Patient blood oxygen self-adaptive monitoring system based on wearable device

Abstract

The invention belongs to the field of blood oxygen self-adaptive monitoring, and particularly discloses a patient blood oxygen self-adaptive monitoring system based on wearable equipment, which comprises a bandage, a display screen, a green light diode, a red light diode, an infrared light diode, a photoelectric detector, a processing chip, a battery and a shell. According to the scheme, the green light diode is introduced into the wearable device, the green light PPG signal is used as the time sequence reference of the red light PPG signal and the infrared light PPG signal, the peak and trough periods of the green light signal are accurately extracted through the increment merging segmentation algorithm in combination with the adaptive threshold updating strategy, and the error rate of period detection is reduced; dynamic elimination of baseline drift is realized by using a multi-stage moving average filter, a PPG signal quality dynamic classification and reconstruction mechanism based on multi-dimensional quality parameters is designed, and the data availability of PPG signals in a motion state is improved.

Description

Technical Field

[0001] The present invention relates to the field of blood oxygen adaptive monitoring, and in particular to a patient blood oxygen adaptive monitoring system based on a wearable device. Background Art

[0002] Adaptive blood oxygen monitoring monitors blood oxygen saturation by adaptively adjusting monitoring-related parameters or modes. Adaptive technology can adjust monitoring parameters according to individual differences or human motion status to ensure accurate measurement of blood oxygen saturation. In wearable blood oxygen monitoring, red light and infrared light are mainly used to obtain PPG signals, but due to the influence of human skin color differences, the peak and trough characteristics of red light PPG signals and infrared PPG signals are blurred; at the same time, wearable devices will generate composite noise when the human body is moving or the wearable device is relaxed, including low-frequency baseline drift caused by breathing and high-frequency motion artifacts caused by strenuous exercise. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a patient blood oxygen adaptive monitoring system based on a wearable device. In view of the technical problem that red light and infrared light are mainly used to obtain PPG signals in wearable blood oxygen monitoring, but the peak and trough characteristics of the red light PPG signal and the infrared PPG signal are blurred due to the influence of human skin color differences, the present solution introduces a green light diode into the wearable device. The reflectivity of green light on the skin surface is high, the signal amplitude is stable, and the periodic characteristics are significant. The green light PPG signal is used as the timing reference of the red light PPG signal and the infrared light PPG signal. The incremental merging and segmentation algorithm is combined with an adaptive threshold update strategy to accurately extract the peak and trough period of the green light signal, enhance the anti-motion interference, and reduce the error rate of period detection. In view of the technical problem that wearable devices generate composite noise when the human body is moving or the wearable device is relaxed, including low-frequency baseline drift caused by breathing and high-frequency motion artifacts caused by strenuous exercise, the present solution uses a multi-stage moving average filter to achieve dynamic elimination of baseline drift, and designs a dynamic classification and reconstruction mechanism of PPG signal quality based on multi-dimensional quality parameters to improve the data availability of PPG signals in motion.

[0004] The technical solution adopted by the present invention is as follows: The present invention provides a patient blood oxygen adaptive monitoring system based on a wearable device, the patient blood oxygen adaptive monitoring system based on a wearable device comprises a strap, a display screen, a green light diode, a red light diode, an infrared light diode, a photoelectric detector, a processing chip, a battery and a housing;

[0005] The battery provides power for the display screen, green light-emitting diode, red light-emitting diode, infrared light-emitting diode, photodetector, and processing chip. The display screen shows the blood oxygen concentration. The green light-emitting diode, red light-emitting diode, and infrared light-emitting diode emit green light, red light, and infrared light respectively. The photodetector is used to obtain PPG signals, including green light PPG signal, red light PPG signal, and infrared light PPG signal. The housing and the display screen form a hollow structure, which wraps the green light-emitting diode, red light-emitting diode, infrared light-emitting diode, photodetector, processing chip, and battery in the hollow structure;

[0006] The processing chip determines the periods of the red light PPG signal and the infrared light PPG signal based on the green light PPG signal, performs preprocessing on the red light PPG signal and the infrared light PPG signal, and calculates the blood oxygen concentration after removing the artifacts.

