A thoraco-abdominal complex deformation full view reconstruction and respiration detection system and method

Abstract

The application discloses a chest and abdominal cavity complex deformation full view reconstruction and respiration detection system and method, relates to the technical field of human deformation measurement, and comprises a binocular vision acquisition device, the binocular vision acquisition device is connected with a computer through a connecting line and carries out data transmission, an image processing module, a displacement and strain calculation module, a respiration parameter extraction module and a remote data interaction module are arranged on the computer; the binocular vision acquisition device is arranged on the front sides of the chest and abdomen of a user, and the user wears an elastic fabric speckle shirt on the upper body. The chest and abdominal cavity complex deformation full view reconstruction and respiration detection system and method have the advantages of non-contact, high precision and full cycle, and are suitable for various scenes such as clinical diagnosis, postoperative rehabilitation, home monitoring and vocal music training.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of human body deformation measurement, in particular to a chest and abdominal cavity complex deformation full view reconstruction and respiration detection system and method. BACKGROUND

[0002] Human respiration can drive the chest and abdominal cavity to produce complex three-dimensional dynamic deformation, which contains rich physiological information such as individual vital capacity, respiratory rhythm, and muscle coordination. Capturing and analyzing these information for respiration detection is of great significance. In the scenes of chest and abdominal breathing training, early identification and rehabilitation treatment of respiratory diseases, postoperative functional recovery evaluation, and acoustic instrument training optimization, accurate perception and full view modeling of chest and abdominal cavity deformation are particularly crucial.

[0003] The current common respiration monitoring technology has significant limitations. The contact type measurement devices such as respiration chest belts, strain sensors, and piezoelectric sensors proposed by related inventions can only record local deformation information with extremely limited spatial resolution. These methods cannot capture the complex deformation of the chest and abdominal cavity during respiration, and are difficult to support accurate respiration detection and physiological feature analysis. In addition, the sensing devices designed by the existing mainstream technology methods are usually in direct contact with the skin, which is easily affected by the subjective influence of individual wearing methods and action patterns. Too many sensing units may even interfere with the natural breathing state, which cannot meet the needs of high sensitivity detection scenes.

[0004] Although medical respiration monitors and flow sensors can automatically obtain the respiratory gas volume change of the wearer, they reflect relatively single physiological information and do not support chest and abdominal cavity deformation full view reconstruction and subsequent targeted accurate analysis.

[0005] CT imaging has high precision, but it is a static imaging and cannot meet the needs of continuous detection. In addition, the operation process depends on professional institutions and doctors, and there are challenges such as high cost and complicated operation, which are difficult to adapt to dynamic evaluation or daily use.

[0006] It can be seen that there is currently a lack of a system that can obtain complex deformation information of the chest and abdominal cavity in real time and accurately under the premise of non-contact and not affecting the limb action of the measured object for intelligent respiration analysis. Especially in the full view reconstruction of the overall deformation of the chest and abdominal cavity, the existing technology has obvious limitations in terms of accuracy, speed, and deployability. SUMMARY

[0007] The purpose of the present application is to provide a chest and abdominal cavity complex deformation full view reconstruction and respiration detection system and method to solve the problems raised in the background technology.

[0008] In order to achieve the above object, the present application provides a thoraco-abdominal complex deformation full view reconstruction and respiration detection system, which comprises a binocular vision acquisition device, the binocular vision acquisition device is connected with a computer through a connecting line and performs data transmission, an image processing module, a displacement and strain calculation module, a respiration parameter extraction module and a remote data interaction module are arranged on the computer;

[0009] The binocular vision acquisition device is arranged at the front sides of the chest and abdomen of the user, and the upper body of the user is provided with an elastic fabric speckle shirt.

[0010] Preferably, the elastic fabric speckle shirt is made of elastic fabric material, and the surface of the elastic fabric speckle shirt is provided with a speckle pattern, and the speckle pattern is a non-periodic high-contrast random distribution pattern.

[0011] Preferably, the binocular vision acquisition device comprises a synchronous trigger instrument and a binocular camera system connected with the synchronous trigger instrument, the binocular camera system comprises a left camera and a right camera, and the right camera and the left camera are respectively arranged on two tripods.

[0012] Preferably, the image processing module comprises a calibration unit, an image stereo matching unit and a three-dimensional reconstruction unit.

