WO2020029031A1 - 一种信号处理方法、系统及计算机存储介质 - Google Patents

一种信号处理方法、系统及计算机存储介质 Download PDF

Info

Publication number
WO2020029031A1
WO2020029031A1 PCT/CN2018/099027 CN2018099027W WO2020029031A1 WO 2020029031 A1 WO2020029031 A1 WO 2020029031A1 CN 2018099027 W CN2018099027 W CN 2018099027W WO 2020029031 A1 WO2020029031 A1 WO 2020029031A1
Authority
WO
WIPO (PCT)
Prior art keywords
signal
useful
sampling
slope
sampling signal
Prior art date
Application number
PCT/CN2018/099027
Other languages
English (en)
French (fr)
Inventor
王超
谭曾
麻正宇
Original Assignee
高维度（深圳）生物信息智能应用有限公司
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 高维度（深圳）生物信息智能应用有限公司 filed Critical 高维度（深圳）生物信息智能应用有限公司
Priority to PCT/CN2018/099027 priority Critical patent/WO2020029031A1/zh
Publication of WO2020029031A1 publication Critical patent/WO2020029031A1/zh

Links

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition

Definitions

  • the present invention relates to signal processing, and more particularly, to a signal processing method, system, and computer storage medium.
  • the processing of human-machine biological signals usually adopts modules such as applicable single-chip microcomputers, each input point is calculated, and a primary processing module is used deal with. Each point needs to be calculated to see if it needs to meet the overall needs. In this way, in order to increase the complexity of the algorithm, a large amount of calculation is required, and too much calculation is difficult to complete on a too small MCU.
  • a window energy calculation method is used, that is, the waveform is integrated, and in other processing processes, the peaks and troughs of the signals are judged and then numerically calculated.
  • the amplitude of two consecutive peaks and troughs of the waveform of the biological signal generated by the same action under different people or different contact intensities or different exercise action intensities is different.
  • the value is very large, and the traditional processing method will cause a problem that the biological signal cannot be identified.
  • the technical problem to be solved by the present invention is to provide a signal processing method, a system, and a computer storage medium aiming at the defects in the foregoing signal processing process of the prior art.
  • the technical solution adopted by the present invention to solve its technical problems is to construct a signal processing method, including the following steps:
  • step S4 if the operation result is available, the method further includes:
  • the obtaining a useful sampling signal corresponding to the useful signal through the slope change includes:
  • S3-2A Obtain a first time point when the slope changes from zero at the time point, and a second time point when the zero point passes during the slope change process thereafter;
  • S3-3A Calculate a middle point between the first time point and the second time point, and obtain a first sampling signal within a first specific time period t 1 before the middle point, and a second specific time after the middle point. A second sampling signal within a time period t 2 ; the first sampling signal and the second sampling signal are combined to form a cycle useful sampling signal.
  • the second specific time period t 2 is greater than or equal to three times the first specific time period t 1 .
  • step S3-3A in particular when the length t 1 of the first difference of said intermediate point and the first point in time.
  • step S3 before the obtaining a useful sampling signal corresponding to the useful signal, further performing:
  • the historical data includes operation results corresponding to all useful sampling signals before the any useful sampling signal.
  • the pre-processing the original signal includes: filtering and normalizing the original signal.
  • the invention also constructs a signal processing system, including: a processor, a memory,
  • the memory is used to store program instructions
  • the processor is configured to execute the steps of any one of the methods according to the program instructions stored in the memory.
  • the present invention also constructs a computer-readable storage medium on which a computer program is stored, and the computer program, when executed by a processor, implements the steps of any of the methods described above.
  • the implementation of the signal processing method, system and computer storage medium of the present invention has the following beneficial effects: the processing process is simple and less resources are required.
  • FIG. 1 is a program flowchart of a first embodiment of a signal processing method according to the present invention
  • FIG. 2 is a schematic diagram of an original signal
  • FIG. 3 is a schematic diagram of comparison between an original signal and a first signal
  • FIG. 4 is a program flowchart of a second embodiment of a signal processing method according to the present invention.
  • the original signal sent by the signal extraction unit and pre-process the original signal to obtain a first signal containing a useful signal.
  • the obtained original signal includes the normal human body in abnormal conditions. Signals generated in the action mode that do not require analysis can be understood as useless signals that do not contain useful information.