[0007] Furthermore, the processing chip uses a blood oxygen adaptive monitoring method to calculate the blood oxygen concentration. The blood oxygen adaptive monitoring method specifically includes the following steps:

[0008] Step S1: Eliminate the baseline drift in the green light PPG signal. Use a moving average filter to extract and eliminate the baseline drift signal in the green light PPG signal caused by respiration and sensor contact. The formula used is as follows: ; ;

[0009] In the formula, is the baseline drift signal, is the window length of the moving average filter, is the traversal of the moving average filter, is the green light PPG signal, is the length of the green light PPG signal, is the green light PPG signal after eliminating the baseline drift;

[0010] Step S2: Extract the period of the green light PPG signal. Use the incremental merging segmentation algorithm combined with the adaptive threshold update method to detect the peaks and valleys in the green light PPG signal and obtain the period of the green light PPG signal;

[0011] Step S3: Preprocess the red light PPG signal and the infrared light PPG signal. Use the period of the green light PPG signal as the time window of the moving average filter for the red light PPG signal and the infrared light PPG signal. Use a moving average filter to extract and eliminate the baseline drift of the red light PPG signal and the infrared light PPG signal, use a low-pass filter to eliminate the high-frequency noise of the red light PPG signal and the infrared light PPG signal, and delimit the positions of the peaks and valleys of the red light PPG signal and the infrared light PPG signal based on the positions of the peaks and valleys of the green light PPG signal;

[0012] Step S4: Classify the PPG signal quality, classifying the red-light PPG signal and the infrared PPG signal into high-quality signals, low-quality signals, and artifact signals;

[0013] Step S5: Calculate the blood oxygen concentration by classification.

[0014] Further, in Step S2, the period of extracting the green-light PPG signal specifically includes the following steps:

[0015] Step S21: Divide the green-light PPG signal into continuous line segments in chronological order. Each line segment is linearly connected by a starting point and an ending point. Take the maximum amplitude of the first 2 seconds of the green-light PPG signal. Set the initial lower threshold of the green-light PPG signal amplitude to 0.6 times the maximum amplitude, and calculate the mean value of the historical period amplitude of the green-light PPG signal;

[0016] Step S22: Screen the effective line segments based on the lower threshold. If the average amplitude of the current line segment is greater than or equal to the initial lower threshold and greater than or equal to 0.2 times the mean value of the historical period amplitude, update the initial lower threshold. The formula used is as follows: ;

[0017] In the formula, is the updated lower threshold, is the initial lower threshold, is the average amplitude of the current line segment, is the attenuation coefficient;

[0018] Step S23: Preset the effective detection times threshold to , if times of detection do not screen out effective line segments, attenuate and adjust the lower threshold. The formula used is as follows: ;

[0019] In the formula, is the function for attenuating and adjusting the lower threshold, is the average amplitude of the current line segment, is the step attenuation factor, is the effective detection times threshold;

[0020] Step S24: Calculate the slope of adjacent effective line segments. If the positive and negative signs are the same, merge the line segments. Otherwise, start detecting new line segments. Obtain the peaks and valleys of the green-light PPG signal according to the slope. Define the peak as the starting point where the slope changes from positive to negative, and the valley as the starting point where the slope changes from negative to positive. If the amplitude difference between adjacent peaks and valleys is less than 0.2 times the mean value of the historical period amplitude, mark it as a false peak and remove it;

[0021] Step S25: Two adjacent wave valleys of the green-light PPG signal form a period, and the average period is calculated every 4 periods, which is recorded as the period of the green-light PPG signal.

[0022] Further, in step S4, the classification of the PPG signal quality specifically includes the following steps:

[0023] Step S41: Calculate the quality parameters of the red-light PPG signal and the infrared PPG signal within each period. The quality parameters include kurtosis, entropy, high-frequency segment signal-to-noise ratio, and correlation coefficient.

