44 results about "Counting respirations" patented technology

A respiratory motion plane alignment system and method based on ultrasound images is disclosed. The system acquires three-dimensional ultrasound images of respiratory organs at at least two different respiratory phases, calculates the motion direction of the respiratory organs based on the three-dimensional ultrasound images, controls and adjusts an ultrasound direction adjustment mechanism to align the scanning plane of the ultrasound probe with the motion direction of the respiratory organs, acquires a three-dimensional model image of the respiratory organs and multiple frames of real-time two-dimensional ultrasound images, performs registration between the multiple frames of real-time two-dimensional ultrasound images of the respiratory organs, and / or registers the three-dimensional model image and the multiple frames of real-time two-dimensional ultrasound images, acquires three-dimensional position data of a target point during respiratory motion, determines the tracking path of the ultrasound probe based on the three-dimensional position data of the target point during respiratory motion, and controls the ultrasound probe to move along the tracking path, thereby achieving long-term stable observation or tracking of target points in the respiratory organs.

The invention relates to the technical field of remote monitoring, in particular to a chronic respiratory disease patient remote monitoring system based on the Internet of Things, which comprises a physiological baseline initialization module, an event response delay measurement and calculation module, a cardiopulmonary coupling phase quantization module and a decompensation risk early warning module. According to the method, an individual blood oxygen and cardiopulmonary coupling health baseline is established to provide accurate reference for subsequent analysis, then the blood oxygen recovery time after apnea is measured and calculated, the physiological response delay characteristic of an individual is quantified, and the characteristic is used for setting a dynamic analysis window to continuously calculate the phase difference between a breathing mode and a blood oxygen signal, so that the phase difference between the breathing mode and the blood oxygen signal is calculated. By continuously comparing the current phase difference with the individual health baseline and triggering early warning when the difference value continuously exceeds a specific range, the early sign of the instability of the cardiopulmonary system can be effectively captured, the conversion from passive response type alarm to active predictive early warning is realized, and a precious time window is provided for preventing decompensation.

The invention relates to an anesthesia puncture visual navigation system combining augmented reality and three-dimensional reconstruction, and belongs to the technical field of anesthesia puncture visual navigation. The anesthesia puncture visual navigation system comprises a physiological movement monitoring unit used for simultaneously collecting respiratory movement cycle data and soft tissue elastic deformation information; a comprehensive motion state data packet is generated by calculating the correlation coefficient of the breathing amplitude and the tissue elasticity modulus change rate, the space deviation caused by breathing motion and soft tissue deformation of a patient to puncture navigation can be effectively compensated, the puncture precision is remarkably improved compared with a traditional static navigation system under the normal breathing state of the patient, and the puncture accuracy is improved. The puncture failure rate and the complication risk are effectively reduced, the problem that precision is further deteriorated due to the fact that correction parameters are still used blindly when registration quality is deteriorated in a traditional system is solved, dependence on personal experience of doctors is reduced, and the problem that the system is completely invalid possibly caused by a single correction mode is solved.

The application discloses a kind of multi-camera-based animal vital sign monitoring method, medium, equipment, it is related to animal vital sign monitoring technical field, method includes: using DLC posture estimation network to train animal multi-view video, obtain key point information;According to the key point information of main visual angle, the centroid motion of animal is analyzed and compared with motion threshold, determine animal state, in the continuous image sequence of relative static state time period, according to key point information, obtain region of interest;Corner point is detected in region of interest, and corresponding motion signal is acquired by tracking corner point;Correlation analysis is carried out to motion signal, and sorting is carried out based on information gain, and vital sign signal is obtained by fusing motion signal higher than set threshold;Power spectrum is used to analyze vital sign signal, obtain the energy distribution of vital sign signal, and the value of respiratory rate and heart rate is calculated.The application can effectively reduce artificial intervention, and has the role of promoting to animal experiment method innovation and data analysis.

Provided is a respiratory rate measurement device provided with: a patient connection unit for acquiring biological information for detecting respiration from a user; a conversion unit for converting the biological information detected by the patient connection unit into an electric signal; a respiratory rate calculation unit for processing the converted electrical signal and calculating a respiratory rate; and a signal output unit that outputs the calculated respiratory rate by means of at least one of voice, screen display, and communication signals. And a respiratory rate calculation unit that calculates respiratory accompanying information including a respiratory interval for each respiration detected on the basis of the electrical signal or an amplitude of a detected respiratory waveform, calculates an index relating to reliability of the calculated respiratory rate on the basis of the respiratory accompanying information, and outputs the index by means of the signal output unit. The respiratory rate measuring device is configured to calculate the respiratory rate and calculate an index related to the stability of the respiration, thereby determining the condition in which the respiration can be stably detected.

