WO2018166136A1 - 波形信号处理的方法及装置 - Google Patents

波形信号处理的方法及装置 Download PDF

Info

Publication number
WO2018166136A1
WO2018166136A1 PCT/CN2017/095073 CN2017095073W WO2018166136A1 WO 2018166136 A1 WO2018166136 A1 WO 2018166136A1 CN 2017095073 W CN2017095073 W CN 2017095073W WO 2018166136 A1 WO2018166136 A1 WO 2018166136A1
Authority
WO
WIPO (PCT)
Prior art keywords
line segment
signal
waveform
data
signal line
Prior art date
Application number
PCT/CN2017/095073
Other languages
English (en)
French (fr)
Inventor
董辰
张�杰
吕超
陈宜欣
许培达
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to CN201780009428.4A priority Critical patent/CN108701219B/zh
Priority to US16/494,573 priority patent/US11321561B2/en
Publication of WO2018166136A1 publication Critical patent/WO2018166136A1/zh

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A61B5/346—Analysis of electrocardiograms
    • A61B5/349—Detecting specific parameters of the electrocardiograph cycle
    • A61B5/366—Detecting abnormal QRS complex, e.g. widening
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • A61B5/318—Heart-related electrical modalities, e.g. electrocardiography [ECG]
    • A61B5/346—Analysis of electrocardiograms
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/24—Detecting, measuring or recording bioelectric or biomagnetic signals of the body or parts thereof
    • A61B5/316—Modalities, i.e. specific diagnostic methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction
    • G06F2218/10—Feature extraction by analysing the shape of a waveform, e.g. extracting parameters relating to peaks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching

Definitions

  • the present application relates to the field of signal processing, and in particular, to a method and an apparatus for processing a waveform signal.
  • ECG signal, the photoelectric easy pulse wave signal or the motion signal detected by the sensor on the wearable device are usually waveform signals with periodicity and continuity, so the user's physical health can be perceived by monitoring the waveform signal. You can find functions such as the step counting period.
  • the waveform signal detected by the sensor of the wearable device is easily affected by factors such as the wearing manner of the wearable device or the behavior of the user, and the waveform signal includes a large noise, causing the waveform signal to be unstable, or the signal period is unstable. Such an abnormal phenomenon, the quality of the waveform signal is poor, and the reliability is low.
  • the embodiment of the present application provides a method and a device for processing a waveform signal, which can extract a signal line segment feature of a waveform signal, improve the recognition efficiency of a waveform signal period, and enhance the applicability of the waveform signal.
  • the first aspect provides a method for processing a waveform signal, which firstly performs processing such as cleaning and filtering on the original waveform signal to obtain a waveform signal after the filtering process.
  • the waveform signal may be labeled as K signal line segments according to monotonicity, K is an integer greater than or equal to 2, and line segment data of each signal line segment is extracted.
  • the waveform signal is segmented, and the feature data of each line segment is extracted in stages, which can improve the signal processing precision of the waveform signal and improve the feature recognition accuracy of the waveform signal.
  • the above method may further determine a line segment matching template of the waveform signal according to the line segment data of each signal line segment, wherein the line segment matching template includes M consecutive signal line segments, M is an integer smaller than K; K signal line segments obtained by dividing the waveform signal Each of the signal line segments is matched with the M signal line segments included in the line segment matching template, and the target wave group of the waveform signal is determined according to the matching result of each signal line segment; and the waveform signal is determined according to the line segment data of the target wave group.
  • Periodic signal data is performed by a line segment matching template, wherein the line segment matching template includes M consecutive signal line segments, M is an integer smaller than K; K signal line segments obtained by dividing the waveform signal.
  • the embodiment of the present application may determine, according to the extracted feature data of each line segment, a line segment matching template for identifying a period of the waveform signal, and the line segment matching template is closely related to the waveform signal.
  • the target wave group in a single cycle of the waveform signal is first identified by the line segment matching template, and the period of the waveform signal is recognized by the target wave group, and the cycle recognition accuracy of the waveform signal is higher, and the waveform signal is improved. Signal processing quality, more applicability.
  • the waveform signal is divided into a monotonous upward signal line segment or a monotonous downward signal line segment according to monotonicity, and each signal line segment is sequentially labeled according to time continuity to obtain a sequence label.
  • Line segment In the embodiment of the present application, the waveform signal can be divided into multiple signal line segments according to the monotonicity of the line segment, and the monotonicity of each signal line segment is unique, and then each signal line segment can be sequentially labeled according to time continuity to better extract.
  • the characteristic data of each signal line segment can integrate the feature data of each signal line segment according to time continuity to better determine the line segment matching template, improve the recognition efficiency of the target wave group of the waveform signal, and improve the accuracy of the period identification of the waveform signal. Rate, more applicability.
  • the embodiment of the present application may extract the line segment length Xi and the line segment width Yi of the signal line segment i, and perform differential stretching on the Xi and Yi according to the preset length and the preset width to obtain a normalized signal.
  • Line segment j extracting the line segment data of the signal line segment j; wherein the line segment data of the signal line segment j includes: the position of the middle point of the line segment, the curvature of the starting point position of the line segment, the curvature of the end point position of the line segment, the ray tangent point of the line segment starting point position and the line segment tangent, and the line segment At least one of the ray tangent points at which the end position is tangent to the line segment.
  • the embodiment of the present application can extract individual features of each signal line segment, and normalize each signal line segment to extract features of each signal line segment, thereby better identifying characteristics of each signal line segment to be processed according to characteristics of each signal line segment.
  • Waveform signal which in turn improves the quality of the waveform signal
  • the embodiment of the present application may further derivate or differentiate the signal line segment j to obtain a derivative curve of the signal line segment j, and extract line segment data of the derivative curve; wherein the line segment data of the derivative curve includes: the number of inflection points, each At least one of an inflection point position, an intermediate point position between adjacent inflection points, and a peak inflection point position.
  • each signal line segment obtained by dividing the waveform signal may be subjected to derivation or differential processing, and line segment data of the derivative curve after the derivation or differential processing may be extracted, and the characteristics of each signal line segment divided by the waveform signal may be further identified, thereby improving
  • the extraction precision of the line segment feature data of the waveform signal improves the processing quality of the waveform signal.
  • the M target signal line segments whose line segment data is smaller than the preset line segment data threshold are selected from the line segment data of each signal line segment, and the M target signal line segments are combined according to the preset line segment combination manner.
  • the embodiment of the present application can obtain the line segment matching template of the waveform signal according to the line segment data of each signal line segment of the waveform signal, and enhance the correlation tightness between the line segment matching template and the waveform signal, thereby improving the period of identifying the waveform signal by the line segment matching template. Accuracy and applicability.
  • the embodiment of the present application may match each of the K signal line segments with the line segment data of each of the M target signal line segments included in the line segment matching template, from the K signal line segments. Determining N effective signal line segments, N is an integer less than M, and the characteristic difference between the line segment data of the valid signal line segment and the line segment data of any target line segment included in the line segment matching template is not less than a preset threshold; if the N valid signals are The line segment includes M valid signal line segments consecutively numbered, and the M effective signal line segments respectively have a characteristic difference value of the M target signal line segments included in the line segment matching template is less than a preset threshold, and the M valid signals are used. The position of the line segment is determined as the target wave group position.
  • the embodiment of the present application can identify the position of the target wave group included in the waveform signal from each signal line segment of the waveform signal by using the line segment matching template of the waveform signal and a preset data error threshold, and further, according to the position of the target wave group.
  • the period of the waveform signal is recognized, the cycle recognition accuracy of the waveform signal is improved, the processing quality of the waveform signal is improved, and the applicability is stronger.
  • the line segment data of the target wave group position includes line segment data of M valid signal line segments at the target wave group position, and line segment data of J adjacent signal line segments consecutive to the sequential label of the M valid signal line segments.
  • the embodiment of the present application can determine the waveform width of the wave group composed of the M effective signal line segments and the J adjacent signal line segments according to the line data of the M valid signal line segments and the J adjacent signal line segments, and the waveform amplitude of the wave group.
  • the wave group characteristic data of the wave group is periodic signal data of the waveform signal; wherein the periodic check data includes: a periodic waveform width, a periodic waveform amplitude, At least one of a cycle start waveform height and a cycle end waveform height.
  • the embodiment of the present application may determine, according to the periodic signal data of the waveform signal, a plurality of waveform periods included in the waveform signal, and delete the abnormal period data included in the waveform signal.
  • the abnormal cycle data is line segment data of a wave group that does not include the periodic signal data in the waveform signal.
  • the embodiment of the present application can eliminate the abnormal period data in the waveform signal, improve the signal quality of the waveform signal, and improve the applicability of the waveform signal.
  • the waveform signal described in the embodiment of the present application may include at least one of an electrocardiogram ECG, a photoelectric volume pulse wave PPG, a pressure gauge signal, a magnetometer signal, and an accelerometer sensor waveform signal.
  • the method provided by the embodiment of the present application is applicable to processing multiple types of waveform signals, has flexible operation, and has wide adaptability.
  • the second aspect provides an apparatus for waveform signal processing, which may include:
  • An acquiring unit configured to obtain a waveform signal after filtering
  • a segmentation unit configured to mark the waveform signal acquired by the acquiring unit as K signal line segments according to monotonicity, where K is an integer greater than or equal to 2;
  • An extracting unit configured to extract line segment data of each signal line segment processed by the segmentation unit
  • a processing unit configured to determine, according to line segment data of each signal line segment extracted by the extracting unit, a line segment matching template of the waveform signal, where the line segment matching template includes M consecutive signal line segments, where the M is less than K Integer
  • the processing unit is further configured to match each of the K signal line segments processed by the segmentation unit with M signal line segments included in the line segment matching template, and according to the signals The matching result of the line segment determines the target wave group of the waveform signal;
  • the processing unit is further configured to determine, according to the line segment data of the target wave group, periodic signal data of the waveform signal acquired by the acquiring unit.
  • the segmentation unit is configured to:
  • the waveform signal is monotonically divided into a monotonous upward signal line segment or a monotonically downward signal line segment, and each signal line segment is sequentially labeled according to time continuity to obtain a line segment carrying the sequential label.
  • the extracting unit is configured to:
  • the line segment data of the signal line segment j includes: a middle point position of the line segment, a curvature of the starting point position of the line segment, a curvature of the end point position of the line segment, a ray tangent point where the starting point of the line segment is tangent to the line segment, and a ray cut tangent to the end position of the line segment and the line segment. At least one of the points.
  • the extracting unit is further configured to:
  • the line segment data of the derivative curve includes at least one of a number of inflection points, each inflection point position, an intermediate point position between adjacent inflection points, and a maximum inflection point position of the peak.
  • the processing unit is configured to:
  • the M target signal line segments are consecutive line segments.
  • the processing unit is configured to:
  • the threshold determines the position of the M valid signal line segments as the target wave group position.
  • the line segment data of the target wave group position includes line segment data of M valid signal line segments at the target wave group position, and J adjacent signals consecutive to the sequential number of the M valid signal line segments.
  • the processing unit is used to:
  • Determining, according to the M valid signal line segments and the line segment data of the J adjacent signal line segments, a waveform width of the group of the M valid signal line segments and the J adjacent signal line segments, and a waveform of the wave group At least one of the wave group characteristic data of the amplitude, the starting point height of the wave group, or the end point fluctuation of the wave group;
  • the characteristic difference between the group characteristic data and the period check data is less than a preset period threshold, determining that the group is one period of the waveform signal, and the group characteristic data of the group is the waveform Periodic signal data of the signal;
  • the period check data includes at least one of a period waveform width, a period waveform amplitude, a period start waveform height, and a period end waveform height.
  • processing unit is further configured to:
  • the abnormal period data is line segment data of a wave group that does not include the periodic signal data in the waveform signal.
  • the waveform signal comprises at least one of an electrocardiogram ECG, a photoelectric volume pulse wave PPG, a pressure gauge signal, a magnetometer signal, and an accelerometer sensor waveform signal.
  • a third aspect provides a terminal device, which can include: a processor, a memory, a transceiver, and a bus system;
  • the memory, the processor and the transceiver are connected by the bus system;
  • the memory is for storing a set of program codes
  • the processor and the transceiver are configured to invoke program code stored in the memory to perform the method provided by the first aspect above.
  • the embodiment of the present application provides a computer storage medium for storing computer software instructions used by the terminal device, which includes a program designed to execute the first aspect.
  • the embodiment of the present application further provides a chip, which is coupled to a transceiver in the terminal device, and is used to implement the technical solution of the first aspect of the embodiment of the present application.
  • a chip which is coupled to a transceiver in the terminal device, and is used to implement the technical solution of the first aspect of the embodiment of the present application.
  • the embodiment of the present application can extract individual line segment data of the signal line segment from the waveform signal, and can also perform processing such as derivation of the signal line segment and extract line segment data of the derivative curve, and construct a waveform for identifying the waveform according to the combined feature data of the extracted line segment data.
  • the line segment matching template can be used to identify the period of the waveform signal to eliminate noise.
  • the line segment matching template provided by the embodiment of the present application can be constructed by the line segment data of the waveform signal, and the line segment matching template can be different due to different feature data of the waveform signal, thereby improving the association between the line segment matching template and the line segment data of the waveform signal.
  • the cycle identification accuracy of the waveform signal is improved, the noise elimination accuracy is improved, and the processing quality of the waveform signal can be improved, and the applicability is strong.
  • FIG. 1 is a schematic diagram of a waveform signal provided by an embodiment of the present application.
  • FIG. 2 is a schematic flowchart of a method for processing a waveform signal according to an embodiment of the present application
  • FIG. 3 is another schematic diagram of a waveform signal provided by an embodiment of the present application.
  • FIG. 4 is another schematic diagram of a waveform signal provided by an embodiment of the present application.
  • FIG. 5 is a schematic diagram of a derivative curve of a signal line segment provided by an embodiment of the present application.
  • FIG. 6 is a schematic diagram of an ECG signal provided by an embodiment of the present application.
  • FIG. 7 is a schematic diagram of a PPG signal provided by an embodiment of the present application.
  • FIG. 8 is a schematic diagram of signals of an accelerometer sensor waveform provided by an embodiment of the present application.
  • FIG. 9 is a schematic structural diagram of an apparatus for processing a waveform signal according to an embodiment of the present application.
  • FIG. 10 is a schematic structural diagram of a communication device according to an embodiment of the present application.
  • the signal processing method and apparatus provided by the embodiments of the present application are applicable to signal processing of any one-dimensional waveform with periodicity, including but not limited to ECG signals collected by an electrocardiogram (ECG) sensor, and photoelectric volume pulse waves ( Photoplethysmography, PPG)
  • ECG electrocardiogram
  • PPG photoelectric volume pulse waves
  • the embodiments of the present application are also applicable to the processing of aperiodic signals, and can identify line segment combinations in aperiodic signals, such as magnetometer signals and the like. It can be determined according to the actual application scenario, and no limitation is imposed here.
  • the periodic one-dimensional waveform may be a periodically repeated waveform signal, such as a heart rate signal recorded by an ECG signal.
  • the heartbeat itself has a periodicity.
  • Each heartbeat has a corresponding waveform signal. How many heartbeat cycles are there in the heartbeat?
  • the waveform signal of each heartbeat cycle can be recorded by an ECG sensor to obtain a one-dimensional waveform with a periodicity.
  • the senor such as the ECG sensor may be a sensor disposed on a terminal device such as a wearable device. Due to the influence of the wearing method of the wearable device, the fixing method of the sensor, and the behavior of the user, the waveform signal detected by the sensor includes a large noise, and the waveform signal is unstable, or the signal period is unstable. unusual phenomenon.
  • FIG. 1 is a schematic diagram of a waveform signal provided by an embodiment of the present application.
  • the waveform signal shown in Figure 1 is an ECG signal with a length of one minute and a sampling frequency of 500 Hz.
  • the waveform signal shown in Fig. 1 above is a filtered waveform signal including a noise portion and a smooth signal (normal signal) portion.
  • the noise part is chaotic, the peak height is different, and the sharpest part of the peak is much higher than other peaks.
  • the front and back of the noise part are normal signals, and the waveform signal changes periodically.
