A real-time monitoring system for rehabilitation nursing of patients with movement disorders

Abstract

The application provides a real-time monitoring system for rehabilitation nursing of patients with movement disorders, which integrates wearable devices, monitoring cameras, video processing, environmental monitoring modules, and cloud data analysis functions, and realizes comprehensive tracking and evaluation of the rehabilitation process. The system continuously monitors the patient's heart rate, blood pressure, and activity level through wearable devices, combines environmental parameters and video analysis, and accurately captures the patient's rehabilitation activities. The video processing module uses calibration points for image correction and accurately tracks joint movement trajectories. Combined with deep learning models (LSTM, CNN) for advanced feature extraction and fusion of physiological, movement, and environmental data, the system can classify rehabilitation stages, determine rehabilitation types, and predict rehabilitation scores. The system not only improves the quantitative evaluation of rehabilitation effects, but also provides real-time monitoring, alarm, and remote management functions through user terminals, promoting doctor-patient interaction and the implementation of personalized rehabilitation plans, and bringing innovation to the rehabilitation management of patients with movement disorders.

Description

Technical Field

[0001] This disclosure relates to the field of medical rehabilitation technology, and in particular to a real-time monitoring system for rehabilitation care of patients with mobility impairments. Background Technology

[0002] With the rapid development of medical technology, rehabilitation medicine, as an important component of the healthcare system, is becoming increasingly prominent. Especially when dealing with patients with mobility impairments, traditional rehabilitation methods often focus on on-site supervision and manual recording. This not only consumes a large amount of human resources but also has significant limitations in terms of the continuity, accuracy, and personalized feedback of monitoring. For example, monitoring of patients' physiological parameters may be limited to periodic clinical examinations, failing to capture real-time changes in daily activities; assessment of activity capacity relies on periodic functional tests, making it difficult to reflect rehabilitation progress or regression in real time; and the impact of environmental factors, such as temperature and humidity, on patient rehabilitation is often overlooked.

[0003] In recent years, with the rise of the Internet of Things (IoT), big data, artificial intelligence (AI), and cloud computing technologies, the healthcare field has ushered in unprecedented opportunities for transformation. The application of these technologies has made it possible to achieve intelligent and precise rehabilitation care. For example, the development of wearable devices has made continuous, non-invasive monitoring of patients' vital signs a reality; advances in video analytics and computer vision technologies have provided new avenues for accurately capturing and analyzing patients' movement behaviors; and cloud platforms and big data analytics can integrate multi-source heterogeneous data for in-depth analysis, providing a scientific basis for evaluating rehabilitation effectiveness and developing personalized rehabilitation plans.

[0004] However, despite significant technological advancements, the current market lacks a comprehensive real-time monitoring system that fully integrates the advantages of these technologies to provide patients with mobility impairments with a range of services, from physiological parameter monitoring and environmental adaptation assessment to motor function analysis and personalized rehabilitation recommendations. Some existing systems may focus only on a single dimension, such as physiological parameter monitoring or providing only simple video monitoring services, failing to form a closed-loop rehabilitation management solution. Furthermore, these systems often neglect efficient interaction with patients, families, and medical teams, lacking effective remote monitoring and immediate feedback mechanisms, thus limiting the optimization of rehabilitation outcomes and the effective utilization of medical resources.

[0005] Therefore, developing a real-time monitoring system for rehabilitation nursing of patients with mobility impairments that integrates wearable devices, high-definition video surveillance, environmental perception, cloud data analysis and decision support is of great significance for promoting the modernization of rehabilitation medicine, improving the quality and efficiency of rehabilitation treatment, and meeting the diversified and personalized rehabilitation needs of patients. Summary of the Invention

[0006] To address the above problems, the present invention provides a real-time monitoring system for rehabilitation nursing of patients with mobility impairments, comprising:

[0007] Monitoring system host: As the core of the system, it is responsible for receiving, processing and coordinating data transmission from various modules, and communicating with cloud modules to achieve remote data storage and analysis.

[0008] User terminal module: Installed on medical staff's workstations or family members' mobile devices, it provides real-time monitoring data display, alarm reception, and remote operation interface.

