A non-contact electrocardiogram-based respiratory disorder evaluation system and method

Abstract

The application provides a respiratory disorder evaluation system and method based on non-contact electrocardio, wherein the system comprises a non-contact electrocardio acquisition device, a non-contact electrocardio signal quality evaluation module, a respiratory signal low-frequency energy threshold discrimination module and a non-contact electrocardio respiratory signal extraction and respiratory disorder evaluation module. The system can acquire the electrocardio and respiratory coupling signal of a patient through a layer of clothes. Compared with the traditional contact wet electrode measurement method, the non-contact measurement method has the advantages of long-term monitoring and high comfort. The signal characteristics can be extracted according to the characteristics of the non-contact electrocardio signal, and the method has the advantage of high accuracy. The method uses different respiratory signal extraction methods for two cases. After the respiratory signal extraction is completed, the respiratory disorder event is judged. Based on the characteristics of the non-contact electrocardio, the signal processing related tasks can be efficiently and accurately completed.

Description

Technical Field

[0001] This invention belongs to the field of respiratory disorder monitoring, and relates to non-contact electrocardiogram and respiratory disorder monitoring technology, physiological signal processing technology, especially a non-contact respiratory disorder monitoring method that is imperceptible, specifically involving a respiratory disorder assessment system and method based on non-contact electrocardiogram. Background Technology

[0002] Currently, people's understanding of respiratory disorders is gradually deepening. Respiratory disorders refer to the patient's feeling of insufficient air, difficulty breathing, rapid and deep breathing movements, and in severe cases, open-mouth breathing, nasal flaring, and cyanosis. Clinically, based on the pathogenesis, they are divided into three categories: 1. Inspiratory dyspnea, characterized by difficulty inhaling, with the presence of three-recession signs (recession in the suprasternal notch, supraclavicular fossa, subxiphoid process, and even intercostal spaces), accompanied by dry cough and high-pitched inspiratory stridor; 2. Expiratory dyspnea, characterized by difficulty exhaling, prolonged expiratory phase, often accompanied by dry rales or wheezing, mainly due to weakened alveolar elastic recoil and widespread narrowing of small airways (inflammation or spasm), mainly seen in lung diseases; 3. Mixed dyspnea, mainly caused by left and / or right heart failure, with obstruction in both the upper and lower respiratory tracts, is relatively rare, mainly seen in older patients with concurrent systemic diseases, including not only upper respiratory tract lesions but also lower respiratory tract lesions such as COPD. For example, sleep apnea is an abnormal breathing event that occurs during sleep, which can cause various cardiovascular and cerebrovascular diseases and is also related to metabolic syndrome, seriously affecting patients' quality of life and lifespan. Currently, approximately 100 million people worldwide suffer from sleep apnea syndrome. Data shows that there are approximately 50 million people with sleep apnea in my country. Therefore, developing a respiratory disorder assessment system is particularly important. The main pathophysiological changes in sleep apnea include recurrent apnea during sleep, hypoxemia and / or hypercapnia caused by hypoventilation, and changes in sleep structure. These changes cause a series of clinical manifestations and damage to multiple organ functions. With the development of wearable technology, more and more people are using wearable devices to monitor cardiopulmonary function. Electrocardiography (ECG) is a mainstream measurement technique used to monitor cardiovascular diseases. In addition, research on extracting respiratory data from ECGs using cardiopulmonary coupling technology has also attracted much attention from researchers. However, long-term ECG monitoring using traditional wet electrodes can lead to problems such as skin redness and swelling in patients. With the development of science and technology, non-contact ECG acquisition using capacitive coupling technology has been widely studied. Non-contact ECG systems can collect ECG signals through a layer of clothing, thus solving the problem of skin allergies caused by long-term acquisition of traditional contact ECG signals. Due to the principle of capacitive coupling, non-contact ECG has stronger respiratory-motor coupling compared to traditional contact ECG, thus showing good application prospects for respiratory monitoring. However, there is no research on cardiopulmonary coupling technology for non-contact ECG.