  • the waveform shape of each period is similar, the peak height is stable, and the period length is fixed.
  • the method and device provided by the embodiments of the present application can identify a normal signal portion (ie, a period of a waveform signal) in a waveform signal, thereby removing a noise portion signal in the waveform signal, obtaining a waveform signal after removing the noise, and improving the waveform signal. Signal quality.
  • the method and apparatus for waveform signal processing provided by the embodiments of the present application will be described below with reference to FIG. 2 to FIG.
  • the method for processing a waveform signal in the embodiment of the present application may be performed by a function module such as a processor in a terminal device such as a wearable device, and may be determined according to a hardware structure of a terminal device such as a wearable device in an actual application scenario. No restrictions.
  • the terminal device may further include: a smart phone, a tablet computer, a personal computer assistant, a portable electrocardiograph, and an electrocardiograph, etc., and are not limited herein.
  • a smart phone a tablet computer
  • a personal computer assistant a portable electrocardiograph
  • an electrocardiograph etc.
  • FIG. 2 is a schematic flowchart of a method for processing a waveform signal provided by an embodiment of the present application.
  • the method provided in the embodiment of the present application includes the following steps:
  • the terminal device may collect waveform data in a preset time length (for example, 1 minute) according to a preset acquisition frequency (for example, 500 Hz, etc.), and then perform processing such as cleaning on the collected data to exclude obviously invalid data.
  • the data cleaning may refer to data that is incomplete in the collected data, or data that does not have periodic changes, or invalid data that does not meet expected requirements (eg, non-one-dimensional signals) to obtain expected demand.
  • Waveform signal eg ECG signal
  • a person's heartbeat rate is uniform and the heartbeat cycle is stable, so the collected data can be cleaned using the human heartbeat characteristics to obtain a waveform signal for characterizing the heartbeat, such as an ECG signal. As shown in the waveform signal of FIG.
  • the waveform signal shown in FIG. 1 includes a large noise portion, the noise is periodically and periodically signaled, so that the waveform signal can be left by data cleaning. If the waveform signal described in FIG. 1 does not have a periodic stationary signal portion, but is disorderly and irregularly searchable, it can be discarded.
  • the terminal device may further filter the data obtained by the cleaning to obtain the filtered waveform data.
  • the filtering may be based on the difference of the sensors for collecting data and the data processing requirements of different usage purposes of the waveform data, and selecting different filters and corresponding parameters to filter the data obtained by the cleaning.
  • the cleaning and filtering operations of the foregoing waveform signals may be referred to the data cleaning and filtering methods of the existing waveform signals.
  • the data cleaning or filtering method may be determined according to the actual application scenario, and is not limited herein.
  • the waveform signal is marked as K signal line segments according to monotonicity.
  • the waveform signal after the filtering process may be segmented.
  • the above-mentioned operation of segmenting the waveform signal does not cut the waveform signal into line segments, but marks the waveform signals as a plurality of signal line segments according to monotonicity.
  • the waveform signal may be marked as a monotonic upward and monotonous downward monotonicity as K signal line segments, and K is an integer greater than or equal to 2.
  • the monotonicity of each signal line segment is unique.
  • each of the K signal line segments may be sequentially numbered according to time continuity to obtain a line segment carrying the sequential label.
  • FIG. 3 is another schematic diagram of a waveform signal provided by an embodiment of the present application.
  • the abscissa is the waveform generation time
  • the ordinate is the waveform amplitude.
  • Inflection points 1, 2, and 3 may be included in the segment of the waveform signal, and each inflection point is a monotonic change node of the line segment. For example, the monotony of the left line segment of the inflection point 1 is monotonous downward, and the monotonicity of the right line segment of the inflection point 1 is monotonous downward, so the left line segment and the right line segment of the inflection point 1 can be marked as two different respectively.
  • Signal line segment is another schematic diagram of a waveform signal provided by an embodiment of the present application.
  • the abscissa is the waveform generation time
  • the ordinate is the waveform amplitude.
  • Inflection points 1, 2, and 3 may be included in the segment of the waveform signal, and each inflection point is a monotonic change node of the line segment.
  • the monotonicity of the signal line segment on the left side of the inflection point 1 is monotonously downward, and the output time of the signal line segment is higher than the right line segment of the inflection point 1, so the signal line segment can be marked as 204.
  • the monotonicity of the signal line segment between the inflection point 1 and the inflection point 2 is monotonously upward, and the time generated by the signal line segment is after the inflection point 1, and therefore, the signal line segment between the inflection point 1 and the inflection point 2 can be marked as time continuity by 205.
  • the monotonicity of the signal line segment between the inflection point 2 and the inflection point 3 is monotonously downward, so the signal line segment can be marked as 206.
  • the signal line segments appearing before the signal line segment 204 may be marked as signal line segments 203, 202, 201, ..., 1, 0, etc., and signals appearing after the signal line segment 206.
  • the line segments may be labeled as signal line segments 207, 208, 209, etc., and are not limited herein.
  • the waveform signal is a wave-shaped curve. Therefore, the waveform signal must have a monotonous downward line segment connected to the monotonous upward line segment, and then the monotonous downward line segment, and the line segment connection point is the inflection point.
  • the monotonic one-to-one line segment is phase-to-phase
  • the line segment contact point is the trough or peak of the waveform signal, and the valley and the peak are connected.
  • the terminal device can mark a monotonous and unique signal line segment between any valley (or peak) and one adjacent peak (or trough) as a signal line segment according to the variation law of the trough and the peak.
  • the inflection point 1 may be a trough of the waveform signal
  • the inflection point 2 is a peak of the waveform signal
  • the inflection point 3 is another trough of the waveform signal.
  • the terminal device may mark a monotonously upward line segment between the valley corresponding to the inflection point and its adjacent peak (inflection point 2) as one signal line segment.
  • the terminal device may also mark the collected segment of the waveform signal as a plurality of monotonous and unique signal line segments according to more segmentation marking manners, which may be determined according to the actual application scenario, and is not limited herein.
  • the terminal device may extract the line segment length Xi and the line segment width Yi of the signal line segment i, and respectively perform differential stretching on the Xi and Yi according to the preset length and the preset width to obtain a normalized signal line segment (markable Is the signal line segment j). Further, the line segment data of the signal line segment j can be extracted.
  • the line segment data of the signal line segment may include, but is not limited to, a midpoint position of the line segment, a starting point and/or an ending position of the line segment, a curvature of the starting point position of the line segment, a curvature of the end point position of the line segment, a ray tangent point of the starting point of the line segment and the line segment, and a signal
  • the end point is the tangent point tangent to the line segment.
  • the line segment width may be the length of time between the start point and the end point of the line segment.
  • the length of the line segment described above may be the magnitude of the change between the start and end points of the line segment.
  • the intermediate point position of the above line segment is a point on the waveform signal line segment corresponding to the intermediate point of the width of the line segment.
  • FIG. 4 is another schematic diagram of a waveform signal provided by an embodiment of the present application.
  • the terminal device can stretch the difference between the line segment length Xi of the signal line segment i and the line segment width Yi to a preset gauge having a length and width of 1*100.
  • the grid size in turn, can perform line segment data extraction on the signal line segment after the difference is stretched.
  • the extracted line segment data may include a coordinate position of the intermediate point O of the line segment, a coordinate position of the starting point A of the line segment and the end point B of the line segment, a curvature of the starting point A of the line segment, a curvature of the end point B of the line segment, a starting point of the line segment A, and a tangent point of the line segment tangent.
  • the terminal device may further derive the signal line segment j to obtain a derivative curve of the signal line segment j, and extract line segment data of the derivative curve.
  • the line segment data of the derivative curve includes data such as the number of inflection points, the position of each inflection point, the position of the intermediate point between the adjacent inflection points, and the position of the inflection point with the largest peak, and is not limited herein.
  • FIG. 5 is a schematic diagram of a derivative curve of a signal line segment provided by an embodiment of the present application.
  • the derivative curve shown in FIG. 5 is a schematic diagram of the derivative curve of the signal line segment shown in FIG. Since the signal line segment shown in Fig. 4 is a curve, the curve is still obtained after the curve is derived.
  • the terminal device can extract the line segment data of the derivative curve of the signal line segment shown in FIG. 4. Further, according to the waveform variation characteristic of the derivative curve, a part of the line segment that satisfies the predefined requirement can be intercepted from the signal line segment shown in FIG. 4. For example, corresponding to the ECG signal, according to medical knowledge, in the ECG signal, there is a straight line between the Q wave and the R wave, and there should be no turning. Therefore, if the curve shown in Fig. 4 is a signal for detecting the Q wave and the R wave, For the line segment, only the straight line portion of the right side of the signal line segment (ie, the right side of the point (60, 0.3)) in the signal line segment shown in FIG. 4 is used, and is used for detecting the Q wave or the R wave, etc., thereby improving the detection accuracy of the waveform signal.
  • the curve shown in Fig. 4 is a signal for detecting the Q wave and the R wave
  • the waveform signal may be first marked as a plurality of signal line segments according to monotonicity, and the line segment data of the signal line segment may be extracted, and further, the line segment data of the derivative curve of the signal line segment may be extracted after the signal line segment is derived.
  • the line segment feature data of the waveform signal is extracted at multiple levels.
  • the embodiment of the present application can identify the characteristics of each signal line segment by using multiple levels of line segment feature data to identify the waveform signal mark, thereby better identifying the waveform feature, identifying the period of the waveform signal, etc., the signal processing quality is higher, and the signal period identification is more accurate.
  • the terminal device extracts feature data of each of the K signal line segments, including monotonicity of the signal line segment, line segment data of the signal line segment, and line segment data of the derivative curve of the signal line segment, and the like. Then, a line segment matching template of the waveform signal can be constructed according to the feature data of each signal line segment. Since each period is a similar signal characteristic (which can be defined as a periodic signal characteristic), the embodiment of the present application can construct an identification template of the periodic signal feature for identifying the waveform signal. Similar signal characteristics in each cycle.
  • the identification template of the periodic signal feature may be the line segment matching template provided by the embodiment of the present application.
  • the terminal device may select, from the line segment data of each signal line segment, M target signal line segments whose line segment data is smaller than a preset line segment data threshold, and perform M target signal line segments according to a preset line segment combination manner. Combine to get the line segment matching template of the waveform signal.
  • the line segment matching template may include M consecutive signal line segments, that is, temporally continuous monotonic one-to-one signal line segments.
  • the target signal line segment may be selected from each signal line segment of the waveform signal by using the preset waveform signal prior information.
  • the ECG signal as an example, see Figure 6, which is a schematic diagram of the ECG signal.
  • a QRS complex, a P wave, a T wave, a U wave, and a PR segment and an ST segment may be included in one periodic signal data of the ECG signal.
  • the QRS complex consists of monotonous down and monotonous upwards, followed by monotonically downward multiple signal segments.
  • the R point in the QRS complex is the highest peak in the cycle, the sharpest waveform peak, R and S. The time interval between them is 0.06 seconds.
  • ECG waveform cycle diagram Basically, it consists of a P wave, a QRS complex, a T wave, and a transition period. In addition, sometimes a small U wave appears after the T wave.
  • P wave also called the atrium to the polar wave, reflecting the potential change of the left and right atrium depolarization process.
  • the waveform is generally round and smooth, lasting 0.08-0.11 seconds, and the amplitude does not exceed 0.25mV.
  • the change in potential generated by the repolarization of the two atriums is called a Ta wave, which usually overlaps with the PR segment, the QRS complex or the ST segment, and the amplitude is very low, which is not easily identifiable on the electrocardiogram.
  • PR interval is the time interval between the start of the P wave and the start of the QRS complex, reflecting the period from the beginning of the atrial depolarization to the beginning of the ventricular depolarization.
  • the normal adult PR interval is 0.12-0.20 seconds. If it exceeds 0.205 seconds, it usually indicates the occurrence of atrioventricular block.
  • the length of the PR interval is related to factors such as age and heart rate, and there is no restriction here.
  • QRS complex reflects the potential change of the two-ventricular depolarization process.
  • a typical QRS complex consists of three closely connected potential fluctuations: the first downward wave is called the Q wave, followed by the upward and high and sharp R waves, and finally the downward S wave. In different leads, these three waves do not necessarily appear, and the amplitude of each wave changes greatly, which lasts about 0.06-0.105 seconds.
  • ST segment refers to the line segment between the end point of the QRS complex and the start of the T wave, generally flush with the zero potential baseline. During this period, since all parts of the ventricle have entered the depolarization state, but the repolarization has not yet begun, there is no potential difference between the various parts of the ventricle, and the ECG curve returns to the baseline level. However, if coronary artery insufficiency or myocardial infarction occurs, the ST segment often deviates from the baseline and exceeds a certain range of amplitude.
  • T wave reflects the potential change of the repolarization process of the two ventricles.
  • the waveform of the T wave is blunt, the lifting branch is not completely symmetrical, the front branch of the waveform is long and the rear branch is short, which takes about 0.05-0.255 seconds.
  • the T wave direction should be consistent with the main wave direction of the QRS complex.
  • the amplitude of the T wave should be no less than 1/10 of the R wave of the lead.
  • QT interval refers to the time from the start of the QRS complex to the end of the T wave, which represents the time required for the ventricle to begin depolarization until the completion of all repolarization.
  • the length of this interval is closely related to the heart rate, the faster the heart rate, the shorter the QT interval. On the contrary, the longer the QT interval.
  • the normal QT interval can vary depending on factors such as heart rate, age and gender. When the heart rate is 75 beats/min, the QT interval is 0.30-0.405 seconds. Analysis of changes in QT interval, early diagnosis of the disease and analysis of the effects of antiarrhythmic drugs on the heart can play a certain auxiliary role.
  • U wave A low-width U wave that coincides with the direction of the T wave may appear 0.02-0.04 seconds after the T wave.
  • the above periodic signal data may be a priori information of the waveform signal.
  • the terminal device can construct a line segment matching template for identifying the QRS complex in the waveform signal. Specifically, the terminal device can use the a priori information of the waveform signal such as the highest peak in one cycle of the R point in the QRS complex to find the sharpest one in each preset time interval (for example, 1.5 seconds) from the waveform signal. Waveform peaks are sorted according to sharpness according to the acquired waveform peaks. Remove the sharpest part, such as removing 20%, and use the remaining peaks to make a line matching template.
  • the line segment matching template may include the characteristics of the three lines of the left side of the peak, the right side down, the right side, the back side, and the right side up, and the median of each line segment data of the surrounding peak correlation feature is taken, and the obtained median is utilized. Build a line segment matching template.
  • the terminal device may first search for a small number of target signal line segments, then construct a template, and then use the template to take more target signal line segments, and update the template according to more target signal line segments.
  • the more target signal line segments selected by the terminal device the higher the accuracy of the line segment matching template corrected by the line segment data of each signal line segment, and finally the line segment matching template with higher similarity to the QRS group of the waveform signal can be determined.
  • the line segment matching template may be a group of data as a matching standard, where the data includes but is not limited to the length of the line segment, the width of the line segment, the X-axis and the Y-axis coordinate position of the starting point of the line segment, and the starting point or the ending point of the line segment.
  • the local slope nearby ie, the local derivative of the signal line segment
  • the tangent slope of the line segment's start or end point tangent to the line segment such as the line segment shown in Figure 4, the line segment starting point A (0,0) to point C (60,0.3) the ray slope
  • the body signals of human individuals have large differences, so the waveform signals detected by ECG sensors or PPG sensors may vary from person to person.
