A multi-variable fusion detection method for neonatal respiratory distress monitoring

Abstract

This invention discloses a multivariate fusion detection method for neonatal respiratory distress monitoring, belonging to the field of variable detection technology. It addresses the problem of inaccurately reflecting changes in true respiratory rhythm under scenarios with sudden changes in light intensity or frequent occlusion. By setting the light source observation time and dividing it into luminous and non-luminous phases, the method extracts the photoelectric response voltage to achieve temporal separation of ambient light and detection light. A photoelectric receiving occupancy level is constructed based on probe contact pressure, and the driving current of the light-emitting diode is compensated and adjusted based on this occupancy level. Ambient light irradiance data is collected under compensated driving conditions. Furthermore, the frequency of ambient light occlusion is analyzed by varying irradiance, and a joint judgment index is constructed by combining this with the respiratory waveform amplitude. This adaptively selects between conventional monitoring and compensated maintenance processing, achieving stable acquisition and quality improvement of respiratory signals under complex lighting and contact conditions.

Description

Technical Field

[0001] This invention relates to the field of variable detection technology, and more specifically, to a multivariate fusion detection method for monitoring neonatal respiratory distress. Background Technology

[0002] In the field of neonatal intensive care, continuous monitoring of neonatal respiratory distress syndrome usually relies on photoplethysmography (PPG) optical sensing technology. This technology uses light-emitting diodes and photodetectors to construct reflective or transmissive detection structures to collect and analyze weak blood oxygen or respiratory-related signals.

[0003] The existing technology has the following shortcomings:

[0004] Currently, existing technologies use a continuous driving method to acquire photoelectric response signals during photoelectric signal acquisition, without dividing the emission stage and the ambient light interaction stage into time sequences. This results in the ambient light component and the detection light component being superimposed and mixed within the same sampling window, making it difficult to effectively isolate ambient light disturbances and establish a stable baseline reference. Consequently, in scenarios with sudden changes in light intensity or frequent changes in occlusion, the respiratory waveform amplitude drifts or even becomes distorted, and it cannot accurately reflect the actual changes in respiratory rhythm. Furthermore, the lack of quantitative modeling of the effective light receiving area in conjunction with changes in probe contact pressure leads to a lack of basis for adjusting the driving current, further amplifying signal fluctuations and monitoring instability. Therefore, a multivariate fusion detection method for neonatal respiratory distress monitoring is proposed. Summary of the Invention

[0005] To overcome the aforementioned deficiencies in the prior art, embodiments of the present invention provide a multivariate fusion detection method for neonatal respiratory distress monitoring. This method employs a photoelectric response extraction mechanism based on emission timing separation, a photoelectric receiver occupancy level modeling mechanism constructed by combining probe contact pressure, and an adaptive compensation mechanism for LED driving current based on occupancy level. Furthermore, it integrates a joint determination mechanism of occupancy frequency analysis based on light irradiance changes and respiratory waveform amplitude to adaptively switch monitoring strategies, thereby solving the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a multivariate fusion detection method for monitoring neonatal respiratory distress, comprising the following steps:

[0007] Step S1: Set the light source observation time, read the LED driving information of the photodetector in the monitored area under test during the light source observation time, divide the light source observation time using the LED driving information and obtain the photoelectric response voltage of the photodetector.

[0008] Step S2: Evaluate the ambient light baseline characteristics of the photodetector based on the photoelectric response voltage, collect the probe contact pressure of the photodetector, and generate the photoelectric receiving occupancy level of the monitored area by combining the probe contact pressure and the ambient light baseline characteristics.

[0009] Step S3: Detect the peak value of the diode current of the photodetector, calculate the occupancy callback coefficient using the photodetector occupancy level and adjust the peak value of the diode current, set the drive compensation period and collect the light irradiance data of the monitored area.

[0010] Step S4: Based on the light irradiance data, the frequency of ambient light occlusion is counted, and an irradiation occlusion index is generated according to the frequency of ambient light occlusion. The amplitude of the respiratory waveform is obtained through the respiratory monitoring channel, and the conventional respiratory monitoring treatment or the irradiation maintenance treatment is selected in combination with the irradiation occlusion index.

[0011] In a preferred embodiment, in step S1, a light source observation time is preset. During the light source observation time, the light-emitting diode driving information of the light-emitting diode is read through the drive control interface of the photoelectric monitoring probe, including the driving pulse turn-on time and the driving pulse turn-off time.

[0012] The driving pulse turn-on and driving pulse turn-off times recorded during the observation time of the light source are sorted in chronological order, and the time interval between each pair of adjacent driving pulse turn-on and driving pulse turn-off times is taken as the emission stage.

[0013] The time interval formed between the turn-off moment of the driving pulse after the end of the light emission phase and the turn-on moment of the next driving pulse is defined as the non-light emission phase;

[0014] The photoelectric response voltage is obtained through the photoelectric signal output terminal of the photodetector.

[0015] In a preferred embodiment, in step S2, the photoelectric response voltages corresponding to the luminescence stage and the non-luminescence stage are combined in time order to form a luminescence response voltage sequence and a non-luminescence response voltage sequence.

