Intelligent helmet capable of detecting safety and health of human body

Abstract

The application discloses an intelligent helmet capable of detecting human safety and health, relates to the technical field of intelligent safety protection, and acquires photoelectric reflection pulse wave data and ambient light change data of a forehead area of a human body; forms ambient light transient impact density characteristics according to the ambient light change data; forms skin photo-thermal hysteresis coupling characteristics according to a time delay relationship between the photoelectric reflection pulse wave data and temperature change of the forehead area; forms micro-blood flow phase locking characteristics based on rhythm stability and phase continuous change relationship between continuous pulse periods; combines the ambient light transient impact density characteristics, the skin photo-thermal hysteresis coupling characteristics and the micro-blood flow phase locking characteristics to judge external strong stray light interference, and carries out frozen compensation and dynamic sampling adjustment on abnormal pulse wave data; the application can reduce the influence of strong sunlight, electric welding arc light and the like on the detection result of vital signs, and improve the stability and reliability of human safety and health detection in a complex operation environment.

Description

Technical Field

[0001] This invention relates to the field of intelligent safety protection technology, specifically to an intelligent helmet that can detect human safety and health. Background Technology

[0002] With the increasing demand for intelligent safety management in high-risk industries such as construction, mining, power maintenance, ship welding, and high-altitude operations, smart helmets with human vital sign detection functions are gradually being widely used. Existing smart helmets typically integrate photoplethysmography (PPG) technology, which emits specific wavelengths of light and receives reflected signals from human blood to achieve real-time monitoring of human health parameters such as heart rate and blood oxygenation, thereby improving the safety and health management of workers.

[0003] However, existing vital sign detection methods based on photoelectric reflection are easily interfered with by strong external light in complex industrial environments. Especially in environments with strong outdoor sunlight, welding arc light, metal reflection light, and high-frequency industrial flicker light sources, stray light can easily enter the helmet's optical detection area, causing saturation, drift, or distortion of the photoelectric reflection signal. This can lead to abnormal heart rate fluctuations, inaccurate blood oxygen data, and misjudgments of vital signs. In severe cases, the back-end monitoring system may misjudge the worker's condition based on erroneous detection results, resulting in personnel at risk continuing to work, posing a significant safety hazard.

[0004] Existing technologies typically employ methods such as light-shielding structures, gain adjustment, or filtering to reduce ambient light interference. However, most of these methods only address single changes in light intensity and are insufficient to effectively distinguish between genuine human pulse waves and pseudo-pulse signals formed by strong stray light. Especially in environments with continuous arc light interference or transient high-energy light impact, the accuracy of vital sign detection is still prone to problems. Therefore, it is necessary to propose a smart helmet that can effectively identify strong light interference and dynamically correct abnormal vital sign data. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent helmet that can detect human safety and health, thereby addressing the shortcomings of the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a smart helmet capable of detecting human safety and health, comprising: The photoelectric data acquisition module is used to acquire photoelectric reflected pulse wave data of the forehead area of ​​the human body after the smart helmet is worn on the human head, and simultaneously acquire ambient light change data around the helmet's optical window; The ambient light impact analysis module is used to extract the number of abnormal light intensity abrupt changes and the discrete change characteristics of light intensity per unit time based on ambient light change data, and form the transient impact density characteristics of ambient light. The photothermal coupling analysis module is used to form the skin photothermal hysteresis coupling characteristics based on the time delay relationship between photoelectric reflected pulse wave data and temperature changes in the forehead area. The phase-locking analysis module is used to form micro-blood flow phase-locking characteristics based on the relationship between rhythm stability and continuous phase changes between consecutive pulse cycles. The interference determination module is used to perform correlation analysis on the transient impact density characteristics of ambient light, the photothermal hysteresis coupling characteristics of skin, and the microblood flow phase-locking characteristics to determine whether the current photoelectric reflection pulse wave data is affected by strong stray light from the outside. The dynamic compensation module is used to freeze abnormal pulse wave amplitude data when the judgment result is that there is interference, and to dynamically compensate for the current vital signs change trend based on the phase continuity characteristics of historical stable pulse cycles. The dynamic adjustment module is used to adjust the sampling timing and exposure time of photoelectric reflection pulse wave data based on the compensated trend of vital signs changes, so as to reduce the impact of strong external stray light on the vital signs detection results. The status output module is used to output the adjusted vital sign detection results and generate corresponding human safety and health status information.

[0007] Preferably, the acquisition of photoelectric reflection pulse wave data and ambient light change data includes: After detecting that the helmet has formed a stable fit with the forehead area of ​​the human body, photoelectric reflection acquisition and ambient light acquisition are initiated; synchronous time markers are added to the photoelectric reflection pulse wave data and ambient light change data respectively; the photoelectric reflection pulse wave data and ambient light change data are stored in correspondence according to the synchronous time markers.

[0008] Preferably, the formation of the transient impact density characteristics of ambient light includes the following steps: The ambient light variation data is divided into continuous and partially overlapping time segments, and the light intensity transition changes between adjacent time segments are extracted. Light intensity transitions that exceed the preset fluctuation range and last for less than the preset transient duration are marked as abnormal light impact data. Based on the frequency, location, and time interval dispersion of abnormal light impact data per unit time, the transient impact density characteristics of ambient light are formed.

[0009] Preferably, the formation of the skin photothermal hysteresis coupling feature includes the following steps: Using the time stamp of photoelectric reflection pulse wave data as a reference, the temperature change data of the forehead area is correlated with time. Extract the peak change time and the onset time of temperature rise in the forehead region corresponding to the effective peak, and form a continuous time series; Based on the temperature rise response lag interval, pulse wave amplitude fluctuation degree, and temperature rise duration in the continuous time sequence, the skin photothermal hysteresis coupling characteristics are formed.

[0010] Preferably, the formation of microflow phase-locking features includes the following steps: The continuous photoelectric reflection pulse wave data is divided into multiple pulse cycles according to the effective peak and adjacent trough, and the phase anchor point is extracted at the rising edge of each pulse cycle. Based on the positional changes of the phase anchor point within adjacent pulse cycles, a phase migration sequence and rhythm continuity relationship are formed. When a sudden jump, discontinuity, or reverse drift occurs at the phase anchor point, the phase lockout state is marked, and microflow phase lockout characteristics are formed based on the number of occurrences and duration of the phase lockout state.

[0011] Preferably, the determination of strong external stray light interference includes the following steps: Correlation segments are established based on the transient impact density characteristics of ambient light, the photothermal hysteresis coupling characteristics of skin, and the phase-locking characteristics of microblood flow within the same time interval; when the change trend of transient impact density of ambient light and the change trend of photothermal hysteresis of skin satisfy the synchronous offset relationship within the same or adjacent correlation segments, a photothermal synchronous perturbation state is formed. When the photothermal synchronization disturbance state and the phase lockout state occur together in a continuous associated segment, the composite interference state is marked, and the strong stray light interference from the outside is judged based on the cumulative duration and recurrence of the composite interference state.

[0012] Preferably, dynamic compensation includes the following steps: When subjected to strong stray light interference from the outside, the pulse wave amplitude data that exceeds the preset amplitude fluctuation range within the interference range is marked as frozen data, and the phase arrangement order of the continuous stable pulse cycles before the interference range is retained. A rhythmic continuation trajectory is formed based on the phase shift interval between continuous and stable pulse cycles, and the predicted pulse occurrence time within the interference interval is determined by the rhythmic continuation trajectory; the predicted pulse occurrence time is matched with the unfrozen pulse wave change data, and the missing pulse wave amplitude change trend is compensated for in a time sequence.

