High-precision pulse wave period cutting algorithm

Abstract

The invention discloses a high-precision pulse wave period cutting algorithm. The algorithm comprises the following steps of 1, signal preprocessing, wherein moving average filtering preprocessing is conducted on an original pulse wave signal; step 2, improved extreme value detection: adopting an AMPD algorithm and combining with a maximum value and a minimum value in an adhesion check detection signal; step 3, dynamic secondary subdivision strategy: based on a preset time interval threshold value and local peak detection, performing dynamic secondary subdivision and combination on the suspected multi-cut region; step 4, baseline drift correction and time normalization: performing baseline drift correction and time normalization processing on each period obtained by cutting; according to the invention, high-precision pulse wave period cutting is realized, and the problem of excessive cutting is avoided; the accuracy of the period boundary can be kept under the condition of obvious noise interference and baseline drift; and stable and standardized periodic data are provided for subsequent physiological signal feature extraction and intelligent diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical field of physiological signal processing, in particular to a high-precision pulse wave cycle cutting algorithm. Background Art

[0002] Currently, no standard algorithm for precisely segmenting pulse wave cycles has been found in publicly available literature. However, within the specialized field of physiological signal processing, researchers typically employ a range of techniques, such as filtering, extreme value detection, and trend analysis, to effectively segment pulse wave signals. While these traditional methods have a proven track record in signal processing, they often struggle to simultaneously meet the dual requirements of high-precision segmentation and strong anti-interference performance. This is particularly true when faced with complex situations such as high signal noise or baseline drift, which can lead to inaccurate segmentation of pulse wave cycles, compromising subsequent signal analysis and processing. Summary of the Invention

[0003] The purpose of the present invention is to provide a high-precision pulse wave cycle cutting algorithm to solve the problems raised in the above background technology.

[0004] To achieve the above object, the present invention provides the following technical solution: a high-precision pulse wave cycle cutting algorithm, comprising the following steps:

[0005] Step 1: Signal preprocessing: Perform moving average filtering on the original pulse wave signal;

[0006] Step 2: Improve extreme value detection: Use the AMPD algorithm combined with adhesion inspection to detect the maximum and minimum values in the signal;

[0007] Step 3: Dynamic secondary segmentation strategy: Based on the preset time interval threshold and local peak detection, the suspected multi-cut areas are dynamically subdivided and merged;

[0008] Step 4: Baseline drift correction and time normalization: perform baseline drift correction and time normalization on each cycle obtained by cutting.

[0009] Preferably, after the moving average filtering preprocessing in step 1, key signal features are retained.

[0010] Preferably, the AMPD algorithm locates the extreme points in the signal, performs morphological operations on the signal by constructing a structural element of length, and identifies local maxima and minima, where the value range is 1 / 3-1 / 2 of the average period of the pulse wave; the adhesion check is specifically to calculate the time interval between adjacent extreme points. If the preset minimum interval threshold, preferably 350ms, is exceeded, it is determined to be an adhesion extreme value and needs to be corrected through subsequent dynamic subdivision strategy.

[0011] Preferably, in step three, specifically, in the case where the time interval between adjacent extreme points is too short, the suspected multi-cut areas are subdivided and merged by detecting the local peak distribution and slope changes to ensure the integrity of each pulse cycle; the dynamic secondary subdivision strategy is based on the local rising or falling slope and peak distribution.

[0012] Preferably, the step four specifically includes performing baseline compensation on each cutting cycle to eliminate the influence of low-frequency drift, and normalizing the cycle data to facilitate subsequent feature extraction and analysis.

[0013] Compared with the prior art, the present invention has the following beneficial effects:

[0014] The present invention achieves high-precision pulse wave cycle cutting, avoiding the problem of over-cutting; can maintain the accuracy of cycle boundaries under conditions of noise interference and obvious baseline drift; and provides stable and standardized cycle data for subsequent physiological signal feature extraction and intelligent diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0016] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] See also Figure 1 The present invention provides a high-precision pulse wave cycle cutting algorithm, comprising the following steps:

[0018] Step 1: Signal preprocessing: Perform moving average filtering on the original pulse wave signal while retaining key signal features;

[0019] Step 2: Improve extreme value detection: The AMPD algorithm is combined with adhesion detection to detect the maximum and minimum values in the signal. The AMPD algorithm locates the extreme points in the signal and performs morphological operations on the signal by constructing a structure element of length , identifying local maxima and minima, where the value range is 1 / 3-1 / 2 of the average pulse wave period. The adhesion detection specifically calculates the time interval between adjacent extreme points. If the preset minimum interval threshold, preferably 350ms, is exceeded, it is determined to be an adhesion extreme value and needs to be corrected through subsequent dynamic subdivision strategies.

