric system to different wavelengths. The threshold is set based on the fact that when the absorbance of the first wavelength approaches saturation or the absorption peak shifts due to color deepening, the detection linearity of the first wavelength will decrease significantly, while the second wavelength is still within a good linear response range.

[0079] In some embodiments of this application, for blood glucose detection, the concentration switching threshold is set to 10 mmol / L. The processor determines whether the first concentration value is greater than or equal to this threshold by comparison logic, thereby deciding which optical path data to use as the final result, thus realizing automatic range switching at the algorithm level.

[0080] S104. If the first concentration value is less than the concentration switching threshold, determine the concentration detection result of the first analyte in the sample based on the first concentration value.

[0081] In this embodiment, when the first concentration value is less than the concentration switching threshold, it indicates that the sample concentration is in a low concentration range. Within this range, the colorimetric reaction corresponding to the first wavelength is not saturated, the light signal has high sensitivity to concentration changes, and no obvious absorption peak shift occurs. Therefore, the first concentration value has high accuracy and reliability.

[0082] In some embodiments of this application, the system directly outputs the first concentration value as the concentration detection result of the first analyte in the sample to the display module of the dual-wavelength dry biochemical measurement system. This processing method fully utilizes the high signal-to-noise ratio advantage of the first wavelength at low concentrations, ensuring measurement accuracy in the low blood glucose range, which meets the requirements of medical testing for high sensitivity at low values.

[0083] S105. If the first concentration value is greater than or equal to the concentration switching threshold, the second reaction stationary point signal is input into the second concentration quantification model corresponding to the first analyte to obtain the second concentration value. Based on the second concentration value, the concentration detection result of the first analyte in the sample is determined.

[0084] In this embodiment, when the first concentration value exceeds the threshold, it means the sample is in a high concentration range. At this time, the signal of the first wavelength may be distorted due to the nonlinear response of the sensor caused by excessively dark colors. The second concentration quantification model is a mathematical model specifically constructed for the optical characteristics of high-concentration samples at the second wavelength. Since the second wavelength has better penetration and a wider linear dynamic range against a dark background, its calculation results can truly reflect the actual content of high-concentration samples.

[0085] In some embodiments of this application, the processor activates the second calculation channel, calls the parameters of the second concentration quantification model, substitutes the second reaction stationary point signal into the calculation to obtain the second concentration value, and designates this second concentration value as the final concentration detection result. This process effectively solves the problem of "flat-top effect" or low measured value at high concentrations in traditional single-wavelength detection. Without changing the chemical reagent formulation, the detection linear range is significantly extended through the combination of optics and algorithms, for example, from the conventional 0-18 mmol / L to 0-30 mmol / L or even higher.

[0086] As can be seen, the embodiments of this application adopt a dual-wavelength segmented detection strategy, combined with prediction logic, to use a high-sensitivity wavelength in the low-concentration region and automatically switch to a wide dynamic range wavelength in the high-concentration region. This effectively overcomes the nonlinear errors caused by absorption spectrum drift of high-concentration samples and sensor saturation in dry chemical detection. While ensuring low-value accuracy, it significantly expands the detection linear range and improves the measurement accuracy across the entire range.

[0087] Building upon the wide-range, high-precision detection achieved through the aforementioned dual-wavelength and segmented detection logic, to further improve the accuracy of the acquired reaction stationary point signal and avoid under-reaction or over-reaction errors caused by the fixed-time detection method, some embodiments of this application also provide an adaptive signal acquisition scheme based on time-series statistical testing, used to accurately capture the thermodynamic equilibrium moment of a chemical reaction. Specifically, acquiring the first reaction stationary point signal of the first analyte reaction region of the sample under test at the first wavelength includes:

[0088] S201. Collect the continuous light signal of the first test substance reaction zone of the sample under the first wavelength, and convert the continuous light signal into a discrete time series signal.

[0089] In this embodiment, the reaction between the sample and the reagent causes a color change, which in turn leads to a dynamic change in the light intensity passing through the reaction zone. A continuous optical signal refers to an analog voltage or current signal continuously received and output by a photoelectric sensor, the amplitude of which changes continuously with the reaction time. A discrete-time series signal refers to a set of digital signals arranged in chronological order after sampling and quantization.

[0090] In some embodiments of this application, the photodiode (PD) in the optical signal receiving unit of the dual-wavelength dry biochemical measurement system monitors the intensity of reflected light from the reaction zone in real time and outputs an analog electrical signal. An analog-to-digital converter (ADC) samples this analog electrical signal at a preset sampling frequency (e.g., 10 Hz or 100 Hz) and converts it into a series of digital voltage values. The processor stores these digital voltage values ​​in a buffer according to timestamps, forming a discrete-time series signal for subsequent mathematical analysis.

[0091] S202. Within the preset sliding time window, perform augmented Dickey-Fuller test on the discrete time series signal to obtain the t-statistic value.

[0092] In this embodiment, the sliding time window refers to a fixed-length time period, such as 10 seconds. This window shifts forward on the timeline as real-time data is acquired, always including the latest data from the current moment and a past period. The Augmented Dickey-Fuller test (ADF test) is a statistical method used to test whether time series data has a unit root, determining whether the series is stationary. The t-statistic is a key output parameter in the ADF test process, used for subsequent hypothesis testing.

[0093] S203. Obtain the probability value corresponding to the t-statistic by looking up the table.

[0094] In this embodiment, the probability value (p-value) represents the probability of observing the current t-statistic value or a more extreme case under the condition that the null hypothesis (i.e., the sequence is non-stationary and has a unit root) is true. Table lookup refers to the processor searching for the probability value corresponding to the calculated t-statistic value in a standard statistical distribution table pre-loaded in internal memory.

