Atrial fibrillation signal detection method and device

Abstract

The present application discloses an atrial fibrillation signal detection method and device, including: obtaining an original electrocardiogram signal, preprocessing the original electrocardiogram signal to obtain an electrocardiogram signal to be analyzed; determining a set of QRS wave starting points and a set of end points in the electrocardiogram signal to be analyzed based on a preset waveform recognition method; reconstructing the electrocardiogram signal to be analyzed according to the set of QRS wave starting points and the set of end points based on a preset reconstruction method to obtain an electrocardiogram reconstruction signal; determining a characteristic value of the electrocardiogram reconstruction signal based on a delay algorithm, and judging whether there is an atrial fibrillation signal in the electrocardiogram reconstruction signal based on the characteristic value and a preset judgment threshold. The present invention determines the characteristic value of the reconstructed signal based on a delay algorithm, and detects the atrial fibrillation signal based on the characteristic value and the preset judgment threshold, which can reduce the delay time of atrial fibrillation signal detection, reduce the amount of calculation in the processing process, and realize real-time detection of atrial fibrillation signals.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and in particular to an atrial fibrillation signal detection method and device. Background Art

[0002] In recent years, the incidence of cardiovascular diseases has increased year by year, and its high disability and mortality rates seriously endanger human health. Atrial fibrillation is a common arrhythmia disease in clinical practice, which can easily lead to serious complications such as atrial thrombus and ischemic stroke. The electrocardiogram waveform of atrial fibrillation often shows characteristics such as RR interval disorder and P wave disappearance. Accurate detection of atrial fibrillation signals as early as possible is of great significance to reducing the mortality caused by atrial fibrillation.

[0003] With the continuous development of nonlinear dynamics technology, it has been widely used in the research of ECG signals. At present, when nonlinear dynamics technology is used to identify atrial fibrillation, it is mainly through intelligent detection by combining phase space reconstruction method with neural network. The cardiac pulsation is regarded as a nonlinear dynamic system, and the phase space reconstruction method is used to map the one-dimensional time series into a high-dimensional phase space. Under the premise of ensuring the same dynamic behavior as the original sequence, the rhythm features implied in the time series that characterize the atrial fibrillation process are mined, and then a suitable convolutional network model is constructed to classify the rhythm features, thereby realizing the recognition of atrial fibrillation. However, this method often requires the selection of multiple dimensions m to reconstruct the m-dimensional space, and also requires the construction of a neural network model for atrial fibrillation detection. There are problems such as large computational complexity and low efficiency in atrial fibrillation recognition, which is not suitable for real-time detection of atrial fibrillation.

[0004] Therefore, it is necessary to propose an atrial fibrillation signal detection method to solve the problem in the prior art that when atrial fibrillation detection is performed based on electrocardiogram signals using nonlinear dynamic technology, the calculation amount is large and the efficiency is low, resulting in the inability to detect atrial fibrillation signals in real time. Summary of the invention

[0005] In view of this, it is necessary to provide an atrial fibrillation signal detection method and device to solve the technical problems in the prior art of using nonlinear dynamic technology to detect atrial fibrillation based on electrocardiogram signals, which are large in calculation amount, low in efficiency, and unable to detect atrial fibrillation signals in real time.

[0006] In order to solve the above problems, the present invention provides an atrial fibrillation signal detection method, comprising:

[0007] Acquire an original electrocardiogram signal, and preprocess the original electrocardiogram signal to obtain an electrocardiogram signal to be analyzed;

[0008] Based on a preset waveform recognition method, determining a QRS wave starting point set and an end point set in the electrocardiogram signal to be analyzed;

[0009] Reconstructing the electrocardiogram signal to be analyzed according to the QRS wave starting point set and the end point set based on a preset reconstruction method to obtain an electrocardiogram reconstruction signal;

[0010] The characteristic value of the electrocardiographic reconstruction signal is determined based on a time delay algorithm, and according to the characteristic value and a preset judgment threshold, it is judged whether there is an atrial fibrillation signal in the electrocardiographic reconstruction signal.