The invention discloses an apnea syndrome evaluation method based on a smart watch. According to the method, a photoelectric volume pulse wave signal and a three-axis acceleration signal in the sleep period of a user are collected through a built-in sensor of the smart watch, and filtering and window segmentation preprocessing is carried out on the signals. Extracting multi-dimensional features such as heart rate, heart rate variability, body movement index and respiratory rate from each analysis window, and constructing a feature vector sequence; and analyzing the feature vector by using a pre-trained machine learning classification model, outputting the apnea event probability and performing binary classification. And finally, quantitatively analyzing the event sequence, calculating an apnea hypopnea index and dividing risk levels. According to the method, the signal processing technology and the machine learning algorithm are combined, intelligent identification and risk assessment of the sleep apnea event are achieved, and an effective solution is provided for portable monitoring.

The application discloses a kind of self-adapting oxygen supply control equipment, belong to oxygen supply control equipment technical field.It includes: acquisition module is used to obtain and time stamp alignment user breathing state signal;Calculation module is used to preprocess respiratory flow and pressure signal, calculate signal quality index;Identification module is used to establish basic parameter set, constructs feature vector, and identifies breathing phase;Determination module is used to calculate respiratory demand score, and determines target oxygen supply interval;Correction module and control module are used to output target oxygen supply parameter vector in target oxygen supply interval by oxygen supply target prediction model, after phase differentiation correction and safety constraint, generate final execution control quantity, and convert into instruction oxygen supply equipment execution.This equipment is through the alignment of multiple source breathing signals, feature extraction and closed-loop safety correction, realizes the accurate adaptation of oxygen supply output and real-time physiological demand.

The invention provides a non-contact cardiopulmonary coupling analysis method and system based on millimeter wave perception. The method comprises the steps that millimeter wave frequency modulation signals are continuously sent to a subject, initial echo signals are received and converted to obtain converted echo signals, and the distance position to a target is positioned from the converted echo signals; calculating a first phase for the conversion echo signal at each time point of the distance position of the target, calculating a second phase based on each time point and the first phase at a time point before the time point, and calculating a radial displacement signal based on the second phase; calculating a breathing displacement signal and a heartbeat displacement signal based on the radial displacement signal, calculating a breathing instantaneous phase of each time point based on the breathing displacement signal, and selecting at least one target time period based on the breathing instantaneous phase; a heartbeat accompanying change value of the target time period is calculated based on the heartbeat displacement signal of the target time period, and a modulation index is calculated based on the heartbeat accompanying change values of all the target time periods.

There is provided a breath sensing system including: a breath sensor which senses exhaled air, the breath sensor having a light emission unit which emits light that is directed toward a path through which the exhaled air passes, and a light receiving unit which receives at least a part of the light, and which outputs a light receiving signal; a respiration information calculation unit which calculates respiration information; a reliability calculation unit which calculates reliability level information; an output unit which outputs the respiration information and the reliability level information; a receiving unit which receives the respiration information and the reliability level information; a movable unit which is provided with the breath sensor, and which has a variable position; and a notification unit which provides a notification of information related to the movable unit, according to the reliability level information.

This invention discloses a fine-grained respiratory monitoring method and device based on a diffusion generation model using millimeter-wave radar. The method includes: acquiring echo signals received by the millimeter-wave radar and calculating a one-dimensional range image; locating the range cell containing the human chest cavity and extracting continuous phase information; constructing a diffusion chain from phase observation to the respiratory truth value by using the phase as input to the forward diffusion process and the difference between the respiratory truth value and the phase as progressively added noise; designing a radar diffusion denoising device based on an attention mechanism and a feedforward neural network; training the radar diffusion denoising device according to the reverse diffusion process of the diffusion chain, iteratively optimizing the denoising device parameters by reducing the difference between the respiratory waveform output by the denoising device and the respiratory truth value; inputting the signal to be denoised into the radar diffusion denoising device to complete respiratory waveform denoising and calculate the respiratory frequency. This invention improves the accuracy of respiratory waveform reconstruction and the precision of respiratory frequency estimation.