[0013] Preferably, the adjustment of sampling timing and exposure time includes the following steps: Based on the compensated trend of vital signs, the stable period of the continuous pulse cycle is extracted, and the rising edge region of the pulse wave within the stable period is determined as the priority sampling interval. Based on the transient impact density characteristics of ambient light, the concentrated area of ​​abnormal light impact is determined, and the sampling start time is avoided in the concentrated area of ​​abnormal light impact. The exposure time is gradually shortened according to the continuity of pulse wave amplitude changes within the priority sampling interval, and the current sampling sequence and exposure time are maintained when the phase continuity consistency is restored to the preset stable range.

[0014] Preferably, the output of vital sign detection results includes the following steps: The adjusted vital sign detection results are integrated in a continuous time sequence, and the continuity of heart rate changes and the stability of blood oxygen fluctuations are extracted. Based on the continuity of heart rate changes, the stability of blood oxygen fluctuations, the state of composite interference, and low-confidence markers, a confidence interval for vital signs is formed. Based on the confidence level of the continuous vital sign confidence interval, information on the safety and health status of normal or abnormal human beings is output.

[0015] Preferably, the storage of human safety and health status information includes the following steps: Add the start time, end time, composite interference status marker, sampling time sequence marker, and exposure duration marker to normal and abnormal human safety and health status information; merge and store normal human safety and health status information with the same confidence level and time interval not exceeding the preset interval; store abnormal human safety and health status information independently and associate it with the corresponding vital sign detection results and abnormality source marker.

[0016] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention does not rely solely on changes in light intensity to determine external strong light interference. Instead, it simultaneously constructs transient impact density characteristics of ambient light, skin photothermal hysteresis coupling characteristics, and microblood flow phase-locking characteristics. The temporal correlation between these three factors is used to identify the impact of strong stray light on photoelectric reflection pulse wave data. Specifically, by jointly analyzing the temporal dispersion of abnormal light impact data, the temperature rise lag relationship in the forehead region, and the continuous migration state of the phase anchor point within the pulse cycle, it can distinguish between genuine changes in human microblood flow and pseudo-pulse waves formed by external arc light and strong reflected light. This solves the problem in existing technologies where filtering or shading alone is insufficient to identify persistent strong stray light interference. Based on these techniques, the temporal stability and data continuity of human vital sign detection can be maintained even in complex environments such as strong sunlight and welding arc light, reducing the occurrence of misjudgments and missed judgments.

[0017] This invention, upon detecting strong stray light interference, employs a freeze-compensation and dynamic sampling adjustment method based on phase continuity to freeze abnormal pulse wave amplitude data. A rhythmic continuation trajectory is formed based on the phase shift intervals between continuous stable pulse cycles before freezing, and the missing pulse waves are then compensated for temporally using the predicted pulse occurrence time. Simultaneously, the sampling sequence and exposure duration are dynamically adjusted according to the compensated vital sign change trend, actively avoiding the concentrated area of ​​abnormal light impact at the sampling start time. Compared to the passive data correction method after interference in existing technologies, this invention can maintain continuous pulse cycle acquisition during continuous light interference and reduce the cumulative time of strong light entering the optical window, thereby directly improving the reliability and stability of vital sign detection results in complex industrial environments. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0019] Figure 1 This is a flowchart of a module of an intelligent helmet that can detect human safety and health according to the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0021] For examples, please refer to Figure 1 As shown in the figure, the smart helmet capable of detecting human safety and health described in this embodiment includes: The photoelectric data acquisition module is used to acquire photoelectric reflected pulse wave data of the forehead area of ​​the human body after the smart helmet is worn on the human head, and simultaneously acquire ambient light change data around the helmet's optical window.

[0022] In this embodiment of the invention, after the smart helmet is worn, the wearing sensing unit located on the inside of the helmet first detects whether the helmet forms a stable fit with the human head. When the forehead contact pressure reaches a preset threshold, the photoelectric data acquisition module is activated. The photoelectric data acquisition module includes a light-emitting unit, a light-receiving unit, and an ambient light acquisition unit. The light-emitting unit periodically emits detection light of a preset wavelength towards the forehead area of ​​the human body. The detection light can be red light, infrared light, or a combination of both. After the subcutaneous blood undergoes periodic volume changes with pulse changes, it dynamically reflects the detection light. The light-receiving unit continuously collects the reflected light to form corresponding photoelectric reflected pulse wave data.

[0023] To ensure the continuity of subsequent vital sign analysis, the optical receiving unit samples the reflected light intensity in real time according to a preset sampling period and generates raw pulse wave data corresponding to the time series. The raw data includes at least peak change data, trough change data, waveform period data, and baseline change data.

[0024] Simultaneously, an ambient light acquisition unit is positioned around the helmet's optical window to synchronously detect changes in ambient light. During photoelectric reflection pulse wave data acquisition, this unit continuously acquires the ambient light intensity around the helmet and records the trend of ambient light changes at preset time intervals to generate corresponding ambient light change data. Furthermore, the ambient light acquisition unit can continuously record transient enhancements, periodic flickering, and high-frequency light intensity fluctuations in ambient light, thus providing a data foundation for subsequent identification of strong stray light interference.

[0025] In this embodiment, photoelectric reflection pulse wave data and ambient light change data are stored in a synchronous time stamping manner to ensure that a temporal correlation between human pulse changes and ambient light changes can be established in subsequent analysis.

[0026] The ambient light impact analysis module is used to extract the number of abnormal light intensity abrupt changes and the discrete change characteristics of light intensity per unit time based on ambient light change data, and form the transient impact density characteristics of ambient light.

[0027] In the ambient light impact analysis process, the ambient light change data is first divided into continuous time segments. The ambient light change data is continuously acquired by the ambient light acquisition unit at a fixed sampling frequency, preferably between 200 Hz and 500 Hz, to ensure the capture of transient light change signals generated by welding arc light and high-frequency flicker sources. The acquired ambient light change data is arranged in chronological order to form a light intensity sequence, denoted as: I = {i1, i2, i3…in}; where i represents the ambient light intensity value corresponding to a single sampling moment, and n represents the total number of samples within the current sampling period.

[0028] Subsequently, the light intensity sequence is segmented and buffered according to continuous time segments. The duration of each time segment is preferably set to 20 to 50 milliseconds, and adjacent time segments are buffered in an overlapping manner, with the overlap ratio preferably being 30% to 50%, to avoid missed detection when transient arc light is located at the boundary of the time segment.

[0029] After dividing the time segments, the trend of light intensity transitions between adjacent time segments is extracted. For any adjacent time segments, the average light intensity value within the corresponding time segment is calculated as follows: Where Ak represents the average light intensity value of the k-th time segment, and m represents the number of sampling points contained in the current time segment.

[0030] Then, the change in light intensity between adjacent time segments was calculated: When ΔA exceeds the preset fluctuation range, it is determined that there is abnormal light impact data in the current time segment. The preset fluctuation range is determined based on the ambient light reference fluctuation value under static environment. It is preferred to use 2 to 4 times the average fluctuation value obtained by continuous sampling in static environment for 10 seconds as the judgment threshold, and the preferred threshold is 35 lux to 80 lux.

