Intelligent robot wounded consciousness state interpretation method based on voiceprint recognition

Abstract

The invention discloses a wounded consciousness state interpretation method of an intelligent robot based on voiceprint recognition, relates to the technical field of voiceprint recognition, and solves the problems of multi-band frequency spectrum partitioning, stable feature segment screening and invalid clutter elimination of voice data. Sound frequencies of different wave crests in each frequency spectrum band are deeply analyzed through a unique stable frequency spectrum band selection mode; in the invalid clutter elimination process, wave bands are strictly screened according to frequency change characteristics, and invalid clutters with the frequencies not obviously changed or the change degree deviating greatly are accurately removed. The accuracy of judging the consciousness state of the wounded person is greatly improved through a series of rigorous processing procedures; compared with a traditional method for judging the consciousness state by means of simple voice volume and intonation, the method can effectively avoid misjudgment and missed judgment caused by factors such as environmental noise interference and different pronunciation habits of the wounded.

Description

Technical Field

[0001] This invention relates to the field of voiceprint recognition technology, specifically a method for intelligent robots to determine the consciousness state of wounded patients based on voiceprint recognition. Background Technology

[0002] In disaster relief and emergency medical scenarios, the rapid and accurate assessment of the injured's level of consciousness is a core prerequisite for subsequent rescue and treatment, directly impacting rescue efficiency and the safety of the injured. Traditional methods for assessing the injured's consciousness often rely on close observation by rescuers on-site, achieved through shouting, tapping, or simple physiological response tests. However, this approach has significant limitations: firstly, in complex rescue environments, rescuers cannot quickly reach all injured individuals, and close contact may pose a risk of secondary disasters; secondly, some injured individuals, due to physical weakness, traumatic pain, or other reasons, are unable to produce obvious limb responses and can only utter weak sounds. Traditional methods struggle to effectively assess their level of consciousness through such vocal information, easily leading to misjudgments or omissions.

[0003] With the widespread application of intelligent robot technology in rescue operations, remote casualty monitoring via robots equipped with sensors has become a new development direction. Voiceprint recognition technology, as a non-contact biometric identification method, offers advantages such as requiring no active cooperation from the casualty and enabling remote data collection and analysis, providing a new technological path for assessing the casualty's state of consciousness. However, the application of existing voiceprint recognition technology in casualty scenarios still faces challenges: complex background noise at rescue sites easily interferes with voice signal acquisition; casualty voice signals are often unstable, making it difficult for traditional spectral analysis methods to accurately extract effective features; and the lack of standardized spectral feature association models for different states of consciousness makes it difficult to meet rescue requirements in terms of accuracy.

[0004] Against this backdrop, there is an urgent need for a technical method that can adapt to complex rescue environments, accurately extract the spectral features of the injured person's speech, and efficiently determine their state of consciousness. By performing multi-band spectral partitioning, stable feature segment screening, and invalid clutter removal on the speech data, combined with comprehensive feature quantification analysis, the limitations of traditional methods can be effectively overcome. This provides reliable technical support for intelligent robots to achieve rapid and accurate determination of the injured person's state of consciousness, further improving emergency response capabilities and the success rate of injured person treatment in rescue scenarios. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method for judging the consciousness state of wounded patients based on voiceprint recognition by intelligent robots, which solves the problems of multi-band spectrum partitioning, stable feature segment screening, and invalid clutter removal of voice data.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for judging the consciousness state of wounded soldiers based on voiceprint recognition by an intelligent robot, comprising the following steps:

[0007] Step 1: Perform spectral analysis on the collected voice data about the injured person, generate the sound spectrum associated with the voice data, and confirm the frequency of the sound spectrum, identifying the low-frequency, mid-frequency, and high-frequency bands present within the sound spectrum. The specific method is as follows:

[0008] During the sound spectrum confirmation process, a sound spectrum belonging to the corresponding speech data is generated based on the different sound frequency data associated with different times within the corresponding speech data. According to the preset frequency range, the sound frequencies associated with different peaks within the sound spectrum are confirmed, and the frequency range to which they belong are determined. Spectrum segments belonging to the same frequency range are recorded as the same interval segment.

[0009] The spectrum range includes the low-frequency range, the mid-frequency range, and the high-frequency range. The spectrum segment belonging to the low-frequency range is called the low-frequency spectrum segment, the spectrum segment belonging to the mid-frequency range is called the mid-frequency spectrum segment, and the spectrum segment belonging to the high-frequency range is called the high-frequency spectrum segment.