  • the embodiment of the present application can construct a line segment matching template of the waveform signal according to the feature data of each waveform signal, thereby improving the processing quality of the waveform signal and improving the cycle recognition accuracy of the waveform signal.
  • the line segment matching template can be used to identify the target wave group of the waveform signal from the waveform signal, such as the QRS wave group in the ECG signal.
  • the terminal device can match each of the K signal line segments obtained by the waveform signal marking with the signal line segments included in the line segment matching template, and search for valid signal line segments from the K signal line segments.
  • the valid signal line segment refers to a signal line segment whose matching degree with any line segment included in the line segment matching template is greater than or equal to a preset similarity threshold. That is, the line difference between the signal line segments whose feature difference is smaller than the preset threshold.
  • the signal characteristics of the group represented by the N valid signal line segments obtained by the search may be matched with the signal characteristics of the wave group presented by the signal line segments included in the line segment matching template. If the characteristic difference of the signal characteristics of the wave group is less than a preset threshold, it may be determined that the wave group similar to the line segment matching template among the groups of the N effective signals is the target wave group. For example, if a group of the three signal line segments of the N effective signal line segments is similar to the QRS group composed of the line segment matching template, it can be determined that the wave group composed of the three effective signal line segments is the target wave group. That is, a QRS complex.
  • an ECG signal that is one minute in length can be labeled as approximately 1500 signal segments after filtering.
  • the number of the foregoing signal line segments is related to the filtering strength of the waveform signal, and may be determined according to an actual application scenario, which is merely an example.
  • target signal line segments included in the line segment matching template, for example, signal line segments QR, RS, and SS' (monotonically rising line segments after point S, which can be labeled SS', as shown in Figure 6).
  • each signal line segment matches the above three target signal line segments, that is, attempts to match 4500 times.
  • the width of the target signal line segment QR is QR.L. If the line segment width of the signal line segment to be verified is within the width range of 50% to 200% of QR.L, the width of the signal line segment can be determined to pass, and then the length of the signal line segment is similarly detected, and then the starting point of the signal line segment is detected. Or the local slope near the end point, the start or end point of the signal line segment, and the ray slope of the line segment. If all parameters match successfully, a valid signal line segment can be determined. Further, the group characteristics of the plurality of consecutive signal line segments in the effective signal line segment can be detected, thereby determining the target wave group.
  • the effective signal line segment further includes the right signal line segment except the line segment on the left side of the turn. That is, the above includes a turn
  • the signal line segment does not conform to the characteristics of the QR line segment, and the formation of the curve may come from errors in data acquisition, or other environmental factors. If the matching result of the valid signal line segment occurs, the three consecutive valid signal line segments in the time sequence are consecutively matched with the QR, RS and SS' in the line segment matching template, respectively, and the three valid signal line segments are determined to be presented.
  • the wave group is a QRS complex, which is the target wave group.
  • the period of the waveform signal may be identified according to the target wave group.
  • the terminal device can also determine the period of the waveform signal according to the peripheral waveform signal of the target wave group and the preset period check data.
  • the peripheral waveform signal of the target wave group may include line segment data of a plurality of signal line segments adjacent to the effective signal line segment of the target wave group. These peripheral waveform signals may present peaks or troughs outside the target wave group, for example, T waves, U waves, or P waves in the ECG.
  • the above-mentioned period check data includes: a period waveform width, a period waveform amplitude, a period start waveform height, and a period end waveform height, and the like, and is not limited herein.
  • the P waveform amplitude and the cycle start and end point height have a limited amplitude or height. These may be referred to as period check data or check data of peripheral waveform signals of the target wave group.
  • the T wave in the ECG signal is a bulge for a period of time after the QRS complex.
  • the peak height of the T wave is larger. If the T wave in the ECG signal is inverted, the individual has signs such as angina. Therefore, by detecting whether there is a T wave and a T wave width after the QRS complex, whether the height is appropriate or not can control the detection quality of the waveform signal, and improve the cycle recognition accuracy of the waveform signal.
  • the waveform signal of the abnormal period in the waveform signal or the waveform signal of the abnormal period may be repaired according to the information obtained by the identification.
  • the heart rate cycle can be found.
  • the identified heartbeat interval can be 1 second, 1 second, 1 second, 2 seconds, 1 second, and the like. It can be determined that a waveform period is missing in the middle, and it may be that the waveform signal of the period is not matched by the periodic signal data.
  • the terminal device can re-match the waveform signal by lowering the data threshold standard such as the preset period preset of the period check data, and remedy one cycle by the period identification. Conversely, if the heartbeat interval occurs for 1 second, 1 second, 0.3 second, 0.7 second, and 1 second, it can be determined that the identified heartbeat period is more than another, and the noise may be recognized as the heart rate period, and thus the waveform may be The signal is re-identified to eliminate noise.
  • the data threshold standard such as the preset period preset of the period check data
  • the periodic signal data recognized by the waveform signal is an isolated period, and the adjacent signals of the period are all noises, the period may also be noise, and the signal data of the period is also not desirable, so throw away.
  • the terminal device can identify each period in the waveform signal according to the determined periodic signal data of the waveform signal, thereby determining characteristic data between the periods, removing the abnormal periodic signal in the waveform signal, and improving the detection accuracy of the waveform signal.
  • FIG. 7 is a schematic diagram of a PPG signal provided by an embodiment of the present application.
  • the terminal device can identify the periodic signal data of a single period of the PPG signal according to the implementation described in the above steps, and can also identify the waveform similarity between the period and the period.
  • the abnormal periodic signal in the waveform signal can be eliminated.
  • the four periodic waveforms on the right side of FIG. 7 can be sorted into waveform 1, waveform 2, waveform 3, and waveform 4 from top to bottom.
  • Waveform 1 and Waveform 2 may represent abnormal periodic signals
  • waveform 3 and wave 4 may represent normal periodic signals.
  • FIG. 8 is a schematic diagram of signals of an accelerometer sensor waveform provided by an embodiment of the present application.
  • an accelerometer sensor waveform can be identified by the terminal device according to the implementation described in the above steps.
  • the waveform signal characteristic (or the waveform signal prior information) of the accelerometer sensor may include a W waveform and a V waveform interphase. Among them, the W waveform is the characteristic of the front foot landing, and the V waveform is the characteristic of the back foot landing.
  • the line segment data of the signal line segment may be extracted from the waveform signal, and the signal line segment may be derivatively processed and the line segment data of the derivative curve may be extracted, and a line segment for identifying the periodic signal data of the waveform signal may be constructed according to the extracted data.
  • the line segment matching template can be used to identify the period of the waveform signal to eliminate noise.
  • the line segment matching template provided by the embodiment of the present application can be constructed by the line segment data of the waveform signal, and the line segment matching template can be different due to different feature data of the waveform signal, thereby improving the association between the line segment matching template and the line segment data of the waveform signal. Improve the cycle recognition accuracy of the waveform signal, improve the noise elimination accuracy, and thus improve the processing quality of the waveform signal.
  • FIG. 9 is a schematic structural diagram of an apparatus for processing a waveform signal according to an embodiment of the present application.
  • the apparatus for processing the waveform signal provided by the embodiment of the present application may be specifically the terminal device provided by the embodiment of the present application, where the terminal device may include:
  • the obtaining unit 91 is configured to obtain a waveform signal after the filtering process.
  • the segmentation unit 92 is configured to mark the waveform signal acquired by the acquiring unit as K signal line segments according to monotonicity, and the K is an integer greater than or equal to 2.
  • the extracting unit 93 is configured to extract line segment data of each signal line segment processed by the segmentation unit.
  • the processing unit 94 is configured to determine, according to the line segment data of each signal line segment extracted by the extracting unit, a line segment matching template of the waveform signal, where the line segment matching template includes M consecutive signal line segments, where the M is smaller than An integer of K.
  • the processing unit 94 is further configured to match each of the K signal line segments processed by the segmentation unit with M signal line segments included in the line segment matching template, and according to the The result of the matching of the signal line segments determines the target wave group of the waveform signal.
  • the processing unit 94 is further configured to determine, according to the line segment data of the target wave group, periodic signal data of the waveform signal acquired by the acquiring unit.
  • the segmentation unit 92 is configured to:
  • the waveform signal is monotonically divided into a monotonous upward signal line segment or a monotonically downward signal line segment, and each signal line segment is sequentially labeled according to time continuity to obtain a line segment carrying the sequential label.
  • the foregoing extracting unit 93 is configured to:
  • the line segment data of the signal line segment j includes: a middle point position of the line segment, a curvature of the starting point position of the line segment, a curvature of the end point position of the line segment, a ray tangent point where the starting point of the line segment is tangent to the line segment, and a ray cut tangent to the end position of the line segment and the line segment. At least one of the points.
  • the foregoing extracting unit 93 is further configured to:
  • the line segment data of the derivative curve includes: the number of inflection points, the position of each inflection point, and the middle between adjacent inflection points At least one of a point position and a peak point of the largest peak.
  • processing unit 94 is configured to:
  • the M target signal line segments are consecutive line segments.
  • processing unit 94 is configured to:
  • the threshold determines the position of the M valid signal line segments as the target wave group position.
  • the line segment data of the target wave group position includes line segment data of M valid signal line segments at the target wave group position, and J adjacent signals consecutive to the sequential number of the M valid signal line segments.
  • the above processing unit 94 is used to:
  • Determining, according to the M valid signal line segments and the line segment data of the J adjacent signal line segments, a waveform width of the group of the M valid signal line segments and the J adjacent signal line segments, and a waveform of the wave group At least one of the wave group characteristic data of the amplitude, the starting point height of the wave group, or the end point fluctuation of the wave group;
  • the characteristic difference between the group characteristic data and the period check data is less than a preset period threshold, determining that the group is one period of the waveform signal, and the group characteristic data of the group is the waveform Periodic signal data of the signal;
  • the period check data includes at least one of a period waveform width, a period waveform amplitude, a period start waveform height, and a period end waveform height.
  • processing unit 94 is further configured to:
  • the abnormal period data is line segment data of a wave group that does not include the periodic signal data in the waveform signal.
  • the waveform signal comprises at least one of an electrocardiogram ECG, a photoelectric volume pulse wave PPG, a pressure gauge signal, a magnetometer signal, and an accelerometer sensor waveform signal.
  • the foregoing terminal device may perform the implementation manners of the terminal device in the foregoing embodiment by using the built-in various units.
  • the foregoing terminal device may perform the implementation manners of the terminal device in the foregoing embodiment by using the built-in various units.
  • the terminal device may extract the line segment data of the signal line segment from the waveform signal, further perform processing such as deriving the signal line segment, and extract line segment data of the derivative curve, and construct a line signal for identifying the waveform signal according to the extracted data.
  • the line segment matching template of the periodic signal data Further, the line segment matching template can be used to identify the period of the waveform signal to eliminate noise.
  • the line segment matching template provided by the embodiment of the present application can be constructed by the line segment data of the waveform signal, and the line segment matching The template may be different due to different characteristic data of the waveform signal, thereby improving the correlation between the line segment matching template and the line segment data of the waveform signal, thereby improving the cycle recognition accuracy of the waveform signal, improving the noise elimination accuracy, and thereby improving the waveform.
  • the quality of the signal processing can be constructed by the line segment data of the waveform signal, and the line segment matching The template may be different due to different characteristic data of the waveform signal, thereby improving the correlation between the line segment matching template and the line segment data of the waveform signal, thereby improving the cycle recognition accuracy of the waveform signal, improving the noise elimination accuracy, and thereby improving the waveform.
  • FIG. 10 is a schematic structural diagram of a communication device according to an embodiment of the present application.
  • the communication device 100 provided by the embodiment of the present application includes a processor 101, a memory 102, a transceiver 103, and a bus system 104.
  • the processor 101, the memory 102 and the transceiver 103 are connected by a bus system 104.
  • the above memory 102 is used to store programs.
  • the program can include program code, the program code including computer operating instructions.
  • the memory 102 includes, but is not limited to, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read only memory (EPROM), or Portable disc read-only memory (CD-ROM). Only one memory is shown in Fig. 10. Of course, the memory can also be provided in plurality as needed.
  • the memory 102 may also be a memory in the processor 101, which is not limited herein.
  • the memory 102 stores the following elements, executable modules or data structures, or subsets thereof, or their extended sets:
  • Operation instructions include various operation instructions for implementing various operations.
  • Operating system Includes a variety of system programs for implementing various basic services and handling hardware-based tasks.
  • the processor 101 controls the operation of the communication device 100.
  • the processor 101 may be one or more central processing units (CPUs).
  • CPUs central processing units
  • the CPU may be a single core CPU. It can also be a multi-core CPU.
  • bus system 104 which may include, in addition to the data bus, a power bus, a control bus, a status signal bus, and the like.
  • bus system 104 may include, in addition to the data bus, a power bus, a control bus, a status signal bus, and the like.
  • bus system 104 may include, in addition to the data bus, a power bus, a control bus, a status signal bus, and the like.
  • bus system 104 for clarity of description, various buses are labeled as bus system 104 in FIG. For ease of representation, only the schematic drawing is shown in FIG.
  • Processor 101 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method may be completed by an integrated logic circuit of hardware in the processor 101 or an instruction in a form of software.
  • the processor 101 may be a general-purpose processor, a digital signal processing (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or Other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components.
  • DSP digital signal processing
  • ASIC application specific integrated circuit
  • FPGA field-programmable gate array
  • the methods, steps, and logical block diagrams disclosed in the embodiments of the present application can be implemented or executed.
  • the general purpose processor may be a microprocessor or the processor or any conventional processor or the like.
  • the steps of the method disclosed in the embodiments of the present application may be directly implemented by the hardware decoding processor, or may be performed by a combination of hardware and software modules in the decoding processor.
  • the software module can be located in a conventional storage medium such as random access memory, flash memory, read only memory, programmable read only memory or electrically erasable programmable memory, registers, and the like.
  • the storage medium is located in the memory 102, and the processor 101 reads the information in the memory 102 in conjunction with its hardware to perform the method steps of the terminal device described in the various embodiments above.
  • the computer program is instructed to execute the associated hardware, and the program can be stored in a computer readable storage medium, which, when executed, can include the flow of an embodiment of the methods described above.
  • the storage medium may be a magnetic disk, an optical disk, a ROM, a RAM, or the like.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Cardiology (AREA)
  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Molecular Biology (AREA)
  • Public Health (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Biophysics (AREA)
  • Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Pathology (AREA)
  • Veterinary Medicine (AREA)
  • Theoretical Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Signal Processing (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Measurement And Recording Of Electrical Phenomena And Electrical Characteristics Of The Living Body (AREA)