[0009] Wearable device module: including heart rate detection unit, blood pressure detection unit and activity level detection unit, which transmits the patient's heart rate, blood pressure and activity level data to the monitoring system host via a wireless communication module (such as Bluetooth or Wi-Fi) in encryption.

[0010] Monitoring camera module: Installed in the patient's activity area, it captures images of the patient's movement, and combines the images with calibration points for image correction and motion analysis, providing visual evidence for rehabilitation assessment.

[0011] Calibration point module: Equipped with visual feature encoding information for image correction, ensuring the accuracy of motion capture and the precision of spatial positioning.

[0012] Video processing module: Utilizes image recognition technology to process images captured by monitoring cameras, extracts the patient's movement trajectory, and provides data support for rehabilitation effect evaluation.

[0013] Environmental monitoring module: Monitors the temperature and humidity of the patient's environment to ensure the suitability of the rehabilitation environment.

[0014] Data storage module: Located on a local server, it stores all monitoring data and analysis results, facilitating long-term tracking and analysis.

[0015] Cloud module: Deployed in a remote data center, it not only provides data backup and storage but also runs rehabilitation analysis models, classifying rehabilitation stages, determining rehabilitation types, and calculating rehabilitation scores based on comprehensive data. The system details are as follows:

[0016] A real-time monitoring system for rehabilitation nursing of patients with mobility impairments includes a monitoring system host, a user terminal module, a monitoring camera, a wearable device, a video processing module, a calibration point, a cloud module, an environmental monitoring module, and a data storage module; the monitoring system host is connected to the user terminal module, the monitoring camera, the wearable device, the video processing module, the calibration point, the cloud module, the environmental monitoring module, and the data storage module.

[0017] Wearable device modules are worn by patients to collect their heart rate, blood pressure, and activity levels in real time;

[0018] The monitoring camera module is installed in the patient's activity area to capture time-series images of the patient and send them to the detection system host.

[0019] The calibration point is fixed at a fixed position in the patient's activity area to provide spatial coordinate reference and improve the accuracy of motion capture and positioning;

[0020] The video processing module is used to process the time-stream images captured by the monitoring camera module and send the processed images to the monitoring system host.

[0021] The environmental monitoring module is installed in different corners of the patient's activity area to detect the ambient temperature T and humidity Rh in real time;

[0022] The storage module is installed on the local server to store all monitoring data and analysis results for easy review and utilization later.

[0023] The cloud module is deployed in a remote data center to provide cloud data storage and cloud data analysis functions for the monitoring system host;

[0024] The user terminal module is installed on the workstations of medical staff, wards, or mobile devices of family members to display real-time monitoring data, receive alarm information, and enable remote operation and management.

[0025] The monitoring system host acquires data from the wearable device module, monitoring camera module, and environmental monitoring module, and sends the data to the cloud module. The cloud module runs the rehabilitation analysis model, which calculates and determines the patient's rehabilitation stage and type, and provides the patient's rehabilitation score. The specific rehabilitation stage, type, and score are then sent to the user terminal module.

[0026] It includes a heart rate detection unit, configured to continuously collect and digitize the patient's electrocardiogram signals in real time, accurately calculate and upload the patient's heart rate parameter HR to the monitoring system host;

[0027] The blood pressure monitoring unit uses non-invasive or minimally invasive techniques to measure the patient's blood pressure parameters in real time, including systolic blood pressure (SBP) and diastolic blood pressure (DBP), and sends the measurement results to the monitoring system host.

[0028] The activity level detection unit integrates an accelerometer and a gyroscope to record and quantify the patient's activity status information, including steps S, activity intensity I, and movement trajectory D.

[0029] The wireless communication module, using Bluetooth or Wi-Fi, encrypts the processed patient heart rate parameters (HR), blood pressure parameters (SBP, DBP), and activity parameters (S, I, D) and transmits them to the monitoring system host in real time.