[0024] The kurtosis reflects the shape information of the PPG signal, and the formula used is as follows: ;

[0025] In the formula, is the kurtosis, is the total number of periods of the red-light PPG signal and the infrared PPG signal, is the traversal of , is the PPG signal, is the mean value of the PPG signal, is the standard deviation of the PPG signal;

[0026] The entropy reflects the uncertainty in the PPG signal, and the formula used is as follows: ;

[0027] In the formula, is the entropy;

[0028] The high-frequency segment signal-to-noise ratio reflects the influence of the background noise of this blood oxygen adaptive monitoring system, and the formula used is as follows: ;

[0029] In the formula, is the high-frequency segment signal-to-noise ratio, is the power spectral density of the current PPG signal, is the power spectral density of the original current PPG signal, is the background noise of this blood oxygen adaptive monitoring system;

[0030] The correlation coefficient uses the green-light PPG signal as a reference to calculate the Pearson correlation coefficients of the red-light PPG signal and the infrared PPG signal with the green-light PPG signal. The formula used is as follows: ;

[0031] In the formula, is the Pearson correlation coefficient, represents the red-light PPG signal and the infrared PPG signal, represents the green PPG signal, is the average value of the red PPG signal and the infrared PPG signal, is the average value of the green PPG signal, is the standard deviation of the red PPG signal and the infrared PPG signal, is the standard deviation of the green PPG signal, represents the expectation;

[0032] Step S42: Preset the quality classification rules for the red PPG signal and the infrared PPG signal, and classify the red PPG signal and the infrared PPG signal into high-quality signals, low-quality signals, and artifact signals based on the quality classification rules;

[0033] Further, in step S5, the blood oxygen concentration is calculated by classification, which specifically includes the following steps:

[0034] Step S51: For high-quality signals, use the optical model based on the Lambert-Beer law to directly calculate the blood oxygen concentration;

[0035] Step S52: For low-quality signals, the photoelectric detector stops capturing the red PPG signal and the infrared PPG signal, the battery increases the current allocated to the red diode and the infrared diode, and after the photoelectric detector recaptures the PPG signal, the post-processing chip analyzes it again;

[0036] Step S53: After segmental reconstruction of the artifact signal, use the optical model based on the Lambert-Beer law to calculate the blood oxygen concentration. The segmental reconstruction specifically includes the following steps:

[0037] Step S531: Obtain the global average red PPG signal and infrared PPG signal, extract the periods of all high-quality signals from the entire red PPG signal and infrared PPG signal, align the peaks of the high-quality signals in each period, and calculate the global high-quality signal mean value after normalization by time;

[0038] Step S532: Obtain the local average red PPG signal and infrared PPG signal, extract 4 periods of high-quality signals before and after the artifact signal respectively, align the peaks of the high-quality signals in each period, and calculate the local high-quality signal mean value after normalization by time;

[0039] Step S533: Perform weighted reconstruction on the artifact signal, and the formula used is as follows: ;

[0040] In the formula, is the reconstruction function of the artifact signal, is the weight of the global high-quality signal mean value, is the global high-quality signal mean value, is the local high-quality signal mean.

[0041] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0042] (1) In wearable blood oxygen monitoring, red light and infrared light are mainly used to obtain PPG signals. However, due to the influence of human skin color differences, the peak and trough characteristics of red light PPG signals and infrared PPG signals are blurred. This solution introduces a green light diode into the wearable device. The reflectivity of green light on the skin surface is high, the signal amplitude is stable, and the periodic characteristics are significant. The green light PPG signal is used as the timing reference of the red light PPG signal and the infrared light PPG signal. The incremental merging and segmentation algorithm is combined with an adaptive threshold update strategy to accurately extract the peak and trough period of the green light signal, enhance the anti-motion interference, and reduce the error rate of period detection.

[0043] (2) To address the technical problem that wearable devices generate complex noise when the human body is exercising or relaxing, including low-frequency baseline drift caused by breathing and high-frequency motion artifacts caused by strenuous exercise, this solution uses a multi-stage moving average filter to dynamically eliminate baseline drift and designs a dynamic classification and reconstruction mechanism for PPG signal quality based on multi-dimensional quality parameters to improve the data availability of PPG signals in motion. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A module connection diagram of a patient blood oxygen adaptive monitoring system based on a wearable device provided by the present invention;

[0045] Figure 2 A flow chart of the steps of a blood oxygen adaptive monitoring method provided by the present invention.