[0095] In some embodiments of this application, the system stores a mapping table that records the correspondence between different t-statistic values ​​and probability values. The processor retrieves the corresponding probability value from the table based on the real-time t-statistic value calculated in S202, using linear interpolation or direct indexing. This probability value intuitively reflects the degree to which the response signal deviates from a "non-stationary state" within the current time window; the smaller the probability value, the stronger the evidence to reject the null hypothesis, meaning the signal tends towards stationarity.

[0096] S204. The discrete time series signal at the moment when the probability value is less than the preset significance level threshold is taken as the first reaction stationary point signal.

[0097] In this embodiment, the significance level threshold is a pre-set critical standard for statistical inference, such as 0.05 (i.e., 5%) or 0.01 (i.e., 1%). When the probability value is less than this threshold, statistically, it is considered that the null hypothesis of the existence of a unit root can be rejected at the corresponding confidence level, that is, it is determined that the data sequence within the current sliding time window is stationary and the chemical reaction has reached thermodynamic equilibrium.

[0098] In some embodiments of this application, in step S204, the system monitors the probability value calculated at each moment in real time. Once the probability value is detected to be lower than 0.05 for the first time, a stop acquisition command is immediately triggered or the current moment is marked as the reaction endpoint. The system reads the discrete signal value corresponding to the endpoint moment (or the average value within the time window), determines it as the first reaction stationary point signal, and uses it for subsequent concentration calculations. This process achieves adaptive intelligent identification of the reaction endpoint, ensuring that the signal value involved in the calculation is the true value after the chemical reaction has fully stabilized, rather than a transient value at a fixed time point.

[0099] As can be seen, the embodiments of this application utilize time series statistical analysis to replace the traditional fixed-time detection method. By using the ADF test to quantify the stability of the reaction curve in real time, the moment when the chemical reaction reaches thermodynamic equilibrium is accurately determined. This method effectively eliminates the impact of reaction rate fluctuations caused by sample differences and varying environmental temperatures on detection accuracy. It avoids errors caused by under-reaction and prevents signal drift caused by over-reaction, significantly improving the accuracy and stability of dry chemistry detection results.

[0100] Having clarified the overall process of using the ADF test to determine signal stationarity, in order to make this statistical method practically applicable in portable devices with limited computing resources, it is necessary to clarify the calculation details of its core parameter—the t-statistic. By constructing a regression equation that conforms to the kinetic characteristics of dry chemical reactions, abstract signal fluctuations can be transformed into quantifiable statistical indicators. The following embodiments of this application will elaborate on the specific mathematical path of this calculation logic. Specifically, within a preset sliding time window, the ADF test is performed on the discrete time series signal to obtain the t-statistic value, including:

[0101] S301. Obtain the regression equation that includes the displacement term, time trend term, and random error term.

[0102] In this embodiment, the regression equation is a mathematical model used to describe the variation of a discrete-time series signal over time, and its structure is pre-stored in the system's data analysis module. The drift term represents the constant intercept of the sequence, reflecting the fundamental offset of the signal. The time trend term captures the deterministic trend of the sequence's evolution over time; in dry chemical reactions, this typically corresponds to the kinetic process of gradual color deepening in the early stages of the reaction. The random error term represents white noise interference that cannot be explained by the model, such as circuit noise or optical perturbations.

[0103] In some embodiments of this application, the specific mathematical expression of the regression equation is set as follows:

[0104]

[0105] in, This represents the first difference at the current time. For displacement terms, For time trend items, Includes the coefficient of existence of unit root. This represents the random error term. The equation establishes a linear relationship between the current change in the reaction curve and its historical state and time trend, providing a mathematical basis for determining whether the reaction has stopped changing.

[0106] S302. Using the discrete time series signal within the sliding time window, perform regression analysis on the regression equation.

[0107] In this embodiment, regression analysis refers to the statistical process of determining the unknown parameters in the regression equation by minimizing the sum of squared errors. The discrete time series signal within the sliding time window constitutes a dataset containing a specific number of data points (e.g., 100 sampling points), which serves as the sample input for the regression analysis.

[0108] In some embodiments of this application, in step S302, the processor uses ordinary least squares to solve the regression equation established in S301. The system constructs an observation matrix and a design matrix, substituting the optical signal values ​​within the window into the matrix operations, aiming to find a set of optimal parameter estimates that minimize the sum of squared residuals between the model predictions and the actual observed optical signal values. This step converts the variation characteristics of the physical optical signal into mathematical parameters in the regression equation.

[0109] S303. Determine the estimated value of the coefficient of existence of the unit root in the regression equation, and the standard deviation of the estimated value.

[0110] In this embodiment, the estimated value of the unit root existence coefficient (typically corresponding to the coefficient in the above equation) The coefficient (σ) is the core output of regression analysis, and its value reflects the degree to which the series deviates from a stationary state. If the coefficient is significantly less than zero, it indicates that the series is stationary. The standard error of the estimate reflects the accuracy or dispersion of the coefficient estimate; the smaller the standard error, the more reliable the estimate.

[0111] In some embodiments of this application, the data analysis module extracts the corresponding lag terms from the result vector of ordinary least squares. The coefficient is used as an estimate of the unit root existence coefficient. Simultaneously, based on the variance of the regression residuals and the diagonal elements of the inverse matrix of the design matrix, the variance corresponding to this coefficient is calculated, and then the standard deviation of the estimate is obtained by taking the square root. These two parameters respectively characterize the trend characteristics of the optical signal over time and the dispersion of the signal.

[0112] S304. Determine the t-statistic value based on the ratio of the estimated value to the standard deviation.

[0113] In the embodiments of this application, the t-statistic is a dimensionless numerical value used in hypothesis testing to measure the degree of deviation between the estimated value and the hypothetical value (usually 0), and is standardized in units of standard deviation.

[0114] In some embodiments of this application, the processor performs a division operation, dividing the estimated value of the unit root existence coefficient by its corresponding standard deviation, and the result is the t-statistic value. The formula is expressed as follows:

[0115]

[0116] This calculation process essentially normalizes the coefficient estimates, eliminating the influence of the absolute signal amplitude and allowing the results to be compared with standard statistical distribution tables. This step condenses the complex morphological characteristics of the response curve into a single statistical indicator, providing a standardized input basis for subsequent probability determination.