[0010] Step 2: After confirming all different spectral segments within the sound spectrogram, perform feature verification on the sound frequencies associated with different peaks within each different spectral segment. Select a stable spectral segment from the corresponding spectral segments and use the selected stable spectral segment as the feature segment of the corresponding spectral segment. The specific method is as follows:

[0011] The associated spectral segments of different spectral intervals are identified, and the sound frequencies associated with different peaks within a single spectral interval are denoted as P. i-k , where i represents different frequency ranges and k represents different peak points;

[0012] A set of sound frequencies associated with a random set of peaks are denoted as the intermediate frequency ZJ. A selection range is confirmed based on a preset threshold Y1, which is [ZJ-Y1, ZJ+Y1]. Other peaks belonging to this selection range are denoted as the subordinate peaks of the selected peak. The peak duration associated with the selected peak and each subordinate peak is determined. Valley points are determined within the waveform segments on both sides of the peak. The waveform trend of the front end of the valley point is upward and the waveform trend of the back end is downward. The waveform trend of the front end of the valley point is downward and the waveform trend of the back end is upward. The time length associated between two valley points is denoted as the time span of the current peak. Several sets of time spans associated with the selected peak and each subordinate peak are summed to confirm the total time span. The confirmed total time span is denoted as the subordinate feature of the selected peak.

[0013] Different peaks are selected in sequence, and the associated features of each selected peak are confirmed in sequence. From the confirmed sets of associated features, the maximum value is selected. The selected peak associated with the maximum value is recorded as the determined peak, and the spectral segment associated with the determined peak and the associated peak is recorded as the stable spectral segment of the current spectral interval.

[0014] Step 3: After determining the characteristic segments associated with each spectral interval, determine the frequency variation characteristics of the individual bands associated with each characteristic segment. Based on the determined frequency variation characteristics within the characteristic segment, eliminate invalid clutter within the characteristic segment. The specific method is as follows:

[0015] For each characteristic segment, the individual bands are sequentially identified. Based on the peaks marked on the characteristic segment and the two sets of valleys associated with the peaks, the bands between the peaks and the two sets of valleys are confirmed. Within the identified bands, the highest and lowest frequencies are determined. Based on the highest and lowest frequencies, the frequency variation characteristics associated with the corresponding peaks are confirmed, where the frequency variation characteristic = highest frequency - lowest frequency. The different frequency variation characteristics associated with different peaks within the characteristic segment are denoted as T. k , where k represents different wave crests;

[0016] The several sets of frequency change features T associated within the feature segment k The frequency sequence is confirmed by sorting the values ​​from smallest to largest. A frequency segment is randomly selected from the frequency sequence, and the variance of several sets of frequency change characteristics of the selected frequency segment is confirmed. The characteristic variance of this frequency segment is locked. If the characteristic variance is ≤ Y2, where Y2 is a preset value, this frequency segment is recorded as the calibration segment. If the characteristic variance is > Y2, no marking is performed.

[0017] From the marked calibration segments, identify the total number of frequency variation features included in the corresponding calibration segment, and denote it as G. q Where q represents different calibration segments, and then from the determined G... q In the process, select the maximum value and denote the calibration segment associated with the maximum value as the determined segment;

[0018] The frequency variation characteristics associated with a defined segment are identified, and other invalid clutter within the segment that is not marked as a defined characteristic is removed. If the maximum value G q If there are multiple groups, the calibration segment associated with the maximum value of a group is randomly selected and recorded as the determined segment;

[0019] Step 4: Determine the comprehensive characteristics associated with different spectral intervals, confirm the state of consciousness based on the comprehensive characteristics, and generate corresponding proportional coefficients for display. The specific method is as follows:

[0020] The characteristic segments that have completed invalid clutter removal in each frequency range are confirmed. From the confirmed characteristic segments, the sound frequencies of several confirmed peaks are averaged and the average frequency is confirmed as the first feature Z1 of the corresponding frequency range. Then, the time spans associated with the corresponding bands of several peaks are summed and used as the second feature Z2 of the corresponding frequency range.

[0021] The comprehensive characteristic QJ of the corresponding frequency range is determined by the formula: QJ = (Z2 × C1) ÷ (Z1 × C2), where C1 and C2 are preset fixed coefficient factors.