[0011] Further, the original ECG signal is preprocessed to obtain a signal to be analyzed, including:

[0012] Performing low-pass filtering on the original ECG signal to obtain a first filtered ECG signal;

[0013] Performing high-pass filtering on the first filtered ECG signal to obtain a second filtered ECG signal;

[0014] The second filtered ECG signal is subjected to power frequency filtering to obtain a signal to be analyzed.

[0015] Further, based on a preset reconstruction method, the electrocardiogram signal to be analyzed is reconstructed according to the QRS wave starting point set and the end point set to obtain an electrocardiogram reconstruction signal, including:

[0016] Eliminate the QRS wave in the electrocardiographic signal to be analyzed according to the QRS wave starting point set and the end point set to obtain a plurality of SQ interval signals;

[0017] Perform signal complementation on each of the SQ interval signals using a preset complementation method to obtain a complementation signal corresponding to each of the SQ interval signals;

[0018] Performing high-pass filtering and smoothing processing on the completed signal to obtain a signal to be spliced;

[0019] Two adjacent signals to be spliced ​​are spliced ​​to obtain an ECG reconstruction signal.

[0020] Further, each of the SQ interval signals is complemented by a preset complement method to obtain a complement signal corresponding to each of the SQ interval signals, including:

[0021] For any SQ interval signal SQ[n], if SQ[n][1] is the amplitude of the first sampling point of the SQ interval signal SQ[n], then a waveform with an amplitude of SQ[n][1] and a preset completion length is added to the head of SQ[n] to obtain the completion signal corresponding to SQ[n].

[0022] Further, the completed signal is subjected to high-pass filtering and smoothing to obtain a signal to be spliced, including:

[0023] Performing high-pass filtering on the complement signal to obtain a filtered complement signal;

[0024] The waveform of the preset completion length at the header of the filtered completion signal is deleted to obtain the signal to be spliced.

[0025] Furthermore, two adjacent signals to be spliced ​​are spliced ​​to obtain an ECG reconstruction signal, including:

[0026] For the adjacent first signal to be spliced ​​SQ2[n] and the second signal to be spliced ​​SQ2[n+1], determine whether there is a sampling point whose amplitude difference is less than a preset search threshold within a preset splicing length range between the tail of SQ2[n] and the head of SQ2[n+1];

[0027] When there is a sampling point whose amplitude difference between the tail of SQ2[n] and the head of SQ2[n+1] is less than the preset search threshold within the preset splicing length range, SQ2[n] and SQ2[n+1] are spliced ​​with the preset splicing length to obtain an ECG reconstruction signal;

[0028] When there are no sampling points whose amplitude difference between the tail of SQ2[n] and the head of SQ2[n+1] is less than the preset search threshold within the preset splicing length range, the tail of SQ2[n] and the head of SQ3[n+1] are fitted respectively and the splicing is completed.

[0029] Further, determining the characteristic value of the ECG reconstruction signal based on a time delay algorithm includes:

[0030] Extracting a signal of a preset duration at the tail of the electrocardiographic reconstruction signal to obtain a target signal;

[0031] Dividing the target signal into a first characteristic signal and a second characteristic signal;

[0032] The first characteristic signal and the second characteristic signal are processed based on a delay algorithm to obtain a first characteristic value and a second characteristic value.

[0033] Further, judging whether there is an atrial fibrillation signal in the electrocardiogram reconstruction signal according to the characteristic value and a preset judgment threshold includes:

[0034] According to the first characteristic value, the second characteristic value and the preset judgment threshold, it is judged whether the target signal, the first characteristic signal and the second characteristic signal are atrial fibrillation signals respectively.

[0035] Furthermore, the preset waveform recognition method is a differential threshold method.

[0036] The present invention also provides an atrial fibrillation signal detection device, comprising:

[0037] A data acquisition module is used to acquire the original ECG signal and pre-process the original ECG signal to obtain the ECG signal to be analyzed;

[0038] A waveform recognition module, used to determine a set of QRS wave starting points and a set of end points in the electrocardiogram signal to be analyzed based on a preset waveform recognition method;

[0039] A signal reconstruction module, used to reconstruct the electrocardiogram signal to be analyzed according to the QRS wave starting point set and the end point set based on a preset reconstruction method to obtain an electrocardiogram reconstruction signal;

[0040] The detection module is used to determine the characteristic value of the electrocardiographic reconstruction signal based on a time delay algorithm, and to determine whether an atrial fibrillation signal exists in the electrocardiographic reconstruction signal according to the characteristic value and a preset judgment threshold.