[0031] After detecting abnormal light impact data, the location and duration of the abnormal light impact data within a unit of time are recorded. The location corresponds to the time coordinate of the abnormal light impact data in the current sampling sequence, and the duration corresponds to the duration for which the abnormal light impact data continuously exceeds a preset fluctuation range.

[0032] To avoid misinterpreting slow changes in normal ambient light as abnormal light impacts, the abnormal light impact data undergoes further transient screening. When the duration of abnormal light impact data is less than 500 milliseconds, it is determined to be a transient light impact; when the duration exceeds 500 milliseconds, it is determined to be a change in background ambient light and is not included in subsequent transient impact analysis.

[0033] After extracting the anomalous optical shock data, the time intervals between each anomalous optical shock data point were analyzed. Let the time interval between adjacent anomalous optical shock data points be: Where tj represents the occurrence time of the j-th anomalous optical shock data. Then, the dispersion between consecutive time intervals is calculated as follows: Where Tavg represents the average value of all time intervals within the current analysis period, and q represents the total number of time intervals. A smaller dispersion D indicates that the abnormal light impact data exhibits a periodic clustered distribution; a larger dispersion D indicates that the abnormal light impact data exhibits a random discrete distribution. Since welding arc light typically has high-frequency irregular flickering characteristics, its corresponding dispersion will be significantly higher than that of ordinary natural light.

[0034] Furthermore, the transient light impact clustering state is extracted based on the temporal clustering state of the anomalous light impact data. The clustering state is represented by the density of anomalous light impacts within a unit time window, which is preferably set to 1 second.

[0035] Within a unit time window, the number of abnormal light impact data N is counted, and the ambient light transient impact density value is calculated in combination with the corresponding dispersion degree D. The calculation method is as follows: P = N × D; where P represents the ambient light transient impact density value.

[0036] When the transient impact density value of ambient light continuously exceeds a preset density threshold, it is determined that there is a risk of strong stray light interference in the current environment. The preset density threshold is preferably obtained by calibration based on normal indoor environment, outdoor sunlight environment, and welding environment. The transient impact density value of ambient light in ordinary indoor environment is usually less than 20, that in outdoor solar reflection environment is usually between 20 and 60, and that in welding arc light environment is usually over 100. Therefore, it is preferable to set the preset density threshold to 80.

[0037] To improve the stability of transient light impact recognition, continuous judgment is performed within multiple consecutive unit time windows. When the ambient light transient impact density value exceeds the preset density threshold within three consecutive unit time windows, the ambient light transient impact density feature is output and transmitted to the subsequent correlation analysis process to determine whether the current photoelectric reflection pulse wave data is affected by strong external stray light interference.

[0038] The photothermal coupling analysis module is used to form the skin photothermal hysteresis coupling characteristics based on the time delay relationship between photoelectric reflected pulse wave data and temperature changes in the forehead area.

[0039] In the photothermal coupling analysis, the photoelectric reflection pulse wave data and the forehead region temperature change data are time-stamped using the same clock source, with a preferred time-stamping accuracy of 1 to 5 milliseconds. The sampling frequency of the forehead region temperature change data is preferably 20 Hz to 100 Hz, and the sampling frequency of the photoelectric reflection pulse wave data is preferably 100 Hz to 500 Hz. When the sampling frequencies of the two types of data are inconsistent, the forehead region temperature change data is time-proximity matched based on the time axis of the photoelectric reflection pulse wave data, so that data near the same moment can correspond.

[0040] When extracting the peak change time from photoelectric reflection pulse wave data, the continuously sampled data is first smoothed, with a smoothing window of 5 to 11 sampling points preferred. Then, the point with the highest amplitude between adjacent troughs is selected as a candidate peak. When the time interval between the candidate peak and the previous peak is within the range of 300 milliseconds to 1500 milliseconds, and the amplitude difference between the candidate peak and the adjacent trough is not less than 40% of the average amplitude difference during the current 30-second stable sampling phase, the candidate peak is determined as a valid peak, and the peak change time corresponding to the valid peak is recorded.

[0041] When extracting the temperature rise initiation time from the forehead area temperature change data, the average temperature within the first 5 seconds before the start of detection is selected as the temperature baseline. During the detection process, if the temperature increment of three consecutive temperature sampling points relative to the temperature baseline is not less than 0.08 degrees Celsius, and the cumulative temperature increment within the subsequent second is not less than 0.15 degrees Celsius, the time corresponding to the first sampling point that meets the conditions is determined as the temperature rise initiation time. The preferred value of 0.08 degrees Celsius is suitable for low-speed temperature change scenarios in the forehead contact area; when the ambient temperature fluctuates by more than 2 degrees Celsius per minute, this value is set to 3 times the average temperature fluctuation of the 10 seconds before the start of detection.

[0042] After obtaining the peak change time and the temperature rise start time, adjacent peak change times and temperature rise start times are combined into a time series, and the time difference between them is calculated. The time difference is obtained by subtracting the corresponding peak change time from the temperature rise start time. If the time difference is positive, it indicates that the temperature rise change lags behind the pulse wave change; if the time difference is negative, it indicates that the temperature rise change precedes the pulse wave change. Five to ten consecutive time series are combined into one analysis group, preferably eight time series are combined into one analysis group.

[0043] Within each analysis group, all time differences are arranged from smallest to largest. After removing the largest and smallest time differences, the average of the remaining time differences is taken to obtain the center value of the temperature rise response lag interval. Then, the absolute difference between each remaining time difference and this center value is calculated, and the average of the absolute differences is taken to obtain the fluctuation value of the temperature rise response lag interval. The preferred allowable range for the center value of the temperature rise response lag interval is 100 milliseconds to 1200 milliseconds; the preferred allowable range for the fluctuation value is 0 milliseconds to 180 milliseconds. The aforementioned ranges can be set by having the wearer continuously collect data for 60 seconds in an environment without strong light interference. A 30% fluctuation above or below the center value corresponding to the stable phase is taken as the allowable range for the center value, and twice the fluctuation value corresponding to the stable phase is taken as the allowable range for the fluctuation value.

[0044] While recording the temperature rise response lag interval, the corresponding pulse wave amplitude change trend is extracted. The amplitude change of each effective peak is obtained by subtracting the amplitude of its preceding adjacent trough from the effective peak amplitude; after arranging the amplitude changes corresponding to consecutive effective peaks in sequence, the absolute value of the difference between adjacent amplitude changes is calculated, and then the absolute values ​​of the differences are averaged to obtain the degree of pulse wave amplitude fluctuation. Preferably, when the degree of pulse wave amplitude fluctuation exceeds 2.5 times the degree of pulse wave amplitude fluctuation in the stable phase, and the center value or fluctuation value of the temperature rise response lag interval exceeds the corresponding allowable range, it is determined that there is an abnormal photothermal coupling state in the forehead region.

[0045] When forming skin photothermal hysteresis coupling characteristics, the center value of the temperature rise response hysteresis interval, the fluctuation value of the temperature rise response hysteresis interval, the fluctuation degree of pulse wave amplitude, and the duration of temperature rise under abnormal photothermal coupling conditions are recorded in combination. The duration of temperature rise is determined from the moment the temperature rises to the moment the temperature falls back to within 0.05 degrees Celsius of the temperature reference value. If the duration of temperature rise exceeds 3 seconds and the fluctuation degree of pulse wave amplitude simultaneously exceeds the corresponding threshold, the skin photothermal hysteresis coupling characteristic is marked as a strong photothermal interference characteristic, which is used for subsequent correlation analysis with ambient light transient impact density characteristics and microblood flow phase-locking characteristics.