[0003] Richardsons was one of the earliest researchers to conduct non-contact ECG research. He designed an insulated electrode based on the principle of capacitive coupling, consisting of an alumina metal plate and a follower, which is placed directly on the skin for measurement. However, due to the electrode's insulating properties, it differs fundamentally from traditional wet electrode ECG monitoring technology. Researchers at the University of Sussex in the UK have built a non-contact ECG measurement electrode model with a high input impedance capacitor and a low output impedance follower. The US Center for Quantum Applied Sciences and Research has developed a novel non-contact electrode made of a low-conductivity material, which can measure ECG signals through clothing. Researchers including Ko Keun Kim from Seoul National University in South Korea designed a non-contact ECG signal acquisition toilet. A square non-contact ECG signal acquisition electrode is placed on the toilet, with a 0.1 mm thick polytetrafluoroethylene (PTFE) film attached to the electrode. The ECG signal is coupled to the electrode through the PTFE film. Researchers at the University of California integrated a high input impedance follower, a shield, a differential amplifier, and a probe layer onto a PCB board. The probe layer, composed of copper sheets coated with insulating material, couples out the human body's electrocardiogram (ECG) signals. This chip enables non-contact ECG measurement and obtains high-fidelity ECG signals. Shuhei Ishida et al. at Ritsumeikan University in Japan designed a non-contact ECG signal acquisition system. This system places a set of strip-shaped non-contact ECG signal acquisition electrodes on a bed, allowing for relatively ideal ECG signal acquisition when the person is lying supine. Researchers at Daimler AG's R&D center in Germany, including BK Chamadiya, designed a non-contact ECG signal acquisition system mounted on a car seat. This system can acquire ECG signals while driving, but the bumps and vibrations during vehicle movement cause significant interference noise.

[0004] In recent years, many domestic universities and research institutions have also conducted research on non-contact ECG measurement technology. Researchers at the Chinese University of Hong Kong designed a novel non-contact ECG measurement chair. This system measures ECG signals via electronic fabric coupling on the seat cushion while the subject remains seated, without interfering with driving, office work, or other activities. Zhou Ping, Wang Feng, and others at Southeast University developed a non-contact ECG signal acquisition electrode based on a printed circuit board, which can acquire relatively ideal ECG signals. Yang Bin and others at Tsinghua University, based on a non-contact ECG electrode based on a printed circuit board, studied the use of conductive fibers and conductive rubber as electrode surfaces. Experimental results showed that using conductive fibers and conductive rubber as electrode surfaces resulted in lower quality ECG signals compared to using copper-plated or tin-plated electrodes, but still yielded important ECG information. Researchers at Chongqing University studied the electromagnetic interference problem of non-contact ECG acquisition systems, and their research results showed that the proposed anti-interference measures achieved good shielding effects.

[0005] Chinese patent CN110710957A discloses a non-contact real-time electrocardiogram (ECG) monitoring device, including capacitively coupled non-contact ECG electrodes, an ECG signal processing module, an analog-to-digital conversion module, and a wireless transmission module. It can collect ECG signals and display them in real-time on a smart terminal via wireless transmission. However, this method does not include the extraction and processing of respiratory signals and cannot handle respiratory disturbance events.

[0006] Chinese patent CN110742585A discloses a sleep staging method based on BCG signals. This method uses heart rate, pulse, and respiratory signals extracted from BCG signals to perform sleep staging by extracting correlation features of heart rate variability and cardiopulmonary coupling power spectrum. However, this method only considers sleep staging and does not utilize respiratory signals to determine respiratory disturbances. Furthermore, it uses oscillation signals, which lack the accuracy of non-contact electrocardiography.