[0022] From the different comprehensive features QJ associated with different frequency ranges, the maximum value is selected, the frequency range associated with the maximum value is recorded as the fixed range, and the consciousness state associated with the corresponding fixed range is recorded as the fixed state. All of these consciousness states are preset states.

[0023] Summing up the comprehensive features associated with the other two spectral intervals, confirming the total sum, then comparing the comprehensive feature QJ of the determined interval with the total sum to confirm the proportionality coefficient, which is calculated as QJ ÷ total sum. The confirmed proportionality coefficient is then displayed.

[0024] This invention provides a method for judging the consciousness state of wounded soldiers based on voiceprint recognition using intelligent robots. Compared with existing technologies, it has the following advantages:

[0025] This invention, through detailed multi-band analysis of the sound spectrum, divides the sound spectrum into low-frequency, mid-frequency, and high-frequency bands, comprehensively capturing the frequency characteristics in speech data.

[0026] In the feature verification stage, a unique stable spectrum segment selection method is used to conduct in-depth analysis of the sound frequencies of different peaks within each spectrum segment;

[0027] During the invalid clutter removal process, the bands are strictly screened based on the frequency change characteristics to accurately remove invalid clutter whose frequency has not changed significantly or whose degree of change has shifted greatly.

[0028] This series of rigorous procedures greatly improves the accuracy of assessing the consciousness level of the injured. Compared with traditional methods that rely on simple voice volume and tone to judge consciousness level, this method can effectively avoid misjudgments and omissions caused by environmental noise interference, different pronunciation habits of the injured, etc., and provide rescuers with more reliable information on the consciousness level of the injured. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0030] 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.

[0031] Please see Figure 1 This application provides a method for judging the consciousness state of wounded soldiers based on voiceprint recognition by intelligent robots, including the following steps:

[0032] Step 1: Perform spectrum verification on the collected voice data about the wounded, generate the sound spectrum associated with the voice data, and verify the frequency of the sound spectrum. Verify the low-frequency, mid-frequency, and high-frequency spectrum segments that exist within the sound spectrum. Specifically, different partitions are associated with different frequency characteristics, and each frequency characteristic is associated with a different numerical range. The numerical range is a preset range, which is determined in advance by relevant personnel to complete the corresponding frequency partition verification process.

[0033] The specific method for frequency confirmation of the sound spectrum is as follows:

[0034] During the sound spectrum confirmation process, based on the different sound frequency data associated with different times within the corresponding speech data, a sound spectrum belonging to the corresponding speech data is generated (its horizontal axis is the time line, and the vertical axis is the sound frequency; the method of generating the sound spectrum associated with the corresponding speech is quite common in existing technologies, and will not be elaborated on here; that is, the corresponding sound spectrum diagram). According to the preset frequency range, the sound frequencies associated with different peaks within the sound spectrum are confirmed, and the frequency range to which they belong is determined. Spectral segments belonging to the same frequency range are recorded as the same interval segment.

[0035] The spectrum range includes the low-frequency range, the mid-frequency range, and the high-frequency range. The spectrum segment belonging to the low-frequency range is called the low-frequency spectrum segment, the spectrum segment belonging to the mid-frequency range is called the mid-frequency spectrum segment, and the spectrum segment belonging to the high-frequency range is called the high-frequency spectrum segment.

[0036] Specifically, within the corresponding frequency band, there are several peak points. Each peak point is a convex turning point. The trend of the line segment before the peak point is upward, and the trend of the line segment after the peak point is downward. Therefore, the corresponding peak point is a convex starting point. Within the corresponding frequency spectrum, the frequency characteristics associated with the corresponding peak point can be quickly identified.

[0037] Step 2: After confirming all different frequency bands in the sound spectrum, perform feature verification on the sound frequencies associated with different peaks in each different frequency band, select a stable frequency band from the corresponding frequency band, and use the selected stable frequency band as the feature segment of the corresponding frequency band.