[0046] The phase-locking analysis module is used to form micro-blood flow phase-locking characteristics based on the rhythm stability and continuous phase change relationship between continuous pulse cycles.

[0047] During phase-locked analysis, continuous photoelectric reflection pulse wave data are processed sequentially over time. The sampling frequency of the photoelectric reflection pulse wave data is preferably between 100 Hz and 500 Hz, with 200 Hz being more preferred. Before dividing the pulse cycle, a moving average of 5 to 9 sampling points is used to remove single-point spike noise, with 7 sampling points being more preferred. The r-th sample value after moving average is the sum of the seven consecutive sample values ​​centered at the r-th sampling point, divided by 7.

[0048] After smoothing, multiple pulse cycles are divided based on the effective peaks and adjacent troughs. Each pulse cycle begins with the adjacent trough before one effective peak and ends with the adjacent trough before the next effective peak. The time interval between adjacent effective peaks is preferably limited to the range of 300 to 1500 milliseconds; if the time interval is less than 300 milliseconds or greater than 1500 milliseconds, the corresponding cycle will not participate in phase-locked analysis. Adjacent troughs are obtained by finding the lowest amplitude point within a preset search interval on both sides of the effective peak, preferably 200 milliseconds before and after the effective peak.

[0049] The phase anchor point with the most stable amplitude change is extracted at the rising edge of each pulse cycle. The rising edge is the data segment between adjacent troughs and effective peaks. First, the amplitude difference between adjacent sampling points within the rising edge is calculated as: Amplitude difference = Amplitude of the next sampling point - Amplitude of the previous sampling point. Then, using three consecutive sampling points as local detection windows, the average value and deviation value of the amplitude difference within each local detection window are calculated. The deviation value is the average of the absolute values ​​of the differences between each amplitude difference and the average value within the local detection window. The center sampling point of the local detection window with the lowest deviation value and a positive average value is determined as the phase anchor point. If there are more than two local detection windows with the same deviation value, the center sampling point of the local detection window closest to the midpoint of the rising edge is selected as the phase anchor point.

[0050] The position of the phase anchor point within the corresponding pulse cycle is represented by a position ratio, calculated as: Position Ratio = Time Difference between the Phase Anchor Point Time and the Start Time of the Pulse Cycle ÷ Pulse Cycle Duration. The position ratio ranges from 0 to 1. A position ratio is obtained for each pulse cycle, and the position ratios of consecutive pulse cycles are arranged in chronological order to form the phase transition sequence.

[0051] The rhythm continuity is determined based on the variation in duration between adjacent pulse cycles. The duration of the k-th pulse cycle is denoted as cycle length k. The rhythm variation between adjacent cycles is calculated as: Rhythm variation = Cycle length k - Cycle length k - 1. Five consecutive pulse cycles are considered as one analysis segment, preferably five. The average absolute value of all rhythm variations within the analysis segment is calculated as the rhythm continuity offset value. Under stable wearing conditions, the rhythm continuity offset value is preferably no more than 80 milliseconds. This threshold can be obtained by sampling continuously for 60 seconds under conditions without strong light interference. Twice the average value of the rhythm continuity offset value in this stage is used as the individualized threshold, with an upper limit not exceeding 120 milliseconds.

[0052] When continuously comparing the phase transition sequence with the rhythm continuity, the phase anchor point is judged to have abrupt changes, discontinuities, or reverse drift. Abrupt changes are determined as follows: the absolute value of the position ratio difference between two adjacent pulse cycles exceeds 0.18, and the rhythm continuity offset exceeds 80 milliseconds in the same time period. Discontinuities are determined as follows: a phase anchor point meeting the conditions cannot be extracted in two consecutive pulse cycles, or the deviation value of the local detection window where the phase anchor point is located exceeds three times the average deviation value of the stable phase. Reverse drift is determined as follows: after three consecutive pulse cycles of unidirectional position ratio movement, a position ratio change in the opposite direction with an amplitude exceeding 0.12 occurs in the next pulse cycle.

[0053] When any of the following conditions is met—sudden jump, intermittent, or reverse drift—the corresponding pulse cycle is marked as a phase-locked state. To avoid misjudgment due to a single sampling error, if a phase-locked state occurs more than twice within five consecutive pulse cycles, or if a phase-locked state continues for more than two consecutive pulse cycles, the system is confirmed to have entered the phase-locked interval.

[0054] The micro-blood flow phase-locking feature consists of the number of occurrences, duration, and density of phase-locked states. The number of occurrences is the total number of phase-locked states within the current analysis segment; the duration is the number of pulse cycles covered by consecutive phase-locked states; the density of phase-locked states is calculated as: Degree of phase-locked states = Number of occurrences ÷ Total number of pulse cycles within the analysis segment. Preferably, when the density of phase-locked states reaches 0.4 and the duration is at least 2 periods, the micro-blood flow phase-locking feature is recorded as a phase-locked feature; when the density of phase-locked states is below 0.2 and there are no consecutive phase-locked states, the micro-blood flow phase-locking feature is recorded as a locked feature; and when it falls between these two, it is recorded as a transitional feature. The formed micro-blood flow phase-locking feature corresponds with the ambient light transient impact density feature and the skin photothermal hysteresis coupling feature within the same time interval, and is used to subsequently determine whether the current photoelectric reflection pulse wave data is interfered with by strong external stray light.

[0055] The interference determination module is used to perform correlation analysis on the transient impact density characteristics of ambient light, the photothermal hysteresis coupling characteristics of skin, and the microblood flow phase-locking characteristics to determine whether the current photoelectric reflection pulse wave data is affected by strong stray light interference from the outside.

[0056] During interference determination, the transient impact density characteristics of ambient light, the photothermal hysteresis coupling characteristics of skin, and the microblood flow phase-locking characteristics are all time-stamped using the same clock source, with a time-stamping accuracy preferably between 1 and 5 milliseconds. Before performing correlation analysis, the three types of features are arranged in order of their time stamps, and corresponding correlation segments are established within the same time interval. The duration of the correlation segment is preferably 1 second, and the overlap ratio of adjacent correlation segments is preferably 50% to avoid segmentation when strong stray light interference occurs at the segment boundary. If there is a deviation between the time stamps of the three types of features, the allowable deviation is preferably no more than 20 milliseconds; data exceeding 20 milliseconds will not be included in the current correlation segment but will be included in the next adjacent correlation segment for re-matching.

[0057] Within each associated segment, the ambient light transient impact density value corresponding to the ambient light transient impact density feature is extracted, and the trend of ambient light transient impact density change is calculated. The trend of ambient light transient impact density change is obtained by subtracting the ambient light transient impact density value of the previous associated segment from the ambient light transient impact density value of the current associated segment. When the current difference is not less than 20 and the ambient light transient impact density value of the current associated segment is not less than 80, it is recorded as an ambient light impact rising state; when the current difference is less than 20, it is not recorded as an ambient light impact rising state. 80 is used as the preferred density threshold, which can be set according to 4 times the average ambient light transient impact density obtained in 60 consecutive seconds under no strong light interference, and the minimum value is not less than 60.