[0016] The median of each photoelectric response voltage in the non-luminescent response voltage sequence was selected as the ambient light baseline value, and the maximum value of each photoelectric response voltage in the luminescent response voltage sequence was selected as the detection superposition peak value.

[0017] The ratio of the ambient light baseline value to the detected superimposed peak value is used as the ambient light baseline feature.

[0018] In a preferred embodiment, in step S2, the probe bonding pressure is collected based on the flexible pressure sensor in the photoelectric monitoring probe, and the bonding pressures of each probe are combined in time order to form a probe bonding pressure sequence.

[0019] The pressure change is obtained by performing adjacent difference on the bonding pressure sequence. The center value of pressure change and the center value of pressure deviation are calculated based on the median and deviation of the pressure change.

[0020] The upper limit and lower limit of the pressure change threshold are calculated by combining the center value of pressure change and the center value of pressure deviation.

[0021] The pressure changes that fall within the upper and lower limits of the pressure change threshold are marked, and the probe contact pressure corresponding to the marked pressure changes is also marked.

[0022] The bonding pressure of adjacent probes is combined into candidate stable bonding intervals, and the candidate stable bonding interval with the longest duration is selected as the stable bonding interval.

[0023] Within the stable bonding range, the median of the probe bonding pressure is selected as the bonding reference pressure value.

[0024] In a preferred embodiment, in step S2, the reference bonding pressure value is retrieved from the probe calibration library, and the bonding comparison coefficient is calculated by combining the bonding reference pressure value and the reference bonding pressure value.

[0025] The bonding mapping factor is calculated based on the bonding comparison coefficient, and the product of the bonding mapping factor and the ambient light baseline feature is used as the receiving occupancy correction feature.

[0026] The photoelectric receiving occupancy correction feature is compared with the preset first occupancy threshold and the preset second occupancy threshold to generate the photoelectric receiving occupancy level of the monitored area to be tested;

[0027] If the received occupancy correction feature is less than or equal to the preset first occupancy threshold, a low occupancy received layer is generated.

[0028] If the received occupancy correction feature is greater than the preset first occupancy threshold and less than the preset second occupancy threshold, then a medium occupancy receiving layer is generated.

[0029] If the receive occupancy correction feature is greater than or equal to the preset second occupancy threshold, a high occupancy receive layer is generated.

[0030] In a preferred embodiment, in step S3, under the current driving conditions, the instantaneous current in the LED driving branch is continuously sampled by a current acquisition device to form a current time series.

[0031] Extract the maximum value of the current time series to obtain the diode current peak value;

[0032] A piecewise continuous function is used to assign corresponding callback coefficient calculation functions to different photoelectric receiver occupancy levels, thereby generating occupancy callback coefficients.

[0033] The peak value of the diode current is calculated by multiplying the peak value of the diode current based on the placeholder callback coefficient, and the compensated target drive current peak value is obtained.

[0034] After obtaining the target drive current peak value, the drive control parameters are updated so that the light-emitting diodes can output pulses according to the compensated target drive current peak value in subsequent drive cycles.

[0035] In a preferred embodiment, in step S3, after the drive parameter adjustment is completed, a drive compensation period is set. The drive compensation period refers to the time interval during which the compensation drive state is maintained to evaluate the compensation effect after the drive current compensation is performed.

[0036] During the drive compensation period, ambient light is collected in real time by light irradiance sensing units deployed in the monitoring area to form light irradiance data, and a light irradiance sequence is constructed in chronological order.

[0037] Irradiance data is the intensity of ambient light energy received per unit area.

[0038] In a preferred embodiment, in step S4, the difference between the light irradiance data at adjacent sampling times in the light irradiance sequence is calculated and the absolute value is taken to obtain the light irradiance change and construct a light irradiance change sequence.

[0039] The median of the sequence of changes in light irradiance was selected as the center value of the change.

[0040] Calculate the absolute value of the deviation between the change in light irradiance and the center value of the change and form a deviation sequence. Take the median of the deviation sequence to obtain the center value of the deviation. Use the sum of the center value of the deviation and the center value of the change as the threshold for judging the change in ambient light.

[0041] The sampling points in the light irradiance change sequence that satisfy the value greater than the ambient light change judgment threshold are marked as occlusion trigger points;

[0042] Each occlusion trigger point corresponds to one ambient light occlusion event. The total number of occlusion trigger points is counted during the drive compensation period to obtain the ambient light occlusion frequency.

[0043] The mean and standard deviation of the irradiance sequence during the driving compensation period were further calculated to obtain the average irradiance and the standard deviation of irradiance.

[0044] In a preferred embodiment, in step S4, an illumination shading index is constructed based on the ambient light shading frequency, average light irradiance, and standard deviation of light irradiance.

[0045] The respiratory waveform is obtained from the photoelectric response signal through the respiratory monitoring channel, and the respiratory waveform is bandpass filtered to extract the respiratory frequency band component to obtain the respiratory waveform sequence.

[0046] Peak-valley analysis is performed on the respiratory waveform sequence to extract the peak and trough values ​​in each respiratory cycle. The peak value is subtracted from the trough value to obtain the respiratory amplitude of each cycle.