  • the signals generated in the abnormal action mode can be filtered out through preprocessing actions. Retaining biological signals generated under normal human movements can be understood as useful signals with useful information, forming a first signal that does not include useless signals.
  • the original signals or biological signals herein may include muscle current waveforms, EEG waveforms, ECG waveforms, etc. of various parts of the human body.
  • sampling rate is controlled to ensure that no useful information is lost during the adoption process. At the same time, it avoids consuming too many resources during the sampling process and increases the processing load in the signal processing process.
  • the type of action is different, and the ratio of the peak or trough of the electrooculogram signal is different.
  • the sampling rate is generally based on the normal action speed, and the frequency range of the signal under a normal action is set to a legal frequency range. When the user If the motion is too slow or too fast, beyond this range, it will be filtered as invalid signal.
  • a constant signal can be understood as a signal whose amplitude changes within an allowable range or a small amplitude change for a period of time, and is not limited to being absolutely constant. This is mainly considering that the resistivity of each person is different, and the contact position may be different. In this way, even if there is no action, the signal obtained from the user is different.
  • the constant level when the user has no action can be taken as the constant signal. For example, it is common to take 5s and average the signal level within 5s.
  • the period of a useful signal is relatively fixed. In order to minimize the influence of clutter, this constant signal can be set as a reference signal, and then the slope of each sampled signal and the reference signal can be calculated, and the change of each slope can be calculated. Change to determine a complete useful signal period, that is, it can correspond to a useful sampling signal in a complete period.
  • any useful sampling signal in turn according to the time sequence, and combine any useful sampling signal with historical data to output an operation result corresponding to any useful sampling signal and determine whether the operation result is available.
  • a historical database is preset, and the useful sampling signals are sequentially processed in chronological order. For example, when the first useful sampling signal is processed, the useful sampling signal is combined with the data in the historical database to generate a set of input data. Then calculate the input data and get the output result after calculation.
  • the running result is to score the degree of matching between the input signal and the possible output signal.
  • the output function will give a score that matches each output signal.
  • the score range can be set to any decimal number between 0 and 1.
  • the input parameter is considered to be the target signal when the matching score reaches 0.6 or more. If the scores of several output products are greater than 0.6, the result with the highest score is selected as the recognition result of the current input data.
  • the output is: the current input data, the degree of matching with the standard centralized type data. When the matching degree is 1, it means that the exact matches are the same.
  • the recognition system of this neural network is a scoring system. Is to calculate the similarity between the input data and various standard data.
  • the waveform of the biological signal generated by the human body is in a complex background waveform, and the shape of the waveform of the biological signal changes greatly, and the corresponding amplitude value changes greatly.
  • the waveform of a biological signal contains both peaks and troughs, the actual measured waveform of the same biological signal is collected multiple times or continuously for a period of time, and the absolute value of the peak and trough amplitudes will vary greatly.
  • the output types can be matched here, and samples that can be identified as such products, such as pictures, arrays, and other data, and products that cannot be identified as an output at all can be included in the historical data. .
  • the useful signal can be identified well, and the judgment result is closer to the judgment result of the human being on the signal.
  • step S4 if an operation result is available, the method further includes: S5: merging the operation result corresponding to any useful sampling signal with historical data to form new historical data. Specifically, in the signal processing process, after obtaining an available operation result, the operation result and its corresponding useful sampling signal are merged with historical data to form new historical data. In this way, the next useful sampling is performed in chronological order. When processing the information of a signal, it is possible to perform a combined calculation based on the new historical data to obtain a more accurate calculation result. It can be understood here that the combination and calculation process is cyclic until all useful sampling signal processing in one signal cycle is completed.
  • step S3 obtaining the useful sampling signal corresponding to the useful signal through the slope change includes:
  • This point in time is the approximate starting point of the useful sampling signal.
  • the slope change can be increased from zero, or it can be reduced from zero to a negative value.