[0030] The calibration point module is fixedly installed at key geometric locations within the patient's activity area, and each calibration point has visual feature encoding information;

[0031] The monitoring camera module sends the acquired time-stream images, including calibration points, to the monitoring system host. The monitoring system host then sends the time-stream images to the video processing module. The video processing module uses image recognition technology and computer vision algorithms to perform calibration and correction processing on the received time-stream images. The specific process of the correction processing is as follows:

[0032] 1) Perform real-time analysis on time-stream image sequences and extract the position information of each calibration point in the image through template matching or feature point detection techniques;

[0033] 2) Based on the principle of perspective transformation, the position of the calibration point in the two-dimensional image coordinate system is transformed to the pre-set three-dimensional world coordinate system, so as to realize the accurate mapping of the image to the real space;

[0034] 3) By comparing the observed position of the calibration point in the image with its known position in actual space, the least squares method is used to correct the internal and external parameters of the camera, thereby achieving effective correction of the camera's field of view and lens distortion.

[0035] The internal parameters include focal length and principal point coordinates, while the external parameters include rotation matrix and translation matrix.

[0036] The calibration point surface is provided with a QR code and edge markers; the QR code corresponds to the unique code of the calibration point, and the edge markers are used to help identify the edge of the calibration point.

[0037] The video processing module identifies the patient's joint movement trajectory based on time-stream images. The specific steps include:

[0038] Foot localization: By analyzing multiple frames of images of the patient's feet in contact with the ground when standing or walking, contour recognition and shape matching algorithms are used to determine the position of each foot in the image. Then, the precise coordinates of the patient's feet in three-dimensional space are calculated and determined through the spatial coordinate system of the calibration points.

[0039] Centroid coordinate determination: Given the spatial coordinates of both feet, image processing and geometric analysis methods are used to identify the patient's body contour, calculate the centroid position, and define the projection point of the centroid on the horizontal plane as the center point of the line connecting the projections of the coordinates of both feet on the horizontal plane, thereby determining the three-dimensional coordinates of the patient's centroid.

[0040] Feature point spatial coordinate determination: Based on foot coordinates and centroid coordinates, the skeleton estimation algorithm, OpenPose algorithm, is used to track and predict other feature points of the patient in continuous time-stream images: ankle joint, knee joint, hip joint, shoulder joint, elbow joint and wrist joint. Through the spatial relationship and kinematic model of these points, their coordinates in three-dimensional space are calculated and determined.

[0041] Determining the spatial coordinates of feature points specifically includes:

[0042] Initial feature point localization: First, based on the determined foot coordinates and centroid coordinates, the initial spatial coordinates of the ankle, knee, hip, shoulder, elbow and wrist joints are estimated using a pre-trained human key point detection model.

[0043] Spatiotemporal continuity optimization: For each feature point, optical flow or continuous inter-frame matching techniques are applied in a temporally continuous image sequence to track the positional changes of these feature points at different time points;

[0044] Motion trajectory generation: Based on the dynamic tracking results of feature point positions in continuous frames, the three-dimensional spatial motion trajectory of each joint of the patient is constructed; for each feature point, the position change in the time dimension is smoothed by interpolation or fitting methods to obtain a continuous and smooth motion path.

[0045] The cloud-based module runs the rehabilitation analysis model as follows:

[0046] Time axis alignment: Unify the timestamps of heart rate parameters (HR), blood pressure parameters (SBP and DBP), activity parameters (S, I, D) acquired in real time by the wearable device module; ambient temperature (T) and humidity (Rh) acquired by the environmental monitoring module; and motion path data of feature points acquired by the monitoring camera module; ensure that the data is standardized according to the unified timestamp and that the time axis is aligned.

[0047] Data preprocessing: missing value imputation, outlier detection and handling, and data standardization are performed on the aligned time series data;

[0048] Feature extraction: LSTM network is used to process time series physiological index data to obtain physiological features, CNN is used to process motion paths to extract motion features, and statistical feature extraction is performed on environmental data;

[0049] Feature fusion: The extracted physiological features, motion features and environmental features are fused to form a comprehensive feature vector, and an attention mechanism is used for weighted fusion;

[0050] Rehabilitation analysis and assessment: Support vector machine was used to classify rehabilitation stages, K-means clustering algorithm was used to determine rehabilitation type, and gradient boosting regression tree was used to calculate rehabilitation score.