[0041] Compared with the prior art, the beneficial effects of the present invention include: first, obtaining the original ECG signal, preprocessing the original ECG signal, and obtaining the ECG signal to be analyzed; second, using a preset waveform recognition method to determine the starting point and end point set of the QRS wave, and reconstructing the ECG signal; finally, determining the characteristic value of the ECG reconstructed signal based on the delay algorithm, and detecting the atrial fibrillation signal according to the characteristic value and the preset judgment threshold. The method of the present invention extracts the QRS wave from the ECG signal, reconstructs the ECG signal using a preset reconstruction method, and then determines the characteristic value of the reconstructed signal based on the delay algorithm, and detects the atrial fibrillation signal according to the characteristic value and the preset judgment threshold. It can reduce the delay time of atrial fibrillation signal detection, reduce the amount of calculation in the processing process, and realize real-time detection of atrial fibrillation signals. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A schematic diagram of a flow chart of an embodiment of an atrial fibrillation signal detection method provided by the present invention;

[0043] Figure 2 A schematic diagram of a flow chart of an embodiment of processing an original electrocardiogram signal and extracting characteristic values ​​provided by the present invention;

[0044] Figure 3 This is a structural schematic diagram of an embodiment of an atrial fibrillation signal detection device provided by the present invention. DETAILED DESCRIPTION

[0045] The preferred embodiments of the present invention are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not used to limit the scope of the present invention.

[0046] Before describing the embodiments, the relevant terms of the present application are first explained.

[0047] F wave: In the atrial fibrillation signal, the P wave of the electrocardiogram disappears, and a fibrillation wave of irregular size and shape appears, which is called the F wave on the electrocardiogram, with a frequency of 350-600 times / minute.

[0048] The time delay algorithm is based on the phase space reconstruction method. The ECG signal x(n) is reconstructed as follows: let the x-axis horizontal coordinate be x(n) and the y-axis vertical coordinate be y(n)=x(n+τ), where τ is the delay time. Draw the ECG signal curve in the phase space, and draw a 40×40 grid in the phase space to calculate the number of grids filled by the signal, and then calculate the eigenvalue d:

[0049]

[0050] Where n1 is the number of filled grids, and n2 is the total number of all grids. The value of the eigenvalue d is used to determine whether the current signal is a chaotic signal. According to the chaotic theory in the electrophysiological mechanism of atrial fibrillation, atrial fibrillation and ventricular fibrillation can both be regarded as chaotic signals. Therefore, if the current signal is a chaotic signal, it can be determined that the signal is an atrial fibrillation signal.

[0051] The inventive concept of the present application is described below.

[0052] In the prior art, when detecting atrial fibrillation signals through nonlinear dynamic technology, the phase space reconstruction method is usually used. The idea of ​​the phase space reconstruction method is to form an m-dimensional space by combining the discrete one-dimensional time series x(t) with x(t+τ),…,x[(t+(m-1)τ], where τ is the time delay and m is the embedding dimension. The real-time quantification of the waveform morphology is achieved through high-dimensional mapping. Under the premise of ensuring the same dynamic behavior as the original sequence, the dynamic information hidden in the time series is mined, which is suitable for processing longer non-stationary data. By selecting the appropriate embedding dimension m and time delay τ for phase space reconstruction, the two-dimensional rhythm features representing the atrial fibrillation signal are extracted, and the extracted two-dimensional rhythm features are classified using the convolutional neural network model, thereby realizing the detection of atrial fibrillation signals. However, in this method, multiple dimensions m are often selected to reconstruct the m-dimensional space, and a CNN neural network model must be built for atrial fibrillation detection. The required amount of calculation is too large and is not suitable for real-time atrial fibrillation detection.

[0053] The present invention proposes an atrial fibrillation signal detection method and device, which eliminates the QRS wave of the original electrocardiogram signal and reconstructs the signal, cuts the reconstructed electrocardiogram signal, and uses a delay algorithm in nonlinear dynamic technology to determine the characteristic value of the cut signal, performs multi-threshold analysis on the characteristic value, and realizes the detection of atrial fibrillation signals, thereby solving the problems of excessive calculation and long delay in the prior art and realizing real-time detection of atrial fibrillation signals.