[0058] Within the same associated segment, skin photothermal hysteresis perturbation values ​​are extracted based on the skin photothermal hysteresis coupling characteristics. The skin photothermal hysteresis perturbation value is obtained as follows: Skin photothermal hysteresis perturbation value = Hysteresis center offset ratio + Hysteresis fluctuation ratio + Pulse wave amplitude fluctuation ratio + Temperature rise duration ratio. Wherein, the hysteresis center offset ratio is the absolute value of the difference between the temperature rise response hysteresis interval center value and the stable phase center value divided by the allowable width of the center value; the hysteresis fluctuation ratio is the temperature rise response hysteresis interval fluctuation value divided by the allowable upper limit of the fluctuation value; the pulse wave amplitude fluctuation ratio is the pulse wave amplitude fluctuation degree divided by 2.5 times the stable phase pulse wave amplitude fluctuation degree; the temperature rise duration ratio is the temperature rise duration divided by 3 seconds. The allowable width of the center value is preferably 30% of the stable phase center value, and the allowable upper limit of the fluctuation value is preferably twice the stable phase fluctuation value.

[0059] Further calculations were performed on the skin photothermal hysteresis trend, which was obtained by subtracting the skin photothermal hysteresis perturbation value of the previous associated segment from the current associated segment's perturbation value. A state of rising skin photothermal hysteresis was defined as the current difference being no less than 0.8 and the current associated segment's perturbation value being no less than 3. If the aforementioned conditions were not met, it was not considered a state of rising skin photothermal hysteresis. A threshold of 3 was used as the preferred threshold, indicating that at least two of the following conditions—temperature rise response lag, continuous temperature change, and pulse wave amplitude fluctuation—met an abnormal condition.

[0060] After obtaining the ambient light impact rise state and the skin photothermal hysteresis rise state, the synchronization offset relationship between them is extracted. The synchronization offset duration is obtained by subtracting the first occurrence time of the ambient light impact rise state from the first occurrence time of the skin photothermal hysteresis rise state. When the synchronization offset duration is between 0 ms and 1200 ms, and the ambient light impact rise state and the skin photothermal hysteresis rise state occur in the same associated segment or one adjacent associated segment, it is determined to be a photothermal synchronous perturbation state. If the synchronization offset duration is negative, it indicates that the skin photothermal hysteresis rise occurs earlier than the ambient light impact rise, which does not conform to the temporal relationship that strong stray light first acts on the optical window and then causes a temperature rise change in the forehead area, and is not marked as a photothermal synchronous perturbation state. If the synchronization offset duration exceeds 1200 ms, it is determined that the temporal correlation between the two is insufficient, and is not marked as a photothermal synchronous perturbation state.

[0061] Subsequently, the photothermal synchronization perturbation state was continuously matched with the microblood flow phase-locking feature. When the microblood flow phase-locking feature was a lost-lock feature, the occurrence time, duration, and density of the lost-lock state were extracted. If the lost-lock state and the photothermal synchronization perturbation state were located in the same associated segment, or if the lost-lock state lagged behind the photothermal synchronization perturbation state for no more than one associated segment, it was marked as a composite perturbation state. To reduce the impact of single, occasional errors, a composite perturbation state was confirmed only if at least two composite perturbation states occurred within three consecutive associated segments, or if the composite perturbation state lasted for at least two consecutive seconds.

[0062] Finally, the current photoelectric reflection pulse wave data is determined to be affected by strong stray light interference based on the cumulative duration and repetition count of the composite interference state in the continuous associated segments. The cumulative duration is the sum of the durations of the associated segments covered by the continuous composite interference state, and the repetition count is the number of times the composite interference state occurs within the same determination period. The determination period is preferably 10 seconds. When the cumulative duration of the composite interference state is not less than 3 seconds within 10 seconds, or the repetition count is not less than 3 times, and the micro-blood flow phase-locking feature corresponding to the most recent composite interference state is still an unlocked feature, the current photoelectric reflection pulse wave data is determined to be affected by strong stray light interference. The determination result is output to the subsequent dynamic compensation processing to freeze the abnormal pulse wave amplitude data and perform compensation based on the phase continuity feature of the historical stable pulse cycle.

[0063] The dynamic compensation module is used to freeze abnormal pulse amplitude data when the judgment result is that there is interference, and to dynamically compensate for the current trend of vital signs based on the phase continuity characteristics of historical stable pulse cycles.

[0064] During dynamic compensation, the interference determination results indicate that after the current photoelectric reflection pulse wave data is interfered with by strong external stray light, the starting time of the associated segment where the composite interference state is first established is taken as the starting point of the interference interval, and the ending time where two consecutive associated segments do not re-establish after the end of the composite interference state is taken as the ending point of the interference interval. The duration of the associated segment is preferably 1 second, and the overlap ratio of adjacent associated segments is preferably 50%. When the interference interval has not yet ended, the data currently entering the interference interval is frozen and compensated in a real-time rolling manner.

[0065] Before freezing the pulse wave amplitude data within the interference interval, a preset amplitude fluctuation range is determined. This preset range is obtained based on the continuous stable pulse cycles before freezing. The preferred continuous stable pulse cycles are the first 10 effective pulse cycles before the start of the interference interval; if there are fewer than 10 effective pulse cycles, at least 5 are used. The amplitude change of the continuous stable pulse cycles is obtained by subtracting the amplitude of the preceding adjacent trough from the effective peak amplitude. The amplitude changes are then sorted by magnitude, and the median value is used as the stable amplitude benchmark.

[0066] The preset amplitude fluctuation range is preferably 40% to 160% of the stable amplitude reference. When the wearer is walking or climbing, the upper limit can be increased to 180% of the stable amplitude reference based on the head movement amplitude, while the lower limit remains at 40% of the stable amplitude reference. The head movement amplitude can be determined by existing motion data within the helmet, without introducing additional vital sign detection features.

[0067] Within the interference range, if any pulse wave amplitude change falls below the lower limit of the preset amplitude fluctuation range or exceeds the upper limit of the preset amplitude fluctuation range, the pulse wave amplitude data is marked as frozen data. Frozen data does not participate in the real-time update of current heart rate and blood oxygen, but its sampling time, original amplitude, and pulse cycle number are retained for subsequent traceability. Data that does not exceed the preset amplitude fluctuation range is retained as unfrozen pulse wave change data. The freezing process also preserves the phase arrangement order of the continuous stable pulse cycles before freezing. The phase arrangement order is formed by arranging the positions of each phase anchor point within the corresponding pulse cycle in chronological order.

[0068] The rhythm continuation trajectory is extracted based on the phase shift intervals between consecutive stable pulse cycles before freezing. The phase shift interval is the time difference between the occurrence times of two adjacent phase anchor points. The phase shift intervals of the consecutive stable pulse cycles before freezing are arranged in chronological order. After removing the largest and smallest intervals, the average of the remaining intervals is calculated to obtain the baseline phase shift interval. Then, the average of the absolute values ​​of the differences between the remaining intervals and the baseline phase shift interval is calculated to obtain the rhythm fluctuation amount. The baseline phase shift interval is preferably within the range of 300 ms to 1500 ms; the rhythm fluctuation amount is preferably no more than 80 ms. If the rhythm fluctuation amount exceeds 80 ms, it is recalculated by extending it forward by two pulse cycles. If it still exceeds 120 ms after extension, no compensation is output; only a flag indicating reduced reliability of the current data is output.