[0051] The rehabilitation analysis model is constructed as follows:

[0052] Collect monitoring parameters from patients with movement disorders of different ages, genders, and etiologies to ensure that the dataset includes patients at different stages of rehabilitation, different types of rehabilitation, and different rehabilitation scores;

[0053] The preprocessed dataset is divided into training, validation, and test sets; monitoring data is used as input; and different rehabilitation stages, rehabilitation types, and rehabilitation scores are used as outputs to train the rehabilitation analysis model.

[0054] The monitoring data includes: heart rate parameters (HR), blood pressure parameters (SBP and DBP), activity parameters (S, I, D) acquired in real time by the wearable device module; ambient temperature (T) and humidity (Rh) acquired by the environmental monitoring module; and motion paths of feature points acquired by the monitoring camera module.

[0055] The rehabilitation stages include initial, intermediate, and final stages; the rehabilitation types include rapid recovery, normal recovery, and slow recovery.

[0056] The rehabilitation score uses a scale of 1-100, where 1 indicates the weakest rehabilitation effect and requires adjustment of the rehabilitation care plan; 100 indicates the best rehabilitation effect and requires no adjustment of the rehabilitation care plan.

[0057] The beneficial effects of this invention are as follows:

[0058] By integrating wearable devices, monitoring cameras, and environmental monitoring modules, this system enables comprehensive and real-time monitoring of physiological parameters (such as heart rate, blood pressure, and activity level), environmental conditions (temperature and humidity), and movement behaviors of patients with mobility impairments. This not only improves the continuity and accuracy of monitoring but also makes it possible to promptly detect health abnormalities and environmental discomfort, ensuring the safety and effectiveness of the rehabilitation process.

[0059] The system employs advanced image recognition and computer vision technologies, combined with human kinematics analysis, to accurately track the patient's joint movement trajectory, providing an intuitive and visual assessment of rehabilitation progress. By incorporating physiological data and environmental factors, the cloud-based rehabilitation analysis model can perform detailed rehabilitation stage classification, rehabilitation type determination, and rehabilitation scoring for patients, providing a scientific basis for developing personalized rehabilitation plans.

[0060] The cloud-based module employs sophisticated data processing algorithms, such as time series analysis, feature extraction, and fusion, to effectively integrate cross-dimensional data, providing strong support for in-depth analysis of rehabilitation outcomes. This intelligent analysis based on big data can not only accurately assess the current rehabilitation status but also predict rehabilitation trends, providing forward-looking decision support for adjusting rehabilitation strategies.

[0061] The establishment of the user terminal module enables medical staff and family members to remotely view the patient's status in real time, receive alarm information, and even participate in rehabilitation management. This strengthens doctor-patient communication, improves nursing efficiency, reduces the burden on patients caused by frequent trips to medical institutions, and enhances the patient's rehabilitation experience and sense of participation.

[0062] Through an intelligent and remote rehabilitation and nursing model, this invention helps optimize the allocation of medical resources, especially in areas with limited or remote medical resources. It enables more patients with mobility impairments to enjoy high-quality rehabilitation services, promoting greater equity and accessibility in healthcare. Through technological integration and innovation, this invention not only improves the quality and efficiency of rehabilitation but also enhances patients' quality of life, representing a significant advancement in the field of modern medical rehabilitation. Attached Figure Description

[0063] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0064] Appendix Figure 1 This is a schematic diagram of the overall architecture of the present invention;

[0065] Appendix Figure 2 This is a schematic diagram of the arrangement of calibration points for this invention. Detailed Implementation

[0066] Example 1:

[0067] See Figure 1 and Figure 2 This invention provides a real-time monitoring system for rehabilitation and nursing of patients with mobility impairments, including a monitoring system host, a user terminal module, a monitoring camera, a wearable device, a video processing module, a calibration point, a cloud module, an environmental monitoring module, and a data storage module; the monitoring system host is connected to the user terminal module, the monitoring camera, the wearable device, the video processing module, the calibration point, the cloud module, the environmental monitoring module, and the data storage module;

[0068] Wearable device modules are worn by patients to collect their heart rate, blood pressure, and activity levels in real time;

[0069] The monitoring camera module is installed in the patient's activity area to capture time-series images of the patient and send them to the detection system host.