[0013] The calibration unit respectively acquires the intrinsic parameters and extrinsic parameters of the left camera and the right camera.

[0014] The image stereo matching unit matches the corresponding relationship and the time sequence change in the respiration process in the left camera and the right camera pictures.

[0015] The three-dimensional reconstruction unit reconstructs the three-dimensional space coordinate points of the chest and abdominal surface according to the matching result.

[0016] Preferably, the displacement and strain calculation module calculates the surface displacement field and the strain field according to the three-dimensional space coordinate points, and generates the three-dimensional reconstruction result.

[0017] Preferably, the respiration parameter extraction module extracts the respiration characteristic parameters based on the displacement field and the strain field, and the respiration characteristic parameters comprise the respiration frequency, the respiration period, the respiration amplitude, the thoraco-abdominal ratio, the respiration phase difference, the left-right symmetry index and the local region delay.

[0018] Preferably, the remote data interaction module comprises a cloud server and a mobile terminal.

[0019] The cloud server receives the three-dimensional reconstruction result and the respiration characteristic parameters and processes them, generates the three-dimensional reconstruction image and the respiration characteristic parameter result.

[0020] The mobile terminal is provided with a graphical user interface, receives and displays the three-dimensional reconstruction image and the respiration characteristic parameter result, and provides the respiration state prompt and the abnormal mode recognition.

[0021] A method for thoraco-abdominal complex deformation full view reconstruction and respiration detection system, comprising the following steps:

[0022] S1, using a binocular camera system to calibrate and record dynamic image sequences of the thoraco-abdominal region:

[0023] Build a binocular camera system, place the left and right cameras on both sides of the front of the user's thoraco-abdominal region and the calibration board;

[0024] Prepare an elastic fabric speckle shirt, which is worn by the user during the collection process;

[0025] Use a synchronous trigger instrument to simultaneously start the left and right cameras, and record the deformation of the thoraco-abdominal region of the user during respiration from two perspectives, respectively;

[0026] S2, obtain the thoraco-abdominal surface displacement field and strain field according to the image sequences collected by the left and right cameras;

[0027] S3, extract the respiration characteristic parameters and perform quantitative analysis and evaluation;

[0028] S4, establish a communication link between the mobile device terminal and the cloud server for remote medical treatment and personalized intervention.

[0029] Preferably, the specific steps of S2 are as follows:

[0030] S21, use Zhang Zhengyou's calibration method to perform camera model calibration calculation, and obtain the intrinsic and extrinsic parameters of the left and right cameras, respectively;

[0031] S22, use a stereo matching algorithm to match the corresponding relationship in the left and right camera images of the observed position of the user's thoraco-abdominal region and the time sequence changes during respiration, and obtain matching information;

[0032] S23, based on the intrinsic and extrinsic parameters obtained in S21 and the matching information obtained in S22, perform three-dimensional reconstruction to calculate the three-dimensional spatial point coordinates of the thoraco-abdominal spatial points;

[0033] S24, based on the plurality of three-dimensional spatial point coordinates obtained in S23, extract the three-dimensional spatial point coordinates at different deformation stages to obtain the displacement field, and perform displacement field fitting based on the displacement information of the plurality of points to obtain the strain field;

[0034] S25, upload the calculation results of the three-dimensional digital image correlation method to the cloud server.

[0035] Preferably, the specific steps of S3 are as follows:

[0036] S31, analyze the displacement, velocity, and strain of the plurality of observation points in the thoraco-abdominal full view reconstruction to extract the respiration characteristic parameters;

[0037] S32, a deep neural network model is constructed, combined with a recurrent neural network and an attention mechanism, to capture long-term dependencies and local features in time series data, and the historical data of the wearer are input into the deep neural network model as respiratory feature parameters, and a specific individual's respiratory mode benchmark is formed through network learning and fitting.

[0038] Therefore, the chest and abdominal cavity complex deformation full view reconstruction and respiration detection system and method has the following beneficial effects:

[0039] (1) Non-contact deformation measurement, without any sensor attachment, suitable for various body types and actions.

[0040] (2) Accurate high spatial resolution full view reconstruction, which can accurately reflect the dynamic deformation of each region of the chest and abdominal cavity surface during the entire respiratory cycle.

[0041] (3) Differentiate different breathing patterns and analyze individualized respiratory states.