[0007] Chinese patent CN105982664B discloses a cardiopulmonary coupling analysis method based on single-lead ECG. This method extracts the RR interval sequence and EDR signal from the single-lead ECG signal and then uses adaptive filtering, empirical mode decomposition, and other methods to perform cardiopulmonary coupling analysis. However, this method uses a contact-type single-lead ECG monitoring device, which, compared to non-contact measurement, has problems such as difficulty in long-term monitoring and insufficient wearing comfort. Summary of the Invention

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] This invention provides a respiratory disorder assessment system based on non-contact electrocardiography, characterized by comprising a non-contact electrocardiogram acquisition device, a non-contact electrocardiogram signal quality assessment module, a respiratory signal low-frequency energy threshold discrimination module, and a non-contact electrocardiogram respiratory signal extraction and assessment module.

[0010] The non-contact ECG acquisition device is used to acquire non-contact ECG signals from the human body for subsequent cardiopulmonary coupling analysis.

[0011] The non-contact ECG signal quality assessment module is used to evaluate the quality of non-contact ECG signals and remove invalid data, including strong power frequency interference, motion artifacts, and other data with no obvious ECG signal.

[0012] The low-frequency energy threshold discrimination module for respiratory signals is used to evaluate the degree of respiratory coupling in non-contact ECG using a time-frequency domain method.

[0013] The non-contact ECG respiratory signal extraction and respiratory disorder assessment module is used to extract respiratory signals from non-contact ECGs using an adaptive respiratory extraction algorithm based on the degree of respiratory coupling of the non-contact ECG, and to assess whether there is a respiratory disorder event.

[0014] Furthermore, the aforementioned non-contact ECG acquisition device is used to acquire non-contact ECG signals for subsequent cardiopulmonary coupling analysis. It comprises a high-impedance analog front-end, fabric electrodes, and a differential instrumentation amplifier. This device can acquire the patient's ECG and respiratory coupling signals through a layer of clothing.

[0015] Furthermore, the non-contact ECG signal quality assessment module uses five ECG signal quality indicators, including but not limited to kSQI, bsSQI, psd_SQI, bsd_SQI, and bSQI, and a random forest method to assess signal quality, remove invalid data with no obvious ECG signal, and retain high-quality non-contact ECG data for subsequent respiratory disorder assessment.

[0016] Furthermore, the respiratory signal low-frequency energy threshold discrimination module includes a signal preprocessing module and a respiratory low-frequency signal threshold evaluation module. The signal preprocessing module is used to preprocess the non-contact ECG signal; the respiratory low-frequency signal threshold evaluation module is used to evaluate the energy of the respiratory coupling waveform in the non-contact ECG.

[0017] Furthermore, the non-contact ECG respiratory signal extraction and respiratory disorder assessment module includes a non-contact ECG respiratory signal extraction module and a respiratory disorder assessment module. The non-contact ECG respiratory signal extraction module is used to adaptively extract respiratory signals from non-contact ECG signals with different degrees of respiratory coupling. The respiratory disorder assessment module is used to assess whether respiratory disorder events exist in the extracted respiratory signals.

[0018] Furthermore, the signal preprocessing module is used to perform preliminary filtering of the ECG signal using an 8th-order bandpass filter of 0.1–35 Hz to remove baseline drift and power line interference, while retaining the ECG and respiratory components. The respiratory low-frequency signal threshold evaluation module uses a 60-second sliding window method to calculate the respiratory low-frequency energy of the non-contact ECG signal. The calculation formula for the respiratory low-frequency energy of the non-contact ECG is measured by the ratio of the respiratory low-frequency energy (0.1–0.8 Hz) to the main energy of the non-contact ECG QRS wave (5–15 Hz). Finally, the non-contact ECG signal was evaluated based on the calculation results and thresholds.

