CN101773392B - Self-adaptive and high-efficiency storage method for dynamic ECG data - Google Patents

Abstract

The invention relates to a self-adaptive high-efficiency storage method for dynamic electrocardiogram data, and provides an electrocardiogram data storage scheme based on extraction of an electrocardiogram main wave position, which can self-adaptively select a data storage mode according to heart rate and electrocardiogram waveform characteristic information. When the heart rate is large (namely the interval between the pre-stored dominant peak point of the cardiac cycle and the dominant peak point of the previous cardiac cycle is smaller than the defined data length of the cardiac cycle), the divided front and rear electrocardiographic waveforms may overlap, the repeatability of the electrocardiographic waveforms is damaged, and all stored data are selected for better recovering the waveforms before storage; when the interval between the dominant wave maximum point of the pre-stored cardiac cycle and the dominant wave maximum point of the previous cardiac cycle is not less than the defined cardiac cycle data length, a corresponding template is established according to the condition of the electrocardiographic waveform, and the same waveform is replaced by a template code number when stored, so that the data storage amount is greatly reduced, the data storage efficiency is improved, and the good waveform reliability is maintained.

Description

Translated fromChinese

动态心电数据自适应高效存储方法Self-adaptive and high-efficiency storage method for dynamic ECG data

【技术领域】【Technical field】

本发明涉及一种动态心电数据存储方法，属于生物医学工程中医学数据存储领域。The invention relates to a dynamic electrocardiographic data storage method, which belongs to the field of medical data storage in biomedical engineering.

【背景技术】【Background technique】

动态心电图(ambulatory electrocardiograph，AECG)是指长时间连续记录的体表心电图，1961年由美国Norman J.Holter发明问世，迄今临床中仍广泛将其称为“Holter”。动态心电图采集的大量心电数据信息通过记录设备进行储存，通过回放分析系统进行心电图形的浏览与分析，用于科学研究或新产品的开发。随着生活水平的提高，科学技术的发展，对心脏疾病的关注程度日益提高，家庭式的远程心电监护产品因可以实时上传心电数据供远程终端分析判断，并提供相应的救助服务而成为研究的一个焦点。能够高效的存储数据，减轻传输数据的压力，是目前研究的一个热点。从另一个角度来讲，心电波形短期时间具有相对稳定性的特点，但对某些心脏疾病而言，因疾病发展状况心电波形在一定时间内会有演变，能高效保存各个时期的心电数据是十分必要的。Ambulatory electrocardiograph (AECG) refers to the body surface electrocardiogram recorded continuously for a long time. It was invented by Norman J. Holter in the United States in 1961. It is still widely called "Holter" in clinical practice so far. A large amount of ECG data information collected by dynamic electrocardiogram is stored by recording equipment, and the ECG graphics are browsed and analyzed through the playback analysis system, which is used for scientific research or new product development. With the improvement of living standards and the development of science and technology, the degree of concern for heart disease is increasing. The family-style remote ECG monitoring product has become a popular choice because it can upload ECG data in real time for remote terminal analysis and judgment, and provide corresponding rescue services. A focus of research. Being able to efficiently store data and reduce the pressure of data transmission is a hot spot in current research. From another point of view, the ECG waveform is relatively stable in the short term, but for some heart diseases, the ECG waveform will evolve within a certain period of time due to the development of the disease, and it can efficiently preserve the ECG in each period. Electrical data is very necessary.

【发明内容】【Content of invention】

本发明基于心电波形的特点，提供一种动态心电数据自适应高效存储方法。Based on the characteristics of electrocardiographic waveforms, the present invention provides an adaptive and efficient storage method for dynamic electrocardiographic data.

上述目的由以下技术方案实现：Above-mentioned purpose is realized by following technical scheme:

一种动态心电数据自适应高效存储方法，包括如下步骤：A method for self-adaptive and efficient storage of dynamic ECG data, comprising the steps of:

(1)采集一定长度的心电数据片段；(1) Collect a certain length of ECG data segment;

(2)对所述心电数据片段进行特征点的提取，特征点至少包括每个心动周期QRS波群中主波极大值点【P_R(i)，V_R(i)】，P_R(i)为位置，V_R(i)为对应的电压幅值；定义每个心动周期数据长度为K秒，数据长度K取0.60S，且以QRS波群中主波极大值点P_R(i)为基准，P_R(i)之前分配0.20S，P_R(i)之后分配0.40S；以每个主波极大值点为基准，根据定义的心动周期数据长度将心电数据片段进行划分；(2) Extracting feature points on the ECG data segment, the feature points at least include the main wave maximum point [P_R (i),_VR (i)] in each cardiac cycle QRS wave group, P_R (i) is the position,_VR (i) is the corresponding voltage amplitude; define the data length of each cardiac cycle as K seconds, the data length K is 0.60S, and the main wave maximum point P_R in the QRS wave group (i) as the benchmark, 0.20S is allocated before_PR (i), and 0.40S is allocated after_PR (i); based on each main wave maximum point, the ECG data segment is divided according to the defined cardiac cycle data length to divide;

(3)选择划分后的某一心动周期的心电数据作为第一模板，保存该第一模板的全部数据及模板代号，并存储第一模板之前的全部心电数据；(3) Select the electrocardiographic data of a certain cardiac cycle after the division as the first template, save all data and template codes of the first template, and store all electrocardiographic data before the first template;