[0054] The embodiment of the present invention provides a method for detecting atrial fibrillation signals. Figure 1 As shown, Figure 1 is a flow chart of the atrial fibrillation signal detection method, comprising:

[0055] Step S101: acquiring an original ECG signal, and preprocessing the original ECG signal to obtain an ECG signal to be analyzed;

[0056] Step S102: determining a set of QRS wave starting points and a set of end points in the electrocardiogram signal to be analyzed based on a preset waveform recognition method;

[0057] Step S103: reconstructing the ECG signal to be analyzed according to the QRS wave starting point set and the QRS wave ending point set based on a preset reconstruction method to obtain an ECG reconstruction signal;

[0058] Step S104: determining a characteristic value of the ECG reconstruction signal based on a time delay algorithm, and judging whether an atrial fibrillation signal exists in the ECG reconstruction signal according to the characteristic value and a preset judgment threshold.

[0059] The atrial fibrillation signal detection method provided in this embodiment first obtains the original ECG signal, pre-processes the original ECG signal, and obtains the ECG signal to be analyzed; secondly, a preset waveform recognition method is used to determine the starting point and end point set of the QRS wave, and the ECG signal is reconstructed; finally, the characteristic value of the ECG reconstructed signal is determined based on the delay algorithm, and the atrial fibrillation signal is detected according to the characteristic value and the preset judgment threshold. The method of this embodiment extracts the QRS wave from the ECG signal, reconstructs the ECG signal using a preset reconstruction method, and then determines the characteristic value of the reconstructed signal based on the delay algorithm, and detects the atrial fibrillation signal according to the characteristic value and the preset judgment threshold. It can reduce the delay time of atrial fibrillation signal detection, reduce the amount of calculation in the processing process, and realize real-time detection of atrial fibrillation signals.

[0060] As a preferred embodiment, in step S101, the original ECG signal is preprocessed to obtain a signal to be analyzed, including:

[0061] Performing low-pass filtering on the original ECG signal to obtain a first filtered ECG signal;

[0062] Performing high-pass filtering on the first filtered ECG signal to obtain a second filtered ECG signal;

[0063] The second filtered ECG signal is subjected to power frequency filtering to obtain a signal to be analyzed.

[0064] As a specific embodiment, the specific process of preprocessing is: low-pass filtering the collected original ECG signal D1 to obtain signal D2, and removing the baseline noise in the signal through low-pass filtering; high-pass filtering the signal D2 to obtain signal D3, and removing the high-frequency noise; performing power frequency filtering on the signal D3 to remove the power frequency noise in the signal and obtain the signal D4 to be analyzed.

[0065] As a preferred embodiment, in step S102, the preset waveform recognition method is a differential threshold method.

[0066] As a preferred embodiment, in step S103, the ECG signal to be analyzed is reconstructed according to the QRS wave starting point set and the end point set based on a preset reconstruction method to obtain an ECG reconstruction signal, including:

[0067] Eliminate the QRS wave in the electrocardiographic signal to be analyzed according to the QRS wave starting point set and the end point set to obtain a plurality of SQ interval signals;

[0068] Perform signal complementation on each of the SQ interval signals using a preset complementation method to obtain a complementation signal corresponding to each of the SQ interval signals;

[0069] Performing high-pass filtering and smoothing processing on the completed signal to obtain a signal to be spliced;

[0070] Two adjacent signals to be spliced ​​are spliced ​​to obtain an ECG reconstruction signal.

[0071] As a preferred embodiment, each of the SQ interval signals is complemented by a preset complement method to obtain a complement signal corresponding to each of the SQ interval signals, including:

[0072] For any SQ interval signal SQ[n], if SQ[n][1] is the amplitude of the first sampling point of the SQ interval signal SQ[n], then a waveform with an amplitude of SQ[n][1] and a preset completion length is added to the head of SQ[n] to obtain the completion signal corresponding to SQ[n].

