A detection and analysis method of a blood oxygen meter
By constructing a piecewise polynomial compensation model, the clamping pressure of the pulse oximeter is adjusted in real time according to the error variation law of different pressure ranges, which solves the problem that the detection accuracy of the pulse oximeter is affected by the clamping pressure, and achieves high-precision detection of ±0.7%-±1.2%.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHENZHEN INSPECTION GRP (ZHEJIANG) QUALITY TECH SERVICE CO LTD
- Filing Date
- 2026-01-29
- Publication Date
- 2026-06-23
Smart Images
Figure CN122250995A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of blood oxygen detection technology, and more specifically, to a detection and analysis method for a pulse oximeter. Background Technology
[0002] Pulse oximeters, as portable medical devices used to detect blood oxygen saturation, are widely used in clinical monitoring, home health management, and other scenarios. Currently, most mainstream pulse oximeters on the market use photoplethysmography (PPG) technology. By emitting red light (660nm) and infrared light (940nm) through the fingertip, the difference in absorption of the two wavelengths of light by oxyhemoglobin and deoxyhemoglobin in the blood is used to calculate the blood oxygen saturation test value.
[0003] However, in actual use, the accuracy of existing pulse oximeters is easily affected by various factors, among which the clamping pressure of the pulse oximeter on the fingertip is one of the key influencing factors. When the clamping pressure is too low, the probe does not make sufficient contact with the fingertip, resulting in low light signal penetration efficiency and susceptibility to interference, leading to increased testing errors. When the clamping pressure is too high, it will compress the blood vessels in the fingertip, causing poor local blood flow and disrupting the normal distribution of hemoglobin in the blood, which will also lead to significant testing errors. Summary of the Invention
[0004] The purpose of this invention is to overcome the shortcomings of the prior art and provide a detection and analysis method for pulse oximeters to improve detection accuracy.
[0005] To achieve the above objectives, the present invention adopts the following technical solution: a pulse oximeter detection and analysis method, comprising the following steps: S1: Obtaining a pulse oximeter dataset ① Obtaining the actual pulse oximeter value of the test subject, denoted as Q; ② In a dark environment, using a pulse oximeter to perform a pulse oximeter test on the test subject's fingertip, while adjusting the pressure of the pulse oximeter on the test subject's fingertip to make the measured pulse oximeter value close to or reach the test subject's true pulse oximeter value, the measured value is denoted as Q1; ③ In a low-light environment, using a pulse oximeter to perform a pulse oximeter test on the test subject's fingertip, obtaining multiple sets of pulse oximeter test values E according to different pressures of the pulse oximeter on the test subject's fingertip, the pressure of the pulse oximeter on the test subject's fingertip is denoted as P, thereby obtaining a dataset: {(P1, E1), (P2, E2), ..., (Pn, En)}; S2: Constructing a compensation model to test the error, denoted as E(P), E(P) = a n P n +a n-1 P n-1 +a n-2 P n-2 +a1P 1 Fit the dataset and solve for the polynomial coefficients a. n-a1, to obtain the fitting polynomial; S3: Implement compensation ① Write the fitted polynomial into the pulse oximeter MCU; ② Perform pulse oximetry tests on multiple test subjects and obtain corrected test values; S4: Compare the corrected test values in S3 with the actual values.
[0006] Furthermore, the pressure P applied by the pulse oximeter to the test subject's fingertip includes two ranges: 0.3N-1.1N and 1.1N-5N. Corresponding fitting polynomials are obtained for each of the different ranges of pressure P.
[0007] Furthermore, the error range of the test values after correction by the compensation model is ±0.7% to ±1.2%.
[0008] Furthermore, the actual blood oxygen value was obtained by sampling blood from the fingertip of the test subject using an arterial blood gas analyzer.
[0009] Furthermore, in step S1, the ambient light intensity of the dark environment is ≤5 lux, and the light intensity range of the low-light environment is 5 lux-50 lux.