(4)判断预存储的心动周期的主波极大值点与其在先心动周期的主波极大值点的间隔是否小于K秒，是则进行(a)，否则进行(b)：(a)除去与其在先心动周期的重叠部分保存预存储心动周期的其他数据；(b)存储预存储心动周期前、其在先心动周期后的数据，并将预存储心动周期的数据与最接近的模板进行特征点吻合判断，如果特征点吻合则存储相应模板代号，如果不吻合则将该预存储心动周期的数据建立为新的模板，并给予新的模板代号；(4) Whether the interval between the main wave maximum point of the pre-stored cardiac cycle and the main wave maximum point of the previous cardiac cycle is less than K seconds, if so, perform (a), otherwise perform (b): (a ) remove the overlapping part of its previous cardiac cycle and save other data of the pre-stored cardiac cycle; The template performs feature point matching judgment. If the feature points match, the corresponding template code is stored. If not, the pre-stored cardiac cycle data is established as a new template, and a new template code is given;

(5)重复步骤(4)；(5) Repeat step (4);

其特征在于：所述特征点还包括每个心动周期的：P波极大值点【P_P(i)，V_P(i)】，T波极大值点【P_T(i)，V_T(i)】，QRS波群极小值点【P_D(i)，V_D(i)】。It is characterized in that: the feature points also include for each cardiac cycle: P wave maximum point [P_P (i), V_P (i)], T wave maximum point [_PT (i), V_T (i)], QRS complex minimum point [P_D (i), V_D (i)].

所述特征点的提取方法具体为：①获取第i个心动周期QRS波群中主波极大值点，位置记为P_R(i)，对应的电压幅值记为V_R(i)，②在主波极大值点P_R(i)之前0.07S到0.20S数据范围内寻找电压幅值最大值点，位置记为P_P(i)，对应的电压幅值记为V_P(i)，③在主波极大值点P_R(i)之后0.07S到0.40S数据范围内寻找电压幅值最大值点，位置记为P_T(i)，对应的电压幅值记为V_T(i)，④在P_P(i)与P_T(i)之间数据范围内，寻找幅值最小值点，位置记为P_D(i)，对应的电压幅值记为V_D(i)。The method for extracting the feature points is specifically: 1. obtain the main wave maximum point in the i-th cardiac cycle QRS complex, the position is denoted as P_R (i), and the corresponding voltage amplitude is denoted as_VR (i), ②Find the maximum voltage amplitude point within the data range of 0.07S to 0.20S before the main wave maximum point P_R (i), record the position as P_P (i), and record the corresponding voltage amplitude as V_P (i ), ③ Find the maximum point of the voltage amplitude within the data range of 0.07S to 0.40S after the main wave maximum point P_R (i), record the position as P_T (i), and record the corresponding voltage amplitude as V_T (i), ④ In the data range between P_P (i) and_PT (i), look for the minimum amplitude point, the position is recorded as P_D (i), and the corresponding voltage amplitude is recorded as V_D (i ).

所述将预存储心动周期的数据与最接近的模板进行特征点吻合判断的方法，具体为：首先，以预存储心动周期数据与最接近的模板的主波极大值点为基准，判断其它特征点位置是否吻合，是则判定初步吻合，否则判定与模板不吻合；与模板初步吻合后，判断四个特征点对应的幅值是否与模板特征点幅值相对一致，若一致，则判断本心动周期波形与模板吻合；若不一致，则判断与模板不吻合。The method for judging the matching of the feature points between the pre-stored cardiac cycle data and the closest template is specifically: first, based on the pre-stored cardiac cycle data and the main wave maximum point of the closest template, judge other Whether the position of the feature points is consistent, if yes, it is judged to be a preliminary match, otherwise it is judged to be inconsistent with the template; after the initial match with the template, it is judged whether the amplitudes corresponding to the four feature points are relatively consistent with the amplitudes of the template feature points, and if they are consistent, the judgment is made. The waveform of the cardiac cycle matches the template; if not, it is judged that it does not match the template.

模板吻合判断前，对心动周期的四个特征点幅值进行矫正，矫正过程如下：Before the judgment of template matching, the amplitudes of the four characteristic points of the cardiac cycle are corrected, and the correction process is as follows:

(1)求取第一模板QRS波群主波极大值点0.10S之前八个数据幅值的平均值，并求取第一模板QRS波群主波极大值点0.10S之后八个数据幅值的平均值，再求两者平均值mean；(1) Calculate the average value of the eight data amplitudes before the maximum point of the main wave of the first template QRS complex 0.10S, and calculate the eight data after the maximum point of the main wave of the first template QRS complex 0.10S The average value of the amplitude, and then calculate the mean value of the two;

(2)求取每个心动周期P_R(i)0.10S之前相邻八个数据幅值的平均值，并求取每个心动周期P_R(i)0.10S之后相邻八个数据幅值的平均值，再求两者平均值mean(i)，以mean(i)与mean的差值作为漂移值，用每个心动周期四个特征点的幅值减去该心动周期对应的漂移值即为矫正后的特征点的幅值。(2) Calculate the average value of the eight adjacent data amplitudes before each cardiac cycle P_R (i)0.10S, and calculate the adjacent eight data amplitudes after each cardiac cycle P_R (i)0.10S Then calculate the mean(i) of the two, take the difference between mean(i) and mean as the drift value, and use the amplitude of the four feature points in each cardiac cycle to subtract the corresponding drift value of the cardiac cycle is the amplitude of the corrected feature points.