[0117] As can be seen, the embodiments of this application construct a regression model that includes time trends and solves it using the least squares method, thereby accurately quantifying the optical changes in the dry chemical reaction process into statistical parameters. The t-statistic is calculated using the ratio of the estimated value to the standard deviation, which effectively eliminates the interference of signal noise on the determination, ensuring high robustness and consistency in identifying the reaction endpoint under different signal intensities and noise levels.

[0118] Through the above steps, the system has acquired the ability to accurately acquire the stationary point signal. However, to convert these dimensionless or voltage-form "light signals" into clinically meaningful "biochemical concentration values," a precise mathematical mapping relationship must be established, namely the aforementioned "first concentration quantification model." The accuracy of this model directly determines the precision of low-concentration sample detection. The following embodiments of this application will describe in detail how to construct this basic model through standardized experimental procedures. Specifically, the first concentration quantification model is generated through the following steps:

[0119] S401. Obtain multiple first standard solutions of the first analyte, wherein the concentration of the first analyte in the different first standard solutions is different.

[0120] In this embodiment, the first standard solution refers to a manually prepared or certified liquid sample containing a known or calibrated concentration of a first analyte, typically using a serum matrix or an aqueous solution matrix. Multiple first standard solutions constitute a concentration gradient series covering the expected detection range. "Acquisition" refers to the process of preparing or procuring these standards.

[0121] In some embodiments of this application, for blood glucose testing, a series of solutions with glucose concentrations ranging from 0 mmol / L to 30 mmol / L in an equally distributed gradient are obtained, for example, including multiple concentration points such as 0, 2, 5, 10, 15, 20, 25, and 30 mmol / L. These solutions serve as the physical benchmark for constructing the mathematical model, ensuring that the model has corresponding measured data to support it across a wide measurement range, especially with sufficient data density near the critical boundary between low and high concentrations.

[0122] S402. Using a pre-set first standard detection device, determine the first standard concentration value of each first standard solution.

[0123] In the embodiments of this application, the first standard detection device refers to a reference measuring instrument recognized in the industry as having high precision and accuracy, such as a large-scale fully automated biochemical analyzer. The first standard concentration value refers to the concentration value obtained by measuring with this reference device and considered as the true value.

[0124] In some embodiments of this application, the first standard solution of each gradient concentration prepared in step S401 is divided into two portions, one of which is sent to a reference-grade biochemical analyzer for detection. The reference device uses a classic wet chemical enzymatic method (such as the hexokinase method) for determination, which is not affected by physical structures such as dry chemical diffusion layers and has extremely high accuracy. The concentration reading output by the reference device is recorded as the first standard concentration value, providing an accurate target variable (Y value) for subsequent model fitting.

[0125] S403. Obtain the first calibration reaction stationary point signal for each first standard solution at the first wavelength.

[0126] In the embodiments of this application, the first calibration reaction stationary point signal refers to the optical signal value collected in the first wavelength channel and determined as the reaction endpoint by ADF verification when testing the standard solution using the dual-wavelength dry chemistry detection system of this application. The calibration process is carried out in a controlled experimental environment, aiming to establish the relationship between the system response and the true concentration.

[0127] In some embodiments of this application, another homologous standard solution from step S401 is added to the dry chemistry test card of this application. The measurement system activates a light source of the first wavelength (e.g., 640 nm) to acquire reaction process data in real time. The system automatically identifies the reaction equilibrium moment using the ADF test algorithm described in the foregoing embodiments and locks the discrete optical signal value at that moment. This operation is repeated for standard solutions of each concentration gradient to obtain a set of optical signal characteristic values ​​(X values) that correspond one-to-one with the first standard concentration value.

[0128] S404. Based on the first calibrated reaction stationary point signal and the corresponding first standard concentration value, generate the first concentration quantification model.

[0129] In this embodiment of the application, generating the first concentration quantification model refers to the process of using a mathematical regression algorithm to find a mathematical expression or set of coefficients that can best describe the quantitative relationship between the light signal (independent variable) and the standard concentration (dependent variable).

[0130] In some embodiments of this application, a standard curve is plotted with the first calibrated stationary point signal as the abscissa and the first standard concentration value as the ordinate. The processor uses the least squares method to perform nonlinear fitting on the data points, constructing a functional relationship C=f(S), where C represents the concentration and S represents the signal value. For the preferred low concentration range of the first wavelength (e.g., 0-10 mmol / L), a higher-order polynomial (e.g., a third-order polynomial) or an exponential function is used for fitting to obtain the best goodness of fit (R-squared value). The fitted parameters (such as slope, intercept, curvature coefficient) are permanently stored in the storage module of the dual-wavelength dry biochemical measurement system, forming a callable first concentration quantification model.

[0131] As can be seen, this embodiment constructs a quantitative model based on the actual physical properties of the reaction by using paired tests of isogradient standard solutions, combined with true value calibration of a high-precision reference device and the acquisition of adaptive endpoint signals by this system. This method ensures that the first concentration quantification model can accurately map the nonlinear relationship between the optical signal and the biochemical concentration, providing a reliable mathematical benchmark and traceability basis for the accurate measurement of low-concentration samples.

[0132] At this point, a full-range, high-precision detection scheme for a single analyte (such as blood glucose) has been completed. Considering that diabetic patients often need to monitor their anemia simultaneously in clinical applications, and that the dual-optical-path hardware architecture of this system inherently possesses multi-channel detection potential, this application further expands the detection dimensions. The following embodiments of this application will describe how to achieve the detection of a second analyte (such as hemoglobin) in parallel using the same optical system without adding operational steps. After determining the first standard concentration value of each first standard solution using a preset first standard detection device, the method further includes:

[0133] S501. Obtain the second calibration reaction stationary point signal of each first standard solution at the second wavelength.