[0073] As a preferred embodiment, high-pass filtering and smoothing are performed on the complemented signal to obtain a signal to be spliced, including:

[0074] Performing high-pass filtering on the complement signal to obtain a filtered complement signal;

[0075] The waveform of the preset completion length at the header of the filtered completion signal is deleted to obtain the signal to be spliced.

[0076] As a preferred embodiment, two adjacent signals to be spliced ​​are spliced ​​to obtain an ECG reconstruction signal, including:

[0077] For the adjacent first signal to be spliced ​​SQ2[n] and the second signal to be spliced ​​SQ2[n+1], determine whether there is a sampling point whose amplitude difference is less than a preset search threshold within a preset splicing length range between the tail of SQ2[n] and the head of SQ2[n+1];

[0078] When there is a sampling point whose amplitude difference between the tail of SQ2[n] and the head of SQ2[n+1] is less than the preset search threshold within the preset splicing length range, SQ2[n] and SQ2[n+1] are spliced ​​with the preset splicing length to obtain an ECG reconstruction signal;

[0079] When there are no sampling points whose amplitude difference between the tail of SQ2[n] and the head of SQ2[n+1] is less than the preset search threshold within the preset splicing length range, the tail of SQ2[n] and the head of SQ3[n+1] are fitted respectively and the splicing is completed.

[0080] The specific process of reconstructing the above-mentioned electrocardiographic signal with gaps is described below with a specific embodiment.

[0081] As a specific embodiment, according to the starting point set A1 and the end point set A2 of the QRS wave in the signal to be analyzed, the QRS wave in the signal to be analyzed is eliminated to obtain the remaining SQ interval signal, and SQ[n] is recorded as the nth SQ interval waveform signal, and SQ[n][1] is the amplitude of the first sampling point of the nth SQ interval waveform signal.

[0082] In order to prevent the filter stabilization stage at the beginning of the signal from affecting the waveform of SQ[n] when SQ[n] is subsequently filtered, a waveform with an amplitude of SQ[n][1] and a length of K (preset completion length) is added to the beginning of the signal SQ[n] (n=1,2,...) to obtain the signal SQ1[n] (n=1,2,...).

[0083] The signal SQ1[n] (n=1, 2, ...) is high-pass filtered, and the waveform with a length of K at the beginning of the filtered signal is deleted to obtain the signal SQ2[n] (n=1, 2, ...), thereby achieving the purpose of smoothing the baseline of the SQ interval signal.

[0084] The preset splicing length range is L4 seconds, then within the tail range L4 of SQ2[n] and the head range L4 of SQ2[n+1], the search is conducted for sampling points idx[n] and idx[n+1] whose amplitude difference is less than the preset search threshold ∈; where ∈ should be a sufficiently small threshold, and the specific value is related to the amplitude unit of the ECG digital signal.

[0085] When there is a sampling point whose amplitude difference between the tail of SQ2[n] and the head of SQ2[n+1] is less than the preset search threshold within the preset splicing length range, SQ2[n] and SQ2[n+1] are spliced ​​with the preset splicing length to obtain the ECG reconstruction signal SQ3;

[0086] When there is no sampling point whose amplitude difference between the tail of SQ2[n] and the head of SQ2[n+1] is less than the preset search threshold within the preset splicing length range, the tail of SQ2[n] and the head of SQ3[n+1] are fitted respectively, and the splicing is completed to obtain the ECG reconstruction signal SQ3.

[0087] As a preferred embodiment, in step S104, determining the characteristic value of the ECG reconstruction signal based on a time delay algorithm includes:

[0088] Extracting a signal of a preset duration at the tail of the electrocardiographic reconstruction signal to obtain a target signal;

[0089] Dividing the target signal into a first characteristic signal and a second characteristic signal in chronological order; wherein the duration of the second characteristic signal is shorter than the duration of the first characteristic signal;

[0090] The first characteristic signal and the second characteristic signal are processed based on a delay algorithm to obtain a first characteristic value and a second characteristic value.

[0091] As a preferred embodiment, judging whether there is an atrial fibrillation signal in the electrocardiogram reconstruction signal according to the characteristic value and a preset judgment threshold includes:

[0092] According to the first characteristic value, the second characteristic value and the preset judgment threshold, it is judged whether the target signal, the first characteristic signal and the second characteristic signal are atrial fibrillation signals respectively.