判断预存储心动周期的特征点是否与模板的特征点吻合，具体判断过程如下：Judging whether the feature points of the pre-stored cardiac cycle coincide with the feature points of the template, the specific judgment process is as follows:

(1)判断预存储心动周期中除主波极大值点外的其他特征点位置与模板中相对应特征点位置之差的绝对值是否都小于相应的特定的阈值，即|(P_P(i)-P_P(模板))-(P_R(i)-P_R(模板))|＜th1，|(P_T(i)-P_T(模板))-(P_R(i)-P_R(模板))|＜th2，且|(P_D(i)-P_D(模板))-(P_R(i)-P_R(模板))|＜th3，其中，th1、th2、th3为判断的阈值，是则判定此心动周期与模板初步吻合，否则判定此心动周期与模板不吻合；(1) Determine whether the absolute values of the differences between the positions of other feature points in the pre-stored cardiac cycle except for the maximum point of the main wave and the positions of the corresponding feature points in the template are all smaller than the corresponding specific threshold, that is |(P_P ( i)-P_P (template))-(P_R (i)-P_R (template))|<th1, |(P_T (i)-P_T (template))-(P_R (i)-P_R (template))|<th2, and |(P_D (i)-PD (template_{))-(P R(} i)-P_R (template))|<th3, where th1, th2, th3 are The threshold of judgment, if it is, it is judged that the cardiac cycle is initially consistent with the template, otherwise it is judged that the cardiac cycle is not consistent with the template;

(2)判定与模板初步吻合后，判断|(V_P(i)-V_P(模板))|*a+|(V_T(i)-V_T(模板))|*b+|(V_D(i)-V_D(模板))|*c+|(V_R(i)-V_R(模板))|是否小于某个特定的阈值th4，其中，a、b、c分别为加权系数，是则判断此心动周期与模板吻合，否则判断此心动周期与模板不吻合。(2) After judging the initial match with the template, determine |(V_P (i)-V_P (template))|*a+|(V_T (i)-V_T (template))|*b+|(V_D ( i)-V_D (template))|*c+|(V_R (i)_-VR (template))| is less than a certain threshold th4, where a, b, and c are weighting coefficients respectively, then It is judged that the cardiac cycle matches the template, otherwise it is judged that the cardiac cycle does not coincide with the template.

所述选择最接近的模板的判断方法是：以V_R(i)、V_T(i)和V_D(i)的和值作为参考，哪一模板的该三个幅值和与预存储心动周期的该三个幅值和接近，则选取该模板作为最接近的模板。The judging method for selecting the closest template is: taking the sum of_VR (i), V_T (i) and V_D (i) as a reference, which template has the same three amplitude sums as the pre-stored heart rate If the three amplitudes of the cycle are close to each other, this template is selected as the closest template.

本发明的有益效果在于：它可以根据心率以及心电波形特征信息自适应的选择数据存储方式，从而提供了一种在心电主波位置提取基础上的心电数据存储方案。具体地，当心率较大时(即预存储心动周期的主波极大值点与在先心动周期的主波极大值点的间隔小于定义的心动周期数据长度)，划分的前、后心电波形可能会发生重叠，破坏心电波形的重复性，为了更好的恢复存储前的波形，选择全部存储数据；当预存储心动周期的主波极大值点与在先心动周期的主波极大值点的间隔不小于定义的心动周期数据长度时，根据心电波形情况建立相应的模板，相同的波形存储时用模板代号代替，这样大大减少了的数据的存储量，提高了数据的存储效率，并且保持了良好的波形可信度。本发明提供的方法还便于操作，而且压缩失真率较低。The beneficial effect of the present invention is that it can self-adaptively select the data storage mode according to the heart rate and the characteristic information of the electrocardiogram waveform, thereby providing an electrocardiogram data storage scheme based on the extraction of the main wave position of the electrocardiogram. Specifically, when the heart rate is large (that is, the interval between the main wave maximum point of the pre-stored cardiac cycle and the main wave maximum point of the previous cardiac cycle is less than the defined cardiac cycle data length), the divided anterior and posterior cardiac cycle Electrical waveforms may overlap, destroying the repeatability of ECG waveforms. In order to better restore the waveforms before storage, select all stored data; when the main wave maximum point of the pre-stored cardiac cycle When the maximum point interval is not less than the defined cardiac cycle data length, the corresponding template is established according to the ECG waveform, and the same waveform is stored with the template code instead, which greatly reduces the data storage capacity and improves the data efficiency. Storage efficiency, and maintain good waveform reliability. The method provided by the invention is also convenient to operate, and the compression distortion rate is low.

【附图说明】【Description of drawings】

图1为包含若干心动周期的特定长度的心电数据片段的波形图；Fig. 1 is the waveform chart of the electrocardiographic data segment of the specific length that comprises several cardiac cycles;

图2为某心动周期四个特征点信息提取流程图；Fig. 2 is a flow chart of information extraction of four feature points of a certain cardiac cycle;

图3为某个心动周期四个特征点信息提取结果的相应波形图；Fig. 3 is the corresponding waveform diagram of the four feature point information extraction results of a certain cardiac cycle;

图4为心电数据自适应高效存储流程图。Fig. 4 is a flowchart of self-adaptive and efficient storage of ECG data.

【具体实施方式】【Detailed ways】

下面结合附图对动态心电数据自适应高效存储方法的最佳实施例进行详细说明。The best embodiment of the method for self-adaptive and high-efficiency storage of dynamic electrocardiographic data will be described in detail below in conjunction with the accompanying drawings.