[0042] (4) Support continuous long-term monitoring and remote data transmission, convenient for clinical promotion and home use.

[0043] (5) Extensible for respiratory pattern analysis, postoperative rehabilitation monitoring, vocal music generation training and other application scenarios.

[0044] The technical solutions of the present application will be further described in detail below with the help of the accompanying drawings and examples. BRIEF DESCRIPTION OF DRAWINGS

[0045] Figure 1 The system structure diagram of the chest and abdominal cavity complex deformation full view reconstruction and respiration detection system and method embodiment of the present application;

[0046] Figure 2 The calibration diagram of the binocular camera system of the present application;

[0047] Figure 3 The schematic diagram of the user wearing the elastic fabric speckle shirt of the present application;

[0048] Figure 4 The structure schematic diagram of the binocular vision acquisition device of the present application;

[0049] Figure 5 The calculation flowchart of the three-dimensional digital image correlation method of the present application;

[0050] Figure 6 The data storage and parameter calculation flowchart of the cloud server of the present application;

[0051] Figure 7The device graphical interface visualization and intervention reminding diagram of the mobile terminal of the present application. DETAILED DESCRIPTION

[0052] The technical solutions of the present application are further described below through the drawings and examples.

[0053] Unless otherwise defined, technical terms or scientific terms used in the present application shall have the usual meaning understood by a person with ordinary skill in the art to which the present application belongs. The terms "first", "second" and similar words used in the present application do not represent any order, number or importance, but are only used to distinguish different components. The terms "include" or "contain" and similar words mean that the elements or objects appearing before the words cover the elements or objects listed after the words and their equivalents, without excluding other elements or objects. The terms "connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. The terms "up", "down", "left", "right" and the like are only used to represent relative positional relationships, which may change accordingly when the absolute position of the described object changes.

[0054] EMBODIMENT

[0055] Referring to Figures 1-7 The present application provides a chest and abdominal cavity complex deformation full view reconstruction and respiration detection system, which comprises a binocular vision acquisition device, the binocular vision acquisition device is connected with a computer through a connecting line and performs data transmission, and the computer is provided with an image processing module, a displacement and strain calculation module, a respiration parameter extraction module and a remote data interaction module.

[0056] The binocular vision acquisition device comprises a synchronous trigger instrument and a binocular camera system connected with the synchronous trigger instrument, the binocular camera system comprises a left camera and a right camera, and the right camera and the left camera are respectively arranged on two tripods. The binocular vision acquisition device is arranged on the front sides of the chest and abdomen of a user, and the user wears an elastic fabric speckle shirt on the upper body. The elastic fabric speckle shirt is made of elastic fabric material, and the surface of the elastic fabric speckle shirt is distributed with a speckle pattern, and the speckle pattern is a non-periodic high-contrast random distribution pattern.

[0057] The image processing module adopts a stereo matching algorithm and a three-dimensional digital image correlation method to perform sub-pixel level matching on binocular images, and completes high spatial resolution three-dimensional full view reconstruction. Specifically, the image processing module comprises a calibration unit, an image stereo matching unit and a three-dimensional reconstruction unit. The calibration unit respectively acquires the intrinsic parameters and extrinsic parameters of the left camera and the right camera. The image stereo matching unit matches the corresponding relationship and the time sequence change in the respiration process in the left camera and right camera pictures. The three-dimensional reconstruction unit reconstructs the three-dimensional spatial coordinate points of the chest and abdominal surface according to the matching results.

[0058] The displacement and strain calculation module compares multiple frames of three-dimensional space coordinate points of multiple feature points of the chest and abdomen, calculates the displacement change, and obtains a continuous strain field through fitting, to finally obtain a surface displacement field and a strain field, and generate a three-dimensional reconstruction result.

[0059] The respiratory parameter extraction module extracts respiratory feature parameters based on the displacement field and the strain field, and the respiratory feature parameters include respiratory frequency, respiratory period, respiratory amplitude, chest-to-abdomen ratio, respiratory phase difference, left-right symmetry index, and local region delay.

[0060] The remote data interaction module has an abnormal respiratory pattern recognition function, which is used to identify abnormal respiratory states and generate alarm information. Specifically, the remote data interaction module includes a cloud server and a mobile terminal. The cloud server receives the three-dimensional reconstruction result and the respiratory feature parameters and processes them to generate a three-dimensional reconstruction image and a respiratory feature parameter result. The mobile terminal is configured with a graphical user interface, which receives and displays the three-dimensional reconstruction image and the respiratory feature parameter result, and provides respiratory state prompts and abnormal pattern recognition.