[0038] The specific method for selecting stable spectrum bands is as follows:

[0039] The associated spectral segments of different spectral intervals are identified, and the sound frequencies associated with different peaks within a single spectral interval are denoted as P. i-k , where i represents different frequency ranges and k represents different peak points;

[0040] A set of sound frequencies associated with a random set of peaks are denoted as the intermediate frequency ZJ. A selection range is confirmed based on a preset threshold Y1, which is [ZJ-Y1, ZJ+Y1]. Other peaks belonging to this selection range are denoted as the subordinate peaks of the selected peak. The peak duration associated with the selected peak and each subordinate peak is determined. Valley points are determined within the waveform segments on both sides of the peak. The waveform trend of the front end of the valley point is upward and the waveform trend of the back end is downward. The waveform trend of the front end of the valley point is downward and the waveform trend of the back end is upward. The time length associated between two valley points is denoted as the time span of the current peak. Several sets of time spans associated with the selected peak and each subordinate peak are summed to confirm the total time span. The confirmed total time span is denoted as the subordinate feature of the selected peak.

[0041] Different peaks are selected in sequence, and the associated features of each selected peak are confirmed in sequence. From the confirmed sets of associated features, the maximum value is selected. The selected peak associated with the maximum value is recorded as the determined peak, and the spectral segment associated with the determined peak and the associated peak is recorded as the stable spectral segment of the current spectral interval.

[0042] Specifically, each spectral interval is associated with different spectral bands. Within the corresponding spectral bands, there are some bands with large frequency changes and some bands with small frequency changes. In order to effectively identify relatively stable frequency bands, it is necessary to verify and analyze the stability characteristics between different frequency bands. According to the specific verification and analysis process, the frequency selection range associated with the corresponding peak point is determined. Then, different spectral bands are selected according to the frequency selection range. Based on several selection and processing processes, the corresponding stable spectral bands can be quickly and effectively identified, which facilitates subsequent feature assessment and identification of the consciousness status of the corresponding injured persons.

[0043] Step 3: After determining the characteristic segments associated with each spectral interval, determine the frequency variation characteristics of the individual bands associated with each characteristic segment. Based on the determined frequency variation characteristics within the characteristic segment, eliminate invalid clutter within the characteristic segment.

[0044] For each characteristic segment, the individual bands are sequentially identified. Based on the peaks marked on the characteristic segment and the two sets of valleys associated with each peak, the bands between the peaks and the two sets of valleys are confirmed. Within the identified bands, the highest frequency (peak) and the lowest frequency are determined. Based on the highest and lowest frequencies, the frequency variation characteristics associated with the corresponding peaks are confirmed, and the frequency variation characteristic = highest frequency - lowest frequency. The different frequency variation characteristics associated with different peaks within the characteristic segment are denoted as T. k , where k represents different wave crests;

[0045] The several sets of frequency change features T associated within the feature segment k The frequencies are sorted from smallest to largest to confirm the frequency sequence. Frequency segments are randomly selected from the frequency sequence (the original sorting method remains unchanged, and the frequency segments are different numerical segments within the frequency sequence). The variance of several sets of frequency change characteristics of the selected frequency segments is confirmed, and the characteristic variance of this frequency segment is locked. If the characteristic variance is ≤ Y2, where Y2 is a preset value, the specific value of which is determined by the operator based on experience, then this frequency segment is recorded as the calibration segment. Otherwise, no marking is performed.

[0046] From the marked calibration segments, identify the total number of frequency variation features included in the corresponding calibration segment, and denote it as G. q Where q represents different calibration segments, and then from the determined G... q In the process, the maximum value is selected, and the calibration segment associated with the maximum value is recorded as the determined segment. If there are multiple sets of maximum values, a set of calibration segments associated with the maximum value is randomly selected and recorded as the determined segment.

[0047] The frequency change characteristics associated with the defined segment are identified, and other invalid clutter within the segment that is not marked as a defined characteristic is removed.

[0048] Specifically, invalid clutter refers to clutter whose frequency band does not change significantly or whose frequency changes by a large margin. In such cases, invalid clutter is removed from the corresponding frequency band, while valid clutter is retained. This facilitates subsequent comprehensive verification of the characteristic bands to ensure the accuracy of the determination.

[0049] Step 4: Determine the comprehensive characteristics associated with different spectral intervals, confirm the state of consciousness based on the comprehensive characteristics, and generate corresponding proportional coefficients for display:

[0050] The characteristic segments that have completed invalid clutter removal in each frequency range are confirmed. From the confirmed characteristic segments, the sound frequencies of several confirmed peaks are averaged and the average frequency is confirmed as the first feature Z1 of the corresponding frequency range. Then, the time spans associated with the corresponding bands of several peaks are summed and used as the second feature Z2 of the corresponding frequency range.