参见图1，首先采集包含若干心动周期的特定长度的心电数据片段，如果存在基线漂移影响，则要对心电数据片段进行截止频率为0.05Hz的数字高通滤波，滤除基线漂移的影响。当然，如果不存在基线漂移或已经通过硬件消除基线漂移，则该步骤可以省略。把心电数据片段依照心动周期进行划分，依经验值，每个心动周期定义数据长度为0.60s，且以各个QRS波群中主波极大值点P_R(i)为基准，P_R(i)之前分配0.20S，P_R(i)之后分配0.40S。Referring to Figure 1, first collect ECG data segments of a certain length including several cardiac cycles, and if there is a baseline drift effect, perform digital high-pass filtering on the ECG data segments with a cutoff frequency of 0.05 Hz to filter out the impact of baseline drift. Of course, if there is no baseline drift or the baseline drift has been eliminated by hardware, this step can be omitted. Divide the ECG data segments according to the cardiac cycle. According to empirical values, each cardiac cycle defines a data length of 0.60s, and based on the main wave maximum point P_R (i) in each QRS complex,_PR ( 0.20S is allocated before i) and 0.40S is allocated after_PR (i).

参见图2，接下来对各个心动周期分别进行特征点的提取，以第i个心动周期为例，具体方法为：①提取第i个心动周期QRS波群中主波极大值点，位置记为P_R(i)，对应的电压幅值记为V_R(i)，②在主波极大值点P_R(i)之前0.07S到0.20S数据范围内(0.13S内)寻找电压幅值最大值点，位置记为P_P(i)，对应的电压幅值记为V_P(i)，③在主波极大值点P_R(i)之后0.07S到0.40S数据范围内(0.33S内)寻找电压幅值最大值点，位置记为P_T(i)，对应的电压幅值记为V_T(i)，④在P_P(i)与P_T(i)之间数据范围内，寻找幅值最小值点，位置记为P_D(i)，对应的电压幅值记为V_D(i)，提取的特征点如图3所示。由上可知，每个心动周期的特征点包括QRS波群中主波极大值点【P_R(i)，V_R(i)】，P波极大值点【P_P(i)，V_P(i)】，T波极大值点【P_T(i)，V_T(i)】，QRS波群极小值点【P_D(i)，V_D(i)】。Referring to Fig. 2, the feature points are extracted for each cardiac cycle respectively. Taking the i-th cardiac cycle as an example, the specific method is as follows: ① extract the maximum point of the main wave in the QRS complex of the i-th cardiac cycle, and record the position P_R (i), the corresponding voltage amplitude is recorded as_VR (i), ② Find the voltage amplitude within the data range of 0.07S to 0.20S (within 0.13S) before the main wave maximum point P_R (i) The maximum value point, the position is recorded as P_P (i), and the corresponding voltage amplitude is recorded as V_P (i), ③ within the data range of 0.07S to 0.40S after the main wave maximum point P_R (i) ( 0.33S) to find the maximum point of the voltage amplitude, the position is recorded as_PT (i), and the corresponding voltage amplitude is recorded as V_T (i), ④The data between P_P (i) and_PT (i) Within the range, find the minimum value point of the amplitude, the position is recorded as_PD (i), and the corresponding voltage amplitude is recorded as V_D (i). The extracted feature points are shown in Figure 3. It can be seen from the above that the characteristic points of each cardiac cycle include the main wave maximum point [P_R (i), V_R (i)] in the QRS complex, the P wave maximum point [P_P (i), V_P (i)], T wave maximum point [P_T (i), V_T (i)], QRS complex minimum point [P_D (i), V_D (i)].

结合图1所示，保存第2个QRS波群主波极大值点P_R(2)位置之前0.2s到之后0.4s的心电数据(即划分的第二个心动周期)作为第一模板，保存该第一模板的全部数据及模板代号，它包含有四个特征点对应的位置信息和幅值信息以及0.60S的心电幅值数据，用此0.60S的模板心电数据作为滑动窗口进行模板吻合判断(下文详述)。选择第二个心动周期作为模板的原因主要是第一个心动周期的心电数据可能不稳定，当然也可以选择其他的心动周期作为第一模板。此外，还存储第一模板之前的全部心电数据。As shown in Figure 1, save the ECG data from 0.2s before to 0.4s after the position of the second QRS complex main wave maximum point P_R (2) (that is, the second divided cardiac cycle) as the first template , save all the data and template code of the first template, it contains the position information and amplitude information corresponding to the four feature points and the 0.60S ECG amplitude data, use the 0.60S template ECG data as the sliding window Perform template matching judgment (detailed below). The reason for choosing the second cardiac cycle as the template is mainly that the electrocardiographic data of the first cardiac cycle may be unstable, and of course other cardiac cycles may also be selected as the first template. In addition, all ECG data before the first template are also stored.