[0134] In this embodiment, the second calibration reaction plateau point signal refers to the optical signal value collected under the second wavelength channel of the dual-wavelength dry chemistry detection system of this application during the calibration experiment for the same batch of first standard solutions, and which is determined to be the point where the reaction has reached a plateau. Since the optical characteristics of the second wavelength (e.g., 520 nm) are different from those of the first wavelength (e.g., 640 nm), even for the same sample, the intensity of reflected light and the shape of the reaction curve will differ, therefore independent acquisition is required.

[0135] In some embodiments of this application, the dual-wavelength dry biochemical measurement system operates using time-division multiplexing or parallel acquisition when detecting the first standard solution. Simultaneously with or within a very short time interval of acquiring the first wavelength signal, the system activates a second light source to acquire a continuous light signal at the second wavelength. Similarly, the ADF test algorithm is used to analyze the signal sequence of the second wavelength in real time, pinpoint the reaction endpoint, and read the signal value at that moment as the second calibration reaction stationary point signal. For high-concentration samples, even if the first wavelength signal has become saturated or nonlinear, the second wavelength signal can still maintain good gradient discrimination.

[0136] S502. Based on the second calibration reaction station point signal and the corresponding first standard concentration value, generate a second concentration quantification model.

[0137] In this embodiment, generating the second concentration quantization model refers to the process of constructing a mathematical model to map the optical signal at the second wavelength to the actual concentration in the high-concentration range. Since the second wavelength is mainly used to solve the linearity problem at high concentrations, the construction of this model focuses on the data fitting accuracy of high-concentration samples.

[0138] In some embodiments of this application, the processor performs regression modeling using the second calibrated stationary point signal as the independent variable and the corresponding first standard concentration value as the dependent variable. During the fitting process, the full range of data can be fitted, but a weighted method is more preferred, focusing on data points in the high concentration range (e.g., 10 mmol / L to 30 mmol / L). The system calculates the parameters C=g(S) of the best-fit curve, where g represents the functional relationship specific to the second wavelength. These parameters are stored and labeled as the second concentration quantization model, which is only invoked in high-concentration mode. When a threshold switching is triggered during actual measurement, the system invokes this model to accurately convert the second-wavelength signal into a concentration value, thereby obtaining linear measurement results in the high-concentration region.

[0139] As can be seen, the embodiments of this application provide complete data support for the dual-wavelength segmented detection strategy by simultaneously acquiring calibration data of the second wavelength and establishing an independent high-concentration quantization model. This method ensures that the system can automatically adopt a precise calculation path based on the second wavelength when high-concentration samples are overly colored or the first wavelength fails, thereby achieving high linearity detection across the entire measurement range.

[0140] Similar to the model construction logic for the first analyte, to ensure that the detection results of the newly added second analyte (such as hemoglobin) also possess clinical-grade accuracy and traceability, a dedicated "third concentration quantification model" also needs to be established. This process also relies on rigorous standard solution calibration. The following embodiments of this application will specifically illustrate the details of model generation and calibration for the second analyte. Specifically, while performing the steps of obtaining the first reaction stationary point signal of the first analyte reaction region of the test sample at a first wavelength and the second reaction stationary point signal at a second wavelength, the process also includes:

[0141] S601. Obtain the signal of the third reaction station point of the second analyte reaction zone of the sample under the second wavelength.

[0142] In this embodiment, the sample to be tested includes a second analyte of primary interest in addition to the first analyte. The reaction zone of the second analyte is a separate physical area on the test card, independent of the reaction zone of the first analyte, and contains a pre-filled reagent specific to the second analyte. In this embodiment, the second analyte is hemoglobin (HGB). The third reaction plateau signal refers to the light signal value collected from the reaction zone of the second analyte under illumination at a second wavelength (e.g., 520 nm), indicating the endpoint of the reaction.

[0143] In some embodiments of this application, the dual-wavelength dry biochemical measurement system is equipped with both a first photoelectric detection module and a second photoelectric detection module, enabling simultaneous or time-division rapid scanning of the first detection well (corresponding to the first analyte) and the second detection well (corresponding to the second analyte). When the sample is added, the second analyte (e.g., hemoglobin) reacts and develops color in the second reaction zone. Since the characteristic absorption peak of hemoglobin matches the second wavelength, the system uses a second light source to illuminate this area and collects continuous light signals through another set of photoelectric sensors. Similarly, the ADF test algorithm is used to determine whether the reaction is stable; once stable, the signal value at that moment is locked as the third reaction stationary point signal. This process is performed in parallel with the detection of the first analyte (e.g., blood glucose) without interference.

[0144] Correspondingly, dual-wavelength dry chemical detection methods also include:

[0145] S602. Input the third reaction station point signal into the third concentration quantification model corresponding to the second analyte to obtain the third concentration value.

[0146] In this embodiment, the third concentration quantification model is a mathematical model specifically established to describe the relationship between the concentration of the second analyte and the intensity of reflected light at the second wavelength. The third concentration value is the quantified content of the second analyte calculated based on this model.

[0147] In some embodiments of this application, in step S602, for the model parameters of the hemoglobin item (such as slope and intercept), the system can substitute the third reaction stationary point signal obtained in step S601 into the calculation. This model is usually calibrated based on a large number of clinical samples or standard solutions, ensuring that the light signal can be accurately converted into a hemoglobin concentration value (such as grams per liter).

[0148] S603. Based on the third concentration value, determine the concentration detection result of the second analyte in the sample to be tested.

[0149] In the embodiments of this application, determining the concentration detection result means that the calculated value is converted into a final diagnostic indicator after necessary unit conversion or range verification.

[0150] In some embodiments of this application, the system directly uses the third concentration value as the hemoglobin measurement result and displays it on the screen simultaneously with the blood glucose measurement result. For patients with diabetes and anemia, this step provides an indicator for monitoring complications. Through a single sampling and testing process, two important physiological parameters are acquired simultaneously.