Abstract

本申请实施例公开了一种波形信号处理的方法及装置，该方法包括：获取滤波处理后的波形信号；将波形信号按照单调性标记为K个信号线段；提取各信号线段的线段数据，并根据各信号线段的线段数据确定波形信号的线段匹配模板，该线段匹配模板中包括M个连续的信号线段，M为小于K的整数；将K个信号线段中的各信号线段与该线段匹配模板中包括的M个信号线段进行匹配，并根据各信号线段的匹配结果确定出波形信号的目标波群；根据目标波群的线段数据确定出波形信号的周期信号数据。本申请实施例具有可识别波形信号的周期信号数据，提高波形信号的周期识别准确率并能提取相关特征点的优点。

Description

波形信号处理的方法及装置 技术领域
本申请涉及信号处理领域，尤其涉及一种波形信号处理的方法及装置。
背景技术
当前随着可穿戴设备的日益普及，在可穿戴设备中设置传感器，用于检测用户的心电信号、血流信号以及运动信号等体征信号的实现方式也随着兴起。可穿戴设备上的传感器检测到的心电信号、光电容易脉搏波信号或者运动信号等通常为具有周期性和连续性等特点的波形信号，因此通过对波形信号的监测可感知用户的身体健康情况，可以发现计步周期等功能。
然而，可穿戴设备的传感器检测到的波形信号容易因为可穿戴设备的佩戴方式或者用户的行为动作等因素影响，波形信号中包括较大的噪声使得波形信号出现波形不稳定，或者信号周期不稳定等异常现象，波形信号的质量差，可靠性低。
发明内容
本申请实施例提供了一种波形信号处理的方法及装置，可提取波形信号的信号线段特征，提高波形信号的周期的识别效率，增强波形信号的适用性。
第一方面提供了一种波形信号处理的方法，该方法可首先将原始波形信号进行清洗和滤波等处理，获取得到滤波处理之后的波形信号。进而可将所述波形信号按照单调性标记为K个信号线段，K为大于或者等于2的整数，并提取各信号线段的线段数据。本申请实施例将波形信号进行分段处理，分段提取各个线段的特征数据，可提高波形信号的信号处理精度，提高波形信号的特征识别准确率。上述方法还可根据各信号线段的线段数据确定该波形信号的线段匹配模板，该线段匹配模板中包括M个连续的信号线段，M为小于K的整数；将波形信号划分得到的K个信号线段中的各信号线段与线段匹配模板中包括的M个信号线段进行匹配，并根据各信号线段的匹配结果确定出上述波形信号的目标波群；根据目标波群的线段数据确定出上述波形信号的周期信号数据。
本申请实施例可根据提取的各个线段的特征数据，确定出用于识别出波形信号的周期的线段匹配模板，线段匹配模板与波形信号的密切相关。本申请实施例可首先通过线段匹配模板识别出波形信号中单个周期内的目标波群，再通过目标波群识别出波形信号的周期，波形信号的周期识别准确率更高，提高了波形信号的信号处理质量，适用性更强。
可选的，本申请实施例可将上述波形信号按照单调性划分为单调向上的信号线段或者单调向下的信号线段，并按照时间连续性对每个信号线段进行顺序标号以得到携带顺序标号的线段。本申请实施例可按照线段的单调性将波形信号划分为多个信号线段，并且每个信号线段的单调性唯一，进而可按照时间连续性为每个信号线段进行顺序标号，以更好地提取各个信号线段的特征数据，按照时间连续性整合各个信号线段的特征数据可更好地确定线段匹配模板，提高波形信号的目标波群的识别效率，提高波形信号的周期识别的准确 率，适用性更强。
可选的，本申请实施例可提取信号线段i的线段长度Xi和线段宽度Yi，并按照预设长度和预设宽度分别对Xi和Yi进行差值拉伸以得到归一化处理后的信号线段j；提取信号线段j的线段数据；其中，信号线段j的线段数据包括：线段中间点位置、线段起点位置曲率、线段终点位置曲率、线段起点位置与线段相切的射线切点，以及线段终点位置与线段相切的射线切点中的至少一种。本申请实施例可提取各个信号线段的个体特征，将各个信号线段进行归一化处理之后在提取各个信号线段的特征，可更好的识别各个信号线段的特性，以根据各个信号线段的特性处理波形信号，进而可提高波形信号的处理质量
可选的，本申请实施例还可对信号线段j进行求导或者差分以得到信号线段j的导数曲线，并提取导数曲线的线段数据；其中，上述导数曲线的线段数据包括：拐点数量、各拐点位置，相邻拐点之间的中间点位置以及峰值最大的拐点位置中的至少一种。本申请实施例可对波形信号划分得到的各个信号线段进行求导或者差分处理，提取求导或者差分处理之后的导数曲线的线段数据，可进一步识别波形信号划分的各个信号线段的特性，提高了波形信号的线段特征数据的提取精度，提高波形信号的处理质量。
可选的，本申请实施例可从各信号线段的线段数据中选取线段数据小于预设线段数据阈值的M个目标信号线段，并根据预置的线段组合方式将M个目标信号线段进行组合以得到波形信号的线段匹配模板；其中，上述M个目标信号线段为顺序标号连续的线段。本申请实施例可根据波形信号的各个信号线段的线段数据得到该波形信号的线段匹配模板，增强了线段匹配模板与波形信号的关联紧密性，进而可提高通过线段匹配模板识别波形信号的周期的准确性，适用性更高。
可选的，本申请实施例可将K个信号线段中的每个信号线段与线段匹配模板中包括的M个目标信号线段中的每个目标信号线段进行线段数据匹配，从K个信号线段中确定N个有效信号线段，N为小于M的整数，上述有效信号线段的线段数据与线段匹配模板包括的任一目标线段的线段数据的特征差值不小于预置阈值；若上述N个有效信号线段中包括顺序标号连续的M个有效信号线段，并且上述M个有效信号线段分别与上述线段匹配模板中包括的M个目标信号线段的特征差值小于预设阈值，则将上述M个有效信号线段的位置确定为目标波群位置。
本申请实施例可通过该波形信号的线段匹配模板和预先设定的数据误差阈值，从波形信号中各个信号线段中识别出波形信号包括的目标波群的位置，进而可根据目标波群的位置识别出波形信号的周期，提高了波形信号的周期识别准确率，可提高波形信号的处理质量，适用性更强。
可选的，上述目标波群位置的线段数据包括目标波群位置上的M个有效信号线段的线段数据，以及与该M个有效信号线段的顺序标号连续的J个相邻信号线段的线段数据；本申请实施例可根据该M个有效信号线段与J个相邻信号线段的线段数据，确定M个有效信号线段与J个相邻信号线段组成的波群的波形宽度，波群的波形幅度、波群的起点高度或者波群的终点波动中的至少一种波群特征数据；若波群特征数据与周期校验数据的特征差值小于预置周期阈值，则确定该波群为波形信号的一个周期，该波群的波群特征数据为波形信号的周期信号数据；其中，上述周期校验数据包括：周期波形宽度、周期波形幅度、 周期起点波形高度以及周期终点波形高度中的至少一种。本申请可根据识别得到的波形信号的目标波群的位置，以及波形信号的周期校验数据等信息，识别出波形信号的周期，提高波形信号的周期识别便捷性和准确率。
可选的，本申请实施例可根据波形信号的周期信号数据确定出波形信号中包括的多个波形周期，并删除波形信号中包括的异常周期数据。其中，上述异常周期数据为波形信号中不包括上述周期信号数据的波群的线段数据。本申请实施例可剔除波形信号中的异常周期数据，提高波形信号的信号质量，提高波形信号的适用性。
可选的，本申请实施例所描述的波形信号可包括：心电图ECG、光电容积脉冲波PPG、压力计信号、磁力计信号以及加速度计传感器波形信号中的至少一种。本申请实施例提供的方法适用于处理多种类型的波形信号，操作灵活，适应范围广。
第二方面提供了一种波形信号处理的装置，其可包括：
获取单元，用于获取滤波处理后的波形信号；
分段单元，用于将所述获取单元获取的所述波形信号按照单调性标记为K个信号线段，所述K为大于或者等于2的整数；
提取单元，用于提取所述分段单元处理得到的各信号线段的线段数据；
处理单元，用于根据所述提取单元提取的所述各信号线段的线段数据确定所述波形信号的线段匹配模板，所述线段匹配模板中包括M个连续的信号线段，所述M为小于K的整数；
所述处理单元，还用于将所述分段单元处理得到的所述K个信号线段中的各信号线段与所述线段匹配模板中包括的M个信号线段进行匹配，并根据所述各信号线段的匹配结果确定出所述波形信号的目标波群；
所述处理单元，还用于根据所述目标波群的线段数据确定出所述获取单元获取的所述波形信号的周期信号数据。
可选的，所述分段单元用于：
将所述波形信号按照单调性划分为单调向上的信号线段或者单调向下的信号线段，并按照时间连续性对每个信号线段进行顺序标号以得到携带顺序标号的线段。
可选的，所述提取单元用于：
提取所述信号线段i的线段长度Xi和线段宽度Yi，并按照预设长度和预设宽度分别对所述Xi和Yi进行差值拉伸以得到归一化处理后的信号线段j；
提取所述信号线段j的线段数据；
其中，所述信号线段j的线段数据包括：线段中间点位置、线段起点位置曲率、线段终点位置曲率、线段起点位置与线段相切的射线切点，以及线段终点位置与线段相切的射线切点中的至少一种。
可选的，所述提取单元还用于：
对所述信号线段j进行求导或者差分以得到所述信号线段j的导数曲线，并提取所述导数曲线的线段数据；
其中，所述导数曲线的线段数据包括：拐点数量、各拐点位置，相邻拐点之间的中间点位置以及峰值最大的拐点位置中的至少一种。
可选的，所述处理单元用于：
从所述各信号线段的线段数据中选取线段数据小于预设线段数据阈值的M个目标信号线段，并根据预置的线段组合方式将所述M个目标信号线段进行组合以得到所述波形信号的线段匹配模板；
其中，所述M个目标信号线段为顺序标号连续的线段。
可选的，所述处理单元用于：
将所述K个信号线段中的每个信号线段与所述线段匹配模板中包括的M个目标信号线段中的每个目标信号线段进行线段数据匹配，从所述K个信号线段中确定N个有效信号线段，所述N为小于M的整数，所述有效信号线段的线段数据与所述线段匹配模板包括的任一目标线段的线段数据的特征差值不小于预置阈值；
若所述N个有效信号线段中包括顺序标号连续的M个有效信号线段，并且所述M个有效信号线段分别与所述线段匹配模板中包括的M个目标信号线段的特征差值小于预设阈值，则将所述M个有效信号线段的位置确定为目标波群位置。
可选的，所述目标波群位置的线段数据包括所述目标波群位置上的M个有效信号线段的线段数据，以及与所述M个有效信号线段的顺序标号连续的J个相邻信号线段的线段数据；
所述处理单元用于：
根据所述M个有效信号线段与所述J个相邻信号线段的线段数据，确定所述M个有效信号线段与所述J个相邻信号线段组成的波群的波形宽度，波群的波形幅度、波群的起点高度或者波群的终点波动中的至少一种波群特征数据；
若所述波群特征数据与周期校验数据的特征差值小于预置周期阈值，则确定所述波群为所述波形信号的一个周期，所述波群的波群特征数据为所述波形信号的周期信号数据；
其中，所述周期校验数据包括：周期波形宽度、周期波形幅度、周期起点波形高度以及周期终点波形高度中的至少一种。
可选的，所述处理单元还用于：
根据所述波形信号的周期信号数据确定出所述波形信号中包括的多个波形周期，并删除所述波形信号中包括的异常周期数据；
其中，所述异常周期数据为所述波形信号中不包括所述周期信号数据的波群的线段数据。
可选的，所述波形信号包括：心电图ECG、光电容积脉冲波PPG、压力计信号、磁力计信号以及加速度计传感器波形信号中的至少一种。
第三方面提供了一种终端设备，其可包括：处理器、存储器、收发器和总线***；
所述存储器、所述处理器和所述收发器通过所述总线***连接；
所述存储器用于存储一组程序代码；
所述处理器和所述收发器用于调用所述存储器中存储的程序代码执行上述第一方面提供的方法。
第四方面，本申请实施例提供了一种计算机存储介质，用于储存为上述终端设备所用的计算机软件指令，其包含用于执行上述第一方面所设计的程序。
第五方面，本申请实施例还提供了一种芯片，该芯片与终端设备中的收发器耦合，用于执行本申请实施例第一方面的技术方案。应理解，在本申请实施例中“耦合”是指两个部件彼此直接或间接地结合。这种结合可以是固定的或可移动性的，这种结合可以允许流动液、电、电信号或其它类型信号在两个部件之间通信。
本申请实施例可从波形信号中提取信号线段的个体线段数据，还可对信号线段进行求导等处理并提取导数曲线的线段数据，根据提取的各个线段数据的组合特征数据构建用于识别波形信号的周期信号数据的线段匹配模板。进一步的可使用线段匹配模板识别波形信号的周期，排除噪声。本申请实施例提供的线段匹配模板可由波形信号的线段数据构建得到，线段匹配模板可因波形信号的特征数据不同而有所差异，进而可提高线段匹配模板与波形信号的线段数据的关联，可提高波形信号的周期识别准确率，提高噪声的排除准确性，进而可提高波形信号的处理质量，适用性强。
附图说明
为了更清楚地说明本申请实施例中的技术方案，下面将对实施例中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本申请的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本申请实施例提供的波形信号的一示意图；
图2是本申请实施例提供的波形信号处理的方法流程示意图；
图3是本申请实施例提供的波形信号的另一示意图；
图4是本申请实施例提供的波形信号的另一示意图；
图5是本申请实施例提供的信号线段的导数曲线示意图；
图6是本申请实施例提供的ECG信号示意图；
图7是本申请实施例提供的PPG信号示意图；
图8是本申请实施例提供的加速度计传感器波形的信号示意图；
图9是本申请实施例提供的波形信号处理的装置的结构示意图；
图10是本申请实施例提供的一种通信设备的结构示意图。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述。
具体实现中，本申请实施例提供的信号处理方法及装置适用于任一有周期的一维波形的信号处理，包括但不限于心电图(electrocardiogram，ECG)传感器采集的ECG信号、光电容积脉冲波(photoplethysmography，PPG)传感器采集的PPG信号、压力计信号(包括气压计信号)、陀螺仪的角速度传感器以及加速度计传感器采集的波形信号等，在此不做限制。为方便描述，本申请实施例将以ECG信号为了进行说明。
本申请实施例还可适用于非周期信号的处理，可识别非周期信号中的线段组合，例如磁力计信号等。具体可根据实际应用场景确定，在此不做限制。
上述有周期的一维波形通常可为周期性重复的波形信号，例如ECG信号所记录的心率信号。心跳本身是有周期的，每一次心跳都有对应的波形信号，心跳了多少次就有多少个心跳周期，通过ECG传感器等可记录每一个心跳周期的波形信号，得到有周期的一维波形。
具体实现中，上述ECG传感器等传感器具体可为设置于可穿戴设备等终端设备上的传感器。由于可穿戴设备的佩戴方式、传感器的固定方式以及用户的行为动作等客观因素影响，使得传感器检测到的波形信号中包括较大的噪声，进而波形信号出现波形不稳定，或者信号周期不稳定等异常现象。例如，参见图1，是本申请实施例提供的波形信号的一示意图。如图1所示波形信号是一段时间长度为一分钟的ECG信号，采样频率为500赫兹(Hz)。上述图1所示波形信号为滤波之后的波形信号，其中包括噪声部分和平稳信号(正常信号)部分。噪声部分波形混乱，波峰高度不一，最尖锐部分波峰远高于其他波峰。噪声部分的前后均为正常信号，波形信号呈周期性变化，每个周期的波形形状相似，波峰高度稳定，周期长度固定。
本申请实施例提供的方法及装置可识别波形信号中的正常信号部分(即波形信号的周期)，进而可去除波形信号中的噪声部分信号，得到去除噪声之后的波形信号，可提高波形信号的信号质量。下面将结合图2至图10对本申请实施例提供的波形信号处理的方法及装置进行描述。具体实现中，本申请实施例中波形信号处理的方法可由可穿戴设备等终端设备中的处理器等功能模块执行，具体可根据实际应用场景中可穿戴设备等终端设备的硬件结构确定，在此不做限制。其中，上述终端设备还可包括：智能手机、平板电脑、个人计算机助理、便携式心电检测仪以及心电监护仪等，在此不做限制。为方便描述，本申请实施例将以终端设备为例进行说明。
参见图2，是本申请实施例提供的波形信号处理的方法流程示意图。本申请实施例提供的方法，包括步骤：
S21，获取滤波处理后的波形信号。
具体实现中，终端设备可按照预设的采集频率(例如500Hz等)采集预设时间长度(例如1分钟)内的波形数据，进而可对采集到的数据进行清洗等处理以排除明显无效的数据。其中，上述数据清洗可指将采集到的数据中波形不完整的数据，或者不具备周期性变化的数据，或者不符合预期需求(例如非一维信号)等无效数据清除，以得到预期需求的波形信号(例如ECG信号)数据。例如，人的心跳心率齐，心跳周期稳定，因此可利用人的心跳特性对采集到的数据进行清洗以得到用于表征心跳的波形信号，例如ECG信号。如图1所示的波形信号，虽然图1所示的波形信号中包括巨大噪声部分，然而噪声前后都是周期性平稳的信号，因此通过数据清洗可将该波形信号留下。若图1所述的波形信号中不具备周期性平稳的信号部分，而是杂乱无章，无规律可寻的数据，则可将其丢弃。
具体实现中，终端设备还可将清洗得到的数据进行滤波处理，以得到滤波后的波形数据。其中，滤波可依照采集数据的传感器的差异性以及波形数据的不同使用目的等数据处理需求，选择不同的滤波器及其对应参数对清洗得到的数据进行滤波。
可选的，上述波形信号的清洗和滤波等操作可参见现有更多的波形信号的数据清洗和滤波方式，具体可根据实际应用场景确定数据清洗或者滤波方式，在此不做限制。
S22，将所述波形信号按照单调性标记为K个信号线段。
在一些可行的实施方式中，终端设备对采集到的数据进行清洗和滤波等处理之后，可将滤波处理之后的波形信号进行分段处理。其中，上述对波形信号进行分段处理的操作并非将波形信号切断为线段，而是将波形信号按照单调性标记为多个信号线段。具体的，可将波形信号按照单调向上和单调向下的单调性标记为K个信号线段，K为大于或者等于2的整数。其中，每个信号线段的单调性唯一。进一步的，可按照时间连续性对K个信号线段中的每个信号线段进行顺序标号，以得到携带顺序标号的线段。如图3，是本申请实施例提供的波形信号的另一示意图。如图3所示的一段波形信号中横坐标为波形产生时间，纵坐标为波形幅度。该段波形信号中可包括拐点1、2和3，每个拐点为线段的单调性变化节点。例如，拐点1的左侧线段的单调性为单调向下，拐点1的右侧线段的单调性为单调向下，因此可将拐点1的左侧线段和右侧线段分别标记为两个不同的信号线段。其中，拐点1左侧的信号线段的单调性为单调向下，并且该信号线段的产出时间高于拐点1的右侧线段，因此该信号线段可标记为204。拐点1和拐点2之间的信号线段的单调性为单调向上，并且该信号线段产生的时间在拐点1之后，因此，可将拐点1和拐点2之间的信号线段按照时间连续性标记为205。同理，拐点2和拐点3之间的信号线段的单调性为单调向下，因此，可将该信号线段标记为206。更多的图3未示出的信号线段中，出现时间在信号线段204之前的信号线段可标记为信号线段203、202、201、…、1、0等，出现时间在信号线段206之后的信号线段可标记为信号线段207、208、209等，在此不做限制。
可选的，波形信号为波浪形变化的曲线，因此，波形信号中一定有单调向下的线段连接单调向上的线段，再接单调向下的线段，线段连接点即为拐点。波形信号中单调性一上一下的线段相间，线段相接点为波形信号的波谷或者波峰，波谷与波峰相间连接。终端设备可按照波谷和波峰的变化规律，将任一波谷(或者波峰)与其相邻的一个波峰(或者波谷)之间单调性唯一的信号线段标记为一个信号线段。例如，上述拐点1可为波形信号的一波谷，拐点2为波形信号的一波峰，拐点3为波形信号的另一波谷。终端设备可将拐点点对应的波谷及其相邻的波峰(拐点2)之间单调性向上的线段标记为一个信号线段。
可选的，终端设备也可按照更多的分段标记方式，将采集到的一段波形信号标记为多个单调性唯一的信号线段，具体可根据实际应用场景确定，在此不做限制。
S23，提取各信号线段的线段数据。
在一些可行的实施方式中，终端设备将波形信号标记为多个信号线段之后，则可提取每个信号线段的线段数据。下面以任一信号线段i为例，对信号线段的线段数据的提取方式进行描述。具体的，终端设备可提取信号线段i的线段长度Xi和线段宽度Yi，按照预设长度和预设宽度分别对Xi和Yi进行差值拉伸以得到归一化处理后的信号线段(可标记为信号线段j)。进一步的，可提取所述信号线段j的线段数据。其中，信号线段的线段数据可包括但不限于：线段中间点位置、线段起点和/或终点位置、线段起点位置曲率、线段终点位置曲率、线段起点位置与线段相切的射线切点，以及信号终点位置与线段相切的射线切点等。其中，上述线段宽度可为线段的起点和终点之间间隔的时间长度。上述线段长度可为线段的起点和终点之间的变化幅度。上述线段中间点位置为线段的宽度的中间点对应的波形信号线段上的点。例如，图4是本申请实施例提供的波形信号的另一示意图。终端设备可将信号线段i的线段长度Xi和线段宽度Yi进行差值拉伸到长宽为1*100的预设规 格大小，进而可对差值拉伸之后的信号线段进行线段数据提取。其中，提取的线段数据可包括线段中间点O的坐标位置、线段起点A和线段终点B的坐标位置、线段起点A的曲率、线段终点B的曲率、线段起点A和线段相切的射线切点C的坐标位置，以及线段终点B与线段相切的射线切点D的坐标位置等。