第一模板之后各心动周期的心电数据将根据心率以及心电波形特征信息自适应的选择数据存储方式，结合图4及图1说明如下：After the first template, the electrocardiographic data of each cardiac cycle will be adaptively selected according to the heart rate and the characteristic information of the electrocardiographic waveform, and the description is as follows in conjunction with Fig. 4 and Fig. 1:

首先，判断第三心动周期的主波极大值点P_R(3)与第二心动周期的主波极大值点P_R(2)的间隔是否小于0.60秒，如果小于0.60秒(即当前心率大于100次)，说明按照划分的前后心电波形发生重叠，破坏心电波形的重复性，此时不进行模板吻合的判断，为了更好的恢复存储前的波形，除去与第二心动周期的重叠的部分保存第三心动周期的其他数据。如果第三心动周期的主波极大值点P_R(3)与在第二动周期的主波极大值点P_R(2)的间隔不小于0.60秒(即心率小于100次情况下)，则进行模板吻合判断：将第一模板向后移动P_R(3)-P_R(2)个数据长度，即以第三心动周期与第一模板的主波极大值点为基准，由四个特征点对比判断第三心动周期的波形是否与第一模板吻合，如果特征点吻合则存储第一模板代号，存储第二心动周期后、第三心动周期前的数据，如果不吻合则将第三心动周期的数据建立为第二模板，以此类推，进行下一个心动周期数据的判断与存储，并按照此过程完成整个数据片段的存储。First, judge whether the interval between the main wave maximum point P_R (3) of the third cardiac cycle and the main wave maximum point P_R (2) of the second cardiac cycle is less than 0.60 seconds, if less than 0.60 seconds (that is, the current Heart rate is greater than 100 times), indicating that the ECG waveforms overlap before and after the division, destroying the repeatability of the ECG waveform. At this time, the judgment of template matching is not performed. The overlapping part of the save the other data of the third cardiac cycle. If the interval between the main wave maximum point_PR (3) in the third cardiac cycle and the main wave maximum point_PR (2) in the second cardiac cycle is not less than 0.60 seconds (that is, when the heart rate is less than 100 beats) , then the template matching judgment is performed: the first template is moved backward by_PR (3)_-PR (2) data lengths, that is, based on the third cardiac cycle and the main wave maximum point of the first template, by Four feature points are compared to judge whether the waveform of the third cardiac cycle matches the first template. If the feature points match, the first template code is stored, and the data after the second cardiac cycle and before the third cardiac cycle are stored. The data of the third cardiac cycle is established as the second template, and so on, to judge and store the data of the next cardiac cycle, and complete the storage of the entire data segment according to this process.

如果已经存在两个以上模板时，预存储心动周期的判断及存储过程选择与之最接近的模板进行。在有多个心电模板的情况下，依据预存储心动周期的主波幅值V_R、V_T和V_D进行模板的匹配选择，例如以V_R、V_T和V_D的和值作为参考，哪一模板的该三个幅值和与预存储心动周期的该三个幅值和接近，则选取该模板作为最接近的模板，依照最匹配的模板进行是否吻合的判断。If there are more than two templates, the process of judging and storing the pre-stored cardiac cycle selects the template closest to it. In the case of multiple ECG templates, the matching selection of templates is performed according to the pre-stored main wave amplitudes_VR , V_T and V_D of the cardiac cycle, for example, the sum of_VR , V_T and V_D is used as a reference , which template whose three amplitude sums are close to the three amplitude sums of the pre-stored cardiac cycle, selects this template as the closest template, and judges whether it matches or not according to the most matching template.

以第三心动周期和第一模板为例，说明基于特征点来判断预存储的心动周期数据与最接近的模板是否吻合的具体方法：Taking the third cardiac cycle and the first template as examples, the specific method of judging whether the pre-stored cardiac cycle data matches the closest template based on feature points is described:

首先，矫正第三心动周期的四个特征点幅值，减少基线漂移对波形带来的影响。如果不存在基线漂移影响则此步骤可省略。First, the amplitudes of the four characteristic points of the third cardiac cycle are corrected to reduce the influence of baseline drift on the waveform. This step can be omitted if there is no baseline drift effect.

其次，将第一模板向后移动P_R(3)-P_R(2)个数据长度，即以第三心动周期与第一模板的主波极大值点为基准，判断其它特征点位置是否吻合，如果吻合则判定第三心动周期与第一模板初步吻合，进行下一步判断；如果不吻合，则判定第三心动周期与第一模板不吻合。Secondly, the first template is moved backward by_PR (3)_-PR (2) data lengths, that is, based on the third cardiac cycle and the main wave maximum point of the first template, it is judged whether the positions of other feature points match, if they match, it is determined that the third cardiac cycle is initially consistent with the first template, and the next step of judgment is performed; if not, it is determined that the third cardiac cycle does not match the first template.

然后，如果第三心动周期与第一模板初步吻合后，进而判断矫正后四个特征点对应的幅值是否与第一模板特征点幅值相对一致，若一致，则判断第三心动周期波形与第一模板吻合；若不一致，则判断第三心动周期波形与第一模板不吻合。Then, if the third cardiac cycle is preliminarily matched with the first template, it is further judged whether the amplitudes corresponding to the four feature points after correction are relatively consistent with the amplitudes of the first template feature points, and if they are consistent, then it is judged that the waveform of the third cardiac cycle is consistent with the first template. The first template matches; if not, it is judged that the waveform of the third cardiac cycle does not match the first template.

上述方法中，对心动周期波形四个特征点幅值矫正过程如下：In the above method, the amplitude correction process of the four characteristic points of the cardiac cycle waveform is as follows:

(1)求取第一模板中QRS波群主波的位置P_R(2)0.10S之前八个数据幅值的平均值，并求取QRS波群主波的位置P_R(2)0.10S之后八个数据幅值的平均值，再求两者平均值mean。(1) Calculate the average value of the eight data amplitudes before the position P_R (2)0.10S of the main wave of the QRS complex in the first template, and calculate the position P_R (2)0.10S of the main wave of the QRS complex After that, the average value of the eight data amplitudes is calculated, and then the average value of the two is calculated.