  • the useful signal is a peak
  • its slope change is increased
  • the useful signal is a valley
  • its slope change is decreased.
  • the approximate starting point of the useful sampling signal has passed, the slope of the slope due to the existence of the peak or trough will then appear as the second point in time at which the slope passes zero, which can also be understood as the end point of a peak or trough.
  • the middle point between the two time points when the time point is obtained and the middle point between the two time points is obtained, it can be understood that the middle point of the peak or trough or an approximate middle point is obtained, and the middle point is extended forward by a specific time t 1 , And extend backward another specific time period t 2 to obtain the sampling signals between two specific time periods, then all the sampling signals between the two specific time periods can be defined as useful sampling signals, and the sum of the two specific time periods is A cycle of useful signals.
  • the second specific time period t 2 is equal to three times the first specific time period t 1 .
  • the second specific time period t 2 may be set to be greater than or equal to three times the first specific time period t 1 according to a general signal waveform rule.
  • the first specific time period t 1 is the difference between the middle point and the first time point.
  • the difference between the intermediate point and the first time point can usually be defined as a half period of a peak or trough.
  • the specific point t 1 when extended forward through the intermediate point, it can be understood as It is only necessary to extend the half cycle of the crest or trough to the starting point of the crest or trough. Then it can be extended backward for three specific time periods t 1 to form a complete useful signal period.
  • step S3 before acquiring a useful sampling signal corresponding to the useful signal, performing:
  • the period of the useful signal is relatively fixed. In order to minimize the effect of clutter, you can determine whether the signal has reached the peak or trough by changing the slope. For example, when the slope is maximum, it can be determined as the peak of the signal. When the slope is minimum, it can be determined as the valley of the signal. After obtaining the peak or valley of the signal, the difference between the peak or valley amplitude value and the reference signal is determined. Whether this is a peak of a useful signal or a trough of a useful signal. In addition, when multiple peaks or valleys are determined, the difference between the peaks or valleys and the reference signal may be used to determine whether the signals of the multiple peaks or valleys are useful signals of the same type or useful signals of different types.
  • the useful signal A and the useful signal B are useful signals generated by two types of eye movements. It can be seen that due to different reasons, the signals are different.
  • the acquired signals may be The useful sampling signals in the data period of the useful signal A and the useful signal B are identified by the network to obtain the similarity between the useful sampling signal and the two standard signals. Then, the positive samples of the two standard types of signals that are relatively standard, that is, the sampled signals that can be identified as the corresponding relationship, are used for recognition training through network recognition. At the same time, it is also possible to use signals for constraint training that are not two types of signals at all, that is, they cannot be used for correspondence.
  • the historical data includes operation results corresponding to all useful sampling signals before any useful sampling signal. Specifically, when the calculation result of the useful sampling signal is calculated based on historical data, the calculation of any useful sampling signal is based on the calculation result of the previous useful sampling signal. It can also be understood here that In the calculation process of any useful sampling signal, the historical data used in it includes the operation results of all the useful sampling signals before it.
  • preprocessing the original signal includes: filtering and normalizing the original signal. Specifically, the processing of the obtained original signal includes filtering and normalization. After filtering and normalization, the signals that are obviously in the abnormal operation mode can be filtered out.
  • a signal processing system of the present invention includes a processor, a memory, and the memory are configured to store program instructions, and the processor is configured to execute the steps of any of the foregoing methods according to the program instructions stored in the memory.
  • the signal processing system here includes, but is not limited to, a computer and the like.
  • a computer-readable storage medium of the present invention stores a computer program thereon, and the computer program, when executed by a processor, implements the steps of any of the methods described above.
  • the computer-readable storage medium herein may be a tangible device that can hold and store instructions used by the instruction execution device.
  • the computer-readable storage medium may be, for example, but not limited to, an electric storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing.