[0061] S1, a binocular camera system is used to calibrate and record dynamic image sequences of the chest and abdomen. It includes:

[0062] S11, a binocular camera system is set up and calibrated. The binocular camera system includes a left camera and a right camera. The left camera and the right camera are two high-resolution industrial cameras with a resolution of 1920x1080 pixels or above. The specific steps are as follows:

[0063] First, set the image acquisition parameters of the left camera and the right camera to be consistent. The acquisition frame rate of the left camera and the right camera should not be less than 30 fps and should be consistent. The trigger frequency and the exposure time should be set to the same value.

[0064] Then, the left camera and the right camera are securely mounted on an adjustable tripod to ensure that the left camera and the right camera can capture complete calibration board images, as shown in Figure 2 .

[0065] Finally, adjust the aperture and focal length of the left camera and the right camera to ensure that the left camera and the right camera can clearly capture the calibration images. Place the circular ring calibration board at different angles and positions, and take multiple calibration images through the left camera and the right camera.

[0066] S12, prepare a close-fitting, thin, and elastic fabric speckle shirt for acquisition, as shown in Figure 3 .

[0067] The elastic fabric material of the elastic fabric speckle shirt is selected from a high-elasticity fabric, and a random high-contrast speckle pattern is made on the surface of the elastic fabric material by digital printing technology. The diameter of the speckles is generally 3 mm, and the density of the pattern is uniform and arranged in a random manner to meet the requirements of the digital image correlation method for deformation measurement.

[0068] During the acquisition process, the elastic fabric speckle shirt is ensured to be tightly attached to the chest and abdominal region of the user to accurately capture the deformation of the chest and abdominal region of the user during breathing, and to ensure that the speckle pattern has no obvious wrinkles or sliding.

[0069] The user is located in the overlapping area of the field of view of the left camera and the right camera. The distance between the user and the left camera and the right camera is adjusted according to the body type of the user, and is generally set to be between 80-100 cm. The chest and abdomen of the user are ensured to be located in the central area of the field of view of the left camera and the right camera.

[0070] S13, the left camera and the right camera are simultaneously sent with a potential signal by using a synchronous trigger instrument, and the image acquisition of the left camera and the right camera is started.

[0071] The synchronous trigger instrument is provided with a control button and a display screen to facilitate user operation and state display. The synchronous trigger instrument is connected to the left camera and the right camera through connection lines. Specifically, the high-voltage output terminal on the synchronous trigger instrument is connected to the positive input signal terminal of the left camera and the right camera, and the low-voltage output terminal on the synchronous trigger instrument is connected to the negative input signal terminal of the left camera and the right camera. The structure of the synchronous trigger instrument and the connection with the left camera and the right camera are shown in Figure 4 .

[0072] The synchronous trigger instrument itself is powered by an additional power supply module, and the power supply terminal is connected to a high voltage, and the other terminal is connected to the ground.

[0073] The synchronous trigger instrument is started, and the system enters the starting state, and the display screen displays "Startup". The left camera and the right camera are started by pressing the button of the synchronous trigger instrument, and the sequence images of breathing are captured.

[0074] The sequence images captured by the left camera and the right camera are saved to the computer connected thereto, as shown in Figure 2 .

[0075] S2, stereo matching and deformation calculation are performed on the image sequence captured by the left camera and the right camera to obtain the chest and abdominal surface displacement field; the image sequence captured by the left camera and the right camera is registered, and the digital image correlation method is used to calculate the strain field of each frame in combination with the calibration result obtained in S11. The specific calculation steps are shown in Figure 5 .

[0076] S21, camera model calibration is performed by using Zhang Zhengyou calibration method, and the specific steps are as follows:

[0077] Marker point positioning. Sub-pixel accuracy marker point detection algorithm is used to detect the ring marker point coordinates in each image, and the accurate matching of the marker points in the left and right camera images is realized based on the known geometric arrangement relationship of the calibration board.

[0078] The matching completed marker point information obtained in the last step is brought into the camera projection model for iterative optimization. In the camera imaging system, there are four coordinate systems: the world coordinate system, the camera coordinate system, the image coordinate system, and the pixel coordinate system.