This enables a simple and easy assessment of the reliability of respiratory rate, without requiring a complex system configuration. [Solution] A biological information measuring device is configured comprising: a respiratory rate calculation unit 7c that calculates the respiratory rate of a subject by deriving a respiratory waveform, which is a periodic virtual waveform, from a photoplethysmography waveform obtained from the subject; a reliability data generation unit 7d that calculates the variance of the period or the variance of the amplitude of the respiratory waveform derived by the respiratory rate calculation unit 7c, and generates reliability data that quantifies the reliability of the respiratory rate calculation result by the respiratory rate calculation unit 7c based on the calculated variance; and a display output unit 6 that displays and outputs the reliability data generated by the reliability data generation unit 7d.

This invention relates to the field of sleep monitoring data fusion processing technology, and discloses a sleep monitoring method based on the fusion of non-invasive brain oxygen saturation (AHI) and multiple physiological parameters. The method includes: S1, simultaneously acquiring multiple physiological parameter signals during the subject's sleep; S2, extracting and analyzing the AHI signal to generate multiple characterization markers; S3, identifying respiratory events related to the subject's apnea information from the respiratory signal and pulse oximetry signal, calculating the apnea-hypopnea index (AHI); and performing correlation analysis on the identified respiratory events and multiple characterization markers to calculate a correlation coefficient; S4, obtaining a comprehensive sleep monitoring assessment report based on the correlation coefficient and the AHI to determine the severity of the subject's current respiratory disturbance and the risk of brain tissue hypoxia. By improving the effectiveness of AHI in non-invasive brain oxygen saturation monitoring, the confirmatory nature of non-invasive brain oxygen saturation information in the respiratory correlation diagnosis during sleep monitoring is ensured.

The application relates to the technical field of data processing, in particular to a breathing intensity data processing system and method and a breathing training device, which are applied to guiding a user to perform inhale and exhale training; the method comprises the following steps: collecting breathing signals of multiple breathing cycles; identifying a starting point and an ending point of each cycle; calculating breathing characteristics such as breathing air time ratio and waveform symmetry through feature extraction logic; and adjusting training feedback based on the breathing characteristics; adopting a topological modeling and derivative analysis mode to correct cycle boundaries, and improving boundary recognition accuracy and training individual adaptability; the breathing training device comprises an air pressure adjusting unit, an atomization unit and a data processing module, and can dynamically optimize training output according to the user state.

This invention discloses an adaptive control method and system for nebulized flow rate based on respiratory rate detection, relating to the field of medical device nebulization control technology. The method includes: acquiring and filtering upstream and downstream pressure sequences of the nebulization path; calculating a differential pressure short-circuit index and generating an event flag based on the degree of synchronization fluctuation and differential pressure amplitude of the two sequences; determining a respiratory rate estimation signal and calculating the respiratory rate based on the flag; generating a baseline nebulized flow rate based on the respiratory rate and correcting it with the differential pressure short-circuit index to obtain the target nebulized flow rate; and finally determining the nebulization execution amount. This invention enables adaptive and conservative adjustment of the nebulized flow rate, improving delivery stability.

The invention belongs to the technical field of medical equipment, and discloses a respiration detection device and method based on magnetic field intensity, and the device comprises a device host and a permanent magnet which are separately pasted on the chest and abdomen of a human body through an elastic pasting material. A three-axis fluxgate is arranged in the device host and used for sensing space magnetic field intensity change caused by displacement of the permanent magnet due to respiration, and respiration parameters are calculated according to the space magnetic field intensity change. The surface of the host is provided with mutually vertical indicating lamp bands, so that a user can be visually guided to accurately position the permanent magnet to an area with strongest magnetic field induction, and the detection sensitivity is ensured. The device supports a single-permanent-magnet working mode and a double-permanent-magnet working mode, and the double-permanent-magnet working mode can synchronously detect breathing movement in multiple directions by utilizing a magnetic field superposition principle, so that the signal strength and the anti-interference capability are remarkably improved. According to the invention, non-constraint, high-precision and high-comfort respiration monitoring is realized, the operation is simple and convenient, and the method is suitable for clinical and family scenes.