Abstract

一种信号处理方法、系统及计算机存储介质，包括：接收信号提取单元发送的原始信号并对原始信号进行预处理以获取包含有用信号的第一信号（S1）；以一预设采样速率对第一信号进行采样，以获取若干采样信号（S2）；获取第一信号中持续一预设时长的恒定信号，以恒定信号为基准信号监测采样信号的斜率，并通过斜率变化获取有用信号对应的有用采样信号（S3）；依照时间顺序依次获取任一有用采样信号，并将任一有用采样信号与一历史数据进行合并运算，输出与任一有用采样信号对应的运算结果并判断运算结果是否可用（S4）。

Description

一种信号处理方法、系统及计算机存储介质 技术领域
本发明涉及信号处理，更具体地说，涉及一种信号处理方法、系统及计算机存储介质。
 背景技术
现有的信号处理过程中，尤其是在利用生物学的人机交互系统中，对人机生物学信号的处理通常采用适用单片机等模块，对每一个输入点进行计算，采用一级处理模块进行处理。对每一个点都需要拿来进行计算，看是否需满足整体需求。这样在提高算法复杂度下，需要很大运算量，而过多的运算量在太小的MCU上难以完成。还有一些信号处理过程中，采用窗口能量计算方法，也就是对波形进行积分，还有一些处理过程中，对信号的波峰波谷进行判断后采用数值计算方式进行计算。
上面描述的这些方法中，在对生物信号处理过程中，对同一个动作，在不同人或不同接触强度或不同运动动作强度下，产生的生物信号的波形的连续两个波峰和波谷的幅度相差数值很大，采用传统的处理方式，会出现不能识别生物信号的问题。
 技术问题
本发明要解决的技术问题在于，针对现有技术的上述信号处理过程中的缺陷，提供一种信号处理方法、系统及计算机存储介质。
 技术解决方案
本发明解决其技术问题所采用的技术方案是：构造一种信号处理方法，包括以下步骤：
S1、接收信号提取单元发送的原始信号并对所述原始信号进行预处理以获取包含有用信号的第一信号；
S2、以一预设采样速率对所述第一信号进行采样，以获取若干采样信号；
S3、获取所述第一信号中持续一预设时长的恒定信号，以所述恒定信号为基准信号监测所述采样信号的斜率，并通过所述斜率变化获取所述有用信号对应的有用采样信号；
S4、依照时间顺序依次获取任一有用采样信号，并将所述任一有用采样信号与一历史数据进行合并运算，输出与所述任一有用采样信号对应的运算结果并判断所述运算结果是否可用。
优选地，在所述步骤S4中，若所述运算结果可用，所述方法还包括：
S5、将所述任一有用采样信号对应的运算结果同所述历史数据合并，形成新的历史数据。
优选地，在所述步骤S3中，所述通过所述斜率变化获取所述有用信号对应的有用采样信号包括：
S3-1A、获取所述斜率从零开始变化过程中，所述斜率为零时的采样信号对应的时间点；
S3-2A、获取所述时间点中，其中所述斜率从零开始变化时的第一时间点，及其后所述斜率变化过程中经过零点时的第二时间点；
S3-3A、计算所述第一时间点和所述第二时间点的中间点，获取所述中间点之前第一特定时长t 1内的第一采样信号，及所述中间点之后第二特定时长t 2内的第二采样信号；所述第一采样信号和所述第二采样信号合并形成一个周期的有用采样信号。
优选地，所述第二特定时长t 2大于或等于所述第一特定时长t 1的三倍。
优选地，在所述步骤S3-3A中，所述第一特定时长t 1为所述中间点与所述第一时间点的差值。
优选地，在所述步骤S3中，在所述获取所述有用信号对应的有用采样信号之前；还执行：
S3-1B、获取所述斜率从零开始变化过程中，所述斜率为最大值时对应的第三采样信号，和所述斜率为最小值时对应的第四采样信号；
S3-2B、根据所述第三采样信号和所述第四采样信号与所述基准信号的差值确认所述第一信号中包含所述有用信号。
优选地，在所述步骤S4中，所述历史数据包括所述任一有用采样信号之前的所有有用采样信号对应的运算结果。
优选地，在所述步骤S1中，所述对所述原始信号进行预处理包括：对所述原始信号进行滤波和归一化。
本发明还构造一种信号处理系统，包括：处理器、存储器，
所述存储器，用于存储程序指令，
所述处理器，用于根据所述存储器所存储的程序指令执行上面任一所述方法的步骤。
本发明还构造一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现上面任意所述方法的步骤。
 有益效果
实施本发明的一种信号处理方法、系统及计算机存储介质，具有以下有益效果：处理过程简单，需要资源少。
 附图说明
下面将结合附图及实施例对本发明作进一步说明，附图中：
图1是本发明一种信号处理方法第一实施例的程序流程图；
图2为一原始信号示意图；
图3为原始信号与第一信号对比示意图；
图4是本发明一种信号处理方法第二实施例的程序流程图。
 本发明的实施方式
为了对本发明的技术特征、目的和效果有更加清楚的理解，现对照附图详细说明本发明的具体实施方式。
如图1所示，在本发明的一种信号处理方法第一实施例中，包括以下步骤：