[0051] The comprehensive characteristic QJ of the corresponding frequency range is confirmed by the formula: QJ = (Z2 × C1) ÷ (Z1 × C2), where C1 and C2 are preset fixed coefficient factors. Their specific values ​​are determined by the operator based on experience. The larger the total time span and the smaller the frequency, the greater the associated comprehensive characteristic.

[0052] From the different comprehensive features QJ associated with different frequency spectrum intervals, select the maximum value, and record the frequency spectrum interval associated with the maximum value as the determined interval. Record the consciousness state associated with the corresponding determined interval as the determined state. All consciousness states are preset states, which are preset and bound in advance by relevant personnel. Sum the comprehensive features associated with the other two frequency spectrum intervals, confirm the total value, and then compare the comprehensive feature QJ of the determined interval with the total value to confirm the proportional coefficient. The proportional coefficient = QJ ÷ total value. Display the confirmed proportional coefficient.

[0053] Specifically, different frequency ranges are associated with different states of consciousness. The low-frequency range is generally associated with a depressed state, the mid-frequency range with a normal state, and the high-frequency range with an excessive state (which may be caused by shouting, resulting in a high frequency state in the sound spectrum). The states of consciousness are all preset in advance and need to be determined according to the actual assessment process. The proportion coefficients should be displayed to facilitate review by relevant personnel.

[0054] Some of the data in the above formulas are numerical calculations with dimensions removed, and the contents not described in detail in this specification are all prior art known to those skilled in the art.