The invention discloses a respiratory event recognition method and device, a storage medium and a computer program product, and relates to the technical field of medical monitoring, the method comprises the following steps: collecting throat vibration data through a three-axis acceleration sensor of throat snore-ceasing equipment, and generating a vibration data sequence according to the throat vibration data; calculating an environmental noise baseline according to the vibration data sequence, and identifying a breathing emergency based on the environmental noise baseline; and calculating the current breathing amplitude according to the vibration data sequence, and calculating a breathing amplitude ratio based on the current breathing amplitude and a pre-constructed breathing cycle amplitude model so as to identify an apnea event. The throat vibration data is collected through the three-axis acceleration sensor, and the synergistic effect of the environmental noise baseline and the respiratory cycle amplitude model is combined, so that the recognition precision and recognition efficiency of the apnea event and the respiratory emergency are improved.

Apparatus, such as a respiratory therapy device 4000, may generate a signal representing an estimate of a patient breathing pattern when using a flow therapy device with an un-sealed patient interface. The device may include a blower and may generate a flow of air at a controlled flow rate. The apparatus may include a controller. The generation may involve computing an estimate of the breathing pattern as a function of a device flow rate signal; a device pressure signal; a device conductance; a patient conductance; and a coupling parameter that characterizes a degree of coupling between the interface and the patients nares. An estimate of an exit pressure representative of a pressure of a flow of air generated by the therapy device just outside an orifice of the unsealed interface may also be computed. An estimate of a flushing flow rate from the estimated exit pressure may also be computed.

The application provides a method and system for promoting diaphragm recovery by abdominal compression based on respiratory depression recognition, relates to the technical field of respiratory function rehabilitation, and comprises the following steps: acquiring chest and abdominal movement signals and respiratory airflow signals, calculating respiratory state representation parameters to determine respiratory depression, calculating pressure conduction efficiency of a compression point based on the chest and abdominal movement signals and establishing a pressure superposition constraint relationship, generating force coordination control parameters and periodic compression control instructions, and driving a compression device to apply mechanical stress and realize real-time optimization. The method can accurately identify the respiratory depression state, accurately control and adaptively optimize the compression force, and effectively promote diaphragm function recovery.

This invention relates to the interdisciplinary field of computer image detection and biomedical engineering, and discloses a method and system for respiratory motion monitoring based on Kinect. The method includes: simultaneously acquiring Kinect 3D point cloud sequences and bioimpedance time-series data; extracting first and second respiratory signals respectively; evaluating Kinect signal quality in real time, and generating continuous respiratory waveforms using a main signal output or dynamic weighted fusion strategy under non-interference, interference, and transitional states respectively; and calculating respiratory rate, relative rate of change of tidal volume, and rhythm variation coefficient. The system includes a Kinect sensing unit, a bioimpedance measurement unit, a signal processing and fusion unit, and a respiratory parameter calculation unit. This invention significantly improves the robustness and reliability of respiratory monitoring in complex scenarios through dual-modal redundant sensing and adaptive fusion mechanisms.

The invention relates to the technical field of medical monitoring, in particular to an intelligent respiration monitoring device and system.The system comprises an ultrasonic imaging module, a respiration monitoring module, a comprehensive evaluation module and a body position compensation module; the ultrasonic imaging module is used for calculating a judgment ratio of a patient; the ultrasonic imaging module is also used for calculating a superficial fast breathing index of the patient; the comprehensive evaluation module is used for receiving the judgment ratio transmitted by the ultrasonic imaging module, receiving the shallow and fast respiration index transmitted by the respiration monitoring module, calculating a respiratory failure risk index based on the shallow and fast respiration index and the judgment ratio, evaluating the respiratory failure risk and outputting an evaluation result; through the design of the ultrasonic imaging module, the respiration monitoring module and the comprehensive evaluation module, respiration data and respiratory muscle groups of a patient can be synchronously monitored, so that the respiratory function of the patient is comprehensively evaluated, and the comprehensiveness and accuracy of evaluation are improved.

Provided are: a specific region setting section (110) that sets a specific region and a motion vector calculation point, which is a point that becomes a reference for a motion vector, from an image that contains an upper body of a subject; a motion vector calculation section (120) that calculates a motion vector from a movement amount of the motion vector calculation point in each image that is continuously shot; a breathing reference axis calculation section (130) that calculates a breathing reference axis from a breathing center point, which is a starting point of a motion caused by breathing, and the motion vector calculation point; a breathing signal calculation section (140) that calculates a breathing signal from a component of the motion vector in the direction of the breathing reference axis; and a breathing information calculation section (150) that calculates a breathing information estimation result, which is a result of estimating a body motion caused by breathing, from the breathing signal.