S1、接收信号提取单元发送的原始信号并对原始信号进行预处理以获取包含有用信号的第一信号；具体的，如图2所示，在获取到原始信号中，包含了平时人体在非正常动作模式下产生的、不需进行解析的信号，既可以理解为不包含有用信息的无用信号，如图3所示，可以通过预处理动作将该非正常动作模式下产生的信号进行滤除，保留其中与人体正常动作下产生的生物学信号即可以理解为带有有用信息的有用信号，形成不包含无用信号的第一信号。还可以理解，这里的原始信号或者生物学信号可以包括人体各个部位的肌肉电流波形、脑电波形、心电波形等等。
S2、以一预设采样速率对第一信号进行采样，以获取若干采样信号；具体的，对第一信号进行采样，获取满足预设规律的若干采样信号，这里的预设规律可以通过预设采样速率来控制，保证采用过程中不会出现有用信息的丢失，同时也尽量避免采样过程中占用过多的资源，增加信号处理过程中的处理负担。例如在具体过程中，动作类型不同，眼电信号的波峰或波谷的比值不同，这里采样速率一般根据正常的动作速度，设定一个正常动作下面的信号的频率范围为合法的频率范围，当用户的动作太慢或者太快，超过这个范围，都会被当做无效信号滤除。
S3、获取第一信号中持续一预设时长的恒定信号，以恒定信号为基准信号监测采样信号的斜率，并通过斜率变化获取有用信号对应的有用采样信号；具体的，对第一信号进行一段时间的平均数值计算，识别出第一信号中在没有任何有用信号，也没有任何杂波等干扰情况下，信号的幅度数值。这里可以理解该幅度数值应该是一个恒定数值，在以时间为Ｘ轴，采样点信号幅度数值为Ｙ轴的波形图上，这里将恒定数值定义为恒定信号。这里恒定信号可以理解为持续一段时间的幅度在允许范围内变化或者幅度变化很小的信号，而不是局限于绝对恒定。这里主要是考虑到每个人的电阻率不同，接触位置也可能不同，这样即使没有任何动作，从用户获取的信号不一样，这里可以取用户没有任何动作的时候的持续电平为该恒定信号，例如通用的可以取5s，对5s内的信号电平取平均。通常的有用信号的周期比较固定，为了最大化减少杂波的影响，可以将该恒定信号设定为基准信号，然后通过计算各个采样信号与基准信号的斜率，并计算各个斜率的变化，通过斜率的变化来判断一个完整的有用信号周期，即可以对应到一个完整周期中的有用采样信号。
S4、依照时间顺序依次获取任一有用采样信号，并将任一有用采样信号与一历史数据进行合并运算，输出与任一有用采样信号对应的运算结果并判断运算结果是否可用。具体的，预设一个历史数据库，按照时间顺序依次处理有用采样信号，例如当对第一有用采样信号进行处理时，将该有用采样信号与历史数据库中的数据进行合并，生成一组输入数据，然后对该输入数据进行计算，经过计算后得到输出结果。运行结果是对输入信号与可能的输出信号的匹配程度进行打分，输出函数会给出与每一种输出信号匹配的分数，分数范围可以设置为0~1之间的任意小数。我们可以设置，当匹配分数达到0.6以上才认为输入的参数是目标信号，如果输出的几种产品分数都大于0.6，则选择分数最大的那一个结果来作为当前输入数据的识别结果。输出结果是：当前输入的数据，与标准的集中类型数据的匹配度。匹配度为１时，代表完全匹配相同，这个神经网络的识别系统就是一个打分系统。就是会计算出输入数据和各种标准数据之间的相似度数值。
这里可以理解，由于接触差异等等外界因素，导致人体产生的生物信号的波形处在复杂背景波形中，并且该生物信号的波形形状变化很大，相应的幅度数值变化也大。尤其当生物信号的波形同时包含波峰和波谷时，实际测试到的同一个生物信号的波形，在多次采集或者在一段时间的持续采集，获得其波峰和波谷的幅度绝对值差异变化会很大，这里采用传统积分等能量的数值精确计算方式很难准确的识别出有用信号波形。而在这里，这里可以将输出类型很匹配的，能够被识别成这类产品的样本，例如图片，数组等等数据，以及完全无法被识别成某种输出的产品，都可以包含在历史数据中。通过斜率判断，同时加上与历史数据的结合，就可以很好的对有用信号的识别，使判别结果更接近人类自己对信号的判断结果。
进一步的，在所述步骤S4中，若运算结果可用，所述方法还包括：S5、将任一有用采样信号对应的运算结果同历史数据合并，形成新的历史数据。具体的，在信号处理过程中，获取可用的运算结果后，会将该运算结果及其对应的有用采样信号同历史数据合并，形成新的历史数据，这样，在依照时间顺序进行下一个有用采样信号的信息处理时，可以基于新的历史数据进行合并计算，以获取更加准确地计算结果，这里可以理解，合并和计算过程是循环的，直到一个信号周期内的所有有用采样信号处理完成。
进一步的，在步骤S3中，通过斜率变化获取有用信号对应的有用采样信号包括：
S3-1A、获取斜率从零开始变化过程中，斜率为零时的采样信号对应的时间点；具体的，当以基准信号为参考，计算采样信号的斜率时，当没有有用采样信号时，可以计算得到的斜率几乎不变，当开始为有用采样信号时，其斜率开始变化，当在有用采样信号的周期内，会出现多次斜率为零的状态。记录斜率变化过程中，斜率为零时的时间点。