[0079] A parameter model of the two-dimensional coordinates of the marker points on the image plane in the pixel coordinate system and the three-dimensional space coordinates of the world coordinate system on the calibration board is established. The information of multiple groups of marker points is brought into the parameter model, and whether the parameters are reasonable is evaluated by the re-projection error. Levenberg-Marquardt algorithm is used for iterative optimization.

[0080] The respective camera intrinsic parameters of the two cameras, including focal length, pixel scale factor and distortion coefficient, and the extrinsic parameters between the left camera and the right camera, including relative rotation pose and translation, are obtained.

[0081] S22, using a stereo matching algorithm, the observation position of the chest and abdomen is matched in the corresponding relationship between the left camera and the right camera picture and the time sequence change in the breathing process, and the specific matching process is realized by a two-dimensional digital image correlation method, and the specific steps are as follows:

[0082] Establishing an initial reference image and selecting a sub-region, the first frame image of the starting stage of the left and right cameras is selected as the reference image, and the subsequent images in the breathing process are selected as the deformed images.

[0083] The range of full-view reconstruction is determined, i.e. the region of interest of the digital image correlation method. It is artificially selected in the camera reference image.

[0084] Set the interest points for calculation in the region of interest, and set the calculation sub-area as the center of these interest points. In the experiment, a square region with a side length of 51 mm is set.

[0085] Selecting a zero-mean normalized cross-correlation function as the quantitative reference basis for sub-area matching, the reference image and the deformed image sub-area are matched.

[0086] Using a second-order shape function to describe the deformation of the sub-area in the deformed image.

[0087] Through the two stages of integer pixel search and sub-pixel matching, the corresponding shape and position of each sub-area in the deformed image are found.

[0088] The position of each interest point in the reference image in the deformed image is obtained.

[0089] Three two-dimensional digital image correlation method calculations are performed in S22, and the specific corrections are as follows:

[0090] First, the two-dimensional digital image correlation method is calculated for the left camera reference image and the left camera deformed image, respectively, to obtain the deformed image matching.

[0091] Second, the two-dimensional digital image correlation method is calculated for the left camera reference image and the right camera reference image, respectively, to obtain the deformed image matching.

[0092] Third, the two-dimensional digital image correlation method is calculated for the left camera reference image and the right camera deformed image, respectively, to obtain the deformed image matching.

[0093] The above three steps of matching are performed on the sequence images of the left camera and the right camera to obtain the left and right field of view matching information and the time sequence matching information of the left and right cameras, and the stereo matching of the left and right cameras is realized.

[0094] S23, combining the left and right camera internal and external parameters obtained in S21 and the left and right camera stereo matching information obtained in S22 to perform three-dimensional reconstruction. The three-dimensional spatial coordinates of each spatial point are obtained. The full appearance reconstruction with high spatial resolution is realized.

[0095] S24, based on the plurality of three-dimensional spatial point coordinates obtained in S23, extracting the three-dimensional coordinates at different deformation stages to obtain the displacement field, and fitting the displacement field according to the displacement information of multiple points to obtain the strain field. The specific steps are as follows:

[0096] S241, based on the plurality of three-dimensional spatial point coordinates obtained in S23, extracting the three-dimensional coordinates at different deformation stages, and subtracting the three-dimensional coordinates at the initial time to obtain the three-dimensional displacement field.

[0097] S242, based on the three-dimensional morphology reconstructed in S23, determining the normal direction of each point. Using the least squares method to fit the plane of the data around each point to obtain the local surface around each point, and then defining the local coordinate system. Project the displacement of each point in three directions onto the local coordinate system.

[0098] S243, the local strain can be obtained by the least squares method for the projected three-dimensional displacement. Further, the strain field can be obtained by fitting the displacement field according to the displacement information of multiple points.

[0099] S25, uploading the displacement field, strain field and related calculation result data obtained by the three-dimensional digital image correlation method to the cloud server to realize remote storage, analysis and management of data.