进一步的，终点设备还可对上述信号线段j进行求导以得到信号线段j的导数曲线，并提取上述导数曲线的线段数据。其中，上述导数曲线的线段数据包括：拐点数量、各拐点位置，相邻拐点之间的中间点位置以及峰值最大的拐点位置等数据，在此不做限制。参见图5，是本申请实施例提供的信号线段的导数曲线示意图。图5所示的导数曲线为图4所示的信号线段的导数曲线示意图。由于图4所示的信号线段是曲线，曲线求导之后得到的依然是曲线。终端设备可提取图4所示的信号线段的导数曲线的线段数据，进一步的，可根据导数曲线的波形变化特征，从图4所示的信号线段中截取满足预定义需求的部分线段。例如，对应ECG信号，根据医学知识可知，在ECG信号中，Q波和R波之间是直线，不应该有拐弯，因此，若图4所示曲线为用于检测Q波和R波的信号线段，则只需考虑图4所示信号线段中拐弯右侧(即点(60，0.3)右侧)的直线部分，用于检测Q波或者R波等，可提高波形信号的检测准确率。
本申请实施例可首先根据单调性将波形信号标记为多个信号线段，再者可提取信号线段的线段数据，进一步的还可将信号线段求导之后提取信号线段的导数曲线的线段数据，通过多个层次提取波形信号的线段特征数据。本申请实施例可通过多个层次的线段特征数据识别波形信号标记得到各个信号线段的特征，进而可更好地识别波形特征，识别波形信号的周期等，信号处理质量更高，信号周期识别更准确。
S24，根据所述各信号线段的线段数据确定所述波形信号的线段匹配模板。
在一些可行的实施方式中，终端设备提取得到上述K个信号线段中各个信号线段的特征数据，包括信号线段的单调性、信号线段的线段数据以及信号线段的导数曲线的线段数据等数据之后，则可根据各个信号线段的特征数据构建该波形信号的线段匹配模板。由于具有周期性的波形信号中，每个周期都是相似的信号特征(可定义为周期信号特征)，因此，本申请实施例可构建一个周期信号特征的识别模板，用于识别出波形信号的各个周期中相似的信号特征。上述周期信号特征的识别模板可为本申请实施例提供的线段匹配模板。
在一些可行的实施方式中，终端设备可从各信号线段的线段数据中选取线段数据小于预设线段数据阈值的M个目标信号线段，并根据预置的线段组合方式将M个目标信号线段进行组合以得到波形信号的线段匹配模板。具体的，线段匹配模板中可包括M个连续的信号线段，即时间上连续的单调性一上一下相间的信号线段。
可选的，终端设备构建线段匹配模板时，可首先利用预设的波形信号先验信息，从波形信号的各个信号线段中选取目标信号线段。以ECG信号为例，参见图6，是ECG信号示意图。在ECG信号的一个周期信号数据中可包括一个QRS波群，一个P波，一个T波，一个U波，以及PR段和ST段等。QRS波群由单调向下接着单调向上，再接着单调向下的多个信号线段组成，QRS波群中的R点是该周期中的最高波峰，即最尖锐的波形波峰，R点和S点之间的时间间隔是0.06秒。
不同导联所记录的心电图，在波形表现上会有所不同，但一个正常的心电波形周期图 基本上都是由一个P波，一个QRS波群，一个T波以及过渡期所组成。此外，有时在T波后，还会出现一个小的U波。心电信号的这些特征波形和过渡期均代表着一定的生理学意义，现以MLH导联的正常心电图波形为例，如图6所示，对ECG信号的主要组成及其特点进行简要介绍如下：
(1)P波:也叫心房去极波,反映的是左右两心房去极化过程的电位变化。波形一般圆钝光滑，历时0.08-0.11秒，波幅不超过0.25mV。两心房复极化过程所产生的电位变化称为Ta波,它通常与PR段，QRS波群或ST段重叠在一起，且波幅很低，在心电图上不易辨认。
(2)PR间期(或称PQ间期):是P波起点到QRS波群起点之间的时间间隔，反映了自心房除极开始至心室除极开始的一段时间。正常成人的PR间期为0.12-0.20秒。若超过0.205秒，一般表明有房室传导阻滞的发生。PR间期的长短与年龄及心率等因素相关，在此不做限制。
(3)QRS波群:反映两心室去极化过程的电位变化。典型的QRS波群包括三个紧密相连的电位波动:第一个向下的波称为Q波，紧接着是向上并且高而尖峭的R波，最后是向下的S波。在不同导联中，这三个波不一定都出现，各波的幅度变化也较大，历时约0.06-0.105秒。
(4)ST段:指QRS波群终点与T波起点之间的线段，一般与零电位基线平齐。在这段时期内，因心室各部分都已全部进入除极化状态，但尚未开始复极，故心室各部分之间没有电位差存在，心电曲线恢复到基线水平。但若有冠状动脉供血不足或心肌梗死等情况发生时，ST段常会偏离基线，并超过一定的幅度范围。
(5)T波:反映两心室复极化过程的电位变化。T波的波形圆钝，升降支并不完全对称，波形的前支较长而后支较短，占时约0.05-0.255秒。T波方向应与QRS波群的主波方向一致，在以R波为主的导联中，T波的波幅应不低于本导联R波的1/10。
(6)QT间期:指从QRS波群起点到T波终点之间的时间，它代表心室开始去极化到全部复极化完毕所需的时间。这一间期的长短与心率密切相关，心率越快，QT间期越短。反之，则QT间期越长。正常的QT间期可因心率，年龄及性别等因素的不同而有所不同。当心率为75次/分时，QT间期为0.30-0.405秒。分析QT间期的变化，对疾病的早期诊断和分析抗心律失常药物对心脏的影响，可起到一定的辅助作用。
(7)U波:T波后0.02-0.04秒可能会出现一个与T波方向一致的低宽U波。
上述周期信号数据则可为该波形信号的先验信息。终端设备可构建一个线段匹配模板用于识别波形信号中的QRS波群。具体的，终端设备可利用QRS波群中的R点为一个周期中的最高波峰等波形信号的先验信息，从波形信号中找到每个预设时间间隔(例如1.5秒)内的最尖锐的波形波峰，按照获取到的各个波形波峰按照尖锐程度排序。去掉最尖锐的一部分，比如去除20％，用剩下的波峰们制作一个线段匹配模板。或者按照波形波峰的尖锐程度，去除尖锐程度超过预设阈值的部分波形波峰，进而可利用剩下的波峰们构建一个线段匹配模板。制作线段匹配模板可包括波峰左侧向上，右侧向下，右侧向下后右侧向上的三条线的特征，取入围波峰相关特征的各个线段数据的中位数，利用获取的中位数构建线段匹配模板。
具体的，终端设备可首先查找少量的目标信号线段，然后构造模板，再利用模板套取更多的目标信号线段，根据更多的目标信号线段更新模板。终端设备选取的目标信号线段越多，利用各个信号线段的线段数据修正的线段匹配模板的准确性则越高，最后则可确定出与波形信号的QRS波群相似度较高的线段匹配模板。
可选的，上述线段匹配模板具体可为一组作为匹配标准的数据，该组数据中包括但不限于线段的长度，线段的宽度，线段起点的X轴、Y轴坐标位置，线段起点或者终点附近的局部斜率(即信号线段的局部导数)，线段的起点或者终点与线段相切的切线斜率(比如图4所示的线段中，线段起点A(0，0)到点C(60，0.3)的射线斜率)等。
具体实现中，对于ECG信号或者PPG信号等人类体征信号，人类个体的体型信号差异较大，因此ECG传感器或者PPG传感器检测到的波形信号也会因人而异。本申请实施例可根据每个波形信号的特征数据构建该波形信号的线段匹配模板，可提高波形信号的处理质量，提高波形信号的周期识别准确率。
S25，将所述K个信号线段中的各信号线段与所述线段匹配模板中包括的M个信号线段进行匹配，并根据所述各信号线段的匹配结果确定出所述波形信号的目标波群。
S26，根据所述目标波群的线段数据确定出所述波形信号的周期信号数据。
在一些可行的实施方式中，终点获取得到该波形信号的线段匹配模板之后，则可使用线段匹配模板从波形信号中识别出波形信号的目标波群，例如ECG信号中的QRS波群等。具体的，终端设备可将波形信号标记得到的K个信号线段中的每个信号线段与线段匹配模板中包括的信号线段进行一一匹配，从K个信号线段中查找有效信号线段。其中，上述有效信号线段指与线段匹配模板中包括的任一线段的匹配度大于或者等于预设相似度阈值的信号线段。即，信号线段之间的特征差值小于预置阈值的线段。进一步的，可将查找得到的N个有效信号线段所呈现的波群的信号特征与线段匹配模板中包括的信号线段所呈现的波群的信号特征进行匹配。若波群的信号特征的特征差值小于预设阈值，则可确定上述N个有效信号中组成的波群中与线段匹配模板相似的波群为目标波群。例如，若上述N个有效信号线段中的3个信号线段组成的一个波群，与线段匹配模板组成的QRS波群相似，则可确定上述3个有效信号线段组成的波群为目标波群，即，为一个QRS波群。
例如，在一段时间长度为一分钟的ECG信号，经过滤波后可标记为大约1500条信号线段。具体的，上述信号线段的数目和波形信号的滤波强度相关，具体可根据实际应用场景确定，这里仅是举例子说明。线段匹配模板中包括的信号线段(目标信号线段)有三个，例如，信号线段QR，RS和SS’(S点之后的单调上升线段，可标记为SS’，如图6)。待校验的信号线段的1500条信号线段中，每一条信号线段均去匹配上述三条目标信号线段，也就是试图匹配4500次。具体每次执行是，以QR为例，假设目标信号线段QR的宽度为QR.L。如果待校验的信号线段的线段宽度在QR.L的50％到200％的宽度范围内，在可确定信号线段的宽度检测过关，然后类似的检测信号线段的长度，再检测信号线段的起点或者终点附近的局部斜率、信号线段的起点或者终点到线段的射线斜率等各项参数。若各项参数均匹配成功，则可确定出有效信号线段。进一步的，可检测有效信号线段中多个连续的信号线段所组成的波群特征，进而可确定出目标波群。其中，上述有效信号线段的匹配中，有效信号线段还包括上述刨除拐弯左侧的线段之外的右侧信号线段。即，上述包括拐弯的 信号线段不符合QR线段的特征，拐弯的形成可能来自数据采集中的误差，或者其他环境因素等。若有效信号线段的匹配结果中，出现的时间顺序标号连续的三个有效信号线段，与线段匹配模板中的QR，RS和SS’分别匹配成功，则可确定为上述三个有效信号线段呈现的波群为一个QRS波群，即为目标波群。
可选的，终端设备识别出目标波群之后，可根据目标波群识别出波形信号的周期。终端设备还可根据目标波群的周边波形信号以及预设的周期校验数据确定出波形信号的周期。其中，上述目标波群的周边波形信号可包括与目标波群的有效信号线段相邻的多个信号线段的线段数据。这些周边波形信号可呈现出目标波群之外的波峰或者波谷，例如，ECG中的T波、U波或者P波等。上述周期校验数据包括：周期波形宽度、周期波形幅度、周期起点波形高度以及周期终点波形高度等，在此不做限制。
比如ECG信号中T波只有一个，P波不会比R波高，甚至矮比较多。S点之后波形不应该继续大幅度向下。再比如PPG波形幅度和周期起止点高度都有限定幅度或者高度。这些都可称为周期校验数据或者目标波群的周边波形信号的校验数据。
比如ECG信号中的T波，是QRS波群之后一段时间的一个凸起。通常心脏不好的个体的ECG信号中T波的波峰高度会大一些，若ECG信号中的T波倒置，则该个体出现了心绞痛等体征。因此，通过检测QRS波群之后是否有T波以及T波的宽度，高度是不是合适可以控制波形信号的检测质量，提高波形信号的周期识别准确率。
可选的，终端设备识别出目标波群或者波形信号的周期信号数据之后，可根据识别得到的信息剔除波形信号中的异常周期的波形信号或者修补异常周期的波形信号。例如，终端设备识别出目标波群之后，可找到心率周期。例如，如果波形信号中少了一个周期则识别出来的心跳间隔可为1秒、1秒、1秒、2秒、1秒等。由此可确定中间缺少了一个波形周期，此时可能是该周期的波形信号没有通过周期信号数据匹配。终端设备可降低周期校验数据的预置周期预置等数据门限标准对波形信号进行重新匹配，通过周期识别补救一个周期。相反的，若心跳间隔出现1秒、1秒、0.3秒、0.7秒、1秒，则可确定识别得到的心跳周期多另一个，此时可能是噪声被识别为心率周期了，进而可对波形信号进行再次识别等操作，排除噪声。
可选的，如果波形信号识别出来的周期信号数据是个孤立的周期，该周期的相邻信号都是噪声，则可该周期也可能是噪声，该周期的信号数据也不可取，故此可将其丢弃。
终端设备可根据确定出的波形信号的周期信号数据识别波形信号中的各个周期，进而可确定周期之间的特征数据，去除波形信号中的异常周期信号，提高波形信号的检测准确率。
参见图7，是本申请实施例提供的PPG信号示意图。如图7所示的两个相互独立的PPG波形，终端设备可根据上述各个步骤所描述的实现方式识别PPG信号的单个周期的周期信号数据，也可识别出周期与周期之间的波形相似度，进而可剔除波形信号中的异常周期信号。其中，图7右侧的4个周期波形中，从上往下可排序为波形1、波形2、波形3和波形4。波形1和波形2可表示异常周期信号，波形3和波4可表示正常周期信号。
参见图8，是本申请实施例提供的加速度计传感器波形的信号示意图。如图8所示是一个加速度计传感器波形，终端设备可根据上述各个步骤所描述的实现方式识别各个周期 的波形信号数据。其中，加速度计传感器的波形信号特征(或者波形信号先验信息)可包括一个W波形和V波形相间呈现。其中，W波形为前脚落地的特征体现，V波形为后脚落地的特征体现。
本申请实施例可从波形信号中提取信号线段的线段数据，还可对信号线段进行求导等处理并提取导数曲线的线段数据，根据提取的数据构建用于识别波形信号的周期信号数据的线段匹配模板。进一步的可使用线段匹配模板识别波形信号的周期，排除噪声。本申请实施例提供的线段匹配模板可由波形信号的线段数据构建得到，线段匹配模板可因波形信号的特征数据不同而有所差异，进而可提高线段匹配模板与波形信号的线段数据的关联，可提高波形信号的周期识别准确率，提高噪声的排除准确性，进而可提高波形信号的处理质量。
参见图9，是本申请实施例提供的波形信号处理的装置的结构示意图。本申请实施例提供的波形信号处理的装置具体可为本申请实施例提供的终端设备，上述终端设备可包括：
获取单元91，用于获取滤波处理后的波形信号。
分段单元92，用于将所述获取单元获取的所述波形信号按照单调性标记为K个信号线段，所述K为大于或者等于2的整数。
提取单元93，用于提取所述分段单元处理得到的各信号线段的线段数据。
处理单元94，用于根据所述提取单元提取的所述各信号线段的线段数据确定所述波形信号的线段匹配模板，所述线段匹配模板中包括M个连续的信号线段，所述M为小于K的整数。
所述处理单元94，还用于将所述分段单元处理得到的所述K个信号线段中的各信号线段与所述线段匹配模板中包括的M个信号线段进行匹配，并根据所述各信号线段的匹配结果确定出所述波形信号的目标波群。
所述处理单元94，还用于根据所述目标波群的线段数据确定出所述获取单元获取的所述波形信号的周期信号数据。
可选的，上述分段单元92用于：
将所述波形信号按照单调性划分为单调向上的信号线段或者单调向下的信号线段，并按照时间连续性对每个信号线段进行顺序标号以得到携带顺序标号的线段。
可选的，上述提取单元93用于：
提取所述信号线段i的线段长度Xi和线段宽度Yi，并按照预设长度和预设宽度分别对所述Xi和Yi进行差值拉伸以得到归一化处理后的信号线段j；
提取所述信号线段j的线段数据；
其中，所述信号线段j的线段数据包括：线段中间点位置、线段起点位置曲率、线段终点位置曲率、线段起点位置与线段相切的射线切点，以及线段终点位置与线段相切的射线切点中的至少一种。
可选的，上述提取单元93还用于：
对所述信号线段j进行求导或者差分以得到所述信号线段j的导数曲线，并提取所述导数曲线的线段数据；
其中，所述导数曲线的线段数据包括：拐点数量、各拐点位置，相邻拐点之间的中间 点位置以及峰值最大的拐点位置中的至少一种。
可选的，上述处理单元94用于：
从所述各信号线段的线段数据中选取线段数据小于预设线段数据阈值的M个目标信号线段，并根据预置的线段组合方式将所述M个目标信号线段进行组合以得到所述波形信号的线段匹配模板；
其中，所述M个目标信号线段为顺序标号连续的线段。
可选的，上述处理单元94用于：
将所述K个信号线段中的每个信号线段与所述线段匹配模板中包括的M个目标信号线段中的每个目标信号线段进行线段数据匹配，从所述K个信号线段中确定N个有效信号线段，所述N为小于M的整数，所述有效信号线段的线段数据与所述线段匹配模板包括的任一目标线段的线段数据的特征差值不小于预置阈值；
若所述N个有效信号线段中包括顺序标号连续的M个有效信号线段，并且所述M个有效信号线段分别与所述线段匹配模板中包括的M个目标信号线段的特征差值小于预设阈值，则将所述M个有效信号线段的位置确定为目标波群位置。
可选的，所述目标波群位置的线段数据包括所述目标波群位置上的M个有效信号线段的线段数据，以及与所述M个有效信号线段的顺序标号连续的J个相邻信号线段的线段数据；
上述处理单元94用于：
根据所述M个有效信号线段与所述J个相邻信号线段的线段数据，确定所述M个有效信号线段与所述J个相邻信号线段组成的波群的波形宽度，波群的波形幅度、波群的起点高度或者波群的终点波动中的至少一种波群特征数据；
若所述波群特征数据与周期校验数据的特征差值小于预置周期阈值，则确定所述波群为所述波形信号的一个周期，所述波群的波群特征数据为所述波形信号的周期信号数据；
其中，所述周期校验数据包括：周期波形宽度、周期波形幅度、周期起点波形高度以及周期终点波形高度中的至少一种。
可选的，上述处理单元94还用于：
根据所述波形信号的周期信号数据确定出所述波形信号中包括的多个波形周期，并删除所述波形信号中包括的异常周期数据；
其中，所述异常周期数据为所述波形信号中不包括所述周期信号数据的波群的线段数据。
可选的，所述波形信号包括：心电图ECG、光电容积脉冲波PPG、压力计信号、磁力计信号以及加速度计传感器波形信号中的至少一种。
具体实现中，上述终端设备可通过其内置的各个单元执行上述实施例中终端设备所执行的实现方式，具体可参见上述实施例中各个步骤所描述的实现方式，在此不再赘述。
在本申请实施例中，终端设备可从波形信号中提取信号线段的线段数据，还可对信号线段进行求导等处理并提取导数曲线的线段数据，根据提取的数据构建用于识别波形信号的周期信号数据的线段匹配模板。进一步的可使用线段匹配模板识别波形信号的周期，排除噪声。本申请实施例提供的线段匹配模板可由波形信号的线段数据构建得到，线段匹配 模板可因波形信号的特征数据不同而有所差异，进而可提高线段匹配模板与波形信号的线段数据的关联，可提高波形信号的周期识别准确率，提高噪声的排除准确性，进而可提高波形信号的处理质量。
请参见图10，图10是本申请实施例提供的一种通信设备的结构示意图。如图10所示，本申请实施例提供的通信设备100包括处理器101、存储器102、收发器103和总线***104。
其中，上述处理器101、存储器102和收发器103通过总线***104连接。
上述存储器102用于存放程序。具体地，程序可以包括程序代码，程序代码包括计算机操作指令。存储器102包括但不限于是随机存储记忆体(random access memory，RAM)、只读存储器(read-only memory，ROM)、可擦除可编程只读存储器(erasable programmable read only memory，EPROM)、或便携式只读存储器(compact disc read-only memory，CD-ROM)。图10中仅示出了一个存储器，当然，存储器也可以根据需要，设置为多个。存储器102也可以是处理器101中的存储器，在此不做限制。
存储器102存储了如下的元素，可执行模块或者数据结构，或者它们的子集，或者它们的扩展集：
操作指令：包括各种操作指令，用于实现各种操作。
操作***：包括各种***程序，用于实现各种基础业务以及处理基于硬件的任务。
上述处理器101控制通信设备100的操作，处理器101可以是一个或多个中央处理器(central processing unit，CPU)，在处理器101是一个CPU的情况下，该CPU可以是单核CPU，也可以是多核CPU。
具体的应用中，通信设备100的各个组件通过总线***104耦合在一起，其中总线***104除包括数据总线之外，还可以包括电源总线、控制总线和状态信号总线等。但是为了清楚说明起见，在图10中将各种总线都标为总线***104。为便于表示，图10中仅是示意性画出。
上述各个实施例揭示的终端设备的方法可以应用于处理器101中，或者由处理器101实现。处理器101可能是一种集成电路芯片，具有信号的处理能力。在实现过程中，上述方法的各步骤可以通过处理器101中的硬件的集成逻辑电路或者软件形式的指令完成。上述的处理器101可以是通用处理器、数字信号处理器(digital signal processing，DSP)、专用集成电路(application specific integrated circuit，ASIC)、现场可编程门阵列(field－programmable gate array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成，或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器102，处理器101读取存储器102中的信息，结合其硬件执行上述各个实施例所描述的终端设备的方法步骤。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过 计算机程序来指令相关的硬件来完成，所述的程序可存储于计算机可读取存储介质中，该程序在执行时，可包括如上述各方法的实施例的流程。其中，所述的存储介质可为磁碟、光盘、ROM或RAM等。