[0069] After determining the rhythm continuation trajectory, the starting time is taken as the time of the most recent phase anchor point before the start of the interference interval. The reference phase shift intervals are accumulated sequentially to obtain the predicted pulse occurrence time within the interference interval. If there is unfrozen pulse wave change data near the previous predicted pulse occurrence time, and the time difference between the phase anchor point of the unfrozen data and the predicted pulse occurrence time does not exceed 20% of the reference phase shift interval, then the phase anchor point of the unfrozen data is used to correct the next prediction. The correction method is to take the current unfrozen phase anchor point as the new starting time and continue to accumulate the reference phase shift interval. If the time difference exceeds 20%, no correction is performed, and the original rhythm continuation trajectory is continued.

[0070] When matching the predicted pulse occurrence time with unfrozen pulse wave change data within the interference interval, a matching window of 150 milliseconds before and after each predicted pulse occurrence time is used. If unfrozen pulse wave change data exists within the matching window, this data is used as the effective compensation point for the predicted pulse; if no unfrozen data exists within the matching window, compensation is performed based on the time sequence between the previous and next effective compensation points. The trend of continuously missing pulse wave amplitude changes is obtained using a linear transition method, i.e., missing point compensation amplitude = previous effective compensation amplitude + missing point number × difference between adjacent effective compensation amplitudes ÷ total number of missing points plus 1. If the number of consecutive missing points exceeds 3 pulse cycles, amplitude compensation is discontinued, and only the predicted pulse occurrence time is maintained for the continuous temporal expression of vital sign change trends.

[0071] After timing compensation, the continuity of pulse wave amplitude variation is calculated. Amplitude variation continuity is obtained by dividing the average absolute value of the difference between adjacent compensated amplitudes by the stable amplitude reference. If this value does not exceed 0.35, the amplitude variation continuity is considered to meet the compensation requirements. Next, phase continuity consistency is calculated. Phase continuity consistency is obtained by dividing the average absolute value of the difference between the time difference between adjacent predicted pulse occurrence times after compensation and the reference phase shift interval by the reference phase shift interval. If this value does not exceed 0.15, the phase continuity consistency is considered to meet the compensation requirements.

[0072] When both the continuity of amplitude changes and the consistency of phase continuity meet the compensation requirements, the current vital sign trend data is generated. This current vital sign trend data includes the compensated predicted pulse occurrence time sequence, the compensated pulse wave amplitude change sequence, the frozen data timestamp, and the compensation confidence flag. If any requirement is not met, the current vital sign trend data only retains the predicted pulse occurrence time sequence and the frozen data timestamp, and the compensation confidence flag is set to low confidence for subsequent sampling timing and exposure duration adjustments.

[0073] The dynamic adjustment module is used to adjust the sampling timing and exposure time of photoelectric reflection pulse wave data based on the compensated trend of vital signs changes, so as to reduce the impact of strong external stray light on the vital signs detection results.

[0074] During dynamic adjustment, the compensated vital sign trend data is first read. This data includes the predicted pulse occurrence time sequence, the compensated pulse amplitude change sequence, the frozen data time stamp, and the compensation confidence marker. When the compensation confidence marker is low, the sampling timing is adjusted only based on the predicted pulse occurrence time sequence, without widening the exposure time based on the compensated pulse amplitude change sequence, to avoid abnormal amplitudes continuing to affect subsequent acquisitions.

[0075] The stable period of a continuous pulse cycle is extracted based on the time difference between adjacent predicted pulse occurrence times. The time difference between two adjacent predicted pulse occurrence times is recorded as the current cycle duration. Five consecutive cycles are grouped into one judgment group, preferably five cycles. The average of all current cycle durations within the judgment group is calculated, and then the average of the absolute values ​​of the differences between each current cycle duration and the average is calculated to obtain the cycle stability offset value. When the cycle stability offset value does not exceed 15% of the average value, and the continuity of the compensated pulse amplitude change does not exceed 0.35, the time period corresponding to the judgment group is determined as the stable period. If the compensation confidence is marked as low confidence, only the cycle stability offset value is used for judgment, and the threshold for the cycle stability offset value is tightened to 10% of the average value.

[0076] After determining the stable time period, the rising edge region of the pulse wave for each pulse cycle is located within the stable time period. The rising edge region of the pulse wave is the time range between the adjacent trough and the effective peak. When the effective peak in the interference interval is frozen, the position ratio of the phase anchor point in the pulse cycle within the previous stable time period is used as a reference, combined with the predicted pulse occurrence time, to estimate the rising edge region. Preferably, the starting point of the rising edge region is 35% of the current cycle duration before the predicted pulse occurrence time, and the ending point is 5% of the current cycle duration before the predicted pulse occurrence time. If there are unfrozen adjacent troughs and effective peaks, the actual range from the adjacent trough to the effective peak is used. The time range corresponding to the rising edge region in the stable time period is determined as the priority sampling interval.

[0077] When adjusting the sampling timing, the concentrated interval of abnormal light impacts is first determined based on the characteristics of the transient impact density of ambient light. The concentrated interval of abnormal light impacts is formed by the location of abnormal light impact data within a unit of time. If the number of abnormal light impact data points within any 200-millisecond time window is no less than 3, or if the transient impact density value of ambient light within that time window is no less than 80, that time window is marked as the concentrated interval of abnormal light impacts. 80 is the preferred density threshold, which can be set based on 4 times the average value of the transient impact density of ambient light obtained over 60 consecutive seconds under conditions without strong light interference, and must not be less than 60.

[0078] Subsequently, a time-avoidance mechanism is applied between the priority sampling interval and the concentrated area of ​​abnormal light impact. The sampling start time is preferably set within 10 to 30 milliseconds after the beginning of the priority sampling interval. If this sampling start time falls within the concentrated area of ​​abnormal light impact, the sampling start time is shifted backward by 10 milliseconds each time until it leaves the concentrated area of ​​abnormal light impact. If the shifted time exceeds the end of the priority sampling interval, this pulse cycle is marked as an avoidance failure cycle, and the sampling start time is redefined for the next pulse cycle. If three consecutive pulse cycles are avoidance failure cycles, the shortest exposure time is maintained for sampling, and a continuous light interference marker is output.

[0079] The exposure time is adjusted based on the continuity of pulse wave amplitude changes within the continuous priority sampling interval. The continuity of pulse wave amplitude changes is obtained by dividing the average of the absolute values ​​of the differences in pulse wave amplitude changes between adjacent priority sampling intervals by a stable amplitude reference. The preset continuity range is preferably 0 to 0.35; when the continuity of pulse wave amplitude changes does not exceed 0.35, the current exposure time is maintained; when the continuity of pulse wave amplitude changes exceeds 0.35 but does not exceed 0.6, the exposure time is shortened by 1 stop; when the continuity of pulse wave amplitude changes exceeds 0.6, the exposure time is shortened by 2 stops.

[0080] The optimal exposure duration levels are 8 milliseconds, 4 milliseconds, 2 milliseconds, and 1 millisecond, with an initial exposure duration of 4 milliseconds being preferred. When progressively shortening the exposure duration, at least two consecutive pulse cycles must be maintained after each adjustment before the next judgment is made to avoid frequent changes in exposure duration. The cumulative light intensity duration during a single sampling period remains consistent with the exposure duration; if the pulse amplitude variation continuity still exceeds 0.6 after the exposure duration is shortened to 1 millisecond, it will not be shortened further, and the current sampling result will be marked as a low-confidence sampling result.