[0151] As can be seen, in this embodiment, while performing dual-wavelength detection of the first analyte (such as blood glucose), the second analyte (such as hemoglobin) is detected in parallel using the second wavelength channel within the same optical system. This multi-index detection design makes full use of hardware resources, providing more comprehensive health monitoring data for diabetic patients without increasing operational complexity, significantly improving detection efficiency and the clinical application value of the equipment.

[0152] The prerequisite for accurately outputting the third concentration value in step S602 is that the system internally stores a "third concentration quantification model" that can accurately describe the mapping relationship between the second wavelength light signal and the actual concentration of the second analyte. The accuracy of this model directly determines the clinical reference value of the second indicator in the joint testing project. To ensure the traceability and high accuracy of this model, it needs to be constructed using specific standard substances and reference methods before the equipment leaves the factory or during the calibration stage. Based on this, the following embodiments of this application will describe in detail the specific steps for generating the third concentration quantification model. Specifically, the third concentration quantification model is generated through the following steps:

[0153] S701. Obtain multiple second standard solutions of the second analyte, wherein the concentration of the second analyte in the different second standard solutions is different.

[0154] In this embodiment, the second standard solution is a reference material specifically used to calibrate the detection accuracy of a second analyte (such as hemoglobin), typically a pre-defined whole blood quality control or hemoglobin-like solution. Multiple second standard solutions constitute a sample sequence covering the clinical testing range with a gradient-like concentration distribution.

[0155] In some embodiments of this application, a set of standards with hemoglobin concentrations ranging from 0 g / L to 255 g / L are prepared or purchased in the laboratory. For example, seven concentration levels are selected, covering the low-value anemia zone, the normal value zone, and the high-value erythrocytosis zone. These solutions undergo rigorous homogenization to ensure that the physical properties of each sample are stable, realistically simulating the diffusion and color development behavior of clinical blood samples on dry chemical test strips.

[0156] S702. Using a pre-set second standard detection device, determine the second standard concentration value of each second standard solution.

[0157] In this embodiment, the second standard testing device refers to a blood cell analyzer or a measuring device that uses international reference methods (such as the cyanide-methemoglobin method) and is used as a reference standard in the field of clinical testing. The second standard concentration value refers to the actual concentration data obtained by this authoritative device and used as the modeling benchmark.

[0158] In some embodiments of this application, the prepared second standard solution is aliquoted and then sent to a calibrated first-level hematology analyzer for measurement. The hemoglobin concentration reading output by the instrument is recorded as the "true value" for the corresponding sample. This step ensures the traceability of the measurement results of this system and maintains a high degree of consistency with the report results of mainstream hospital testing equipment.

[0159] S703. Obtain the signal of the third calibration reaction station point of each second standard solution at the second wavelength.

[0160] In the embodiments of this application, the third calibration reaction station point signal refers to the photoelectric response value collected under the second wavelength excitation for the reaction zone of the second analyte using the dual-wavelength dry biochemical measurement system of this application, and determined by the algorithm to be the reaction endpoint.

[0161] In some embodiments of this application, a second standard solution is added dropwise into the second detection well (HGB detection well) of the system. The system turns on a second wavelength (e.g., 520 nm) light source and continuously monitors the change in reflected light intensity in the reaction zone. The ADF test is used to determine in real time whether the signal has entered a stationary state. When the stationary condition is met, the system automatically locks and records the discrete signal value at that moment. The test is repeated for each concentration of standard solution to establish a one-to-one correspondence dataset between the "true concentration" and the "measured signal of this system".

[0162] S704. Based on the second standard concentration value and the corresponding third calibration reaction stationary point signal, a third concentration quantification model is generated.

[0163] In this embodiment of the application, generating a third concentration quantification model refers to constructing a functional relationship that can accurately invert the light signal into the concentration of the second analyte through mathematical processing.

[0164] In some embodiments of this application, the data analysis module uses regression analysis to process the data pairs obtained in S702 and S703. Using the third calibrated reaction stationary point signal as the independent variable and the second standard concentration value as the dependent variable, the least squares method is used to fit the hemoglobin concentration curve equation HGB=h(S). For example, a third-order polynomial model is constructed and the best-fit coefficients are calculated to obtain the third concentration quantification model.

[0165] As can be seen, the embodiments of this application establish a traceability relationship for the dry chemical measurement system through a rigorous benchmarking and calibration process, utilizing industry gold standard equipment. The third concentration quantification model constructed by this method can accurately describe the physical mapping between the light signal at the second wavelength and the hemoglobin concentration, ensuring that the detection results of the second analyte also possess clinical-grade accuracy and reliability in multi-index joint detection applications.

[0166] All the foregoing embodiments have detailed the method logic and algorithm flow of dual-wavelength dry chemical detection. However, the method must rely on specific hardware to be implemented in the physical world. In order to make the technical solution of this application complete and closed, some embodiments of this application will introduce a "dual-wavelength dry biochemical measurement system" capable of performing all the above detection steps, and describe in detail the composition and function of its core modules.

[0167] Specifically, the dual-wavelength dry biochemical measurement system includes a first photoelectric detection module and a data analysis module.

[0168] In the embodiments of this application, the dual-wavelength dry biochemical measurement system is typically designed as a handheld portable device, intended for rapid detection of biochemical indicators in human blood samples.

[0169] The first photoelectric detection module includes a first light source corresponding to a first wavelength, a second light source corresponding to a second wavelength, and a light signal receiving unit. The first photoelectric detection module is used to acquire the reaction signal of the first analyte (e.g., blood glucose). The first and second light sources are typically light-emitting diodes (LEDs). The emission wavelength of the first light source is set between 600 nm and 680 nm, preferably 640 nm, to excite the optimal colorimetric response for low-concentration samples. The emission wavelength of the second light source is set between 500 nm and 550 nm, preferably 520 nm, to provide illumination with better penetration and linearity when high-concentration samples show deeper color development. The light signal receiving unit is typically a photodiode or phototransistor, used to receive the light reflected from the reagent card and convert it into an analog current or voltage signal.