Claims (18)

  1. 一种波形信号处理的方法，其特征在于，包括：
    获取滤波处理后的波形信号；
    将所述波形信号按照单调性标记为K个信号线段，所述K为大于或者等于2的整数；
    提取各信号线段的线段数据，并根据所述各信号线段的线段数据确定所述波形信号的线段匹配模板，所述线段匹配模板中包括M个连续的信号线段，所述M为小于K的整数；
    将所述K个信号线段中的各信号线段与所述线段匹配模板中包括的M个信号线段进行匹配，并根据所述各信号线段的匹配结果确定出所述波形信号的目标波群；
    根据所述目标波群的线段数据确定出所述波形信号的周期信号数据。
  2. 如权利要求1所述的方法，其特征在于，所述将所述波形信号按照单调性标记为K个信号线段包括：
    将所述波形信号按照单调性划分为单调向上的信号线段或者单调向下的信号线段，并按照时间连续性对每个信号线段进行顺序标号以得到携带顺序标号的线段。
  3. 如权利要求2所述的方法，其特征在于，所述提取各信号线段的线段数据中提取任一信号线段i的线段数据包括：
    提取所述信号线段i的线段长度Xi和线段宽度Yi，并按照预设长度和预设宽度分别对所述Xi和Yi进行差值拉伸以得到归一化处理后的信号线段j；
    提取所述信号线段j的线段数据；
    其中，所述信号线段j的线段数据包括：线段中间点位置、线段起点位置曲率、线段终点位置曲率、线段起点位置与线段相切的射线切点，以及线段终点位置与线段相切的射线切点中的至少一种。
  4. 如权利要求3所述的方法，其特征在于，所述方法还包括：
    对所述信号线段j进行求导或者差分以得到所述信号线段j的导数曲线，并提取所述导数曲线的线段数据；
    其中，所述导数曲线的线段数据包括：拐点数量、各拐点位置，相邻拐点之间的中间点位置以及峰值最大的拐点位置中的至少一种。
  5. 如权利要求3或4所述的方法，其特征在于，所述根据所述各信号线段的线段数据确定所述波形信号的线段匹配模板包括：
    从所述各信号线段的线段数据中选取线段数据小于预设线段数据阈值的M个目标信号线段，并根据预置的线段组合方式将所述M个目标信号线段进行组合以得到所述波形信号的线段匹配模板；
    其中，所述M个目标信号线段为顺序标号连续的线段。
  6. 如权利要求5所述的方法，其特征在于，所述将所述K个信号线段中的各信号线段与所述线段匹配模板中包括的M个信号线段进行匹配，并根据所述各信号线段的匹配结果确定出所述波形信号的目标波群包括：
    将所述K个信号线段中的每个信号线段与所述线段匹配模板中包括的M个目标信号线段中的每个目标信号线段进行线段数据匹配，从所述K个信号线段中确定N个有效信号线段，所述N为小于M的整数，所述有效信号线段的线段数据与所述线段匹配模板包括的任一目标线段的线段数据的特征差值不小于预置阈值；
    若所述N个有效信号线段中包括顺序标号连续的M个有效信号线段，并且所述M个有效信号线段分别与所述线段匹配模板中包括的M个目标信号线段的特征差值小于预设阈值，则将所述M个有效信号线段的位置确定为目标波群位置。
  7. 如权利要求6所述的方法，其特征在于，所述目标波群位置的线段数据包括所述目标波群位置上的M个有效信号线段的线段数据，以及与所述M个有效信号线段的顺序标号连续的J个相邻信号线段的线段数据；
    所述根据所述目标波群的线段数据确定出所述波形信号的周期信号数据包括：
    根据所述M个有效信号线段与所述J个相邻信号线段的线段数据，确定所述M个有效信号线段与所述J个相邻信号线段组成的波群的波形宽度，波群的波形幅度、波群的起点高度或者波群的终点波动中的至少一种波群特征数据；
    若所述波群特征数据与周期校验数据的特征差值小于预置周期阈值，则确定所述波群为所述波形信号的一个周期，所述波群的波群特征数据为所述波形信号的周期信号数据；
    其中，所述周期校验数据包括：周期波形宽度、周期波形幅度、周期起点波形高度以及周期终点波形高度中的至少一种。
  8. 如权利要求7所述的方法，其特征在于，所述方法还包括：
    根据所述波形信号的周期信号数据确定出所述波形信号中包括的多个波形周期，并删除所述波形信号中包括的异常周期数据；
    其中，所述异常周期数据为所述波形信号中不包括所述周期信号数据的波群的线段数据。
  9. 如权利要求1至8任一项所述的方法，其特征在于，所述波形信号包括：心电图ECG、光电容积脉冲波PPG、压力计信号、磁力计信号以及加速度计传感器波形信号中的至少一种。
  10. 一种波形信号处理的装置，其特征在于，包括：
    获取单元，用于获取滤波处理后的波形信号；
    分段单元，用于将所述获取单元获取的所述波形信号按照单调性标记为K个信号线段，所述K为大于或者等于2的整数；
    提取单元，用于提取所述分段单元处理得到的各信号线段的线段数据；
    处理单元，用于根据所述提取单元提取的所述各信号线段的线段数据确定所述波形信号的线段匹配模板，所述线段匹配模板中包括M个连续的信号线段，所述M为小于K的整数；
    所述处理单元，还用于将所述分段单元处理得到的所述K个信号线段中的各信号线段与所述线段匹配模板中包括的M个信号线段进行匹配，并根据所述各信号线段的匹配结果确定出所述波形信号的目标波群；
    所述处理单元，还用于根据所述目标波群的线段数据确定出所述获取单元获取的所述波形信号的周期信号数据。
  11. 如权利要求10所述的装置，其特征在于，所述分段单元用于：
    将所述波形信号按照单调性划分为单调向上的信号线段或者单调向下的信号线段，并按照时间连续性对每个信号线段进行顺序标号以得到携带顺序标号的线段。
  12. 如权利要求11所述的装置，其特征在于，所述提取单元用于：
    提取所述信号线段i的线段长度Xi和线段宽度Yi，并按照预设长度和预设宽度分别对所述Xi和Yi进行差值拉伸以得到归一化处理后的信号线段j；
    提取所述信号线段j的线段数据；
    其中，所述信号线段j的线段数据包括：线段中间点位置、线段起点位置曲率、线段终点位置曲率、线段起点位置与线段相切的射线切点，以及线段终点位置与线段相切的射线切点中的至少一种。
  13. 如权利要求12所述的装置，其特征在于，所述提取单元还用于：
    对所述信号线段j进行求导或者差分以得到所述信号线段j的导数曲线，并提取所述导数曲线的线段数据；
    其中，所述导数曲线的线段数据包括：拐点数量、各拐点位置，相邻拐点之间的中间点位置以及峰值最大的拐点位置中的至少一种。
  14. 如权利要求12或13所述的装置，其特征在于，所述处理单元用于：
    从所述各信号线段的线段数据中选取线段数据小于预设线段数据阈值的M个目标信号线段，并根据预置的线段组合方式将所述M个目标信号线段进行组合以得到所述波形信号的线段匹配模板；
    其中，所述M个目标信号线段为顺序标号连续的线段。
  15. 如权利要求14所述的装置，其特征在于，所述处理单元用于：
    将所述K个信号线段中的每个信号线段与所述线段匹配模板中包括的M个目标信号线段中的每个目标信号线段进行线段数据匹配，从所述K个信号线段中确定N个有效信号线段，所述N为小于M的整数，所述有效信号线段的线段数据与所述线段匹配模板包括的任一目标线段的线段数据的特征差值不小于预置阈值；
    若所述N个有效信号线段中包括顺序标号连续的M个有效信号线段，并且所述M个有效信号线段分别与所述线段匹配模板中包括的M个目标信号线段的特征差值小于预设阈值，则将所述M个有效信号线段的位置确定为目标波群位置。
  16. 如权利要求15所述的装置，其特征在于，所述目标波群位置的线段数据包括所述目标波群位置上的M个有效信号线段的线段数据，以及与所述M个有效信号线段的顺序标号连续的J个相邻信号线段的线段数据；
    所述处理单元用于：
    根据所述M个有效信号线段与所述J个相邻信号线段的线段数据，确定所述M个有效信号线段与所述J个相邻信号线段组成的波群的波形宽度，波群的波形幅度、波群的起点高度或者波群的终点波动中的至少一种波群特征数据；
    若所述波群特征数据与周期校验数据的特征差值小于预置周期阈值，则确定所述波群为所述波形信号的一个周期，所述波群的波群特征数据为所述波形信号的周期信号数据；
    其中，所述周期校验数据包括：周期波形宽度、周期波形幅度、周期起点波形高度以及周期终点波形高度中的至少一种。
  17. 如权利要求16所述的装置，其特征在于，所述处理单元还用于：
    根据所述波形信号的周期信号数据确定出所述波形信号中包括的多个波形周期，并删除所述波形信号中包括的异常周期数据；
    其中，所述异常周期数据为所述波形信号中不包括所述周期信号数据的波群的线段数据。
  18. 如权利要求10-17任一项所述的装置，其特征在于，所述波形信号包括：心电图ECG、光电容积脉冲波PPG、压力计信号、磁力计信号以及加速度计传感器波形信号中的至少一种。
PCT/CN2017/095073 2017-03-14 2017-07-29 波形信号处理的方法及装置 WO2018166136A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201780009428.4A CN108701219B (zh) 2017-03-14 2017-07-29 波形信号处理的方法及装置
US16/494,573 US11321561B2 (en) 2017-03-14 2017-07-29 Electrocardiogram waveform signal processing method and apparatus