[0047] The respiratory amplitudes are arranged into an amplitude sequence according to time order, and the median is taken as the characteristic value of the respiratory waveform amplitude. The respiratory signal quality coefficient is constructed based on the characteristic value of the respiratory waveform amplitude.

[0048] A joint judgment index was constructed based on the radiation occlusion index and the respiratory signal quality coefficient.

[0049] When the joint judgment index is greater than or equal to the preset judgment threshold, the judgment is selected to perform conventional respiratory monitoring processing. Conventional respiratory monitoring processing refers to the processing method of directly performing feature analysis and status judgment on the respiratory waveform output by the respiratory monitoring channel.

[0050] When the joint judgment index is less than the preset judgment threshold, the judgment selects the compensation irradiation maintenance treatment. The compensation irradiation maintenance treatment refers to the enhanced treatment method adopted to maintain the continuity of monitoring.

[0051] The technical effects and advantages of this invention are as follows:

[0052] This invention achieves temporal separation of ambient light and detection light by setting the light source observation time and dividing it into luminous and non-luminous phases, extracting the photoelectric response voltage, constructing a photoelectric receiving occupancy level based on probe contact pressure, and compensating and adjusting the LED driving current based on the occupancy level. Ambient light irradiance data is collected under compensated driving conditions. Furthermore, the frequency of ambient light obstruction is analyzed by varying irradiance, and a joint judgment index is constructed by combining it with respiratory waveform amplitude. This allows for adaptive selection of conventional monitoring or compensated maintenance processing, achieving stable acquisition and quality improvement of respiratory signals under complex lighting and contact conditions. Attached Figure Description

[0053] Figure 1 This is a flowchart illustrating the implementation of a multivariate fusion detection method for monitoring neonatal respiratory distress according to the present invention.

[0054] Figure 2 This is a schematic diagram illustrating the steps of a multivariate fusion detection method for monitoring neonatal respiratory distress according to the present invention. Detailed Implementation

[0055] 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.

[0056] This invention achieves temporal separation of ambient light and detection light by setting the observation time of the light source and dividing it into luminous and non-luminous phases, extracting the photoelectric response voltage, constructing a photoelectric receiver occupancy level based on the probe's contact pressure, and compensating and adjusting the LED drive current based on the occupancy level. Ambient light irradiance data is collected under compensated drive conditions. Furthermore, the frequency of ambient light obstruction is analyzed by examining changes in irradiance, and a joint judgment index is constructed by combining this with the respiratory waveform amplitude to adaptively select between conventional monitoring and compensated maintenance processing.

[0057] Example 1, as Figures 1 to 2 As shown, a multivariate fusion detection method for monitoring neonatal respiratory distress includes the following steps:

[0058] Step S1: Set the light source observation time, read the LED driving information of the photodetector in the monitored area under test during the light source observation time, divide the light source observation time using the LED driving information and obtain the photoelectric response voltage of the photodetector.

[0059] Step S2: Evaluate the ambient light baseline characteristics of the photodetector based on the photoelectric response voltage, collect the probe contact pressure of the photodetector, and generate the photoelectric receiving occupancy level of the monitored area by combining the probe contact pressure and the ambient light baseline characteristics.

[0060] Step S3: Detect the peak value of the diode current of the photodetector, calculate the occupancy callback coefficient using the photodetector occupancy level and adjust the peak value of the diode current, set the drive compensation period and collect the light irradiance data of the monitored area.

[0061] Step S4: Based on the light irradiance data, the frequency of ambient light occlusion is counted, and an irradiation occlusion index is generated according to the frequency of ambient light occlusion. The amplitude of the respiratory waveform is obtained through the respiratory monitoring channel, and the conventional respiratory monitoring treatment or the irradiation maintenance treatment is selected in combination with the irradiation occlusion index.

[0062] The specific implementation is as follows:

[0063] In step S1, a preset light source observation time is used to observe the working status of the light-emitting unit of the photoelectric monitoring probe and the optical signal response of the photoelectric receiving link;

[0064] During the observation time of the light source, the LED driving information is read through the drive control interface of the photoelectric monitoring probe. The LED driving information refers to the actual light-emitting working state of the LED during the observation time of the light source, including the driving pulse turn-on time and the driving pulse turn-off time.

[0065] The driving pulse turn-on time is the time when the light-emitting diode starts to emit detection light, and the driving pulse turn-off time is the time when the light-emitting diode stops emitting detection light.

[0066] The driving pulse turn-on and driving pulse turn-off times recorded during the observation time of the light source are sorted in chronological order, and the time interval between each pair of adjacent driving pulse turn-on and driving pulse turn-off times is taken as the emission stage.

[0067] The time interval formed between the turn-off moment of the driving pulse after the end of the light emission phase and the turn-on moment of the next driving pulse is defined as the non-light emission phase;

[0068] The light-emitting phase refers to the time period during which the light-emitting diode is in the on state and emits detection light; the non-light-emitting phase refers to the time period during which the light-emitting diode is in the off state.