[0170] In some embodiments of this application, to ensure uniform illumination, the first photoelectric detection module employs a specific spatial arrangement: with the light signal receiving unit at the center, the first and second light sources are symmetrically distributed on both sides. For example, a linear or circular array arrangement of "LED1, LED2, light signal receiving unit, LED1, LED2" is used. This staggered arrangement design ensures that light of the first and second wavelengths can illuminate the same reaction area at similar angles and coverage areas, reducing measurement errors caused by differences in the optical path.

[0171] The data analysis module is used to execute the dual-wavelength dry chemical detection method in any embodiment based on the optical signals at the first and second wavelengths acquired by the optical signal receiving unit. The data analysis module includes an analog-to-digital converter (ADC) and a central processing unit (such as a microcontroller unit). The ADC converts the analog continuous signal output from the optical signal receiving unit into a digital discrete signal. The central processing unit runs a preset algorithm program, first performing an ADF test on the digital signal to identify the stationary point of the reaction, and then intelligently selecting either the first or second wavelength calculation model to output the final concentration based on the comparison between the measurement result at the first wavelength and a preset threshold.

[0172] In some embodiments of this application, the system further includes a storage module, a display module, and a power module. The identification card storage module is used to read and store batch information and specific concentration quantification model data of the accompanying reagent cards. This enables the system to perform calibration updates for different batches of reagent strips. The display module is used to visually present the test values. The power module is responsible for powering the light source, data processing, and display of the entire system.

[0173] In some embodiments of this application, the dual-wavelength dry biochemical measurement system further includes a second photoelectric detection module and a sample holding space. The sample holding space refers to the physical slot or cavity at the front end of the device used for inserting and securing the dry chemistry test card. The test card is inserted into and secured within this space via a detection interface. The dry chemistry test card typically includes a support layer, a diffusion layer distributed on the support layer, a blood filtration layer, a reaction layer, and a colorimetric layer.

[0174] The sidewall of the housing forming the sample containment space has a first detection hole corresponding to the first analyte reaction area of ​​the sample and a second detection hole corresponding to the second analyte reaction area of ​​the sample. The first and second detection holes are optical windows located inside the system housing and facing the reaction surface of the test card. The position of the first detection hole is precisely aligned with the area on the test card where the first analyte (e.g., blood glucose) reagent is pre-placed; the position of the second detection hole is precisely aligned with the area on the test card where the second analyte (e.g., hemoglobin) reagent is pre-placed.

[0175] The first photoelectric detection module faces the first detection aperture, and the second photoelectric detection module faces the second detection aperture. The first photoelectric detection module is fixedly installed above or behind the first detection aperture to ensure that its emitted dual-wavelength beam illuminates only the blood glucose reaction area. The second photoelectric detection module is fixedly installed above or behind the second detection aperture to ensure that its beam illuminates the hemoglobin reaction area. The second photoelectric detection module also includes a light source and a light signal receiving unit; its light source wavelength is typically configured to match the absorption peak of the second analyte (e.g., 520 nm). This structural design enables the simultaneous acquisition of reaction data for different analytes through dual-channel parallel optical paths during a single card insertion operation, without interference.

[0176] As can be seen, the dual-wavelength dry biochemical measurement system provided in this application, through its dual-channel, dual-source hardware architecture design and staggered light source arrangement, combined with built-in identification card parameter calling and adaptive algorithm processing, achieves high-precision joint detection of blood glucose and hemoglobin on a single handheld device. The deep integration of the physical optical path and intelligent algorithm in this system effectively solves the challenge of detecting high-concentration samples, improving the accuracy and practicality of portable medical testing.

[0177] To verify the feasibility and effectiveness of the embodiments of this application, sample preparation and molding were carried out according to the dual-wavelength dry biochemical measurement system in the embodiments of this application. Signals from 40 blood samples were collected and the solutions of the embodiments of this application were implemented. The relevant data processing procedures and results are as follows: Figures 2 to 10 As shown.

[0178] Figure 2 This is a flowchart illustrating the overall technical route of the dual-wavelength dry chemical detection method provided in this application. As shown, the technical route demonstrates how the system processes the detection of a first analyte (e.g., blood glucose (GLU, glucose)) and a second analyte (e.g., hemoglobin) in parallel. Specifically, the system first initiates dual-wavelength (wavelength 1 and wavelength 2) signal acquisition of the sample; then, it performs an ADF test on the acquired time-series signal in real time to identify the stationary point (i.e., the reaction endpoint); after obtaining the stationary point signal, the process splits into two paths: the left path, for the first analyte, acquires the stationary point signals (Sg1end, Sg2end) at the first and second wavelengths respectively, and selects an appropriate quantization model to output the final concentration based on the preliminary calculation results of the first wavelength; the right path, for the second analyte, acquires the stationary point signal (Shend) at the second wavelength and directly quantifies and outputs the concentration. This diagram intuitively demonstrates how this application achieves the synergistic detection of two indicators and two wavelengths through a single process.

[0179] Figure 3This is a flowchart illustrating the dual-wavelength segmented detection logic for a first analyte (such as blood glucose) provided in this application embodiment. The flowchart details the core control logic of the method described in claim 1: First, the system synchronously acquires and stores the first wavelength signal Sg1(t) and the second wavelength signal Sg2(t); the p-value of Sg1(t) is monitored using the ADF test. When the p-value < 0.05, the reaction is determined to have reached equilibrium, acquisition is stopped, and the endpoint values ​​Sg1end and Sg2end are locked. Then, the segmented determination stage begins: first, the initial concentration C is calculated using Sg1end through Model 1; if C < 10 mmol / L (concentration switching threshold), the C value is directly output; if C ≥ 10 mmol / L, the high-concentration processing logic is activated, and the concentration C is recalculated using Sg2end through Model 2 and output. This flowchart clearly illustrates how the system intelligently switches between low and high concentration ranges to solve the range limitation problem at high concentrations.