(2)求取每个心动周期P_R(i)0.10S之前相邻八个数据幅值的平均值，并求取QRS波群主波的位置P_R(i)0.10S之后相邻八个数据幅值的平均值，再求两者平均值mean(i)，以mean(i)与mean的差值作为漂移值，用每个心动周期四个特征点的幅值减去该心动周期对应的漂移值即为矫正后的特征点的幅值。(2) Calculate the average value of the eight adjacent data amplitudes before each cardiac cycle P_R (i)0.10S, and calculate the position of the main wave of the QRS complex after P_R (i)0.10S The average value of the data amplitude, and then find the mean(i) of the two, take the difference between mean(i) and mean as the drift value, and subtract the corresponding value of the four characteristic points of each cardiac cycle from the The drift value is the amplitude of the corrected feature points.

值得说明的是，上述方法中判断预存储心动周期的特征点是否与模板的特征点吻合，并不是指预存储心动周期特征点的位置和幅值与模板的特征点完全相同，而是指预存储心动周期特征点位置和幅值与模板的特征点的差值小于设定即认为吻合，具体判断过程如下：It is worth noting that judging whether the feature points of the pre-stored cardiac cycle match the feature points of the template in the above method does not mean that the position and amplitude of the feature points of the pre-stored cardiac cycle are exactly the same as the feature points of the template, but that the pre-stored The difference between the stored cardiac cycle feature point position and amplitude and the feature point of the template is considered to be consistent if the difference is less than the set value. The specific judgment process is as follows:

(1)以第一模板为例，判断预存储心动周期中四个特征点位置与第一模板中相对应特征点位置之差的绝对值都小于相应的特定的阈值，则判断与模板初步吻合。若|(P_P(i)-P_P(2))-(P_R(i)-P_R(2))|＜th1，|(P_T(i)-P_T(2))-(P_R(i)-P_R(2))|＜th2，且|(P_D(i)-P_D(2))-(P_R(i)-P_R(2))|＜th3，其中，th1、th2、th3为判断的阈值，应根据采样时的误差状况取一个稍大于零的值，如th1取0.004S，th2取0.004S，th3取0.006S，则判定此心动周期与第一模板初步吻合。(1) Taking the first template as an example, it is judged that the absolute value of the difference between the four feature point positions in the pre-stored cardiac cycle and the corresponding feature point positions in the first template is less than the corresponding specific threshold, then the judgment is initially consistent with the template . If |(P_P (i)-P_P (2))-(P_R (i)-P_R (2))|<th1, |(P_T (i)-P_T (2))-(P_R (i)-P_R (2))|<th2, and |(P_D (i)-P_D (2))-(P_R (i)-P_R (2))|<th3, where, Th1, th2, and th3 are the judgment thresholds, and a value slightly greater than zero should be taken according to the error status at the time of sampling. For example, th1 takes 0.004S, th2 takes 0.004S, and th3 takes 0.006S, then it is determined that the cardiac cycle is consistent with the first template. Preliminary match.

(2)判定与模板初步吻合后，判断|(V_P(i)-V_D(2))|*a+|(V_T(i)-V_T(2))|*b+|(V_D(i)-V_D(2))|*c+|(V_R(i)-V_R(2))|是否小于某个特定的阈值th4(其中，a、b、c分别为加权系数，a、b、c的取值可以根据采集对象的生理特征、经验值或者有针对性的要求灵活确定，鉴于V_T与V_P幅值相对较大，因采样等因素不可避免会造成相邻两个心动周期特征值差异的存在，故对其设置相应较低的加权系数，以便客观地进行吻合判断。例如本实施例根据经验将其确定为a＝1.5，b＝1，c＝1.5)，是则判断此心动周期与模板吻合，否则判断此心动周期与模板不吻合。(2) After judging the initial match with the template, judge |(V_P (i)-V_D (2))|*a+|(V_T (i)-V_T (2))|*b+|(V_D ( i)-V_D (2))|*c+|(V_R (i)_-VR (2))| is less than a certain threshold th4 (where a, b, c are weighting coefficients, a, The values of b and c can be flexibly determined according to the physiological characteristics, experience values or targeted requirements of the acquisition object. In view of the relatively large amplitudes of V_T and V_P , it is inevitable that two adjacent heart beats will be caused by factors such as sampling. The existence of periodic eigenvalue difference, so it is set correspondingly lower weighting factor, so that objectively carry out match judgment.For example present embodiment is determined as a=1.5 according to experience, b=1, c=1.5), then It is judged that the cardiac cycle matches the template, otherwise it is judged that the cardiac cycle does not coincide with the template.

在心动周期波形与模板是否吻合判断时，因采集时或其他一些客观因素，相邻两个主波幅值会有差异，而且在P_P(i)与P_T(i)之间最小幅值点位置也会有些差异，故在进行吻合判断时，相对降低其加权系数，以使其更符合实际情况，提高存储的可靠度。When judging whether the cardiac cycle waveform matches the template, due to acquisition or some other objective factors, there will be differences in the amplitudes of two adjacent main waves, and the minimum amplitude between P_P (i) and_PT (i) There will also be some differences in the position of the points, so when making a matching judgment, the weighting coefficient is relatively reduced to make it more in line with the actual situation and improve the reliability of storage.