[0019] Furthermore, the respiratory signal extraction module within the non-contact ECG respiratory signal extraction and respiratory disorder assessment module adaptively extracts respiratory signals based on a low-frequency energy threshold of the non-contact ECG signal. When the low-frequency energy is greater than the threshold, a Butterworth bandpass filter is used for respiratory signal extraction, with the filtering frequency set between 0.1Hz and 0.8Hz. When the low-frequency energy is less than the threshold, a respiratory signal extraction method based on ECG features is used to extract signal features such as QRS complex area and R wave position, obtaining the respiratory signal through cubic spline interpolation and a low-pass filter. The respiratory disorder assessment module assesses whether respiratory disorder events exist in the extracted respiratory signals, primarily utilizing the RR interval and respiratory signal features, and employing a support vector machine for classification.

[0020] Furthermore, a non-contact ECG-based method for assessing respiratory distress includes the following steps:

[0021] Step 1: Use a non-contact ECG acquisition device to acquire non-contact ECG signals. The acquisition device consists of a high-impedance analog front end, fabric electrodes, and a differential instrumentation amplifier, which can acquire the patient's ECG and respiratory coupling signals through a layer of clothing.

[0022] Step 2: Based on the collected non-contact ECG signals, the non-contact ECG signal quality is evaluated using a non-contact ECG signal quality assessment module. Based on five signal quality indicators—kSQI, bsSQI, psd_SQI, bsd_SQI, and bSQI—the ECG signal quality is assessed using a random forest method, and invalid data with no obvious ECG signal are removed.

[0023] Step 3: The respiratory signal low-frequency energy threshold discrimination module is used to preprocess the signal. An 8th-order bandpass filter (0.1–35 Hz) is used to initially filter the ECG signal to remove baseline drift and power line interference, while retaining the ECG and respiratory components. Next, a respiratory low-frequency signal threshold assessment is performed. A 60-second sliding window method is used to calculate the respiratory low-frequency energy of the non-contact ECG signal. The formula for calculating the respiratory low-frequency energy of the non-contact ECG is based on the ratio of the respiratory low-frequency energy (0.1–0.8 Hz) to the main energy of the non-contact ECG QRS wave (5–15 Hz). Finally, the non-contact ECG signal was evaluated based on the calculation results and thresholds.

[0024] Step 4: Using the non-contact ECG-based respiratory signal extraction and respiratory disorder assessment module, respiratory signals are first adaptively extracted based on the low-frequency energy threshold of the non-contact ECG signal. When the low-frequency energy is greater than the threshold, a Butterworth bandpass filter is used to extract the respiratory signal, with the filter frequency set between 0.1Hz and 0.8Hz. When the low-frequency energy is less than the threshold, a respiratory signal extraction method based on ECG features is used to extract signal features such as QRS complex area and R wave position, and the respiratory signal is obtained through cubic spline interpolation and a low-pass filter. Then, using the RR interval and respiratory signal features, a support vector machine is used for classification to determine whether there are respiratory disorder events in the extracted respiratory signals.

[0025] Compared with existing technologies, the present invention has the following advantages:

[0026] (1) The non-contact ECG acquisition device of this system can collect the patient's ECG and respiratory coupling signals through a layer of clothing. Compared with the traditional contact wet electrode measurement method, this non-contact measurement method has the advantages of long-term monitoring and high comfort.

[0027] (2) The non-contact ECG signal quality assessment algorithm of this system extracts signal features based on the characteristics of non-contact ECG signals, and has the advantage of high accuracy.

[0028] (3) The respiratory signal low-frequency energy threshold discrimination module of this system preprocesses the non-contact electrocardiogram signal and evaluates the respiratory low-frequency signal threshold. Combining the characteristics of the non-contact electrocardiogram signal, it considers using the frequency range corresponding to the signal, calculates the power spectral density distribution index of the low-frequency energy full band, and completes the respiratory low-frequency energy evaluation based on the index.

[0029] (4) The non-contact ECG respiratory signal extraction and respiratory disorder assessment module of this system uses an adaptive algorithm. Based on the low-frequency energy threshold of non-contact ECG, the non-contact ECG signal is divided into two categories for processing. Different respiratory signal extraction methods are used for the two cases. After the respiratory signal extraction is completed, the respiratory disorder event is judged. This adaptive algorithm, based on the characteristics of non-contact ECG, can efficiently and accurately complete signal processing related tasks. Attached Figure Description

[0030] Figure 1 This is the overall flowchart of the present invention.