The invention relates to an electrocardiogram detection method of wearable equipment, which comprises the following steps of: acquiring six-lead electrocardiogram signals to obtain richer cardiac electrical activity information, and performing real-time impedance matching, amplification and anti-interference processing on the six-lead electrocardiogram signals through a circuit system; the high signal-to-noise ratio of the six-lead electrocardiosignals in the motion state is ensured from the hardware level; then, a UNet-BiLSTM hybrid neural network model is combined with a relocation algorithm, high-precision R peak recognition is conducted on signals with large motion artifacts and waveform variation, and ultrahigh accuracy of heart rate calculation is achieved; according to the method and the system, the electrocardiosignal is acquired, then the respiration waveform is synchronously extracted from the same electrocardiosignal, and the respiration rate is calculated, so that the electrocardio and respiration double parameters are detected by a single sensor, the equipment complexity and the discomfort of a user are reduced, the leap from single index monitoring to multi-system function collaborative evaluation is realized, and a more comprehensive basis is provided for exercise safety and disease early warning.

The invention provides a ruminant respiration detection method and system, computer equipment and a storage medium, and belongs to the field of vital sign detection.The ruminant respiration detection method includes the steps that radar echo ADC data are collected; performing distance dimension fast Fourier transform on the radar echo ADC data to obtain a distance-snapshot matrix, performing phase extraction, calculating the difference of phase signals, and constructing a vital sign time sequence; performing fast Fourier transform on the vital sign time sequence to obtain fast Fourier spectrum energy distribution; determining breathing candidate signals based on a preset breathing frequency band; empirical mode decomposition is carried out on the breath candidate signals to obtain a plurality of intrinsic mode components, comprehensive scores are calculated based on the frequency matching degree, the energy weight and the period consistency pair, and a breath main component is determined through the intrinsic mode component with the highest comprehensive score; and calculating a power spectrum peak corresponding to the breathing principal component, and determining the number of times of breathing according to the power spectrum peak. According to the method, the robustness of the animal respiration detection method is enhanced.

This application relates to an intelligent respiratory training device for post-thoracic surgery respiratory function rehabilitation training, belonging to the technical field of medical rehabilitation devices. The device includes a housing, a resistance training module, an airflow sensing module, and a main control module. The resistance training module includes an actuator that can be rotated by the respiratory airflow. The airflow sensing module includes a flow obstruction screen and a differential pressure sensor disposed in the airflow channel. The main control module is configured to: calculate the instantaneous velocity of the respiratory airflow based on the differential pressure signal collected in real time by the differential pressure sensor, and generate a control signal based on the real-time deviation between the instantaneous velocity and the preset target velocity to dynamically adjust the resistance torque on the actuator, thereby guiding the instantaneous velocity towards the preset target velocity. This application solves the problems of fixed resistance modes, inability to quantify and evaluate, and lack of personalized guidance in traditional training devices by constructing a closed-loop control system with real-time monitoring and dynamic adjustment, achieving intelligent, precise, and adaptive control of the respiratory training process.

The application discloses a mechanical ventilation patient grading exercise rehabilitation system for weaning preparation and a control method thereof, relates to the technical field of intelligent medical treatment, and comprises a sensor module, a patient parameter initialization module, a real-time data acquisition module, a respiratory load index calculation module, a motion intensity dynamic adjustment module, a motion rehabilitation execution module, a safety processing and alarm module, and a data recording and report generation module. The working process of the grading exercise rehabilitation system is as follows: system initialization and patient parameter setting; real-time data acquisition and preprocessing; calculation of the respiratory load index; real-time calculation of the respiratory load index based on the collected and preprocessed data, and assessment of the respiratory load state of the patient by comparison with an individualized threshold value; automatic adjustment of the current motion intensity level according to the comparison result of the respiratory load index and the threshold value; execution of the motion rehabilitation and continuous monitoring; automatic recording of all data changes collected by the sensor module during the training process, and generation of a data report.