[0093] As a specific embodiment, the signal D5 in the range of the last L1 seconds (preset duration) of the ECG reconstruction signal SQ3 is divided into two segments D6 and D7 with lengths of L2s and L3s; wherein L2>L3, and there is no specific limitation on L1 (for example, L1=10, L2=8, L3=2).

[0094] The signals D6 and D7 are processed using a time delay algorithm to obtain characteristic values ​​M1 and M2 respectively.

[0095] In order to facilitate the understanding of the above method, Figure 2 As shown, Figure 2 The schematic diagram of the process of processing the original ECG signal and extracting the eigenvalues ​​is shown, including the following steps:

[0096] Step S201: low-pass filter the original ECG signal D1 to obtain a signal D2; high-pass filter the signal D2 to obtain a signal D3; power frequency filter the signal D3 to obtain a signal D4 to be analyzed;

[0097] Step S202: Identify the QRS wave starting point set A1 and the end point set A2 of the signal to be analyzed D4, and eliminate the QRS wave to obtain the SQ interval signal SQ[n];

[0098] Step S203: completing the SQ interval signal SQ[n] by inserting a waveform with a length of K (preset completion length) and an amplitude of SQ[n][1] to obtain SQ1[n], performing high-pass filtering on SQ1[n], deleting the waveform with a length of K at the beginning, and obtaining SQ2[n];

[0099] Step S204: performing a splicing operation of a preset splicing length on SQ2[n] and SQ2[n+1] to obtain an ECG reconstruction signal SQ3;

[0100] Step S205: dividing the signal D5 in the range of the last L1 seconds (preset duration) of the ECG reconstruction signal SQ3 into two segments D6 and D7 with lengths of L2s and L3s, and processing the signals D6 and D7 using a time delay algorithm to obtain characteristic values ​​M1 and M2 respectively;

[0101] It should be noted that if the delay algorithm is directly used for the detection of atrial fibrillation signals, since there is an R wave with a larger amplitude than the F wave in the atrial fibrillation signal, the characteristic values ​​calculated for the atrial fibrillation signal and the non-atrial fibrillation signal are There will be no significant difference. In addition, if the input signal is a non-atrial fibrillation signal followed by an atrial fibrillation signal, the time to detect the atrial fibrillation signal will be delayed by several seconds compared to the input time of the normal atrial fibrillation signal, and the real-time detection of the atrial fibrillation signal cannot be achieved. Therefore, in the method of this embodiment, by cutting the reconstructed target signal, the delay time of detecting the F wave of the atrial fibrillation signal can be shortened, providing a basis for the real-time detection of the atrial fibrillation signal.

[0102] As a specific embodiment, according to the first characteristic value, the second characteristic value and the preset judgment threshold, it is judged whether the target signal, the first characteristic signal and the second characteristic signal are atrial fibrillation signals respectively, specifically:

[0103] Assuming that the preset judgment thresholds are T1, T2_1 and T2_2, the target signal D5 within the last L1 seconds of the ECG reconstruction signal SQ3 is divided into two segments D6 and D7 (L2>L3) with lengths of L2 seconds and L3 seconds respectively; the judgment process is as follows:

[0104] If M1>T1 and M2>T2_2, the entire D5 signal is determined to be an atrial fibrillation signal;

[0105] If M1>T1 and M2>T2_1 and M2<T2_2, then the entire D5 signal segment is determined to be an atrial fibrillation signal.

[0106] If M1>T1 and M2<T2_1, then the D6 signal segment is determined to be an atrial fibrillation signal, and the D7 signal segment is determined to be a non - atrial fibrillation signal. That is, it indicates that the historical data is an atrial fibrillation signal and the currently input signal is a non - atrial fibrillation signal.

[0107] If M1<T1 and M2<T2_1, then the entire D5 signal segment is determined to be a non - atrial fibrillation signal.

[0108] If M1<T1 and M2>T2_1 and M2<T2_2, then the entire D5 signal segment is determined to be a non - atrial fibrillation signal.

[0109] If M1<T1 and M2>T2_2, then the D6 signal segment is determined to be a non - atrial fibrillation signal, and the D7 signal segment is determined to be an atrial fibrillation signal. That is, it indicates that the historical data is a non - atrial fibrillation signal and the currently input signal is an atrial fibrillation signal.