[0100] S3, extracting the respiratory feature parameters to realize quantitative analysis and evaluation. The personal database is established and the historical data analysis based on deep learning is completed in the cloud server, such asFigure 6 The specific steps are as follows:

[0101] S31, respiratory feature parameter extraction. The displacement, velocity and strain changes of multiple observation points generated by respiratory motion are captured by reconstructing the data of the chest and abdomen. Specifically, the multiple observation points include the feature regions of the chest and abdomen. By processing the data of these observation points, the following key respiratory feature parameters are calculated and extracted:

[0102] (1) Respiratory frequency: the number of respiratory movements per unit time;

[0103] (2) Respiratory period: the duration of a single complete respiratory movement;

[0104] (3) Respiratory amplitude: the maximum range of displacement of the chest and abdomen in respiratory movement;

[0105] (4) Chest-abdomen ratio: the ratio of chest movement amplitude to abdominal movement amplitude, reflecting the chest-abdominal breathing feature;

[0106] (5) Respiratory phase difference: the time difference between the chest and abdominal movement waveforms, reflecting the respiratory coordination;

[0107] (6) Left-right symmetry index: the degree of symmetry difference between the left and right movement data of the chest and abdomen;

[0108] (7) Local region delay: the delay time of the respiratory movement of a specific chest and abdominal region relative to the overall average movement.

[0109] Through the extraction of the above feature parameters, quantitative analysis of respiratory function can be realized, providing scientific and objective basis for clinical diagnosis.

[0110] S32, respiratory mode historical data analysis. A deep neural network model is constructed, which combines recurrent neural networks and attention mechanisms to capture long-term dependencies and local features in time series data. The respiratory feature parameters of the wearer's historical data are input into the deep neural network model, and through network learning and fitting, the respiratory mode benchmark of a specific individual is formed.

[0111] In the process of continuous monitoring, by comparing the current extracted respiratory feature parameters with the historical benchmark data, the difference index is calculated to evaluate the change trend, abnormal event or potential risk of the wearer's respiratory mode.

[0112] In S4, by establishing a communication link between the mobile device terminal and the cloud server, remote medical service and personalized intervention functions are realized, such as Figure 7 The specific steps are as follows:

[0113] S41, the cloud server transmits the three-dimensional reconstruction result generated after analysis and processing and the respiratory parameter result to the mobile device terminal, the three-dimensional reconstruction result is a three-dimensional deformation atlas, and the respiratory parameter result is a respiratory index curve. The mobile device terminal displays the three-dimensional deformation atlas and the respiratory index curve through a graphical user interface, thereby providing visual auxiliary decision support for remote diagnosis and treatment.

[0114] S42, based on the respiratory mode database constructed by the cloud server and the historical data of the wearer, an abnormal respiratory mode is automatically identified. When an abnormal respiratory mode is detected, the system pushes an intervention reminder to the wearer through the mobile terminal, and the reminding mode includes but is not limited to message prompt, voice broadcast or vibration reminder, thereby guiding the wearer to take corresponding adjustment measures, such as adjusting the body position, taking deep breath or contacting medical personnel, etc.

[0115] Therefore, the present application adopts the above-mentioned chest and abdominal cavity complex deformation full view reconstruction and respiratory detection system and method. Based on the three-dimensional digital image correlation method, the system adopts a non-contact method to obtain three-dimensional dynamic topographic data of the user's chest and abdomen during the respiratory process in real time. The system calculates the deformation field of the chest and abdominal cavity complex deformation by configuring a binocular camera system, an elastic fabric speckle shirt and a three-dimensional digital image method, realizes high spatial resolution reconstruction of the chest and abdominal cavity full view, can extract key respiratory indicators including respiratory frequency, amplitude, period, left-right symmetry, etc., and combines a deep learning algorithm to perform personalized modeling and respiratory mode recognition. The data can be used for remote analysis and health intervention through a mobile terminal device.

[0116] Finally, it should be noted that: the above examples are only used to illustrate the technical solutions of the present application but not to limit it, although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that: it can still modify or equivalently replace the technical solutions of the present application, and these modifications or equivalent replacements also cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present application.