[0031] Figure 2 This is a flowchart of the signal preprocessing process in this invention.

[0032] Figure 3 This is a flowchart of the respiratory signal extraction and respiratory disorder assessment algorithm in this invention.

[0033] Figure 4 This is a block diagram of the respiratory disorder assessment system based on non-contact electrocardiography in this invention.

[0034] Figure 5 It is the result of non-contact ECG and respiratory waveform extraction when the low-frequency respiratory energy is less than the threshold.

[0035] Figure 6 It is the result of non-contact ECG and respiratory waveform extraction when the low-frequency respiratory energy is greater than the threshold.

[0036] Figure 7 The results are based on waveforms for identifying respiratory disorders using non-contact electrocardiography. Specific implementation methods

[0037] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are for illustrative purposes only and are not intended to limit the scope of the invention.

[0038] As shown in the figure, this invention provides a respiratory disorder assessment system based on non-contact electrocardiography (ECG), including a non-contact ECG acquisition device, a non-contact ECG signal quality assessment module, a low-frequency energy threshold discrimination module for respiratory signals, and a non-contact ECG respiratory signal extraction and assessment module. The non-contact ECG acquisition device comprises a high-impedance analog front-end, fabric electrodes, and a differential instrumentation amplifier. This device can acquire the patient's ECG and respiratory coupling signals through a layer of clothing. The non-contact ECG signal quality assessment module uses ECG signal quality indicators and a random forest method to assess signal quality and remove invalid data with no obvious ECG signal. The low-frequency energy threshold discrimination module for respiratory signals includes a signal preprocessing module and a low-frequency respiratory signal threshold assessment module. It preprocesses the non-contact ECG signal before assessing the energy of the respiratory coupling waveform in the non-contact ECG signal. The non-contact ECG respiratory signal extraction and respiratory disorder assessment module includes a non-contact ECG respiratory signal extraction module and a respiratory disorder assessment module. It adaptively extracts respiratory signals from non-contact ECG signals with different degrees of respiratory coupling and determines whether respiratory disorder exists in the respiratory signal.

[0039] The item was evaluated.

[0040] This mainly includes a non-contact ECG acquisition device, a non-contact ECG signal quality assessment module, a low-frequency energy threshold discrimination module for respiratory signals, and a non-contact ECG respiratory signal extraction and assessment module.

[0041] The non-contact ECG acquisition device consists of a high-impedance analog front end, fabric electrodes, and a differential instrumentation amplifier, and can acquire the patient's ECG and respiratory coupling signals through a layer of clothing.

[0042] Secondly, based on the signal quality assessment module for non-contact ECG, the signal quality of the collected non-contact ECG signals is assessed. Based on five signal quality indicators, kSQI, bsSQI, psd_SQI, bsd_SQI, and bSQI, the random forest method is used to complete the ECG signal quality assessment and remove invalid data with no obvious ECG signal.

[0043] kSQI: This index is used to assess the significance of the QRS complex. It is the fourth moment (kurtosis) of the cECG signal distribution, defined as follows: Where, x i It is a cECG signal with N sample points, where μ is the signal mean and σ is the standard deviation.

[0044] bsSQI: This index is primarily used to assess the impact of baseline drift (<1Hz) on cECG, and is defined as follows: Where Rai is the peak-to-peak amplitude of the cECG waveform surrounding each QRS wave (from R-0.07s to R+0.08s), R is the baseline value of the QRS wave, i is the QRS wave detected in the analysis window, and Ba... i It is the peak-to-peak amplitude of the baseline around each QRS wave (from R-1s to R+1s).