Claims (6)

Translated fromChinese

1.一种动态心电数据自适应高效存储方法，包括如下步骤：1. A method for self-adaptive and efficient storage of dynamic ECG data, comprising the steps of:(1)采集一定长度的心电数据片段；(1) Collect a certain length of ECG data segment;(2)对所述心电数据片段进行特征点的提取，特征点至少包括每个心动周期QRS波群中主波极大值点【P_R(i)，V_R(i)】，P_R(i)为位置，V_R(i)为对应的电压幅值；定义每个心动周期数据长度为K秒，数据长度K取0.60S，且以QRS波群中主波极大值点P_R(i)为基准，P_R(i)之前分配0.20S，P_R(i)之后分配0.40S；以每个主波极大值点为基准，根据定义的心动周期数据长度将心电数据片段进行划分；(2) Extracting feature points on the ECG data segment, the feature points at least include the main wave maximum point [P_R (i),_VR (i)] in each cardiac cycle QRS wave group, P_R (i) is the position,_VR (i) is the corresponding voltage amplitude; define the data length of each cardiac cycle as K seconds, the data length K is 0.60S, and the main wave maximum point P_R in the QRS wave group (i) as the benchmark, 0.20S is allocated before_PR (i), and 0.40S is allocated after_PR (i); based on each main wave maximum point, the ECG data segment is divided according to the defined cardiac cycle data length to divide;(3)选择划分后的某一心动周期的心电数据作为第一模板，保存该第一模板的全部数据及模板代号，并存储第一模板之前的全部心电数据；(3) Select the electrocardiographic data of a certain cardiac cycle after the division as the first template, save all data and template codes of the first template, and store all electrocardiographic data before the first template;(4)判断预存储的心动周期的主波极大值点与其在先心动周期的主波极大值点的间隔是否小于K秒，是则进行(a)，否则进行(b)：(a)除去与其在先心动周期的重叠部分保存预存储心动周期的其他数据；(b)存储预存储心动周期前、其在先心动周期后的数据，并将预存储心动周期的数据与最接近的模板进行特征点吻合判断，如果特征点吻合则存储相应模板代号，如果不吻合则将该预存储心动周期的数据建立为新的模板，并给予新的模板代号；(4) Whether the interval between the main wave maximum point of the pre-stored cardiac cycle and the main wave maximum point of the previous cardiac cycle is less than K seconds, if so, perform (a), otherwise perform (b): (a ) remove the overlapping part of its previous cardiac cycle and save other data of the pre-stored cardiac cycle; The template performs feature point matching judgment. If the feature points match, the corresponding template code is stored. If not, the pre-stored cardiac cycle data is established as a new template, and a new template code is given;(5)重复步骤(4)；(5) Repeat step (4);其特征在于：所述特征点还包括每个心动周期的：P波极大值点【P_P(i)，V_P(i)】，T波极大值点【P_T(i)，V_T(i)】，QRS波群极小值点【P_D(i)，V_D(i)】。It is characterized in that: the feature points also include for each cardiac cycle: P wave maximum point [P_P (i), V_P (i)], T wave maximum point [_PT (i), V_T (i)], QRS complex minimum point [P_D (i), V_D (i)].2.根据权利要求1所述的动态心电数据自适应高效存储方法，其特征在于：所述特征点的提取方法具体为：①获取第i个心动周期QRS波群中主波极大值点，位置记为P_R(i)，对应的电压幅值记为V_R(i)，②在主波极大值点P_R(i)之前0.07S到0.20S数据范围内寻找电压幅值最大值点，位置记为P_P(i)，对应的电压幅值记为V_P(i)，③在主波极大值点P_R(i)之后0.07S到0.40S数据范围内寻找电压幅值最大值点，位置记为P_T(i)，对应的电压幅值记为V_T(i)，④在P_P(i)与P_T(i)之间数据范围内，寻找幅值最小值点，位置记为P_D(i)，对应的电压幅值记为V_D(i)。2. the dynamic electrocardiographic data adaptive high-efficiency storage method according to claim 1, is characterized in that: the extracting method of described feature point is specifically: 1. obtain the dominant wave maximum value point in the i-th cardiac cycle QRS wave group , the position is recorded as P_R (i), and the corresponding voltage amplitude is recorded as_VR (i), ②Find the maximum voltage amplitude within the data range of 0.07S to 0.20S before the main wave maximum point P_R (i) value point, the position is recorded as P_P (i), and the corresponding voltage amplitude is recorded as V_P (i). ③ Find the voltage amplitude within the data range of 0.07S to 0.40S after the main wave maximum point P_R (i). The maximum value point, the position is recorded as_PT (i), and the corresponding voltage amplitude is recorded as V_T (i), ④In the data range between P_P (i) and_PT (i), look for the minimum amplitude Value point, the position is recorded as_PD (i), and the corresponding voltage amplitude is recorded as V_D (i).3.根据权利要求1所述的动态心电数据自适应高效存储方法，其特征在于：所述将预存储心动周期的数据与最接近的模板进行特征点吻合判断的方法，具体为：首先，以预存储心动周期数据与最接近的模板的主波极大值点为基准，判断其它特征点位置是否吻合，是则判定初步吻合，否则判定与模板不吻合；与模板初步吻合后，判断四个特征点对应的幅值是否与模板特征点幅值相对一致，若一致，则判断本心动周期波形与模板吻合；若不一致，则判断与模板不吻合。