[0081] After adjusting the sampling timing and exposure duration, the consistency of phase continuity across the adjusted continuous pulse cycles is verified. Phase continuity consistency is calculated by averaging the absolute values ​​of the differences between the times of occurrence of adjacent phase anchor points after adjustment and the baseline phase migration interval, then dividing by the baseline phase migration interval. The preset stability range is preferably 0 to 0.15. If the phase continuity consistency does not exceed 0.15 within five consecutive pulse cycles, and the number of avoidance failure cycles does not exceed one, the phase continuity consistency is considered to have recovered to the preset stability range.

[0082] When the phase continuity consistency returns to a preset stable range, the current sampling sequence and exposure duration are maintained, and the adjusted vital sign detection results are output based on the photoelectric reflection pulse wave data obtained under the current sampling sequence. If the phase continuity consistency still exceeds 0.15 after 5 consecutive pulse cycles, the current shortest effective exposure duration is continued, and a low confidence marker is attached to the adjusted vital sign detection results for subsequent determination of human safety and health status information.

[0083] The status output module is used to output the adjusted vital sign detection results and generate corresponding human safety and health status information.

[0084] During the status output process, the adjusted vital sign detection results are read first. These results include the adjusted heart rate, adjusted blood oxygen saturation, sampling time sequence marker, exposure duration marker, low confidence marker, and corresponding time marker. Status integration is performed sequentially, preferably with a 5-second interval, and adjacent intervals overlapping by 2 seconds. When the wearer is moving rapidly, the integration interval can be extended to 8 seconds to reduce the impact of single motion disturbances on the output results.

[0085] Within each state integration time interval, the continuity of heart rate changes is first extracted. The adjusted heart rate values ​​arranged chronologically within this time interval are recorded as a heart rate sequence. The continuity of heart rate changes is calculated as follows: Heart rate change continuity = Average of the absolute values ​​of the differences between adjacent adjusted heart rate values ​​÷ Average of the adjusted heart rate values ​​within this time interval. The average of the absolute values ​​of the differences between adjacent adjusted heart rate values ​​is obtained by subtracting each adjacent adjusted heart rate value, taking the absolute value, and then averaging all the absolute values. Preferably, when the heart rate change continuity does not exceed 0.12, the heart rate change within this time interval is considered to meet the continuity requirement; when the heart rate change continuity exceeds 0.12 but does not exceed 0.25, the heart rate change within this time interval is considered to have slight fluctuations; when the heart rate change continuity exceeds 0.25, the heart rate change within this time interval is considered to have abnormal fluctuations. The preferred value of 0.12 can be set as twice the average value of the heart rate change continuity over 60 seconds under conditions without strong stray light interference, and the upper limit is not more than 0.15.

[0086] Subsequently, blood oxygen fluctuation stability is extracted. The adjusted blood oxygen values ​​arranged chronologically within this time interval are recorded as a blood oxygen sequence. Blood oxygen fluctuation stability is calculated as follows: Blood oxygen fluctuation stability = average of the absolute values ​​of the differences between each adjusted blood oxygen value and the average blood oxygen value. The average blood oxygen value is the sum of all adjusted blood oxygen values ​​within this time interval divided by the number of adjusted blood oxygen values. Preferably, when the blood oxygen fluctuation stability does not exceed 1.5 percentage points, the blood oxygen fluctuation is considered to meet the stability requirements; when the blood oxygen fluctuation stability exceeds 1.5 percentage points but does not exceed 3 percentage points, the blood oxygen fluctuation is considered to have slight instability; when the blood oxygen fluctuation stability exceeds 3 percentage points, the blood oxygen fluctuation is considered to have abnormal instability. The preferred value of 1.5 percentage points can be set based on twice the average value of blood oxygen fluctuation stability over 60 seconds under conditions without strong stray light interference, and the upper limit is no more than 2 percentage points.

[0087] After extracting the continuity of heart rate changes and the stability of blood oxygen fluctuations, a reliable interval for vital signs is formed by combining the current associated fragment with the corresponding composite interference state. Each state integration time interval corresponds to at least one associated fragment. When the state integration time interval covers multiple associated fragments, the associated fragment with the longest cumulative duration of the composite interference state is taken as the current associated fragment. The reliable interval for vital signs includes a reliable start time, a reliable end time, a reliability level, and an anomaly source marker. The reliability level is determined according to the reliability value, which is calculated as follows: Reliability value = 100 - Heart rate deduction - Blood oxygen deduction - Interference deduction - Low reliability deduction.

[0088] Heart rate deduction is determined based on the continuity of heart rate changes; when the continuity of heart rate changes does not exceed 0.12, the heart rate deduction is 0; when it exceeds 0.12 but does not exceed 0.25, the heart rate deduction is 15; when it exceeds 0.25, the heart rate deduction is 30. Blood oxygen saturation deduction is determined based on the stability of blood oxygen fluctuations; when the stability of blood oxygen fluctuations does not exceed 1.5 percentage points, the blood oxygen deduction is 0; when it exceeds 1.5 percentage points but does not exceed 3 percentage points, the blood oxygen deduction is 15; when it exceeds 3 percentage points, the blood oxygen deduction is 30. Interference deduction is determined based on the state of composite interference; when there is no composite interference in the current associated segment, the interference deduction is 0; when there is composite interference but the cumulative duration is less than 3 seconds, the interference deduction is 20; when the cumulative duration is not less than 3 seconds, the interference deduction is 35. Low confidence deduction is determined based on whether a low confidence marker is attached to the adjusted vital sign detection results; it is 0 when there is no low confidence marker, and 20 when there is a low confidence marker.

[0089] When the confidence value is not lower than 80, the confidence interval for vital signs is recorded as a stable confidence interval; when the confidence value is lower than 80 but not lower than 60, it is recorded as an observation confidence interval; when the confidence value is lower than 60, it is recorded as an abnormal confidence interval. When adjacent confidence intervals for vital signs have the same confidence level and the interval between adjacent intervals does not exceed 2 seconds, the adjacent intervals are merged, and the earliest confidence start time and the latest confidence end time are retained; when adjacent intervals have different confidence levels or the interval exceeds 2 seconds, they are recorded separately.

[0090] The output rules for normal human safety and health status information are as follows: When three consecutive vital sign confidence intervals are all stable confidence intervals, and the adjusted heart rate values ​​within these three confidence intervals are all within the range of 50 to 120 beats per minute, and the adjusted blood oxygen saturation is not lower than 94%, and there is no composite interference state with a cumulative duration of not less than 3 seconds in the currently associated segment, then normal human safety and health status information is output. 50 to 120 beats per minute is the preferred range for general work monitoring, which can be individually set by fluctuating up or down by 20% based on the wearer's registered resting heart rate; 94% is the preferred lower limit for blood oxygen saturation, which can be adjusted to 90% based on a preset benchmark value for high-altitude work scenarios.

[0091] The output rules for abnormal human safety and health status information are as follows: If any vital sign confidence interval is an abnormal confidence interval, and within two consecutive state integration time intervals, any of the following occurs: heart rate change continuity exceeds 0.25, blood oxygen fluctuation stability exceeds 3 percentage points, adjusted heart rate value exceeds the preset range, or adjusted blood oxygen value is lower than the preset lower limit, then abnormal human safety and health status information is output. If an abnormal confidence interval occurs only once within a single state integration time interval, but the next state integration time interval returns to a stable confidence interval, it is recorded as a short-term abnormality, and abnormal human safety and health status information is not output.