[0045] PSD_SQI: This index is the power spectral density distribution of the QRS wave across the entire frequency band, primarily used to assess the interference of PLI on cECG. High-quality cECG signals have a distinguishable QRS wave, with the energy of the QRS wave mainly concentrated in the 5-15Hz range. To assess the impact of PLI on the signal, a frequency range (5-125Hz) is considered. Therefore, PSD_SQI is defined as: Where p(f) is the autoregressive model spectrum, and the Burg algorithm is used for parameter estimation.

[0046] bsd_SQI: This index is the power spectral density distribution index for baseline drift across the entire frequency band, primarily used to assess the interference of baseline drift on cECG. This index can be defined as: Where p(f) is the autoregressive model spectrum, and the Burg algorithm is used for parameter estimation.

[0047] bSQI: This metric is the ratio of the number of R-peaks detected by the PT (Pantompkin) algorithm to the number detected by the WQRS algorithm. For high-quality cECG signals, the number of R-peaks obtained by different QRS wave detection algorithms should be approximately the same; otherwise, there will be significant differences in the number of R-peaks. The PT (Pantompkin) algorithm is insensitive to noise, while the WQRS algorithm is sensitive to noise. Therefore, this metric can be used to evaluate the difference in signal noise sensitivity, defined as follows: Where N PT N represents the number of R peaks detected by the PT (pantompkin) algorithm. wqrs This represents the number of R peaks detected by the wqrs algorithm.

[0048] Random forests can identify key correlation variables from high-dimensional datasets with multiple predictor variables exhibiting complex interactions and weak main effect variables. They are characterized by simple implementation, fast training speed, strong robustness, and high classification accuracy. Combining the aforementioned five signal quality metrics with the random forest algorithm, it is possible to perform electrocardiogram (ECG) signal quality assessment and eliminate invalid data lacking significant ECG signals.

[0049] Then, based on the respiratory signal low-frequency energy threshold discrimination module, the signal is preprocessed. An 8th-order bandpass filter (0.1–35 Hz) is used to initially filter the ECG signal to remove baseline drift and power line interference, while retaining the ECG and respiratory components. A respiratory low-frequency signal threshold assessment is then performed. A 60-second sliding window method is used to calculate the respiratory low-frequency energy of the non-contact ECG signal. The formula for calculating the respiratory low-frequency energy of the non-contact ECG is based on the ratio of the respiratory low-frequency energy (0.1–0.8 Hz) to the main energy of the non-contact ECG QRS wave (5–15 Hz). Finally, the non-contact ECG signal was evaluated based on the calculation results and thresholds.

[0050] Finally, based on the non-contact ECG respiratory signal extraction and respiratory disorder assessment module, respiratory signals are first adaptively extracted according to the low-frequency energy threshold of the non-contact ECG signal. When the low-frequency energy is greater than the threshold, a Butterworth bandpass filter is used to extract the respiratory signal, with the filtering frequency set between 0.1Hz and 0.8Hz. When the low-frequency energy is less than the threshold, a respiratory signal extraction method based on ECG features is adopted. This involves integrating the signal within the QRS wave range of the ECG signal and placing the calculated value at the corresponding R-wave peak position. The feature values ​​of the R-wave peak position in each heartbeat form a new time series, with the horizontal axis representing the R-wave peak position and the vertical axis representing the corresponding QRS area. Then, cubic spline interpolation is used to re-expand the time series signal of the feature values ​​to a signal with the same length and sampling rate as the original signal. Finally, a low-pass filter with a cutoff frequency of 5Hz is used to filter the expanded signal to obtain the cardiopulmonary coupling signal. Next, using the features of the RR interval and respiratory signal, a support vector machine is used for classification to determine whether there are respiratory disorder events in the extracted respiratory signals.