[0020] Step 3: Dynamic secondary segmentation strategy: Based on the preset time interval threshold and local peak detection, the suspected multi-cut areas are dynamically subdivided and merged. Specifically, when the time interval between adjacent extreme points is too short, the suspected multi-cut areas are subdivided and merged by detecting the local peak distribution and slope changes to ensure the integrity of each pulse cycle. The dynamic secondary segmentation strategy is implemented based on the local rising or falling slope and peak distribution.

[0021] Step 4: Baseline drift correction and time normalization: Baseline drift correction and time normalization are performed on each cycle obtained by cutting. Specifically, baseline compensation is performed on each cutting cycle to eliminate the influence of low-frequency drift, and the cycle data is uniformly normalized to facilitate subsequent feature extraction and analysis.

[0022] Example 1:

[0023] Preprocessing and extreme value detection:

[0024] Moving average filtering is applied to the original signal, and then the AMPD algorithm is used to detect local extreme values, and the preliminary detection results are reviewed by adhesion inspection.

[0025] Example 2:

[0026] Dynamic segmentation and baseline correction:

[0027] When multiple extreme values are detected within a cycle, a secondary subdivision strategy is used based on the local time interval and peak distribution to correct the multi-cut phenomenon;

[0028] Baseline drift correction is performed on each segmented cycle, and then the timestamp is normalized to form standard cycle data.

[0029] During specific use, the pre-processed pulse wave signal is input into the high-precision pulse wave cycle cutting algorithm to automatically complete the cutting and correction of the pulse wave cycle. Through the pre-processing and extreme value detection steps of Example 1, the noise in the original signal can be effectively filtered out, and the extreme value points in the pulse wave signal can be accurately identified. For possible adhesion extreme values, correction is performed through adhesion inspection and subsequent dynamic subdivision strategy to ensure the accuracy of the extreme value points. In Example 2, a dynamic subdivision strategy is used to correct the situation where there are multiple extreme values in the cycle, thereby avoiding the occurrence of multiple cutting phenomena. At the same time, baseline drift correction and time normalization processing are performed on each segmented cycle to obtain standard cycle data, which provides a reliable basis for subsequent feature extraction and analysis. The high-precision pulse wave cycle cutting algorithm has the advantages of simple operation and high cutting accuracy, and can be widely used in medical care, health monitoring and other fields.

[0030] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A high-precision pulse wave cycle cutting algorithm, characterized by: The following steps are involved: Step 1: Signal preprocessing: Perform moving average filtering on the original pulse wave signal; Step 2: Improve extreme value detection: Use the AMPD algorithm combined with adhesion inspection to detect the maximum and minimum values in the signal; Step 3: Dynamic secondary segmentation strategy: Based on the preset time interval threshold and local peak detection, the suspected multi-cut areas are dynamically subdivided and merged; Step 4: Baseline drift correction and time normalization: perform baseline drift correction and time normalization on each cycle obtained by cutting.

2. A high-precision pulse wave cycle cutting algorithm according to claim 1, characterized in that: After the moving average filtering preprocessing in step 1, key signal features are retained.

3. The high-precision pulse wave cycle cutting algorithm according to claim 1, characterized in that: The AMPD algorithm locates the extreme points in the signal and performs morphological operations on the signal by constructing a structure element of length L to identify the local maximum M = {m1, m2, ..., m n } and minimum value m={v1,v2,…,v m }, where L ranges from 1 / 3 to 1 / 2 of the average pulse wave period; adhesion inspection is specifically to calculate the time interval between adjacent extreme points Δt = |t i+1 -t i |, if Δt <T min , preset minimum interval threshold, preferably 350ms, is determined as an adhesion extreme value and needs to be corrected through subsequent dynamic segmentation strategy.

4. The high-precision pulse wave cycle cutting algorithm according to claim 1, characterized in that: Specifically, in step three, in the case where the time interval between adjacent extreme points is too short, the suspected multi-cut areas are subdivided and merged by detecting the local peak distribution and slope changes to ensure the integrity of each pulse cycle; the dynamic secondary subdivision strategy is based on the local rising or falling slope and peak distribution.

5. The high-precision pulse wave cycle cutting algorithm according to claim 1, characterized in that: The step 4 specifically includes performing baseline compensation on each cutting cycle to eliminate the influence of low-frequency drift, and normalizing the cycle data to facilitate subsequent feature extraction and analysis.

Images