[0180] Figure 4 This is a schematic diagram of the hardware structure layout of the first and second photoelectric detection modules provided in the embodiments of this application. The first and second photoelectric detection modules adopt an interleaved dual-light source layout design, specifically arranged as follows: second wavelength light source (e.g., LED2, green light 520nm), first wavelength light source (e.g., LED1, red light 640nm), light signal receiving unit (e.g., photodiode), first wavelength light source (LED1), and second wavelength light source (LED2). This physical architecture, with the signal receiving unit as the center and the dual-color light sources symmetrically distributed, can maximize the uniformity of the optical path when light of different wavelengths illuminates the reaction area of ​​the reagent card, reduce measurement errors caused by differences in the position of the light sources, and provide a hardware foundation for the accuracy of the dual-wavelength algorithm and its model.

[0181] Figure 5 This is a graph showing the correspondence between the light signal change curve of the first analyte during the reaction process and the real-time ADF test probability value (p-value) provided in this application embodiment. The horizontal axis represents time, the left vertical axis represents light signal intensity (reflectivity or voltage value), and the right vertical axis represents the p-value calculated by the ADF test. As shown, in the initial stage of the reaction (0-30 seconds), the light signal changes drastically, and the corresponding p-value remains at a high level (much greater than 0.05), indicating that the sequence is non-stationary. As the reaction progresses, the light signal tends to level off, and the measured p-value decreases rapidly. When the p-value falls below the significance level threshold of 0.05 (shown by the dotted line in the graph), the corresponding time point is determined by the system to be the stationary point of the reaction. This graph visually demonstrates the effectiveness and sensitivity of using the ADF statistic as a criterion for the reaction endpoint.

[0182] Figure 6This is a scatter plot showing the discrete-time series signal of the first analyte reaction zone at a first wavelength (e.g., 640 nm) and the standard concentration, provided in an embodiment of this application. The plot illustrates the trend of the reflected light signal (y-axis) as the concentration of the analyte (blood glucose) increases (x-axis) at the first wavelength. It can be seen from the plot that the signal has good discriminative power in the low concentration range (e.g., 0-10 mmol / L); however, in the high concentration range (e.g., >18 mmol / L), the data points become increasingly dense and flat, indicating absorption peak drift or sensor saturation, leading to a decrease in the signal's concentration resolution. This demonstrates the limitations of a single wavelength in high-concentration detection and the necessity of setting a concentration switching threshold.

[0183] Figure 7 This is a scatter plot showing the discrete-time sequence signal of the reaction region of the first analyte at a second wavelength (e.g., 520 nm) and the standard concentration, provided in an embodiment of this application. Figure 6 To create a contrast, Figure 7 The results show that even at high concentrations (e.g., 10-30 mmol / L), the optical signal maintains a good linear gradient distribution at the second wavelength, without any obvious saturation. This figure fully verifies the physical basis for introducing a second wavelength for high-concentration sample measurement in this application, demonstrating that the second wavelength can effectively extend the detection linear range of the system.

[0184] Figure 8 This is a linear regression analysis graph showing the relationship between the measurement results of the first analyte obtained using a single-wavelength detection method in related technologies and the standard values ​​of clinical biochemical instruments. The horizontal axis represents the standard concentration measured by the clinical biochemical analyzer, and the vertical axis represents the concentration measured by the traditional single-wavelength method. As shown in the figure, after the concentration exceeds a certain range (e.g., 15 mmol / L), the data points begin to deviate significantly from the diagonal (y=x), exhibiting a downward bending trend, and the linear correlation coefficient (R0) decreases. 2 The value was only 0.963. This indicates that traditional methods have a large systematic negative bias in the detection of high-concentration samples.

[0185] Figure 9 This is a linear regression analysis graph showing the relationship between the measurement result of the first analyte obtained using the dual-wavelength segmented detection method in this application and the standard value of a clinical biochemical device. (Comparison) Figure 8 , Figure 9 The data points in the study are closely distributed around the diagonal of y=x across the entire range of 0-30 mmol / L, and their linear correlation coefficient (R0) is relatively high. 2 The accuracy was improved to 0.9964. This comparative result shows that the method proposed in this application, based on ADF test and dual-wavelength segmentation model, can reduce nonlinear errors at high concentrations and improve the overall detection accuracy of the system.

[0186] Figure 10 This is a comparison of the relative deviations between the dual-wavelength dry chemical detection method and the single-wavelength method provided in this application for the detection results of high-concentration samples (≥10 mmol / L). The horizontal axis represents different test sample numbers, and the vertical axis represents the percentage of measurement deviation relative to clinical standard values. It is clearly visible from the figure that the single-wavelength method generally exhibits larger deviations in high-concentration samples; while using the dual-wavelength method of this application, the measurement deviation for the vast majority of samples is controlled within a very small range. This figure demonstrates the advantages of the technical solution of this application in improving the accuracy of high-value sample detection.

[0187] In some embodiments of this application, the ADF-based reaction station point identification and dual-wavelength segmented detection method provided in this application has universal scalability and is not limited to the detection of blood glucose and hemoglobin. It can be applied to other dry biochemical detection items with colorimetric reaction characteristics (such as uric acid, cholesterol, etc.).

[0188] For different biochemical detection items, the dual-wavelength dry biochemical measurement system supports parameter configuration adjustment:

[0189] Adjustment for reaction station point discrimination: For analytes with different reaction kinetic rates, the system allows changing the size of the "sliding time window" in the ADF test (e.g., from 10 seconds to 5 seconds or 15 seconds) and the "significance level threshold" (p-value, e.g., adjusted to 0.01 or 0.1) to match the equilibrium characteristics of a specific biochemical reaction, enabling accurate discrimination of the reaction termination point in any dry biochemical test.