S3-2A、获取时间点中，其中斜率从零开始变化时的第一时间点，及其后斜率变化过程中经过零点时的第二时间点；具体的，斜率为零时的时间点，当这个时间点出现在斜率刚开始变化的过程中，这个时间点即为有用采样信号的近似起始点，这里的斜率变化可以是从零开始增加，也可以是从零开始减小到负值，例如，当有用信号是波峰时，其斜率变化是增加，当有用信号是波谷时，其斜率变化是减小。当经过了有用采样信号的近似起始点，其斜率变化中由于波峰或者波谷的存在，会接着出现经过斜率为零的时间点即第二时间点，也可以理解为一个波峰或者波谷的结束点。
S3-3A、计算第一时间点和第二时间点的中间点，获取中间点之前第一特定时长t 1内的第一采样信号，及中间点之后第二特定时长t 2内的第二采样信号；第一采样信号和第二采样信号合并形成一个周期的有用采样信号。具体的，在获取了时间点，求两个时间点的中间点，则可以理解为获取了该波峰或者波谷的中间点或者为近似的中间点，对该中间点向前扩展一个特定时长t 1，并向后扩展另一个特定时长t 2，获取两个特定时长之间的采样信号，那么可以将两个特定时长之间的所有采样信号定义为有用采样信号，两个特定时长的和即为一个有用信号的周期。
进一步的，第二特定时长t 2或等于所述第一特定时长t 1的三倍。具体的，为了使整个有用信号包含在上面获得一个有用信号周期内，可以按照通用的信号波形规律，将第二特定时长t 2设置为大于或等于第一特定时长t 1的三倍。
进一步，在步骤S3-3A中，第一特定时长t 1为中间点与第一时间点的差值。具体的，在理想的情况下，通常可以将中间点与第一时间点的差值定义为一个波峰或波谷的半周期，那么在通过中间点往前扩展一特定时长t 1时，可以理解为只需要扩展该波峰或波谷的半周期，扩展到波峰或波谷的起始点即可。那么其向后扩展三个特定时长t 1即可构成一个完整的有用信号周期。
进一步的，在步骤S3中，在获取有用信号对应的有用采样信号之前；还执行：
S3-1B、获取斜率从零开始变化过程中，斜率为最大值时对应的第三采样信号，和斜率为最小值时对应的第四采样信号； S3-2B、根据第三采样信号和第四采样信号与基准信号的差值确认第一信号中包含有用信号。
具体的，由于有用信号的周期比较固定。为了最大化减少杂波的影响，可以通过斜率的变化来判断信号是否达到波峰或者波谷。例如，当斜率最大时，则可以判定为信号的波峰，当斜率最小时，可以判定为信号的波谷，在获取到信号的波峰或者波谷后，通过该波峰或波谷幅度数值与基准信号的差值这是否是一个有用信号的波峰或者一个有用信号的波谷。此外这里，当判断出多个波峰或者波谷时，还可以通过波峰或者波谷与基准信号的差值来判断这多个波峰或者波谷的信号是同一类别有用信号还是不同类别的有用信号。如图2所示有用信号Ａ和有用信号Ｂ分别为两种眼睛动作生成的有用信号，可以看出其由于产生的原因不同，信号存在差异，在信号处理过程中，需要把获取到的可能是有用信号Ａ和有用信号Ｂ的数据周期内的有用采样信号，进行网络识别，以此来获取该有用采样信号与两种标准信号的相似程度。然后通过网络识别分别将比较标准的两类信号的正样本即能够被识别为对应关系的采样信号进行识别训练。同时也可以将完全不是两类信号即不能被是被为对应关系的采用信号进行约束训练的．
进一步的，在步骤S4中，历史数据包括任一有用采样信号之前的所有有用采样信号对应的运算结果。具体的，在对有用采样信号的运算结果进行计算时，其是基于历史数据进行合并计算，那么任意一个有用采样信号的运算都是基于其前一个有用采样信号的运算结果，这里也可以理解，任一有用采样信号计算过程中，其采用的历史数据是包含了其前面的所有有用采样信号的运算结果的。
进一步的，在步骤S1中，对原始信号进行预处理包括：对原始信号进行滤波和归一化。具体的，对获取的原始信号的处理包括滤波和归一化，通过滤波和归一化之后，可以滤除其中的明显属于非正常动作模式下的信号。
另，本发明的一种信号处理系统，包括：处理器、存储器，存储器用于存储程序指令，处理器用于根据存储器所存储的程序指令执行上面任一所述方法的步骤。这里信号处理系统包括但不局限于计算机等。
另，本发明的一种计算机可读存储介质，其上存储有计算机程序，该计算机程序被处理器执行时实现上面任意所述方法的步骤。这里计算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。计算机可读存储介质例如可以是但不限于电存储设备、磁存储设备、光存储设备、电磁存储设备、半导体存储设备或者上述的任意合适的组合。
  可以理解的，以上实施例仅表达了本发明的优选实施方式，其描述较为具体和详细，但并不能因此而理解为对本发明专利范围的限制；应当指出的是，对于本领域的普通技术人员来说，在不脱离本发明构思的前提下，可以对上述技术特点进行自由组合，还可以做出若干变形和改进，这些都属于本发明的保护范围；因此，凡跟本发明权利要求范围所做的等同变换与修饰，均应属于本发明权利要求的涵盖范围。