[0069] The photoelectric response voltage is obtained through the photoelectric signal output terminal of the photoelectric detector. The photoelectric response voltage is the voltage signal corresponding to the intensity of the incident light received by the photoelectric detector, reflecting the intensity of the light signal received by the photoelectric detector.

[0070] During the luminescent phase, the photoelectric response voltage corresponds to the response voltage generated by the superposition of the received light signal (formed by the detection light emitted by the light-emitting diode after being scattered or reflected by the skin tissue) and the ambient light signal. During the non-luminescent phase, the photoelectric response voltage mainly corresponds to the background response voltage formed by the ambient light in the monitoring environment shining on the photodetector.

[0071] It should be noted that the preset light source observation time can be set according to the driving modulation cycle of the light-emitting diode in the photoelectric monitoring probe and the photoelectric signal sampling frequency; the photoelectric monitoring probe is a monitoring component used to collect photoelectric signals from the monitored area, including a photodetector and a light-emitting unit. The photodetector is a photosensitive device used to receive light signals scattered or reflected by skin tissue and generate photocurrent. The light-emitting unit is a light source component used to emit detection light to the monitored area, including a light-emitting diode and its driving circuit; the driving control interface is used to read the driving status information of the light-emitting diode during operation.

[0072] In step S2, during the acquisition process, the light-emitting unit of the photoelectric monitoring probe emits detection light, and the photodetector receives the returned signal. External ambient light will be superimposed into the receiving channel, causing the signal baseline to rise. At the same time, changes in the probe's bonding state will change the optical path coupling conditions and affect the receiving stability. Therefore, it is necessary to jointly identify the influence of ambient light and the bonding state.

[0073] The photoelectric response voltages corresponding to the luminescence stage and the non-luminescence stage are combined in time order to form luminescence response voltage sequences and non-luminescence response voltage sequences, respectively.

[0074] The median of each photoelectric response voltage in the non-luminescent response voltage sequence is selected as the ambient light baseline value to reflect the background level of ambient light in the photoelectric receiving link under external light source illumination conditions.

[0075] The maximum value of each photoelectric response voltage in the emission response voltage sequence is selected as the detection superposition peak value, which reflects the highest response level that the photoelectric receiving link can achieve under the action of detection light.

[0076] The ambient light baseline characteristic is obtained by calculating the ratio between the ambient light baseline value and the detected superimposed peak value. The larger the value, the greater the degree of baseline rise of the photoelectric receiving link by the ambient light, and the smaller the effective space in the receiving channel that can be used to detect changes in the optical signal. The smaller the value, the lower the background proportion of the ambient light in the photoelectric receiving signal, and the more sufficient the dynamic range of the detected optical signal in the receiving link.

[0077] The probe contact pressure is collected by the flexible pressure sensor in the photoelectric monitoring probe. The probe contact pressure refers to the pressure value formed by the contact force generated by the skin tissue on the probe contact area, which is detected and converted by the flexible pressure sensor.

[0078] The probe contact pressure during the light source observation time is combined in chronological order to form a probe contact pressure sequence.

[0079] In the bonding pressure sequence, the pressure change is obtained by subtracting the bonding pressure of adjacent probes and taking the absolute value. The median of the pressure change is taken as the center value of the pressure change. The absolute value of the difference between each pressure change and the center value of the pressure change is taken as the pressure change deviation value. The product of the median of the pressure change deviation value and the preset pressure correction coefficient is taken as the center value of the pressure deviation.

[0080] The lower limit of the pressure change threshold is obtained by subtracting the center value of the pressure deviation from the center value of the pressure change, and the upper limit of the pressure change threshold is obtained by adding the center value of the pressure change to the center value of the pressure deviation.

[0081] Furthermore, the pressure changes that fall within the lower limit and upper limit of the pressure change threshold are marked, and the probe contact pressure corresponding to the marked pressure changes is also marked.

[0082] The bonding pressure of adjacent probes is combined into candidate stable bonding intervals, and the candidate stable bonding interval with the longest duration is selected as the stable bonding interval.

[0083] The probe bonding pressure within the stable bonding range is statistically processed. Specifically, the bonding pressure of each probe within the stable bonding range is sorted according to its numerical value, and the pressure value at the middle position after sorting is selected as the bonding reference pressure value.

[0084] Access the probe calibration library to retrieve the reference bonding pressure value. The reference bonding pressure value is the target contact pressure value determined by calibration experiments under standard bonding conditions. The bonding reference pressure value and the reference bonding pressure value are subtracted and the absolute value is taken to obtain the bonding pressure deviation value. The ratio of the bonding pressure deviation value to the reference bonding pressure value is used as the bonding comparison coefficient.

[0085] The larger the bonding comparison coefficient, the more obvious the probe bonding is or the more obvious the pressure bonding is, and the higher the sensitivity of the photoelectric receiving link to ambient light; the smaller the bonding comparison coefficient, the closer the probe bonding state is to the standard bonding state, and the better the optical path sealing.

[0086] Calculate the bonding mapping factor based on the bonding comparison coefficient: ,in, To match the comparison coefficient, To preset the mapping weights, To fit the mapping factor, e is the natural constant;

[0087] The product of the bonding mapping factor and the ambient light baseline feature is used as the receiver occupancy correction feature, which reflects the actual effective occupancy of the photoelectric receiver link by the ambient light under the current bonding state.