Claims

1. A respiratory disorder assessment system based on non-contact electrocardiography, characterized in that: It includes a non-contact ECG acquisition device, a non-contact ECG signal quality assessment module, a low-frequency energy threshold discrimination module for respiratory signals, and a non-contact ECG respiratory signal extraction and assessment module; The non-contact ECG acquisition device is used to acquire non-contact ECG signals and perform subsequent cardiopulmonary coupling analysis. The non-contact ECG signal quality assessment module is used to assess the quality of non-contact ECG signals and remove invalid data, including strong power frequency interference and motion artifacts, that do not have obvious ECG signals. The low-frequency energy threshold discrimination module for respiratory signals is used to evaluate the degree of respiratory coupling in non-contact ECG using a time-frequency domain method. The non-contact ECG respiratory signal extraction and respiratory disorder assessment module is used to extract respiratory signals from non-contact ECGs using an adaptive respiratory extraction algorithm based on the degree of respiratory coupling of the non-contact ECG, and to assess whether there is a respiratory disorder event. The respiratory signal extraction module in the non-contact ECG respiratory signal extraction and respiratory disorder assessment module is used to adaptively extract respiratory signals based on the low-frequency energy threshold of the non-contact ECG signal. When the low-frequency energy of the respiratory signal is greater than the threshold, a Butterworth bandpass filter is used to extract the respiratory signal, with the filtering frequency set between 0.1Hz and 0.8Hz. When the low-frequency energy of the respiratory signal is less than the threshold, a respiratory signal extraction method based on ECG characteristics is used to extract the QRS area and R-wave position signal features, and the respiratory signal is obtained through cubic spline interpolation and a low-pass filter. The respiratory disorder assessment module is used to assess whether there are respiratory disorder events in the extracted respiratory signals. It uses support vector machines to classify and judge the events based on the RR interval and the characteristics of the respiratory signals.

2. The respiratory disorder assessment system based on non-contact electrocardiography according to claim 1, characterized in that, The non-contact ECG acquisition device includes a high-impedance analog front end, fabric electrodes, and a differential instrumentation amplifier. This device can acquire the patient's ECG and respiratory coupling signals through a layer of clothing.

3. The respiratory disorder assessment system based on non-contact electrocardiography according to claim 1, characterized in that, The non-contact ECG signal quality assessment module extracts time-frequency domain indicators, including but not limited to five signal quality indicators: kSQI, bsSQI, psd_SQI, bsd_SQI, and bSQI. It uses a random forest method to complete the ECG signal quality assessment, removes invalid data with no obvious ECG signal, and retains the non-contact ECG data with better quality for subsequent respiratory disorder assessment.

4. The respiratory disorder assessment system based on non-contact electrocardiography according to claim 1, characterized in that, The respiratory signal low-frequency energy threshold discrimination module includes a signal preprocessing module and a respiratory low-frequency signal threshold evaluation module; the signal preprocessing module is used to preprocess non-contact electrocardiogram signals. The respiratory low-frequency signal threshold assessment module is used to assess the energy of respiratory coupling waveforms in non-contact electrocardiography.

5. A respiratory disorder assessment system based on non-contact electrocardiography according to claim 1, characterized in that, The non-contact ECG respiratory signal extraction and respiratory disorder assessment module includes a non-contact ECG respiratory signal extraction module and a respiratory disorder assessment module; the non-contact ECG respiratory signal extraction module is used to adaptively extract respiratory signals from non-contact ECG signals with different respiratory coupling degrees. The respiratory distress assessment module is used to assess whether respiratory distress events exist in the extracted respiratory signals.

6. A respiratory disorder assessment system based on non-contact electrocardiography according to claim 4, characterized in that, The signal preprocessing module uses an 8th-order bandpass filter (0.1~35Hz) to perform preliminary filtering on the ECG signal to remove baseline drift and power line interference, while retaining ECG and respiratory components. The respiratory low-frequency signal threshold evaluation module uses a 60-second sliding window method to calculate the respiratory low-frequency energy of the non-contact ECG signal. The calculation formula for the respiratory low-frequency energy of the non-contact ECG is measured by the ratio of the respiratory low-frequency energy to the main energy of the non-contact ECG QRS wave. Finally, the non-contact ECG signal is evaluated based on the calculation results and thresholds.

Images