[0046] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0047] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0048] Example 1: See Figure 1, this embodiment provides a patient blood oxygen adaptive monitoring system based on a wearable device. The patient blood oxygen adaptive monitoring system based on a wearable device includes a strap, a display screen, a green light emitting diode, a red light emitting diode, an infrared light emitting diode, a photodetector, a processing chip, a battery, and a housing;

[0049] The battery provides power for the display screen, the green light emitting diode, the red light emitting diode, the infrared light emitting diode, the photodetector, and the processing chip. The display screen displays the blood oxygen concentration. The green light emitting diode, the red light emitting diode, and the infrared light emitting diode respectively emit green light, red light, and infrared light. The photodetector is used to obtain PPG signals, including a green light PPG signal, a red light PPG signal, and an infrared light PPG signal. The housing and the display screen form a hollow structure, which wraps the green light emitting diode, the red light emitting diode, the infrared light emitting diode, the photodetector, the processing chip, and the battery in the hollow structure;

[0050] The processing chip determines the periods of the red light PPG signal and the infrared light PPG signal based on the green light PPG signal, preprocesses the red light PPG signal and the infrared light PPG signal, and calculates the blood oxygen concentration after removing the artifacts.

[0051] Embodiment 2: Refer to Figure 1 and Figure 2 , this embodiment is based on the above embodiment. The processing chip uses a blood oxygen adaptive monitoring method to calculate the blood oxygen concentration. The blood oxygen adaptive monitoring method specifically includes the following steps:

[0052] Step S1: Eliminate the baseline drift in the green light PPG signal. Use a moving average filter to extract and eliminate the baseline drift signal in the green light PPG signal caused by breathing and sensor contact. The formula used is as follows: ; ;

[0053] In the formula, is the baseline drift signal, is the window length of the moving average filter, is the traversal of the moving average filter, is the green light PPG signal, is the length of the green light PPG signal, is the green light PPG signal after eliminating the baseline drift;

[0054] Step S2: Extract the period of the green light PPG signal. Use the incremental merging segmentation algorithm combined with the adaptive threshold update method to detect the peaks and valleys in the green light PPG signal and obtain the period of the green light PPG signal;

[0055] Step S3: Preprocess the red-light PPG signal and the infrared PPG signal. Use the period of the green-light PPG signal as the time window of the moving average filter for the red-light PPG signal and the infrared PPG signal. Extract and eliminate the baseline drift of the red-light PPG signal and the infrared PPG signal using the moving average filter, and eliminate the high-frequency noise of the red-light PPG signal and the infrared PPG signal using a low-pass filter. Based on the positions of the peaks and valleys of the green-light PPG signal, delimit the positions of the peaks and valleys of the red-light PPG signal and the infrared PPG signal;

[0056] Step S4: Classify the quality of the PPG signal. Classify the red-light PPG signal and the infrared PPG signal into high-quality signals, low-quality signals, and artifact signals;

[0057] Step S5: Calculate the blood oxygen concentration by classification.

[0058] Embodiment 3: Refer to Figure 1 and Figure 2 . Based on the above embodiment, in step S2, the extraction of the period of the green-light PPG signal specifically includes the following steps:

[0059] Step S21: Divide the green-light PPG signal into continuous line segments in chronological order. Each line segment is linearly connected by a starting point and an ending point. Take the maximum amplitude of the first 2 seconds of the green-light PPG signal. Set the initial lower threshold of the amplitude of the green-light PPG signal to 0.6 times the maximum amplitude, and calculate the mean value of the historical period amplitudes of the green-light PPG signal;

[0060] Step S22: Screen the valid line segments based on the lower threshold. If the average amplitude of the current line segment is greater than or equal to the initial lower threshold and greater than or equal to 0.2 times the mean value of the historical period amplitudes, update the initial lower threshold. The formula used is as follows: ;

[0061] In the formula, is the updated lower threshold, is the initial lower threshold, is the average amplitude of the current line segment, is the attenuation coefficient;

[0062] Step S23: Preset the effective detection times threshold to . If no valid line segment is screened after detections, perform attenuation adjustment on the lower threshold. The formula used is as follows: ;