[0010] Furthermore, in step S1③, the pressure adjustment gradient is 0.1N-0.2N, and the test is repeated at least 5 times for each pressure value. The average value of the multiple test values is taken as the test value E at the corresponding pressure.
[0011] Furthermore, in step S2, the polynomial fitting adopts the least squares method, and the order n of the fitted polynomial takes the value of 2-5.
[0012] Furthermore, in step S1②, the initial adjustment pressure of the pulse oximeter on the test subject's fingertip is the minimum pressure threshold required for the test, with a value range of 0.3N-0.5N.
[0013] Furthermore, the pulse oximeter integrates a pressure sensor.
[0014] In summary, the present invention has the following beneficial effects:
[0015] By dividing the pressure range into two core intervals and constructing a compensation model using piecewise polynomial fitting, the error variation patterns in different intervals can be accurately captured. After compensation, the test error is controlled within ±0.7% to ±1.2%. The compensation model is written into the pulse oximeter's MCU, which can collect pressure values in real time and automatically call the corresponding interval compensation algorithm to achieve real-time correction of the test values and thus calibrate the equipment. Attached Figure Description
[0016] Figure 1 The flowchart is for an example. Detailed Implementation
[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0018] This embodiment discloses a method for detecting and analyzing pulse oximeters, including the following equipment: an arterial blood gas analyzer (model: GEM Premier 3000, detection accuracy ±0.1%), a pulse oximeter to be tested (using photoplethysmography technology, red light wavelength 660nm, infrared light wavelength 940nm), a light intensity detector (accuracy ±0.1 lux), and an adjustable dark box (internal dimensions 50cm×50cm×50cm, light adjustment range 0-100 lux, no external light interference). The pulse oximeter to be tested integrates a pressure sensor that can provide feedback on the fingertip pressure exerted by the pulse oximeter on the tester.
[0019] The specific steps are as follows:
[0020] Step S1: Obtain the blood oxygen test dataset
[0021] ① For each test subject, blood was drawn from their fingertip using an arterial blood gas analyzer to obtain the actual blood oxygen value Q, and the Q value of each test subject was recorded;
[0022] ② Adjust the light intensity inside the dark chamber to 3 lux (meeting the requirement of ≤5 lux in a dark environment, monitored in real time by a light intensity detector, with a fluctuation range of ≤±0.2 lux). Fix the pulse oximeter to be tested on the special bracket inside the dark chamber to avoid test deviations caused by displacement. The tester washes and dries their hands (to avoid sweat and dirt on the fingertips affecting the penetration of the light signal), and places their fingertips steadily into the detection chamber of the pulse oximeter. The pressure sensor collects the pressure data of the pulse oximeter on the fingertips in real time and displays it synchronously. Using 0.5N (within the minimum pressure threshold range of 0.3N-0.5N, which ensures sufficient contact between the probe and the fingertip without compressing the blood vessels in the fingertip and hindering blood flow) as the initial adjustment pressure, the pressure is gradually fine-tuned using the pressure adjustment module built into the pulse oximeter (adjustment step size of 0.05N). After each adjustment, the pressure is held stable for 3 seconds before testing, until the difference between the pulse oximeter's three consecutive blood oxygen test values and the test subject's Q value is ≤1.5% (this difference is the acceptable initial deviation under uncompensated conditions). The pressure value and test value (Q1) at this point are recorded.
[0023] A dark environment can greatly reduce the impact of external light on the test and improve the accuracy of the test.