[0055] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A method for judging the consciousness state of a wounded person based on voiceprint recognition of an intelligent robot, characterized in that, Comprise the following steps: Step one, the frequency spectrum of the collected voice data about the wounded is confirmed, the sound spectrum associated with the voice data is generated, and the frequency of the sound spectrum is confirmed, and the low frequency spectrum, the medium frequency spectrum and the high frequency spectrum existing in the sound spectrum are confirmed; Step two, after the different frequency spectrum bands in the sound spectrum graph are confirmed, the sound frequency associated with different wave peaks in each different frequency spectrum band is verified, a stable frequency spectrum band is selected from the corresponding frequency spectrum band, and the selected stable frequency spectrum band is taken as the characteristic band of the corresponding frequency spectrum band; Step three, after the characteristic band associated with each frequency spectrum interval is determined, the frequency variation characteristics of each characteristic band associated with a single wave band are determined, and the invalid clutter existing in the characteristic band is removed according to the frequency variation characteristics determined in the characteristic band; Step four, the comprehensive characteristics associated with different frequency spectrum intervals are determined, and the consciousness state is confirmed according to the comprehensive characteristics, and the corresponding proportion coefficient is generated and displayed. 2.The method of claim 1, wherein the method further comprises: In the step one, the specific way of frequency confirmation of sound spectrum is: In the sound spectrum confirmation process, different sound frequency data associated with different times in the corresponding voice data is generated to generate the sound spectrum belonging to the corresponding voice data, the sound frequency associated with different wave peaks is confirmed from the sound spectrum according to the preset frequency interval, the belonging frequency interval is determined, and the frequency spectrum band belonging to the same interval is recorded as the same interval band; Wherein, the frequency spectrum interval includes low frequency interval, medium frequency interval and high frequency interval, the frequency spectrum band belonging to the low frequency interval is recorded as the low frequency spectrum band, the frequency spectrum band belonging to the medium frequency interval is recorded as the medium frequency spectrum band, and the frequency spectrum band belonging to the high frequency interval is recorded as the high frequency spectrum band. 3.The method of claim 1, wherein the method further comprises: In the step two, the specific way of selecting stable frequency spectrum band is: The sound frequencies associated with different peaks in the spectrum segment of a single spectrum interval are denoted as P i-k where i represents different spectrum intervals and k represents different peak points. Randomly select a group of wave peaks associated with sound frequency as intermediate frequency ZJ, and confirm a group of selection range based on the preset threshold Y1, the selection range is [ZJ-Y1, ZJ+Y1], the other wave peaks belonging to the selection range are recorded as the affiliated wave peaks of the selected wave peak, and the wave peak time associated with the selected wave peak and each affiliated wave peak is determined, the valley points are determined in the wave form band on both sides of the wave peak, the front end wave form trend of the front end valley point of the wave peak is upward, and the rear end wave form trend is downward, the front end wave form trend of the rear end valley point of the wave peak is downward, and the rear end wave form trend is upward, the time length associated between the two valley points is recorded as the time span of the current wave peak, and the time span associated with the selected wave peak and each affiliated wave peak is summed to confirm the total time span, and the confirmed total time span is recorded as the affiliated characteristics of the selected wave peak; The different wave peaks are selected in turn, and the affiliated characteristics of each selected wave peak are confirmed in turn, the maximum value is selected from the confirmed several groups of affiliated characteristics, and the selected wave peak associated with the maximum value is recorded as the determined wave peak, and the part of the frequency spectrum band associated with the determined wave peak and the affiliated wave peak is recorded as the stable frequency spectrum band of the current frequency spectrum interval. 4.The method of claim 1, wherein the method further comprises: In the step three, the specific way of removing invalid clutter is: The single wave band existing in each characteristic section is determined in sequence, the wave peak marked by the characteristic section and the two groups of valley points associated before and after the corresponding wave peak are confirmed, the wave band between the wave peak and the two groups of valley points before and after the corresponding wave peak is confirmed, the highest frequency and the lowest frequency in the determined wave band are confirmed, the frequency change characteristic associated with the corresponding wave peak is confirmed according to the highest frequency and the lowest frequency, and the frequency change characteristic = the highest frequency - the lowest frequency. Different frequency change characteristics associated with different wave peaks in the characteristic section are recorded as T k wherein k represents different wave peak points. The several groups of frequency variation characteristics T k According to the order from small to large, the frequency sequence is confirmed, a frequency segment is randomly selected from the frequency sequence, the variance of the several groups of frequency variation characteristics of the selected frequency segment is confirmed, the characteristic variance of the frequency segment is locked, if the characteristic variance ≤ Y2, wherein Y2 is a preset value, the frequency segment is recorded as a calibration segment; From the marked several calibration segments, the total number of frequency change characteristics included in the corresponding calibration segment is identified and recorded as G q where q represents different calibration segments, and the maximum value is selected from the determined several G q The calibration segment associated with the maximum value is recorded as the determined segment. The frequency variation characteristics associated with the determined segment are determined as the determined characteristics, and other invalid clutter in the characteristic segment that is not marked as a determined characteristic is removed. 5.The method of claim 4, wherein the method further comprises: If the selected maximum value G q If there are multiple groups, randomly select a group of maximum value associated with the calibration section as the determination section. 6.The method of claim 4, wherein the method further comprises: If the characteristic variance > Y2, no marking processing is performed. 7.The method of claim 1, wherein the method further comprises: determining a voiceprint of the wounded person based on the voice signal; and determining the consciousness state of the wounded person based on the voiceprint. In the fourth step, the specific manner of confirming the state of consciousness is: Confirm the characteristic segment in which the invalid clutter removal is completed in each frequency spectrum interval, and the sound frequency of the confirmed wave peak is processed by mean value from the confirmed characteristic segment. The mean frequency is confirmed as the first characteristic Z1 of the corresponding frequency spectrum interval. The time span associated with the wave band corresponding to the wave peak is summed up as the second characteristic Z2 of the corresponding frequency spectrum interval. The comprehensive characteristic QJ of the corresponding frequency spectrum interval is confirmed by using: QJ= (Z2×C1) ÷ (Z1×C2), wherein C1 and C2 are both preset fixed coefficient factors. From different comprehensive characteristics QJ associated with different frequency spectrum intervals, the maximum value is selected, the frequency spectrum interval associated with the maximum value is recorded as a determined interval, and the state of consciousness corresponding to the determined interval is recorded as a determined state, and the state of consciousness is a preset state. 8.The method of claim 7, wherein the method further comprises: determining the consciousness state of the wounded person based on the voiceprint recognition result. In the fourth step, the specific manner of generating the proportionality coefficient is: The comprehensive characteristics associated with the other two groups of frequency spectrum intervals are summed up to confirm the sum value, and the comprehensive characteristic QJ of the determined interval is compared with the sum value to confirm the proportionality coefficient, the proportionality coefficient = QJ ÷ sum value, and the confirmed proportionality coefficient is displayed.

Images