[0088] The photoelectric receiving occupancy correction feature is compared with the preset first occupancy threshold and the preset second occupancy threshold to generate the photoelectric receiving occupancy level of the monitored area to be tested;

[0089] If the received occupancy correction feature is less than or equal to the preset first occupancy threshold, a low occupancy received layer is generated.

[0090] If the received occupancy correction feature is greater than the preset first occupancy threshold and less than the preset second occupancy threshold, then a medium occupancy receiving layer is generated.

[0091] If the received occupancy correction feature is greater than or equal to the preset second occupancy threshold, a high occupancy received layer is generated.

[0092] A low occupancy reception level indicates that the ambient light occupancy is low and the bonding state is stable, and the detected optical signal has sufficient dynamic range; a medium occupancy reception level indicates that the ambient light or bonding deviation has affected the receiving link, and the detection space is compressed but can be compensated for; a high occupancy reception level indicates that the ambient light occupancy or bonding deviation is significant, and the receiving dynamic range is limited.

[0093] It should be noted that the flexible pressure sensor is a surface-contact pressure sensing device constructed based on a flexible substrate material; the preset pressure correction coefficient can be set according to the measurement accuracy of the flexible pressure sensor; the probe calibration library is a set of calibration data used to store reference bonding pressure values ​​corresponding to different monitoring sites and different probe structures; the preset mapping weight can be set according to the calibration results of the probe bonding structure's sensitivity to ambient light; the preset first occupancy threshold and the preset second occupancy threshold can be set according to the correspondence between the degree of photoelectric signal distortion and the ambient light occupancy characteristics in historical monitoring data.

[0094] In step S3, the peak value of the diode current of the photodetector is detected. Specifically, under the current driving conditions, the instantaneous current in the LED driving branch is continuously sampled by a current acquisition device to form a current time series, and the maximum value of the current time series is extracted to obtain the diode current peak value.

[0095] It should be noted that the current acquisition device is a current detection unit set in the LED driving branch, used to measure the driving current during LED conduction in real time and output the current signal.

[0096] Among them, the peak diode current characterizes the maximum luminous intensity of the light-emitting diode under the current driving conditions. The larger the value, the higher the detection light energy emitted to the skin tissue per unit time, and the larger the amplitude of the effective photoelectric response signal formed on the photodetector side.

[0097] After obtaining the photoelectric receiver placeholder level, placeholder callback coefficients are constructed based on the photoelectric receiver placeholder level. Specifically, a piecewise continuous function is used to assign a corresponding callback coefficient calculation function to different photoelectric receiver placeholder levels to generate placeholder callback coefficients. The expression for the callback coefficient calculation function is as follows:

[0098]

[0099] in, This is a placeholder callback coefficient. To receive placeholder correction features, and These are the preset first placeholder threshold and the preset second placeholder threshold, respectively. This is a linear adjustment coefficient, reflecting the growth rate of the drive compensation within the mid-occupancy interval. A larger value indicates a more sensitive response to ambient light rise. This is a nonlinear enhancement coefficient used to amplify the compensation amplitude in high-occupancy regions. The larger its value, the more significant the driving improvement is under severe occupancy conditions.

[0100] The placeholder callback coefficient is used to characterize the required drive compensation ratio under the current degree of occupancy of the receiving channel by ambient light. The larger the value, the more serious the occupancy of the receiving dynamic range by ambient light, and the more powerful the drive compensation needs to be applied to the light-emitting diode to restore the effective detection signal ratio.

[0101] The peak diode current is adjusted based on the placeholder callback coefficient to obtain the compensated target drive current peak value, and its calculation expression is as follows:

[0102] ;

[0103] in, The target drive current peak value after compensation. This represents the peak value of the diode current. This is a placeholder callback coefficient.

[0104] The occupancy callback coefficient represents the target driving intensity required to ensure that the output signal of the photodetector is within the identifiable range under the current ambient light occupancy state. The larger the value, the more necessary it is to enhance the output of the light-emitting unit to increase the proportion of the effective signal relative to the ambient light baseline.

[0105] After obtaining the target drive current peak value, the drive control parameters are updated so that the light-emitting diodes can output pulses according to the compensated target drive current peak value in subsequent drive cycles.

[0106] Meanwhile, to avoid excessive driving current causing the photodetector to enter saturation or increasing the thermal load on the device, upper and lower limits are set for the peak value of the target driving current, and amplitude limiting is performed. The upper limit is the maximum allowable value of the safe operating current of the light-emitting diode, and the larger the value, the stronger the maximum allowable driving capability. The lower limit is the minimum driving requirement to maintain effective detection light output, and the smaller the value, the more basic detection capability can be maintained under low power consumption.

[0107] After completing the drive parameter adjustment, set the drive compensation period. The drive compensation period refers to the time interval during which the compensated drive state is maintained to evaluate the compensation effect after the drive current compensation is performed.