The invention relates to a chest impedance signal measurement and separation method based on double-source differential topology, and belongs to the technical field of physiological parameter monitoring, and the method comprises the following steps: S1, respectively arranging electrodes at the symmetrical positions of the left chest and the right chest of a subject, and arranging a common backflow electrode at the central position of the spine of the back of the subject as a potential reference point of a measurement system; s2, two constant current sources with identical amplitude, frequency and phase are set as injection currents to be injected into the left chest electrode and the right chest electrode respectively, and the currents are converged at the same time and flow to a common backflow electrode; s3, the subject holds breath, and the potential difference of the left chest position and the right chest position relative to the back common point is collected; s4, cardiac pulsation electrical impedance and baseline impedance of the left chest and the right chest are calculated; and S5, the subject restores normal respiration, the voltage signal of the left chest is measured, the pulsation electrical impedance of the left chest and the heart is kept unchanged, and impedance change caused by respiration is calculated.

The application discloses a contactless breathing monitoring method and system for a newborn. Firstly, an infrared temperature image set is generated by a sliding time window, and a temperature change parameter set is read according to the infrared temperature image set, and then a real-time temperature change period model is generated by fitting. Subsequently, extreme value recognition is performed on the real-time temperature change period model, a breathing frequency parameter and a temperature parameter are calculated, the infrared temperature image set is analyzed, a breathing intensity parameter is calculated, and finally a comprehensive breathing score is generated according to the breathing frequency parameter, the temperature parameter and the breathing intensity parameter, and a breathing state monitoring result is output in combination with a standard judgment model. Based on the characteristics of the fixed posture of the newborn, the breathing of the newborn will cause rhythmic changes in temperature in a certain fixed area, and the above rules can be captured by analyzing the infrared temperature images of the above area, and the analysis of the breathing frequency, body temperature and breathing intensity is realized, and further, the contactless monitoring of the breathing of the newborn is realized.

The application discloses a kind of based on millimeter wave radar's motion heart rate breath and high counter warning method, it is related to health warning technical field, including the following steps: step S1, radar acquisition signal is laid out;Step S2, pre-treatment filters out interference;Phase difference is solved displacement in step S3;Step S4, separate breathing heartbeat signal;Step S5, calculate respiratory rate;Step S6, estimate blood oxygen saturation;Step S7, determine high counter level;Step S8, output warning and correction.The application uses the above-mentioned one based on millimeter wave radar's motion heart rate breath and high counter warning method, realizes the non-contact type motion heart rate, breath frequency accurate detection, noninvasive estimation blood oxygen value, classifies and identifies high altitude reaction risk, improves the safety guarantee of high altitude outdoor activity and field operation.

The invention discloses a method and a device for eliminating respiratory artifacts of uterine myoelectricity. The method comprises the following steps: S1, acquiring an abdominal multichannel signal and a respiratory rate RR to obtain a uterine myoelectricity original data sequence X; s2, subtracting the mean value from the original data sequence to obtain zero mean value data; s3, carrying out linear transformation on the zero mean value data; s4, solving zero-mean data after linear transformation to obtain an orthogonal matrix W by searching for a direction in which negentropy or kurtosis is approximately maximized; s5, constructing a two-dimensional comprehensive scoring function, and identifying a respiratory artifact vector VR; s6, calculating a projection component VRX of the respiratory artifact vector VR in the original hybrid matrix X, and removing the projection component from the original signal to obtain a signal matrix X'after preliminary removal of the respiratory artifacts; and S7, carrying out band-pass filtering on the signal matrix X'to obtain a pure uterine electromyographic signal with a high signal-to-noise ratio. The method can effectively eliminate the interference of the respiratory artifacts on the uterine myoelectricity signals, and improves the signal-to-noise ratio of the collected uterine myoelectricity.

The invention relates to the technical field of data processing, in particular to a respiration intensity data processing system and method and a respiration training device, and is applied to guiding a user to perform inspiration and expiration training. The method comprises the following steps: collecting respiratory signals of a plurality of respiratory cycles; identifying a starting point and an ending point of each period; respiratory characteristics such as a respiratory gas duration ratio and waveform symmetry are calculated through characteristic extraction logic; training feedback is adjusted based on the breathing features; a periodic boundary is corrected by adopting a topological modeling and derivative analysis mode, so that the boundary recognition precision and the training individual adaptability are improved; the respiratory training device comprises an air pressure adjusting unit, an atomization unit and a data processing module, and training output can be dynamically optimized according to the state of a user.