[0063] In the formula, is the function for performing attenuation adjustment on the lower threshold, is the average amplitude of the current line segment, is the step attenuation factor, is the threshold of the effective detection times;

[0064] Step S24: Calculate the slope of adjacent effective line segments. If the positive and negative signs are the same, merge the line segments; otherwise, start detecting new line segments. Obtain the peaks and valleys of the green light PPG signal according to the slope. Define the peak as the starting point where the slope changes from positive to negative, and the valley as the starting point where the slope changes from negative to positive. If the amplitude difference between adjacent peaks and valleys is less than 0.2 times the mean of the historical period amplitude, mark it as a pseudo-peak and remove it;

[0065] Step S25: Two adjacent valleys of the green light PPG signal form a period. Calculate the average period every 4 periods, and record it as the period of the green light PPG signal.

[0066] Example 4: Refer to Figure 1 and Figure 2 , based on the above example, in step S4, the PPG signal quality classification specifically includes the following steps:

[0067] Step S41: Calculate the quality parameters of the red light PPG signal and the infrared PPG signal in each period. The quality parameters include kurtosis, entropy, high-frequency band signal-to-noise ratio, and correlation coefficient;

[0068] The kurtosis reflects the shape information of the PPG signal, and the formula used is as follows: ;

[0069] In the formula, is the kurtosis, is the total number of periods of the red light PPG signal and the infrared PPG signal, is the traversal of , is the PPG signal, is the mean of the PPG signal, is the standard deviation of the PPG signal;

[0070] The entropy reflects the uncertainty in the PPG signal, and the formula used is as follows: ;

[0071] In the formula, is the entropy;

[0072] The high-frequency band signal-to-noise ratio reflects the influence of the background noise of this blood oxygen adaptive monitoring system, and the formula used is as follows: ;

[0073] In the formula, is the high-frequency band signal-to-noise ratio, is the power spectral density of the current PPG signal, is the power spectral density of the original current PPG signal, is the background noise of this blood oxygen adaptive monitoring system;

[0074] The correlation coefficient uses the green light PPG signal as a reference, and calculates the Pearson correlation coefficients of the red light PPG signal and the infrared PPG signal with the green light PPG signal. The formula used is as follows: ;

[0075] In the formula, is the Pearson correlation coefficient, represents the red light PPG signal and the infrared PPG signal, represents the green light PPG signal, is the average value of the red light PPG signal and the infrared PPG signal, is the average value of the green light PPG signal, is the standard deviation of the red light PPG signal and the infrared PPG signal, is the standard deviation of the green light PPG signal, represents the expectation;

[0076] Step S42: Preset the quality classification rules for the red light PPG signal and the infrared PPG signal, and classify the red light PPG signal and the infrared PPG signal into high-quality signals, low-quality signals, and artifact signals based on the quality classification rules.

[0077] Example 5: Refer to Figure 1 and Figure 2 , this example is based on the above example. In step S5, the blood oxygen concentration is classified and calculated, which specifically includes the following steps:

[0078] Step S51: For high-quality signals, use the optical model based on Lambert-Beer's law to directly calculate the blood oxygen concentration;

[0079] Step S52: For low-quality signals, the photoelectric detector stops capturing the red light PPG signal and the infrared PPG signal, the battery increases the current allocated to the red light diode and the infrared light diode, and after the photoelectric detector recaptures the PPG signal, the post-processing chip analyzes it again;

[0080] Step S53: After segmental reconstruction of the artifact signal, use the optical model based on Lambert-Beer's law to calculate the blood oxygen concentration. The segmental reconstruction specifically includes the following steps:

[0081] Step S531: Obtain the global average red PPG signal and the infrared PPG signal, extract the periods of all high-quality signals from the entire red PPG signal and the infrared PPG signal, align the peaks of the high-quality signals in each period, and calculate the global high-quality signal mean value after normalization by time;

[0082] Step S532: Obtain the local average red PPG signal and the infrared PPG signal, extract 4 periods of high-quality signals before and after the artifact signal respectively, align the peaks of the high-quality signals in each period, and calculate the local high-quality signal mean value after normalization by time;

[0083] Step S533: Perform weighted reconstruction on the artifact signal, and the formula used is as follows: ;

[0084] In the formula, is the reconstruction function of the artifact signal, is the weight of the global high-quality signal mean value, is the global high-quality signal mean value, is the local high-quality signal mean value.