[0108] During the drive compensation period, ambient light is collected in real time by light irradiance sensing units deployed in the monitoring area to form light irradiance data. The light irradiance sequence is constructed in chronological order. The light irradiance data is the ambient light energy intensity received per unit area. The larger the value, the higher the irradiance intensity of the phototherapy lamp and the more obvious the lifting effect on the receiving baseline of the photodetector.

[0109] It should be noted that the light irradiance sensing unit is an optical sensing component deployed in the monitoring area to detect the intensity of ambient light, and is used to quantitatively collect the ambient light generated by the light therapy lamp in real time.

[0110] In step S4, after obtaining the light irradiance time series during the driving compensation period, the ambient light occlusion behavior is quantitatively identified.

[0111] Specifically, the light irradiance data collected during the driving compensation period are combined into a light irradiance sequence in chronological order. The difference between the light irradiance data at adjacent sampling times is calculated and the absolute value is taken to obtain the light irradiance change amount and construct a light irradiance change amount sequence. The light irradiance change amount is used to reflect the degree of instantaneous change in ambient light intensity within a unit sampling interval. The larger the value, the more drastic the change in ambient light.

[0112] The changes in light irradiance were statistically processed, and the median value was selected as the center value of change to characterize the typical level of ambient light change.

[0113] Further calculate the absolute value of the deviation between the change in light irradiance and the center value of the change and form a deviation sequence. Take the median of the deviation sequence to obtain the center value of the deviation. Use the sum of the center value of the deviation and the center value of the change as the threshold for judging the change in ambient light. The threshold for judging the change in ambient light is used to distinguish between normal fluctuations and sudden changes. The larger the value, the wider the tolerance range for changes in ambient light.

[0114] Sampling points in the irradiance change sequence that satisfy the threshold for ambient light change are marked as occlusion trigger points. Each occlusion trigger point corresponds to one ambient light occlusion event. The total number of occlusion trigger points is counted during the driving compensation period to obtain the ambient light occlusion frequency. The higher the ambient light occlusion frequency, the more times the ambient light is occluded or disturbed per unit time.

[0115] The mean and standard deviation of the irradiance sequence during the driving compensation period were further calculated to obtain the average irradiance and the standard deviation of irradiance. The average irradiance reflects the overall level of ambient light illumination; a larger value indicates stronger continuous ambient light interference. The standard deviation of irradiance reflects the degree of ambient light fluctuation; a larger value indicates stronger ambient light instability.

[0116] An illumination shading index was constructed based on the frequency of ambient light shading, average irradiance, and standard deviation of irradiance.

[0117] ;

[0118] in, The illumination obstruction index, Frequency of ambient light blocking To drive the duration of the compensation period, The standard deviation of light irradiance. The average irradiance, To prevent tiny constants with a denominator of zero.

[0119] The illumination obstruction index comprehensively reflects the frequency and intensity of ambient light obstruction per unit time. The higher the value, the more frequent and unstable the ambient light obstruction.

[0120] Based on this, the respiratory waveform in the photoelectric response signal is obtained through the respiratory monitoring channel, and the respiratory waveform is bandpass filtered to extract the respiratory frequency band component, thus obtaining the respiratory waveform sequence.

[0121] It should be noted that the respiratory monitoring channel is a signal processing path built based on the output signal of the photodetector, used to extract periodic change components related to respiratory activity from the photoelectric response signal.

[0122] Peak-valley analysis is performed on the respiratory waveform sequence to extract the peak and trough values ​​in each respiratory cycle. The peak value is subtracted from the trough value to obtain the respiratory amplitude of each cycle. The respiratory amplitudes are arranged into an amplitude sequence according to time order, and the median is taken as the respiratory waveform amplitude feature value. The respiratory waveform amplitude feature value reflects the overall intensity of the current respiratory signal. The larger the value, the clearer and more recognizable the respiratory signal.

[0123] Further, a respiratory signal quality coefficient is constructed based on the amplitude characteristic values ​​of the respiratory waveform:

[0124] ;

[0125] in, The respiratory signal quality coefficient. The characteristic value of the respiratory waveform amplitude. The average irradiance, To prevent tiny constants with a denominator of zero.

[0126] The respiratory signal quality coefficient is used to characterize the effective intensity of the respiratory signal under the current ambient light background. The larger the value, the more obvious the interference of the respiratory signal with ambient light.

[0127] A joint evaluation index was constructed based on the radiation obstruction index and the respiratory signal quality coefficient.

[0128] ;

[0129] in, As a joint judgment indicator, The respiratory signal quality coefficient. The illumination obstruction index, To prevent tiny constants with a denominator of zero.

[0130] The joint judgment index is used to comprehensively reflect the relative relationship between respiratory signal quality and the degree of ambient light disturbance. The larger the value, the higher the respiratory signal quality and the weaker the ambient light disturbance.

[0131] The joint judgment index is compared with the preset judgment threshold:

[0132] When the joint judgment index is greater than or equal to the preset judgment threshold, the current monitoring environment is determined to meet the stable monitoring conditions, and conventional respiratory monitoring processing is selected. Conventional respiratory monitoring processing refers to the processing method of directly performing feature analysis and status judgment on the respiratory waveform output by the respiratory monitoring channel when the photoelectric receiving link is in a stable working state.