3. The method for self-adaptive and efficient storage of dynamic electrocardiographic data according to claim 1, characterized in that: the method for judging the matching of feature points between the pre-stored cardiac cycle data and the closest template is specifically: firstly, Based on the pre-stored cardiac cycle data and the main wave maximum point of the closest template, judge whether the positions of other feature points are consistent. Whether the amplitude corresponding to each feature point is relatively consistent with the template feature point amplitude, if it is consistent, it is judged that the waveform of this cardiac cycle matches the template; if not, it is judged that it does not match the template.4.根据权利要求3所述的动态心电数据自适应高效存储方法，其特征在于：模板吻合判断前，对心动周期的四个特征点幅值进行矫正，矫正过程如下：4. the dynamic electrocardiographic data self-adaptive high-efficiency storage method according to claim 3, is characterized in that: before the template matches judgment, four characteristic point amplitudes of the cardiac cycle are corrected, and the correction process is as follows:(1)求取第一模板QRS波群主波极大值点0.10S之前八个数据幅值的平均值，并求取第一模板QRS波群主波极大值点0.10S之后八个数据幅值的平均值，再求两者平均值mean；(1) Calculate the average value of the eight data amplitudes before the maximum point of the main wave of the first template QRS complex 0.10S, and calculate the eight data after the maximum point of the main wave of the first template QRS complex 0.10S The average value of the amplitude, and then calculate the mean value of the two;(2)求取每个心动周期P_R(i)0.10S之前相邻八个数据幅值的平均值，并求取每个心动周期P_R(i)0.10S之后相邻八个数据幅值的平均值，再求两者平均值mean(i)，以mean(i)与mean的差值作为漂移值，用每个心动周期四个特征点的幅值减去该心动周期对应的漂移值即为矫正后的特征点的幅值。(2) Calculate the average value of the eight adjacent data amplitudes before each cardiac cycle P_R (i)0.10S, and calculate the adjacent eight data amplitudes after each cardiac cycle P_R (i)0.10S Then calculate the mean(i) of the two, take the difference between mean(i) and mean as the drift value, and use the amplitude of the four feature points in each cardiac cycle to subtract the corresponding drift value of the cardiac cycle is the amplitude of the corrected feature points.5.根据权利要求3或4所述的动态心电数据自适应高效存储方法，其特征在于：判断预存储心动周期的特征点是否与模板的特征点吻合，具体判断过程如下：5. according to claim 3 or 4 described dynamic ECG data self-adaptive high-efficiency storage methods, it is characterized in that: judge whether the feature point of pre-stored cardiac cycle coincides with the feature point of template, concrete judging process is as follows:(1)判断预存储心动周期中除主波极大值点外的其他特征点位置与模板中相对应特征点位置之差的绝对值是否都小于相应的特定的阈值，即|(P_P(i)-P_P(模板))-(P_R(i)-P_R(模板))|＜th1，|(P_T(i)-P_T(模板))-(P_R(i)-P_R(模板))|＜th2，且|(P_D(i)-P_D(模板))-(P_R(i)-P_R(模板))|＜th3，其中，th1、th2、th3为判断的阈值，是则判定此心动周期与模板初步吻合，否则判定此心动周期与模板不吻合；(1) Determine whether the absolute values of the differences between the positions of other feature points in the pre-stored cardiac cycle except for the maximum point of the main wave and the positions of the corresponding feature points in the template are all smaller than the corresponding specific threshold, that is |(P_P ( i)-P_P (template))-(P_R (i)-P_R (template))|<th1, |(P_T (i)-P_T (template))-(P_R (i)-P_R (template))|<th2, and |(P_D (i)-PD (template_{))-(P R(} i)-P_R (template))|<th3, where th1, th2, th3 are The threshold of judgment, if it is, it is judged that the cardiac cycle is initially consistent with the template, otherwise it is judged that the cardiac cycle is not consistent with the template;(2)判定与模板初步吻合后，判断|(V_P(i)-V_P(模板))|*a+|(V_T(i)-V_T(模板))|*b+|(V_D(i)-V_D(模板))|*c+|(V_R(i)-V_R(模板))|是否小于某个特定的阈值th4，其中，a、b、c分别为加权系数，是则判断此心动周期与模板吻合，否则判断此心动周期与模板不吻合。(2) After judging the initial match with the template, determine |(V_P (i)-V_P (template))|*a+|(V_T (i)-V_T (template))|*b+|(V_D ( i)-V_D (template))|*c+|(V_R (i)_-VR (template))| is less than a certain threshold th4, where a, b, and c are weighting coefficients respectively, then It is judged that the cardiac cycle matches the template, otherwise it is judged that the cardiac cycle does not coincide with the template.6.根据权利要求1所述的动态心电数据自适应高效存储方法，其特征在于：选择最接近的模板的判断方法是：以V_R(i)、V_T(i)和V_D(i)的和值作为参考，哪一模板的该三个幅值和与预存储心动周期的该三个幅值和接近，则选取该模板作为最接近的模板。6. dynamic electrocardiographic data self-adaptive high-efficiency storage method according to claim 1 is characterized in that: the judging method of selecting the nearest template is: with V_R (i), V_T (i) and V_D (i ) as a reference, which template whose three amplitude sums are close to the three amplitude sums of the pre-stored cardiac cycle, selects this template as the closest template.

Images