Claims (9)

  1. 一种信号处理方法，其特征在于，包括以下步骤：
    S1、接收信号提取单元发送的原始信号并对所述原始信号进行预处理以获取包含有用信号的第一信号；
    S2、以一预设采样速率对所述第一信号进行采样，以获取若干采样信号；
    S3、获取所述第一信号中持续一预设时长的恒定信号，以所述恒定信号为基准信号监测所述采样信号的斜率，并通过所述斜率变化获取所述有用信号对应的有用采样信号；
    S4、依照时间顺序依次获取任一有用采样信号，并将所述任一有用采样信号与一历史数据进行合并运算，输出与所述任一有用采样信号对应的运算结果并判断所述运算结果是否可用。
  2. 根据权利要求1所述的信号处理方法，其特征在于，在所述步骤S4中，若所述运算结果可用，所述方法还包括：
    S5、将所述任一有用采样信号对应的运算结果同所述历史数据合并，形成新的历史数据。
  3. 根据权利要求1所述的信号处理方法，其特征在于，在所述步骤S3中，所述通过所述斜率变化获取所述有用信号对应的有用采样信号包括：
    S3-1A、获取所述斜率从零开始变化过程中，所述斜率为零时的采样信号对应的时间点；
    S3-2A、获取所述时间点中，其中所述斜率从零开始变化时的第一时间点，及其后所述斜率变化过程中经过零点时的第二时间点；
    S3-3A、计算所述第一时间点和所述第二时间点的中间点，获取所述中间点之前第一特定时长t 1内的第一采样信号，及所述中间点之后第二特定时长t 2内的第二采样信号；所述第一采样信号和所述第二采样信号合并形成一个周期的有用采样信号。
  4. 根据权利要求3所述的信号处理方法，其特征在于，所述第二特定时长t 2大于或等于所述第一特定时长t 1的三倍。
  5. 根据权利要求4所述的信号处理方法，其特征在于，在所述步骤S3-3A中，所述第一特定时长t 1为所述中间点与所述第一时间点的差值。
  6. 根据权利要求1所述的信号处理方法，其特征在于，在所述步骤S3中，在所述获取所述有用信号对应的有用采样信号之前；还执行：
    S3-1B、获取所述斜率从零开始变化过程中，所述斜率为最大值时对应的第三采样信号，和所述斜率为最小值时对应的第四采样信号；
    S3-2B、根据所述第三采样信号和所述第四采样信号与所述基准信号的差值确认所述第一信号中包含所述有用信号。
  7. 根据权利要求1所述的信号处理方法，其特征在于，在所述步骤S4中，所述历史数据包括所述任一有用采样信号之前的所有有用采样信号对应的运算结果。
  8. 8、根据权利要求1所述的信号处理方法，其特征在于，在所述步骤S1中，所述对所述原始信号进行预处理包括：对所述原始信号进行滤波和归一化。
      9.一种信号处理系统，其特征在于，包括：处理器、存储器，
    所述存储器，用于存储程序指令，
    所述处理器，用于根据所述存储器所存储的程序指令执行权利要求1-8中任意一项所述方法的步骤。
  9. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现权利要求1-8任意一项所述方法的步骤。
PCT/CN2018/099027 2018-08-06 2018-08-06 一种信号处理方法、系统及计算机存储介质 WO2020029031A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/099027 WO2020029031A1 (zh) 2018-08-06 2018-08-06 一种信号处理方法、系统及计算机存储介质