[0133] When the joint judgment index is less than the preset judgment threshold, it is determined that the ambient light occlusion disturbance is significant or the respiratory signal quality is insufficient. Compensation irradiation maintenance processing is selected. Compensation irradiation maintenance processing is an enhanced processing method used to maintain the continuity of monitoring when the ambient light occlusion disturbance is significant or the respiratory signal quality is reduced. In compensation irradiation maintenance processing, the driving compensation level of the light-emitting diode is further improved and the time window of the respiratory monitoring channel output signal is extended to enhance the stability and extractability of the respiratory signal.

[0134] It should be noted that the preset judgment threshold is set using a simplified quantitative method based on statistical distribution. Specifically, during the calibration phase, multiple sets of monitoring data under different ambient light interference conditions are collected. The joint judgment index corresponding to each set of data is calculated according to a predetermined method, and simultaneously, it is determined whether the corresponding respiratory waveform meets the basic identifiable condition (i.e., the respiratory waveform amplitude characteristic value is greater than the preset minimum amplitude threshold). The joint judgment index corresponding to the data that meets the condition constitutes a valid sample set. After sorting the valid sample set, the minimum value is selected as the preset judgment threshold.

[0135] Finally, it should be noted that in this paper, relational terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations.

[0136] Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0137] In this document, the singular forms “a,” “an,” and “the” may also include the plural forms unless the context clearly indicates otherwise. It should also be understood that terms such as “comprising / including” or “having” specify the presence of the stated features, integrals, steps, operations, components, parts, or combinations thereof, but do not preclude the possibility of the presence or addition of one or more other features, integrals, steps, operations, components, parts, or combinations thereof. Meanwhile, the term “and / or” as used in this specification includes any and all combinations of the associated listed items.

[0138] The various embodiments in this specification are described in a progressive manner. Each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as needed, and the same or similar parts can be referred to each other.

[0139] The above description of the disclosed embodiments will enable those skilled in the art to make or use various modifications to these embodiments. It will be readily apparent to those skilled in the art that the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A multivariate fusion detection method for monitoring neonatal respiratory distress, characterized in that: Includes the following steps: Step S1: Set the light source observation time, read the LED driving information of the photodetector in the monitored area under test during the light source observation time, divide the light source observation time using the LED driving information and obtain the photoelectric response voltage of the photodetector. Step S2: Evaluate the ambient light baseline characteristics of the photodetector based on the photoelectric response voltage, collect the probe contact pressure of the photodetector, and generate the photoelectric receiving occupancy level of the monitored area by combining the probe contact pressure and the ambient light baseline characteristics. Step S3: Detect the peak value of the diode current of the photodetector, calculate the occupancy callback coefficient using the photodetector occupancy level and adjust the peak value of the diode current, set the drive compensation period and collect the light irradiance data of the monitored area. Step S4: Based on the light irradiance data, the frequency of ambient light occlusion is counted, and an irradiation occlusion index is generated according to the frequency of ambient light occlusion. The amplitude of the respiratory waveform is obtained through the respiratory monitoring channel, and the conventional respiratory monitoring treatment or the irradiation maintenance treatment is selected in combination with the irradiation occlusion index.

2. The multivariate fusion detection method for monitoring neonatal respiratory distress according to claim 1, characterized in that: In step S1, a light source observation time is preset. During the light source observation time, the light-emitting diode driving information, including the driving pulse turn-on time and the driving pulse turn-off time, is read through the driving control interface of the photoelectric monitoring probe. The driving pulse turn-on and driving pulse turn-off times recorded during the observation time of the light source are sorted in chronological order, and the time interval between each pair of adjacent driving pulse turn-on and driving pulse turn-off times is taken as the emission stage. The time interval formed between the turn-off moment of the driving pulse after the end of the light emission phase and the turn-on moment of the next driving pulse is defined as the non-light emission phase; The photoelectric response voltage is obtained through the photoelectric signal output terminal of the photodetector.

3. The multivariate fusion detection method for monitoring neonatal respiratory distress according to claim 2, characterized in that: In step S2, the photoelectric response voltages corresponding to the luminescence stage and the non-luminescence stage are combined in time order to form a luminescence response voltage sequence and a non-luminescence response voltage sequence. The median of each photoelectric response voltage in the non-luminescent response voltage sequence was selected as the ambient light baseline value, and the maximum value of each photoelectric response voltage in the luminescent response voltage sequence was selected as the detection superposition peak value. The ratio of the ambient light baseline value to the detected superimposed peak value is used as the ambient light baseline feature.

4. The multivariate fusion detection method for monitoring neonatal respiratory distress according to claim 1, characterized in that: In step S2, the probe bonding pressure is collected based on the flexible pressure sensor in the photoelectric monitoring probe, and the bonding pressure of each probe is combined into a probe bonding pressure sequence according to the time sequence. The pressure change is obtained by performing adjacent difference on the bonding pressure sequence. The center value of pressure change and the center value of pressure deviation are calculated based on the median and deviation of the pressure change. The upper limit and lower limit of the pressure change threshold are calculated by combining the center value of pressure change and the center value of pressure deviation. The pressure changes that fall within the upper and lower limits of the pressure change threshold are marked, and the probe contact pressure corresponding to the marked pressure changes is also marked. The bonding pressure of adjacent probes is combined into candidate stable bonding intervals, and the candidate stable bonding interval with the longest duration is selected as the stable bonding interval. Within the stable bonding range, the median of the probe bonding pressure is selected as the bonding reference pressure value.