[0092] Both normal and abnormal human safety and health status information are time-stamped. The time stamps include the start and end times of the status, the corresponding vital sign confidence interval number, adjusted heart rate value, adjusted blood oxygen value, composite interference status marker, sampling sequence marker, and exposure duration marker. They are stored in chronological order; when two adjacent normal human safety and health status information entries are within 10 seconds of each other and have the same confidence level, they are merged; abnormal human safety and health status information is not merged with adjacent information, but its corresponding vital sign detection results and abnormality source marker are stored independently to facilitate subsequent tracing of the correspondence between strong stray light interference and human safety and health status.

[0093] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. An intelligent helmet for detecting safety and health of a human body, characterized in that, include: The photoelectric data acquisition module is used to acquire photoelectric reflected pulse wave data of the forehead area of ​​the human body after the smart helmet is worn on the human head, and simultaneously acquire ambient light change data around the helmet's optical window; The ambient light impact analysis module is used to extract the number of abnormal light intensity abrupt changes and the discrete change characteristics of light intensity per unit time based on ambient light change data, and form the transient impact density characteristics of ambient light. The photothermal coupling analysis module is used to form the skin photothermal hysteresis coupling characteristics based on the time delay relationship between photoelectric reflected pulse wave data and temperature changes in the forehead area. The phase-locking analysis module is used to form micro-blood flow phase-locking characteristics based on the relationship between rhythm stability and continuous phase changes between consecutive pulse cycles. The interference determination module is used to perform correlation analysis on the ambient light transient impact density characteristics, skin photothermal hysteresis coupling characteristics and microblood flow phase locking characteristics to determine whether the current photoelectric reflection pulse wave data is affected by strong external stray light interference. The dynamic compensation module is used to freeze abnormal pulse wave amplitude data when the judgment result is that there is interference, and to dynamically compensate for the current vital signs change trend based on the phase continuity characteristics of historical stable pulse cycles. The dynamic adjustment module is used to adjust the sampling timing and exposure time of photoelectric reflection pulse wave data based on the compensated trend of vital signs changes, so as to reduce the impact of strong external stray light on the vital signs detection results. The status output module is used to output the adjusted vital sign detection results and generate corresponding human safety and health status information.

2. The smart helmet capable of detecting human safety and health according to claim 1, characterized in that, Acquisition of photoelectric reflection pulse wave data and ambient light change data includes: After detecting that the helmet has formed a stable fit with the forehead area of ​​the human body, the photoelectric reflection acquisition and ambient light acquisition are initiated; synchronous time markers are added to the photoelectric reflection pulse wave data and ambient light change data respectively; the photoelectric reflection pulse wave data and ambient light change data are stored in correspondence according to the synchronous time markers.

3. The smart helmet capable of detecting human safety and health according to claim 1, characterized in that, The formation of transient impact density characteristics of ambient light includes the following steps: The ambient light variation data is divided into continuous and partially overlapping time segments, and the light intensity transition changes between adjacent time segments are extracted. Light intensity transitions that exceed the preset fluctuation range and last for less than the preset transient duration are marked as abnormal light impact data. Based on the frequency, location, and time interval dispersion of abnormal light impact data per unit time, the transient impact density characteristics of ambient light are formed.

4. The smart helmet capable of detecting human safety and health according to claim 1, characterized in that, The formation of the skin's photothermal hysteresis coupling characteristics involves the following steps: Using the time stamp of photoelectric reflection pulse wave data as a reference, the temperature change data of the forehead area is correlated with time. Extract the peak change time and the onset time of temperature rise in the forehead region corresponding to the effective peak, and form a continuous time series; Based on the temperature rise response lag interval, pulse wave amplitude fluctuation degree, and temperature rise duration in the continuous time sequence, the skin photothermal hysteresis coupling characteristics are formed.

5. A smart helmet capable of detecting human safety and health according to claim 4, characterized in that, The formation of microflow phase-locking features involves the following steps: The continuous photoelectric reflection pulse wave data is divided into multiple pulse cycles according to the effective peak and adjacent trough, and the phase anchor point is extracted at the rising edge of each pulse cycle. Based on the positional changes of the phase anchor point within adjacent pulse cycles, a phase migration sequence and rhythm continuity relationship are formed. When a sudden jump, discontinuity, or reverse drift occurs at the phase anchor point, the phase lockout state is marked, and microflow phase lockout characteristics are formed based on the number of occurrences and duration of the phase lockout state.

6. A smart helmet capable of detecting human safety and health according to claim 5, characterized in that, The determination of strong external stray light interference includes the following steps: Correlation segments are established based on the transient impact density characteristics of ambient light, the photothermal hysteresis coupling characteristics of skin, and the phase-locking characteristics of microblood flow within the same time interval; when the change trend of transient impact density of ambient light and the change trend of photothermal hysteresis of skin satisfy the synchronous offset relationship within the same or adjacent correlation segments, a photothermal synchronous perturbation state is formed. When the photothermal synchronization disturbance state and the phase lockout state occur together in a continuous associated segment, the composite interference state is marked, and the strong stray light interference from the outside is judged based on the cumulative duration and recurrence of the composite interference state.

7. A smart helmet capable of detecting human safety and health according to claim 6, characterized in that, Dynamic compensation includes the following steps: When subjected to strong stray light interference from the outside, the pulse wave amplitude data that exceeds the preset amplitude fluctuation range within the interference range is marked as frozen data, and the phase arrangement order of the continuous stable pulse cycles before the interference range is retained. A rhythmic continuation trajectory is formed based on the phase shift interval between continuous and stable pulse cycles, and the predicted pulse occurrence time within the interference interval is determined by the rhythmic continuation trajectory; the predicted pulse occurrence time is matched with the unfrozen pulse wave change data, and the missing pulse wave amplitude change trend is compensated for in a time sequence.

8. A smart helmet capable of detecting human safety and health according to claim 7, characterized in that, Adjusting the sampling timing and exposure time involves the following steps: Based on the compensated trend of vital signs, the stable period of the continuous pulse cycle is extracted, and the rising edge region of the pulse wave within the stable period is determined as the priority sampling interval. Based on the transient impact density characteristics of ambient light, the concentrated area of ​​abnormal light impact is determined, and the sampling start time is avoided in the concentrated area of ​​abnormal light impact. The exposure time is gradually shortened according to the continuity of pulse wave amplitude changes within the priority sampling interval, and the current sampling sequence and exposure time are maintained when the phase continuity consistency is restored to the preset stable range.

9. A smart helmet capable of detecting human safety and health according to claim 8, characterized in that, The output of vital sign test results includes the following steps: The adjusted vital sign test results are integrated in a continuous time sequence, and the continuity of heart rate changes and the stability of blood oxygen fluctuations are extracted. The confidence intervals of vital signs are formed based on the continuity of heart rate changes, the stability of blood oxygen fluctuations, the state of composite interference, and low confidence markers. Based on the confidence level of the continuous vital signs confidence interval, output information on the safety and health status of normal or abnormal human beings.

10. A smart helmet capable of detecting human safety and health according to claim 9, characterized in that, The storage of human safety and health status information includes the following steps: Add state start time, state end time, composite interference state marker, sampling time sequence marker and exposure duration marker to normal human safety and health status information and abnormal human safety and health status information; merge and store normal human safety and health status information with the same confidence level and time interval not exceeding the preset interval; Abnormal human safety and health status information is stored independently and associated with corresponding vital sign detection results and abnormality source markers.

Images