Claims

1. A complex deformation reconstruction and respiration detection system for thoraco-abdominal cavity, characterized in that: The binocular vision acquisition device is connected with the computer through a connecting line and performs data transmission, and the computer is provided with an image processing module, a displacement and strain calculation module, a respiratory parameter extraction module and a remote data interaction module; The binocular vision acquisition device is arranged on the front sides of the chest and abdomen of the user, and the user wears an elastic fabric speckle shirt on the upper body; The elastic fabric speckle shirt is made of elastic fabric material, and the surface of the elastic fabric speckle shirt is distributed with a speckle pattern, which is a non-periodic high-contrast random distribution pattern; The binocular vision acquisition device comprises a synchronous trigger instrument and a binocular camera system connected with the synchronous trigger instrument, and the binocular camera system comprises a left camera and a right camera, and the right camera and the left camera are respectively arranged on two tripods; The displacement and strain calculation module calculates the surface displacement field and the strain field according to the three-dimensional space coordinate points, and generates a three-dimensional reconstruction result; The respiratory parameter extraction module extracts respiratory characteristic parameters based on the displacement field and the strain field, and the respiratory characteristic parameters comprise a respiratory frequency, a respiratory period, a respiratory amplitude, a chest-to-abdomen ratio, a respiratory phase difference, a left-right symmetry index and a local region delay.

2. The complex thoraco-abdominal deformation full-view reconstruction and respiration detection system according to claim 1, characterized in that: The image processing module comprises a calibration unit, an image stereo matching unit and a three-dimensional reconstruction unit; The calibration unit respectively acquires the intrinsic parameters and the extrinsic parameters of the left camera and the right camera; The image stereo matching unit matches the corresponding relationship and the time sequence change in the respiratory process in the left camera and the right camera pictures; The three-dimensional reconstruction unit reconstructs the three-dimensional space coordinate points of the chest and abdominal surface according to the matching result.

3. The complex thoraco-abdominal deformation full view reconstruction and respiration detection system of claim 2, wherein: The remote data interaction module comprises a cloud server and a mobile terminal; The cloud server receives the three-dimensional reconstruction result and the respiratory characteristic parameters and processes them to generate a three-dimensional reconstruction image and a respiratory characteristic parameter result; The mobile terminal is provided with a graphical user interface, receives and displays the three-dimensional reconstruction image and the respiratory characteristic parameter result, and provides a respiratory state prompt and an abnormal mode recognition.

4. The method of using the complex thoraco-abdominal deformation profile reconstruction and respiration detection system of any of claims 1-3, wherein, The method comprises the following steps: S1. Calibrating and recording a dynamic image sequence of the chest and abdomen using a binocular camera system: Build the binocular camera system, and arrange the left camera and the right camera on the calibration board and on the front sides of the chest and abdomen of the user; Prepare the elastic fabric speckle shirt, which is worn by the user during the collection process; Use the synchronous trigger instrument to simultaneously start the left camera and the right camera, and record the deformation of the chest and abdomen of the user during the breathing process from two perspectives; S2. Obtain the surface displacement field and the strain field of the chest and abdomen according to the image sequence collected by the left camera and the right camera; S3. Extract the respiratory characteristic parameters and perform quantitative analysis and evaluation; S4. Establish a communication link between the mobile terminal and the cloud server for remote medical treatment and personalized intervention.

5. The method of claim 4, wherein, The specific steps of S2 are as follows: S21. Perform camera model calibration calculation using Zhang Zhengyou calibration method to obtain the intrinsic parameters and the extrinsic parameters of the left camera and the right camera respectively; S22. Use a stereo matching algorithm to match the corresponding relationship and the time sequence change in the respiratory process in the left camera and the right camera pictures of the observed position of the chest and abdomen of the user, and obtain matching information; S23. Perform three-dimensional reconstruction based on the intrinsic parameters and the extrinsic parameters obtained in S21 and the matching information obtained in S22 to calculate the three-dimensional space coordinate points of the chest and abdominal space points. S24, based on the plurality of three-dimensional spatial point coordinates obtained in S23, extracting three-dimensional spatial point coordinates at different deformation stages to obtain a displacement field, and fitting the strain field according to the displacement information of the plurality of points; S25, uploading the calculation results of the three-dimensional digital image correlation method to a cloud server.

6. The method of claim 5, wherein, The specific steps of S3 are as follows: S31, by analyzing the displacement, velocity and strain of the plurality of observation points in the chest and abdominal full view reconstruction, extracting the respiratory characteristic parameters; S32, constructing a deep neural network model, combining it with a recurrent neural network and an attention mechanism to capture long-term dependencies and local features in time series data, inputting the respiratory characteristic parameters of the wearer's historical data into the deep neural network model, and forming a specific individual's respiratory pattern benchmark through network learning and fitting.

Images