5. A multivariate fusion detection method for monitoring neonatal respiratory distress according to claim 4, characterized in that: In step S2, the reference bonding pressure value is retrieved from the probe calibration library, and the bonding comparison coefficient is calculated by combining the bonding reference pressure value and the reference bonding pressure value. The bonding mapping factor is calculated based on the bonding comparison coefficient, and the product of the bonding mapping factor and the ambient light baseline feature is used as the receiving occupancy correction feature. The photoelectric receiving occupancy correction feature is compared with the preset first occupancy threshold and the preset second occupancy threshold to generate the photoelectric receiving occupancy level of the monitored area to be tested; If the received occupancy correction feature is less than or equal to the preset first occupancy threshold, a low occupancy received layer is generated. If the received occupancy correction feature is greater than the preset first occupancy threshold and less than the preset second occupancy threshold, then a medium occupancy receiving layer is generated. If the receive occupancy correction feature is greater than or equal to the preset second occupancy threshold, a high occupancy receive layer is generated.

6. The multivariate fusion detection method for neonatal respiratory distress monitoring according to claim 1, characterized in that: In step S3, under the current driving conditions, the instantaneous current in the LED driving branch is continuously sampled by the current acquisition device to form a current time series; Extract the maximum value of the current time series to obtain the diode current peak value; A piecewise continuous function is used to assign corresponding callback coefficient calculation functions to different photoelectric receiver occupancy levels, thereby generating occupancy callback coefficients. The peak value of the diode current is calculated by multiplying the peak value of the diode current based on the placeholder callback coefficient, and the compensated target drive current peak value is obtained. After obtaining the target drive current peak value, the drive control parameters are updated so that the light-emitting diodes can output pulses according to the compensated target drive current peak value in subsequent drive cycles.

7. The multivariate fusion detection method for monitoring neonatal respiratory distress according to claim 1, characterized in that: In step S3, after the drive parameter adjustment is completed, the drive compensation period is set. The drive compensation period refers to the time interval during which the compensation drive state is maintained to evaluate the compensation effect after the drive current compensation is performed. During the drive compensation period, ambient light is collected in real time by light irradiance sensing units deployed in the monitoring area to form light irradiance data, and a light irradiance sequence is constructed in chronological order. Irradiance data is the intensity of ambient light energy received per unit area.

8. A multivariate fusion detection method for monitoring neonatal respiratory distress according to claim 7, characterized in that: In step S4, the difference between the light irradiance data at adjacent sampling times in the light irradiance sequence is calculated and the absolute value is taken to obtain the light irradiance change and construct the light irradiance change sequence. The median of the sequence of changes in light irradiance was selected as the center value of the change. Calculate the absolute value of the deviation between the change in light irradiance and the center value of the change and form a deviation sequence. Take the median of the deviation sequence to obtain the center value of the deviation. Use the sum of the center value of the deviation and the center value of the change as the threshold for judging the change in ambient light. The sampling points in the light irradiance change sequence that satisfy the value greater than the ambient light change judgment threshold are marked as occlusion trigger points; Each occlusion trigger point corresponds to one ambient light occlusion event. The total number of occlusion trigger points is counted during the drive compensation period to obtain the ambient light occlusion frequency. The mean and standard deviation of the irradiance sequence during the driving compensation period were further calculated to obtain the average irradiance and the standard deviation of irradiance.

9. A multivariate fusion detection method for monitoring neonatal respiratory distress according to claim 8, characterized in that: In step S4, an illumination shading index is constructed based on the frequency of ambient light shading, average irradiance, and standard deviation of irradiance. The respiratory waveform is obtained from the photoelectric response signal through the respiratory monitoring channel, and the respiratory waveform is bandpass filtered to extract the respiratory frequency band component to obtain the respiratory waveform sequence. Peak-valley analysis is performed on the respiratory waveform sequence to extract the peak and trough values ​​in each respiratory cycle. The peak value is subtracted from the trough value to obtain the respiratory amplitude of each cycle. The respiratory amplitudes are arranged into an amplitude sequence according to time order, and the median is taken as the characteristic value of the respiratory waveform amplitude. The respiratory signal quality coefficient is constructed based on the characteristic value of the respiratory waveform amplitude. A joint judgment index was constructed based on the radiation occlusion index and the respiratory signal quality coefficient. When the joint judgment index is greater than or equal to the preset judgment threshold, the judgment is selected to perform conventional respiratory monitoring processing. Conventional respiratory monitoring processing refers to the processing method of directly performing feature analysis and status judgment on the respiratory waveform output by the respiratory monitoring channel. When the joint judgment index is less than the preset judgment threshold, the judgment selects the compensation irradiation maintenance treatment. The compensation irradiation maintenance treatment refers to the enhanced treatment method adopted to maintain the continuity of monitoring.

Images