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/099027 WO2020029031A1 (zh) 2018-08-06 2018-08-06 一种信号处理方法、系统及计算机存储介质

Publications (1)

Publication Number Publication Date
WO2020029031A1 true WO2020029031A1 (zh) 2020-02-13

Family

ID=69413923

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/099027 WO2020029031A1 (zh) 2018-08-06 2018-08-06 一种信号处理方法、系统及计算机存储介质

Country Status (1)

Country Link
WO (1) WO2020029031A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102613968A (zh) * 2011-12-31 2012-08-01 深圳市邦健电子有限公司 一种极化电压检测方法和设备
US8891713B2 (en) * 2011-04-06 2014-11-18 Siemens Medical Solutions Usa, Inc. System for adaptive sampled medical signal interpolative reconstruction for use in patient monitoring
CN107080526A (zh) * 2016-02-15 2017-08-22 三星电子株式会社 生物信号处理方法和设备
CN107101984A (zh) * 2017-05-18 2017-08-29 广东顺德工业设计研究院(广东顺德创新设计研究院) 信号波形特征检测方法、装置、存储介质和计算机设备

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8891713B2 (en) * 2011-04-06 2014-11-18 Siemens Medical Solutions Usa, Inc. System for adaptive sampled medical signal interpolative reconstruction for use in patient monitoring
CN102613968A (zh) * 2011-12-31 2012-08-01 深圳市邦健电子有限公司 一种极化电压检测方法和设备
CN107080526A (zh) * 2016-02-15 2017-08-22 三星电子株式会社 生物信号处理方法和设备
CN107101984A (zh) * 2017-05-18 2017-08-29 广东顺德工业设计研究院(广东顺德创新设计研究院) 信号波形特征检测方法、装置、存储介质和计算机设备

Similar Documents

Publication Publication Date Title
Gao et al. An effective LSTM recurrent network to detect arrhythmia on imbalanced ECG dataset
CN109948647B (zh) 一种基于深度残差网络的心电图分类方法及系统
Al-Fahoum et al. Methods of EEG signal features extraction using linear analysis in frequency and time-frequency domains
CN109117730B (zh) 心电图心房颤动实时判断方法、装置、系统及存储介质
Wang et al. Arrhythmia classification algorithm based on multi-head self-attention mechanism
Udhayakumar et al. Approximate entropy profile: a novel approach to comprehend irregularity of short-term HRV signal
WO2019161611A1 (zh) 心电信息处理方法和心电工作站
Guendil et al. Emotion recognition from physiological signals using fusion of wavelet based features
CN116089851A (zh) 一种自适应更新的在线运动单位识别方法
CN109117729A (zh) 心电图室性逸搏实时判断方法、装置、系统及存储介质
CN109063652B (zh) 一种信号处理方法、系统及计算机存储介质
CN115563484A (zh) 一种基于生理唤醒识别的街道绿化品质检测方法
CN116712056B (zh) 心电图数据的特征图像生成与识别方法、设备及存储介质
CN110491504B (zh) 一种心音信号医学指标数据的获取方法
CN108338777A (zh) 一种脉搏信号检测分析方法及装置
WO2020029031A1 (zh) 一种信号处理方法、系统及计算机存储介质
Poddar et al. Heart rate variability: Analysis and classification of healthy subjects for different age groups
CN106778561B (zh) 一种穿戴式设备的身份识别方法及识别装置
Yu et al. ECG identification based on PCA-RPROP
CN109325402B (zh) 一种信号处理方法、系统及计算机存储介质
Wang et al. A comparative study on sign recognition using sEMG and inertial sensors
WO2020029032A1 (zh) 一种信号处理方法、系统及计算机存储介质
CN115177260A (zh) 基于人工神经网络的智能心电心音诊断方法及装置
Lata et al. Disease classification using ECG signals based on R-peak analysis with ABC and ANN
CN114861706A (zh) 一种基于质量评估和深度迁移学习的心电身份识别方法

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: 18929493

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS (EPO FORM 1205A DATED 17.06.2021)

122 Ep: pct application non-entry in european phase

Ref document number: 18929493

Country of ref document: EP

Kind code of ref document: A1