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710151454.2 2017-03-14
CN201710151454 2017-03-14

Publications (1)

Publication Number Publication Date
WO2018166136A1 true WO2018166136A1 (zh) 2018-09-20

Family

ID=63522818

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/095073 WO2018166136A1 (zh) 2017-03-14 2017-07-29 波形信号处理的方法及装置

Country Status (3)

Country Link
US (1) US11321561B2 (zh)
CN (1) CN108701219B (zh)
WO (1) WO2018166136A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111291727A (zh) * 2020-03-17 2020-06-16 乐普（北京）医疗器械股份有限公司 一种光体积变化描记图法信号质量检测方法和装置
CN116302897A (zh) * 2023-05-15 2023-06-23 合肥联宝信息技术有限公司 一种数据集的建立方法、装置、电子设备及存储介质

Families Citing this family (19)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020155169A1 (zh) * 2019-02-03 2020-08-06 华为技术有限公司 房颤筛查的方法和装置
CN109745037B (zh) * 2019-02-14 2022-06-24 广州视源电子科技股份有限公司 一种图像处理方法、装置、设备和存储介质
CN110135516B (zh) * 2019-05-24 2022-04-01 北京天泽智云科技有限公司 一种基于包络线及内积的高频数据模式识别方法
CN110353664B (zh) * 2019-07-25 2022-09-20 山东凯迪泰科智能***有限公司 心电图pqrst波分段方法及***
CN110503069B (zh) * 2019-08-28 2022-10-18 中广核研究院有限公司 电流波形波动起点识别方法及电子设备、可读存储介质
CN112945257B (zh) * 2019-12-11 2024-06-21 瑞昱半导体股份有限公司 计步装置及方法
CN110960203B (zh) * 2019-12-13 2022-11-25 心核心科技(北京)有限公司 一种心血管特性参数检测方法及装置
CN111150387B (zh) * 2020-01-15 2021-06-08 深圳市邦健科技有限公司 心电图模板匹配方法、装置、计算机设备及存储介质
CN111297351A (zh) * 2020-02-13 2020-06-19 苏州百慧华业精密仪器有限公司 一种动态心电图中的运动伪差识别方法及装置
CN113261972B (zh) * 2020-02-17 2023-06-27 华为技术有限公司 心电检测装置、电路及方法
CN112434561B (zh) * 2020-11-03 2023-09-22 中国工程物理研究院电子工程研究所 一种自动判定冲击波信号有效性的方法
CN112765802B (zh) * 2021-01-13 2022-11-29 陕西师范大学 基于高阶水波模型演化水波波形的方法
CN112958840B (zh) * 2021-02-10 2022-06-14 西南电子技术研究所（中国电子科技集团公司第十研究所） 精密零件加工削力信号自动分段方法
CN113349791A (zh) * 2021-05-31 2021-09-07 平安科技（深圳）有限公司 异常心电信号的检测方法、装置、设备及介质
CN113283413B (zh) * 2021-07-26 2021-09-24 枫树谷(成都)科技有限责任公司 脉冲波形模板库的创建方法、***、存储介质及设备
CN114742114B (zh) * 2022-06-09 2022-10-21 毕胜普生物科技有限公司 高频qrs波形曲线分析方法、装置、计算机设备与存储介质
CN114767102B (zh) * 2022-06-20 2022-11-29 天津大学 基于波形比例系数提取的动态光谱数据处理方法
CN116548928B (zh) * 2023-07-11 2023-09-08 西安浩阳志德医疗科技有限公司 一种基于互联网的护理服务***
CN117357130B (zh) * 2023-12-07 2024-02-13 深圳泰康医疗设备有限公司 基于人工智能的心电图数字化曲线分割方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2973998A1 (fr) * 2011-04-15 2012-10-19 Nacira Zegadi Dispostif et procede de suivi d'une activite cardiaque et pulmonaire
CN104102915A (zh) * 2014-07-01 2014-10-15 清华大学深圳研究生院 一种心电异常状态下基于ecg多模板匹配的身份识别方法
CN104161510A (zh) * 2014-08-27 2014-11-26 电子科技大学 一种多级导联心电信号qrs波形识别方法
CN104834921A (zh) * 2015-05-22 2015-08-12 清华大学深圳研究生院 心电正/异常大数据处理方法及装置
CN106137184A (zh) * 2015-04-08 2016-11-23 四川锦江电子科技有限公司 基于小波变换的心电信号qrs波检测方法

Family Cites Families (27)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5271411A (en) * 1990-09-21 1993-12-21 Colin Electronics Co., Ltd. Method and apparatus for ECG signal analysis and cardiac arrhythmia detection
US5301677A (en) * 1992-02-06 1994-04-12 Cardiac Pacemakers, Inc. Arrhythmia detector using delta modulated turning point morphology of the ECG wave
US6684100B1 (en) * 2000-10-31 2004-01-27 Cardiac Pacemakers, Inc. Curvature based method for selecting features from an electrophysiologic signals for purpose of complex identification and classification
US6564091B2 (en) * 2001-05-22 2003-05-13 Biotronik Mess-Und Therapiegerate Gmbh & Co. Ingenieurburo Berlin Method and memory means for storing cardiac rhythm information
US7184818B2 (en) * 2002-03-25 2007-02-27 Cardiac Pacemakers, Inc. Method and system for characterizing a representative cardiac beat using multiple templates
US7376458B2 (en) * 2004-11-29 2008-05-20 Cameron Health, Inc. Method for defining signal templates in implantable cardiac devices
US7477935B2 (en) * 2004-11-29 2009-01-13 Cameron Health, Inc. Method and apparatus for beat alignment and comparison
US7983742B2 (en) * 2006-02-27 2011-07-19 Vito Starc Multi-channel system for beat to beat QT interval variability
US8260406B2 (en) * 2006-11-30 2012-09-04 Aalborg Universitet System and method for analyzing complex curvature of ECG curves
US20090048528A1 (en) * 2007-08-16 2009-02-19 Bruce Hopenfeld System and methods for detecting ischemia with a limited extracardiac lead set
DE602008005841D1 (de) * 2008-03-18 2011-05-12 Biotronik Crm Patent Ag Vorrichtung und computerlesbarer Datenträger zur SVT- und VT-Klassifizierung
CN102379694B (zh) 2011-10-12 2014-07-16 中国科学院苏州纳米技术与纳米仿生研究所 心电图r波检测方法
US20140081162A1 (en) * 2012-09-19 2014-03-20 Pacesetter, Inc. Method and system for st morphology discrimination utilizing reference morphology templates
CN103020472B (zh) 2012-12-27 2015-12-09 中国科学院深圳先进技术研究院 基于约束估计的生理信号质量评估方法和***
CN103190901A (zh) 2013-04-01 2013-07-10 天津工业大学 基于极值域均值模式分解和改进Hilbert包络的R波检测算法
CN103488751B (zh) 2013-09-24 2016-07-13 浙江大学 基于光谱曲线波形相似度的光谱匹配方法
CN103961089B (zh) * 2014-05-27 2015-11-18 山东师范大学 基于分段直线拟合的窦性心率震荡趋势检测方法
CN105590011B (zh) 2014-10-20 2019-04-30 深圳市迈迪加科技发展有限公司 一种基于脉搏回归模型的心电信号数据修正方法及***
EP3892198B1 (en) * 2014-11-14 2024-03-06 ZOLL Medical Corporation Medical premonitory event estimation
CN104523266B (zh) 2015-01-07 2017-04-05 河北大学 一种心电信号自动分类方法
CN104706349B (zh) 2015-04-13 2017-03-22 大连理工大学 一种基于脉搏波信号的心电信号构建方法
CN105125206B (zh) 2015-09-15 2019-01-22 中山大学 一种智能心电监护方法及装置
CN106093467A (zh) 2016-02-23 2016-11-09 河南理工大学 输电网络运动在线监测***及安全评估方法
JP6764956B2 (ja) * 2016-06-27 2020-10-07 カーディアック ペースメイカーズ， インコーポレイテッド 再同期ペーシング管理に皮下で感知されたｐ波を使用する心臓治療法システム
US10426962B2 (en) * 2016-07-07 2019-10-01 Cardiac Pacemakers, Inc. Leadless pacemaker using pressure measurements for pacing capture verification
US10870008B2 (en) * 2016-08-24 2020-12-22 Cardiac Pacemakers, Inc. Cardiac resynchronization using fusion promotion for timing management
CN106473750B (zh) * 2016-10-08 2019-03-26 西安电子科技大学 基于光电容积脉搏波最佳周期波形的身份识别方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2973998A1 (fr) * 2011-04-15 2012-10-19 Nacira Zegadi Dispostif et procede de suivi d'une activite cardiaque et pulmonaire
CN104102915A (zh) * 2014-07-01 2014-10-15 清华大学深圳研究生院 一种心电异常状态下基于ecg多模板匹配的身份识别方法
CN104161510A (zh) * 2014-08-27 2014-11-26 电子科技大学 一种多级导联心电信号qrs波形识别方法
CN106137184A (zh) * 2015-04-08 2016-11-23 四川锦江电子科技有限公司 基于小波变换的心电信号qrs波检测方法
CN104834921A (zh) * 2015-05-22 2015-08-12 清华大学深圳研究生院 心电正/异常大数据处理方法及装置

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111291727A (zh) * 2020-03-17 2020-06-16 乐普（北京）医疗器械股份有限公司 一种光体积变化描记图法信号质量检测方法和装置
CN111291727B (zh) * 2020-03-17 2023-10-03 乐普（北京）医疗器械股份有限公司 一种光体积变化描记图法信号质量检测方法和装置
CN116302897A (zh) * 2023-05-15 2023-06-23 合肥联宝信息技术有限公司 一种数据集的建立方法、装置、电子设备及存储介质
CN116302897B (zh) * 2023-05-15 2023-10-17 合肥联宝信息技术有限公司 一种数据集的建立方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
US11321561B2 (en) 2022-05-03
CN108701219B (zh) 2021-05-04
US20200089948A1 (en) 2020-03-19
CN108701219A (zh) 2018-10-23

Similar Documents

Publication Publication Date Title
WO2018166136A1 (zh) 波形信号处理的方法及装置
Liu et al. Real-time multilead convolutional neural network for myocardial infarction detection
JP7065185B2 (ja) 心電信号の検出
Xu et al. Interpretation of electrocardiogram (ECG) rhythm by combined CNN and BiLSTM
EP3589199B1 (en) Deep learning algorithms for heartbeats detection
JP6811773B2 (ja) フォトプレチスモグラム（ｐｐｇ）信号品質を定量化する方法
EP3366203B1 (en) Method and system for cuffless blood pressure estimation using photoplethysmogram features and pulse transit time
Zhao et al. A robust QRS detection and accurate R-peak identification algorithm for wearable ECG sensors
Banerjee et al. PhotoECG: Photoplethysmographyto estimate ECG parameters
CN103815897A (zh) 一种心电图特征提取方法
Velic et al. Computer aided ECG analysis—State of the art and upcoming challenges
Sedighian et al. Pediatric heart sound segmentation using Hidden Markov Model
CN115470828A (zh) 基于卷积与自注意力机制的多导联心电图分类识别方法
WO2018129822A1 (zh) 一种血压测量方法及设备
CN101919704A (zh) 一种心音信号定位、分段方法
Wu et al. Personalizing a generic ECG heartbeat classification for arrhythmia detection: a deep learning approach
Sharma et al. Novel T-wave detection technique with minimal processing and RR-interval based enhanced efficiency
Martinez et al. Boosted-SpringDTW for comprehensive feature extraction of PPG signals
EP3506314B1 (en) Systems and methods for detecting anomaly in a cardiovascular signal using hierarchical extremas and repetitions
CN106778561B (zh) 一种穿戴式设备的身份识别方法及识别装置
CN110688414B (zh) 时序数据的处理方法、装置和计算机可读存储介质
Martinez et al. Boosted-SpringDTW for Comprehensive Feature Extraction of Physiological Signals
CN113057611B (zh) 用于输出信息的方法、装置、设备以及存储介质
Bai et al. A deep multi-scale convolutional neural network for classifying heartbeats
LIN et al. Myocardial infarction classification using polynomial approximation and principal component analysis

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17900810

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17900810

Country of ref document: EP

Kind code of ref document: A1