[0070] The calibration point is fixed at a fixed position in the patient's activity area to provide spatial coordinate reference and improve the accuracy of motion capture and positioning;

[0071] The video processing module is used to process the time-stream images captured by the monitoring camera module and send the processed images to the monitoring system host.

[0072] The environmental monitoring module is installed in different corners of the patient's activity area to detect the ambient temperature T and humidity Rh in real time;

[0073] The storage module is installed on the local server to store all monitoring data and analysis results for easy review and utilization later.

[0074] The cloud module is deployed in a remote data center to provide cloud data storage and cloud data analysis functions for the monitoring system host;

[0075] The user terminal module is installed on the workstations of medical staff, wards, or mobile devices of family members to display real-time monitoring data, receive alarm information, and enable remote operation and management.

[0076] The monitoring system host acquires data from the wearable device module, monitoring camera module, and environmental monitoring module, and sends the data to the cloud module. The cloud module runs the rehabilitation analysis model, which calculates and determines the patient's rehabilitation stage and type, and provides the patient's rehabilitation score. The specific rehabilitation stage, type, and score are then sent to the user terminal module.

[0077] It includes a heart rate detection unit, configured to continuously collect and digitize the patient's electrocardiogram signals in real time, accurately calculate and upload the patient's heart rate parameter HR to the monitoring system host;

[0078] The blood pressure monitoring unit uses non-invasive or minimally invasive techniques to measure the patient's blood pressure parameters in real time, including systolic blood pressure (SBP) and diastolic blood pressure (DBP), and sends the measurement results to the monitoring system host.

[0079] The activity level detection unit integrates an accelerometer and a gyroscope to record and quantify the patient's activity status information, including steps S, activity intensity I, and movement trajectory D.

[0080] The wireless communication module, using Bluetooth or Wi-Fi, encrypts the processed patient heart rate parameters (HR), blood pressure parameters (SBP, DBP), and activity parameters (S, I, D) and transmits them to the monitoring system host in real time.

[0081] The calibration point module is fixedly installed at key geometric locations within the patient's activity area, and each calibration point has visual feature encoding information;

[0082] The monitoring camera module sends the acquired time-stream images, including calibration points, to the monitoring system host. The monitoring system host then sends the time-stream images to the video processing module. The video processing module uses image recognition technology and computer vision algorithms to perform calibration and correction processing on the received time-stream images. The specific process of the correction processing is as follows:

[0083] 1) Perform real-time analysis on time-stream image sequences and extract the position information of each calibration point in the image through template matching or feature point detection techniques;

[0084] 2) Based on the principle of perspective transformation, the position of the calibration point in the two-dimensional image coordinate system is transformed to the pre-set three-dimensional world coordinate system, so as to realize the accurate mapping of the image to the real space;

[0085] 3) By comparing the observed position of the calibration point in the image with its known position in actual space, the least squares method is used to correct the internal and external parameters of the camera, thereby achieving effective correction of the camera's field of view and lens distortion.

[0086] The internal parameters include focal length and principal point coordinates, while the external parameters include rotation matrix and translation matrix.

[0087] The calibration point surface is provided with a QR code and edge markers; the QR code corresponds to the unique code of the calibration point, and the edge markers are used to help identify the edge of the calibration point.

[0088] The video processing module identifies the patient's joint movement trajectory based on time-stream images. The specific steps include:

[0089] Foot localization: By analyzing multiple frames of images of the patient's feet in contact with the ground when standing or walking, contour recognition and shape matching algorithms are used to determine the position of each foot in the image. Then, the precise coordinates of the patient's feet in three-dimensional space are calculated and determined through the spatial coordinate system of the calibration points.

[0090] Centroid coordinate determination: Given the spatial coordinates of both feet, image processing and geometric analysis methods are used to identify the patient's body contour, calculate the centroid position, and define the projection point of the centroid on the horizontal plane as the center point of the line connecting the projections of the coordinates of both feet on the horizontal plane, thereby determining the three-dimensional coordinates of the patient's centroid.