[0110] An embodiment of the present invention further provides an atrial fibrillation signal detection device, as Figure 3 shown, the atrial fibrillation signal detection device 300 includes:

[0111] A data acquisition module 301, configured to acquire an original electrocardiogram signal, pre - process the original electrocardiogram signal to obtain an electrocardiogram signal to be analyzed;

[0112] A waveform recognition module 302, configured to determine a QRS wave start point set and an end point set in the electrocardiogram signal to be analyzed based on a preset waveform recognition method;

[0113] A signal reconstruction module 303, configured to reconstruct the electrocardiogram signal to be analyzed based on a preset reconstruction method according to the QRS wave start point set and the end point set to obtain a reconstructed electrocardiogram signal;

[0114] A detection module 304, configured to determine a characteristic value of the reconstructed electrocardiogram signal based on a delay algorithm, and determine whether there is an atrial fibrillation signal in the reconstructed electrocardiogram signal according to the characteristic value and a preset judgment threshold.

[0115] An atrial fibrillation signal detection method and device disclosed by the present invention, first, acquire an original electrocardiogram signal, pre - process the original electrocardiogram signal to obtain an electrocardiogram signal to be analyzed; second, use a preset waveform recognition method to determine the start and end point sets of QRS waves and reconstruct the electrocardiogram signal; finally, determine the characteristic value of the reconstructed electrocardiogram signal based on a delay algorithm, and detect the atrial fibrillation signal according to the characteristic value and a preset judgment threshold.

[0116] The method of the present invention extracts the QRS wave in the electrocardiogram signal, reconstructs the electrocardiogram signal using a preset reconstruction method, determines the characteristic value of the reconstructed signal based on a delay algorithm, and detects the atrial fibrillation signal according to the characteristic value and a preset judgment threshold. This can shorten the delay time of atrial fibrillation signal detection, solves the problems of low calculation amount and efficiency in the prior art, and realizes real-time detection of atrial fibrillation signals.

[0117] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A method for detecting atrial fibrillation signals, characterized in that: include: Acquire an original electrocardiogram signal, and preprocess the original electrocardiogram signal to obtain an electrocardiogram signal to be analyzed; Based on a preset waveform recognition method, determining a set of QRS wave starting points and a set of end points in the electrocardiogram signal to be analyzed; Reconstructing the electrocardiogram signal to be analyzed according to the QRS wave starting point set and the end point set based on a preset reconstruction method to obtain an electrocardiogram reconstruction signal; Determine a characteristic value of the ECG reconstruction signal based on a time delay algorithm, and determine whether there is an atrial fibrillation signal in the ECG reconstruction signal according to the characteristic value and a preset judgment threshold; Reconstructing the electrocardiogram signal to be analyzed according to the QRS wave starting point set and the end point set based on a preset reconstruction method to obtain an electrocardiogram reconstruction signal, including: Eliminate the QRS wave in the electrocardiographic signal to be analyzed according to the QRS wave starting point set and the end point set to obtain a plurality of SQ interval signals; Perform signal complementation on each of the SQ interval signals using a preset complementation method to obtain a complementation signal corresponding to each of the SQ interval signals; Performing high-pass filtering and smoothing processing on the completed signal to obtain a signal to be spliced; Splicing two adjacent signals to be spliced ​​to obtain an ECG reconstruction signal; Using a preset complement method to complete each of the SQ interval signals to obtain a complement signal corresponding to each of the SQ interval signals, including: For any SQ interval signal SQ[n], if SQ[n][1] is the amplitude of the first sampling point of the SQ interval signal SQ[n], a waveform with an amplitude of SQ[n][1] and a preset completion length is added to the head of SQ[n] to obtain the completion signal corresponding to SQ[n].

2. The atrial fibrillation signal detection method according to claim 1, characterized in that: Preprocessing the original ECG signal to obtain a signal to be analyzed includes: Performing low-pass filtering on the original ECG signal to obtain a first filtered ECG signal; Performing high-pass filtering on the first filtered ECG signal to obtain a second filtered ECG signal; The second filtered ECG signal is subjected to power frequency filtering to obtain a signal to be analyzed.