[0085] Example Six: Based on the above example, in step S1, the moving average filter is used to eliminate the baseline drift signal caused by breathing and sensor contact in the green PPG signal, and the window size of the moving average filter is set to 0.55 seconds;

[0086] In step S21, when extracting the period of the green PPG signal, is the attenuation coefficient, taking 0.8, which controls the weight of the historical lower threshold, is the step attenuation coefficient, taking 0.1, which is used to gradually relax the threshold.

[0087] Example Seven: Based on the above example, in step S3, based on the positions of the peaks and valleys of the green PPG signal, the positions of the peaks and valleys of the red PPG signal and the infrared PPG signal are delimited, specifically: the positions of the peaks and valleys of the red PPG signal and the infrared PPG signal are the same as those of the green PPG signal.

[0088] Example Eight: Based on the above example, in step S42, a quality classification rule for the red PPG signal and the infrared PPG signal is preset, and based on the quality classification rule, the red PPG signal and the infrared PPG signal are classified into high-quality signals, low-quality signals, and artifact signals;

[0089] The quality classification rule is as follows:

[0090] High-quality signal: kurtosis less than 3.6, entropy not exceeding 3-4 times the historical mean, signal-to-noise ratio in the high-frequency band greater than or equal to 10, correlation coefficient greater than or equal to 0.72;

[0091] Artifact signal (meeting any one condition): kurtosis less than 2 or greater than 4, entropy exceeding 3-4 times the historical mean, correlation coefficient less than 0.5;

[0092] Poor-quality signal: signal-to-noise ratio in the high-frequency band less than or equal to 10, kurtosis greater than 3.6 and less than or equal to 4, correlation coefficient greater than or equal to 0.5 and less than 0.72.

[0093] It should be noted that in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device.

[0094] Although embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made in these embodiments without departing from the principles and spirit of the invention, and the scope of the invention is defined by the appended claims and their equivalents.

[0095] The above describes the present invention and its embodiments, and this description is not restrictive. What is shown in the drawings is only one of the embodiments of the present invention, and the actual structure is not limited thereto. Generally speaking, if those of ordinary skill in the art are inspired by it and design similar structural forms and embodiments without creative efforts without departing from the purpose of the present invention, they shall fall within the protection scope of the present invention.

Claims

1. A patient blood oxygen adaptive monitoring system based on a wearable device, characterized in that, It includes a strap, a display screen, a green light-emitting diode, a red light-emitting diode, an infrared light-emitting diode, a photodetector, a processing chip, a battery, and a housing; The battery provides power for the display screen, the green light-emitting diode, the red light-emitting diode, the infrared light-emitting diode, the photodetector, and the processing chip. The display screen shows the blood oxygen concentration. The green light-emitting diode, the red light-emitting diode, and the infrared light-emitting diode emit green light, red light, and infrared light respectively. The photodetector is used to obtain PPG signals, including a green light PPG signal, a red light PPG signal, and an infrared light PPG signal. The housing and the display screen form a hollow structure that wraps the green light-emitting diode, the red light-emitting diode, the infrared light-emitting diode, the photodetector, the processing chip, and the battery in the hollow structure; The processing chip determines the periods of the red light PPG signal and the infrared light PPG signal based on the green light PPG signal, uses the green light PPG signal as the timing reference for the red light PPG signal and the infrared light PPG signal, combines the incremental merging segmentation algorithm with the adaptive threshold update strategy to extract the peak-valley period of the green light signal, preprocesses the red light PPG signal and the infrared light PPG signal, uses a multi-stage moving average filter to dynamically eliminate baseline drift, and calculates the blood oxygen concentration after removing artifacts.