[0190] High-concentration sample accuracy adjustment: Addressing the absorption spectrum drift or nonlinear response issues that occur with other analytes at high concentrations, the system allows for accurate high-value sample measurements of biochemical parameters other than blood glucose by changing the wavelength configurations of the first and second light sources (e.g., changing the wavelength combination to 600nm and 500nm) and retraining the segmented concentration quantification model. This design enables the same hardware device to meet the wide-range detection needs of various biochemical indicators by adjusting software parameters.

[0191] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A dual-wavelength dry chemical detection method, characterized in that, The dual-wavelength dry chemical detection method includes: Acquire the first reaction station point signal of the first analyte reaction region of the sample under test at the first wavelength, and the second reaction station point signal at the second wavelength; The first reaction station point signal is input into the first concentration quantification model corresponding to the first analyte to obtain the first concentration value; The first concentration value is compared with the preset concentration switching threshold. If the first concentration value is less than the concentration switching threshold, the concentration detection result of the first analyte in the sample to be tested is determined based on the first concentration value. If the first concentration value is greater than or equal to the concentration switching threshold, the second reaction stationary point signal is input into the second concentration quantification model corresponding to the first analyte to obtain the second concentration value. Based on the second concentration value, the concentration detection result of the first analyte in the sample to be tested is determined.

2. The dual-wavelength dry chemical detection method as described in claim 1, characterized in that, The acquisition of the first reaction station point signal of the first analyte reaction region of the sample under the first wavelength includes: The continuous light signal of the first analyte reaction zone of the sample under test at a first wavelength is acquired, and the continuous light signal is converted into a discrete time series signal. Within a preset sliding time window, the discrete time series signal is subjected to an augmented Dickey-Fuller test to obtain the t-statistic value. The probability value corresponding to the t-statistic is obtained by looking up a table. The discrete time series signal at the moment when the probability value is less than the preset significance level threshold is taken as the first reaction stationary point signal.

3. The dual-wavelength dry chemical detection method as described in claim 2, characterized in that, Within a preset sliding time window, the discrete time series signal is subjected to an augmented Dickey-Fuller test to obtain the t-statistic value, including: Obtain the regression equation that includes displacement, time trend, and random error terms; Regression analysis is performed on the regression equation using the discrete time series signal within the sliding time window; Determine the estimated value of the coefficient of existence of the unit root in the regression equation, and the standard deviation of the estimated value; The t-statistic value is determined based on the ratio of the estimated value to the standard deviation.

4. The dual-wavelength dry chemical detection method as described in claim 1, characterized in that, The first concentration quantification model is generated through the following steps: Multiple first standard solutions of the first analyte are obtained, and the concentration of the first analyte in the different first standard solutions is different; Using a pre-set first standard detection device, determine the first standard concentration value of each first standard solution; Acquire the first calibration reaction stationary point signal of each first standard solution at the first wavelength; Based on the first calibrated reaction stationary point signal and the corresponding first standard concentration value, the first concentration quantification model is generated.

5. The dual-wavelength dry chemical detection method as described in claim 4, characterized in that, After determining the first standard concentration value of each first standard solution using a preset first standard detection device, the method further includes: Acquire the second calibration reaction stationary point signal of each of the first standard solutions at the second wavelength; Based on the second calibrated reaction stationary point signal and the corresponding first standard concentration value, the second concentration quantification model is generated.

6. The dual-wavelength dry chemical detection method as described in claim 1, characterized in that, While performing the steps of acquiring the first reaction stationary point signal of the first analyte reaction region of the sample under a first wavelength and the second reaction stationary point signal under a second wavelength, the method further includes: Obtain the signal of the third reaction station point of the second analyte reaction region of the sample under the second wavelength; The dual-wavelength dry chemical detection method also includes: The third reaction station point signal is input into the third concentration quantification model corresponding to the second analyte to obtain the third concentration value; Based on the third concentration value, the concentration detection result of the second analyte in the sample to be tested is determined.

7. The dual-wavelength dry chemical detection method as described in claim 6, characterized in that, The third concentration quantification model is generated through the following steps: Multiple second standard solutions of the second analyte are obtained, and the concentration of the second analyte in the different second standard solutions is different; The second standard concentration value of each second standard solution is determined using a pre-set second standard detection device. Acquire the third calibration reaction stationary point signal of each second standard solution at the second wavelength; Based on the second standard concentration value and the corresponding third calibration reaction stationary point signal, the third concentration quantification model is generated.

8. The dual-wavelength dry chemical detection method according to any one of claims 1 to 7, characterized in that, The first test substance is blood glucose, and the second test substance is hemoglobin.

9. A dual-wavelength dry biochemical measurement system, characterized in that, The dual-wavelength dry biochemical measurement system includes a first photoelectric detection module and a data analysis module. The first photoelectric detection module includes a first light source corresponding to a first wavelength, a second light source corresponding to a second wavelength, and a light signal receiving unit. The data analysis module is used for: The dual-wavelength dry chemical detection method according to any one of claims 1 to 8 is executed based on the optical signal at the first wavelength and the optical signal at the second wavelength collected by the optical signal receiving unit.

10. The dual-wavelength dry biochemical measurement system as described in claim 9, characterized in that, The dual-wavelength dry biochemical measurement system also includes a second photoelectric detection module and a sample holding space. The sample accommodating space is used to accommodate the sample to be tested. A first detection hole corresponding to the first analyte reaction zone of the sample to be tested and a second detection hole corresponding to the second analyte reaction zone of the sample to be tested are provided on the side wall of the shell forming the sample accommodating space. The photoelectric detection direction of the first photoelectric detection module is towards the first detection hole, and the photoelectric detection direction of the second photoelectric detection module is towards the second detection hole.

Images