3. The atrial fibrillation signal detection method according to claim 1, characterized in that: The completed signal is subjected to high-pass filtering and smoothing to obtain a signal to be spliced, including: Performing high-pass filtering on the complement signal to obtain a filtered complement signal; The waveform of the preset completion length at the header of the filtered completion signal is deleted to obtain the signal to be spliced.

4. The atrial fibrillation signal detection method according to claim 1, characterized in that: Splicing two adjacent signals to be spliced ​​to obtain an ECG reconstruction signal, including: For the adjacent first signal to be spliced ​​SQ2[n] and the second signal to be spliced ​​SQ2[n+1], determine whether there is a sampling point whose amplitude difference is less than a preset search threshold within a preset splicing length range between the tail of SQ2[n] and the head of SQ2[n+1]; When there is a sampling point whose amplitude difference between the tail of SQ2[n] and the head of SQ2[n+1] is less than the preset search threshold within the preset splicing length range, SQ2[n] and SQ2[n+1] are spliced ​​with the preset splicing length to obtain an ECG reconstruction signal; When there are no sampling points whose amplitude difference between the tail of SQ2[n] and the head of SQ2[n+1] is less than the preset search threshold within the preset splicing length range, the tail of SQ2[n] and the head of SQ3[n+1] are fitted respectively and the splicing is completed.

5. The atrial fibrillation signal detection method according to claim 1, characterized in that: Determining the characteristic value of the electrocardiographic reconstruction signal based on a time delay algorithm includes: Extracting a signal of a preset duration at the tail of the electrocardiographic reconstruction signal to obtain a target signal; Dividing the target signal into a first characteristic signal and a second characteristic signal in chronological order; wherein the duration of the second characteristic signal is shorter than the duration of the first characteristic signal; The first characteristic signal and the second characteristic signal are processed based on a delay algorithm to obtain a first characteristic value and a second characteristic value.

6. The atrial fibrillation signal detection method according to claim 5, characterized in that: Judging whether there is an atrial fibrillation signal in the electrocardiogram reconstruction signal according to the characteristic value and a preset judgment threshold includes: According to the first characteristic value, the second characteristic value and the preset judgment threshold, it is judged whether the target signal, the first characteristic signal and the second characteristic signal are atrial fibrillation signals respectively.

7. The atrial fibrillation signal detection method according to claim 1, characterized in that: The preset waveform recognition method is a differential threshold method.

8. An atrial fibrillation signal detection device, characterized in that: include: A data acquisition module is used to acquire the original ECG signal and pre-process the original ECG signal to obtain the ECG signal to be analyzed; A waveform recognition module, used to determine a set of QRS wave starting points and a set of end points in the electrocardiogram signal to be analyzed based on a preset waveform recognition method; A signal reconstruction module, used to reconstruct the electrocardiogram signal to be analyzed according to the QRS wave starting point set and the end point set based on a preset reconstruction method to obtain an electrocardiogram reconstruction signal; A detection module, used to determine a characteristic value of the ECG reconstruction signal based on a time delay algorithm, and determine whether an atrial fibrillation signal exists in the ECG reconstruction signal according to the characteristic value and a preset judgment threshold; Reconstructing the electrocardiogram signal to be analyzed according to the QRS wave starting point set and the end point set based on a preset reconstruction method to obtain an electrocardiogram reconstruction signal, including: Eliminate the QRS wave in the electrocardiographic signal to be analyzed according to the QRS wave starting point set and the end point set to obtain a plurality of SQ interval signals; Perform signal complementation on each of the SQ interval signals using a preset complementation method to obtain a complementation signal corresponding to each of the SQ interval signals; Performing high-pass filtering and smoothing processing on the completed signal to obtain a signal to be spliced; Splicing two adjacent signals to be spliced ​​to obtain an ECG reconstruction signal; Using a preset complement method to complete each of the SQ interval signals to obtain a complement signal corresponding to each of the SQ interval signals, including: For any SQ interval signal SQ[n], if SQ[n][1] is the amplitude of the first sampling point of the SQ interval signal SQ[n], a waveform with an amplitude of SQ[n][1] and a preset completion length is added to the head of SQ[n] to obtain the completion signal corresponding to SQ[n].

Images