2. The patient blood oxygen adaptive monitoring system based on a wearable device according to claim 1, wherein, The processing chip uses a blood oxygen adaptive monitoring method to calculate the blood oxygen concentration. The blood oxygen adaptive monitoring method specifically includes the following steps: Step S1: Eliminate the baseline drift in the green light PPG signal, and use a moving average filter to extract and eliminate the baseline drift signal in the green light PPG signal caused by breathing and sensor contact; Step S2: Extract the period of the green light PPG signal, and use the incremental merging segmentation algorithm combined with the adaptive threshold update method to obtain the period of the green light PPG signal; Step S3: Preprocess the red light PPG signal and the infrared light PPG signal. Use the period of the green light PPG signal as the time window of the moving average filter for the red light PPG signal and the infrared light PPG signal. Use a moving average filter to extract and eliminate the baseline drift of the red light PPG signal and the infrared light PPG signal. Use a low-pass filter to eliminate the high-frequency noise of the red light PPG signal and the infrared light PPG signal. Based on the positions of the peaks and valleys of the green light PPG signal, delimit the positions of the peaks and valleys of the red light PPG signal and the infrared light PPG signal; Step S4: Classify the quality of the PPG signals, and classify the red light PPG signal and the infrared light PPG signal into high-quality signals, low-quality signals, and artifact signals; Step S5: Calculate the blood oxygen concentration by classification.

3. The patient blood oxygen adaptive monitoring system based on a wearable device according to claim 2, characterized in that, In step S2, the extraction of the period of the green light PPG signal specifically includes the following steps: Step S21: Divide the green light PPG signal into continuous line segments in chronological order. Each line segment is linearly connected by a starting point and an ending point. Take the maximum amplitude of the first 2 seconds of the green light PPG signal. Set the initial lower threshold of the amplitude of the green light PPG signal to 0.6 times the maximum amplitude, and calculate the mean value of the historical period amplitude of the green light PPG signal; Step S22: Screen the valid line segments based on the lower threshold. If the average amplitude of the current line segment is greater than or equal to the initial lower threshold and greater than or equal to 0.2 times the mean value of the historical period amplitude, then update the initial lower threshold; Step S23: Preset the effective detection times threshold to be , if times of detection fail to screen out effective line segments, then decay and adjust the lower threshold; Step S24: Calculate the slopes of adjacent valid line segments. If the positive and negative signs are the same, merge the line segments; otherwise, start detecting new line segments. Obtain the peaks and valleys of the green-light PPG signal based on the slopes. If the amplitude difference between adjacent peaks and valleys is less than 0.2 times the mean of the historical period amplitudes, mark it as a false peak and remove it. Step S25: Two adjacent valleys of the green-light PPG signal form a period. Calculate the average period every 4 periods, which is recorded as the period of the green-light PPG signal.

4. The patient blood oxygen adaptive monitoring system based on a wearable device according to claim 3, wherein In step S4, the PPG signal quality classification specifically includes the following steps: Step S41: Calculate the quality parameters of the red-light PPG signal and the infrared PPG signal within each period. The quality parameters include kurtosis, entropy, high-frequency segment signal-to-noise ratio, and correlation coefficient. Step S42: Preset the quality classification rules for the red-light PPG signal and the infrared PPG signal. Based on the quality classification rules, classify the red-light PPG signal and the infrared PPG signal into high-quality signals, low-quality signals, and artifact signals.

5. The patient blood oxygen adaptive monitoring system based on a wearable device according to claim 4, characterized in that, In step S5, calculate the blood oxygen concentration by classification, specifically including the following steps: Step S51: For high-quality signals, use the optical model based on Lambert-Beer's law to directly calculate the blood oxygen concentration. Step S52: For low-quality signals, the photoelectric detector stops capturing the red-light PPG signal and the infrared PPG signal, the battery increases the current allocated to the red-light diode and the infrared-light diode, and after the photoelectric detector recaptures the PPG signal, the post-processing chip analyzes it again. Step S53: After segmental reconstruction of the artifact signal, use the optical model based on Lambert-Beer's law to calculate the blood oxygen concentration. The segmental reconstruction specifically includes the following steps: Step S531: Calculate the global high-quality signal mean. Step S532: Calculate the local high-quality signal mean. Step S533: Perform weighted reconstruction on the artifact signal.

Images