[0024] ③ Adjust the light intensity inside the dark chamber to 25 lux (within the low-light environment range of 5-50 lux). A low-light environment minimizes the impact of external light on the test and simulates the real measurement environment. Using a pressure adjustment gradient of 0.15 N (within the range of 0.1 N-0.2 N, this gradient ensures sufficient data density while maintaining testing efficiency, fully covering the error variation characteristics under different pressures), gradually adjust the pulse oximeter pressure on the fingertip from 0.3 N to 5 N, completely covering the two core pressure ranges of 0.3 N-1.1 N and 1.1 N-5 N (the dividing point of 1.1 N is the experimentally verified inflection point of error variation; the error variation pattern before and after this pressure value differs significantly, and separate modeling can improve compensation accuracy). Within the pressure range of 0.3 N-1.1 N, the core influencing factor is the "sufficient contact between the probe and the fingertip," which is easily affected by external light within this range. Within the pressure range of 1.1N-5N, the core influencing factor is the "degree of compression of the fingertip blood vessels". Increased blood vessel compression leads to poor blood flow, affecting the detection. The error increases rapidly with increasing pressure. When the pressure is ≥2N, the blood vessel compression reaches a certain level, and the increase in error tends to level off.
[0025] Repeat the test 5 times at each pressure value, with a 10-second interval between each test (to ensure that the fingertip blood vessels return to normal between two tests and to avoid vascular fatigue caused by continuous testing affecting the data). Each test lasts 5 seconds, and blood oxygen test value and pressure value are recorded simultaneously.
[0026] Preferably, the two test pressure ranges are selected as 0.5N-1.1N and 1.1N-1.8N to narrow the detection area and improve the fitting degree within this detection range. Meanwhile, the pulse oximeter's operating instructions specify that the clamping pressure on the fingertip should be between 0.6N and 1.6N to further improve detection accuracy.
[0027] For each stress value, the arithmetic mean of 5 valid test values is taken (after removing the maximum and minimum values, the average is taken to further reduce random error), which is used as the final test value E under that stress. Finally, the complete dataset of each group of test subjects is obtained: {(P1, E1), (P2, E2), ..., (Pn, En)}, where n is the number of stress adjustments.
[0028] Step S2: Constructing the compensation model
[0029] Based on the dataset of each tester, the test error E(P) = EQ is calculated for each pressure value (E is the final test value under that pressure, and Q is the actual blood oxygen value of the tester (approximately equal to)).
[0030]
[0031] Table 1: Blood oxygenation test error (%) under uncompensated conditions
[0032] To further verify the accuracy of the error calculation, the consistency of five original test values (before removing extreme values) under the same pressure for a single tester was tested, and the coefficient of variation (CV) was used to evaluate the data dispersion. The formula is: CV = SD_original / E_original, where SD_original is the standard deviation of the original test values and E_original is the arithmetic mean of the speed measurement values.
[0033] The least squares method was used to perform polynomial fitting on the error data for the two pressure ranges of 0.3N-1.1N and 1.1N-5N, respectively. Preferably, the two pressure ranges were 0.5N-1.1N and 1.1N-1.8N.
[0034] The order n of the fitting polynomial is 3 (within the range of 2-5; when n=3, the fitted curve has the highest fit to the actual error data, which can fully capture the error variation pattern and avoid the decline in model generalization ability caused by overfitting of high-order polynomials). The fitting formula is E(P)=a3P. 3 +a2P 2 +a1P. Using MATLAB software, the error data for the two intervals are fitted and calculated separately. The polynomial coefficients for each interval are then obtained, yielding the fitted polynomials for the two pressure intervals, which constitute the final compensation model.
[0035]
[0036] Table 2: Polynomial coefficients (%) of one type of pulse oximeter in two pressure ranges
[0037] The goodness-of-fit R² values were all above 0.95, indicating that the fitted curves closely matched the actual error data and could accurately reflect the error variation patterns under different pressures.
[0038] If a single polynomial model is used to fit the error data across the entire range of 0.3N-5N, it cannot accurately capture the distinct error variation patterns of the two ranges simultaneously, leading to large deviations in the model's fit to certain pressure segments and poor compensation. Therefore, a piecewise fitting method avoids the fitting defects of a single model and improves detection accuracy.
[0039] Step S3: Implement compensation
[0040] The compensation model (fitting polynomial) corresponding to the two pressure ranges is programmed into the MCU (microcontroller unit) of the pulse oximeter under test. The pressure detection module built into the MCU can collect the pressure value of the pulse oximeter on the tester's fingertip in real time, and automatically determine the range to which the pressure value belongs. The compensation model of the corresponding range is called to correct the original test value in real time. The correction formula is: Corrected test value = Original test value - E(P) (E(P) is the fitting error value under the current pressure).
[0041] Step S4: Comparison and Verification
[0042] The corrected test values were compared one by one with the actual blood oxygen value Q obtained by the arterial blood gas analyzer, and the absolute error, relative error and error distribution range of the corrected test values were statistically analyzed.
[0043]
[0044] Table 3: Blood oxygenation test error (%) under tester compensation state
[0045] After compensation, the overall error was significantly reduced, and the error of all test data was controlled within the range of ±0.7%–±1.2%, meeting the requirements for high-precision blood oxygen detection.
[0046] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.
Claims
1. A method of detecting analysis of a blood oxygen meter, characterized by, Includes the following steps: S1: Obtain the blood oxygen test dataset ① Obtain the actual blood oxygen value of the test subject, denoted as Q; ② In a dark environment, use a pulse oximeter to test the blood oxygen level of the test subject's fingertip. At the same time, adjust the pressure of the pulse oximeter on the test subject's fingertip so that the measured blood oxygen test value is close to or reaches the test subject's true blood oxygen value. The measured value is recorded as Q1. ③ In a low-light environment, a pulse oximeter is used to test the blood oxygen level of the test subject's fingertip. Based on the different pressures applied by the pulse oximeter to the test subject's fingertip, multiple sets of blood oxygen test values E are obtained. The pressure applied by the pulse oximeter to the test subject's fingertip is recorded as P, thus obtaining the dataset: {(P1, E1), (P2, E2), ..., (Pn, En)}; S2: Component Compensation Model The test error is denoted E(P), E(P) = a n P n +a n-1 P n-1 +a n-2 P n-2 +a1P 1 , the data set is fitted to solve the polynomial coefficients a n -a1, to obtain the fitted polynomial; S3: Implement compensation ① Write the fitted polynomial into the pulse oximeter MCU; ② Perform blood oxygenation tests on multiple test subjects and obtain corrected test values; S4; Compare the corrected test value from S3 with the actual value.
2. The method of claim 1, wherein the blood oxygenation instrument is a pulse oximeter. The pressure P exerted by the pulse oximeter on the test subject's fingertip includes two ranges: 0.3N-1.1N and 1.1N-5N. Corresponding fitting polynomials are obtained for each of the different ranges of pressure P.
3. The method of claim 1, wherein the blood oxygenation instrument is a pulse oximeter. The error range of the test values after correction by the compensation model is ±0.7% to ±1.2%.
4. The method for detecting and analyzing pulse oximeter according to claim 1, characterized in that, The actual blood oxygen value is obtained by sampling blood from the fingertip of the test subject using an arterial blood gas analyzer.
5. The method for detecting and analyzing pulse oximeter according to claim 1, characterized in that, In step S1, the ambient light intensity of the dark environment is ≤5 lux, and the light intensity range of the low light environment is 5 lux-50 lux.
6. The method for detecting and analyzing pulse oximeter according to claim 1, characterized in that, In step S1③, the pressure adjustment gradient is 0.1N-0.2N, and the test is repeated at least 5 times for each pressure value. The average value of the multiple test values is taken as the test value E at the corresponding pressure.
7. The method for detecting and analyzing pulse oximeter according to claim 1, characterized in that, In step S2, the least squares method is used for polynomial fitting, and the order n of the fitted polynomial takes the value of 2-5.
8. The method for detecting and analyzing pulse oximeter according to claim 1, characterized in that, In step S1②, the initial adjustment pressure of the pulse oximeter on the test subject's fingertip is the minimum pressure threshold required for the test, which ranges from 0.3N to 0.5N.
9. The method for detecting and analyzing pulse oximeter according to claim 1, characterized in that, The pulse oximeter integrates a pressure sensor.