[0091] Feature point spatial coordinate determination: Based on foot coordinates and centroid coordinates, the skeleton estimation algorithm, OpenPose algorithm, is used to track and predict other feature points of the patient in continuous time-stream images: ankle joint, knee joint, hip joint, shoulder joint, elbow joint and wrist joint. Through the spatial relationship and kinematic model of these points, their coordinates in three-dimensional space are calculated and determined.

[0092] Determining the spatial coordinates of feature points specifically includes:

[0093] Initial feature point localization: First, based on the determined foot coordinates and centroid coordinates, the initial spatial coordinates of the ankle, knee, hip, shoulder, elbow and wrist joints are estimated using a pre-trained human key point detection model.

[0094] Spatiotemporal continuity optimization: For each feature point, optical flow or continuous inter-frame matching techniques are applied in a temporally continuous image sequence to track the positional changes of these feature points at different time points;

[0095] Motion trajectory generation: Based on the dynamic tracking results of feature point positions in continuous frames, the three-dimensional spatial motion trajectory of each joint of the patient is constructed; for each feature point, the position change in the time dimension is smoothed by interpolation or fitting methods to obtain a continuous and smooth motion path.

[0096] The cloud-based module runs the rehabilitation analysis model as follows:

[0097] Time axis alignment: Unify the timestamps of heart rate parameters (HR), blood pressure parameters (SBP and DBP), activity parameters (S, I, D) acquired in real time by the wearable device module; ambient temperature (T) and humidity (Rh) acquired by the environmental monitoring module; and motion path data of feature points acquired by the monitoring camera module; ensure that the data is standardized according to the unified timestamp and that the time axis is aligned.

[0098] Data preprocessing: missing value imputation, outlier detection and handling, and data standardization are performed on the aligned time series data;

[0099] Feature extraction: LSTM network is used to process time series physiological index data to obtain physiological features, CNN is used to process motion paths to extract motion features, and statistical feature extraction is performed on environmental data;

[0100] Feature fusion: The extracted physiological features, motion features and environmental features are fused to form a comprehensive feature vector, and an attention mechanism is used for weighted fusion;

[0101] Rehabilitation analysis and assessment: Support vector machine was used to classify rehabilitation stages, K-means clustering algorithm was used to determine rehabilitation type, and gradient boosting regression tree was used to calculate rehabilitation score.

[0102] The rehabilitation analysis model is constructed as follows:

[0103] Collect monitoring parameters from patients with movement disorders of different ages, genders, and etiologies to ensure that the dataset includes patients at different stages of rehabilitation, different types of rehabilitation, and different rehabilitation scores;

[0104] The preprocessed dataset is divided into training, validation, and test sets; monitoring data is used as input; and different rehabilitation stages, rehabilitation types, and rehabilitation scores are used as outputs to train the rehabilitation analysis model.

[0105] The monitoring data includes: heart rate parameters (HR), blood pressure parameters (SBP and DBP), activity parameters (S, I, D) acquired in real time by the wearable device module; ambient temperature (T) and humidity (Rh) acquired by the environmental monitoring module; and motion paths of feature points acquired by the monitoring camera module.

[0106] The rehabilitation stages include initial, intermediate, and final stages; the rehabilitation types include rapid recovery, normal recovery, and slow recovery.

[0107] The rehabilitation score uses a scale of 1-100, where 1 indicates the weakest rehabilitation effect and requires adjustment of the rehabilitation care plan; 100 indicates the best rehabilitation effect and requires no adjustment of the rehabilitation care plan.

[0108] In the rehabilitation analysis model running in the cloud module of this invention, deep learning technology is used to construct a multi-layered neural network structure, the specific layers of which include, but are not limited to, the following key components:

[0109] Input layer:

[0110] The system receives and standardizes various data sets: heart rate (HR), blood pressure (SBP & DBP), activity parameters (steps S, activity intensity I, motion trajectory D) from wearable devices, environmental monitoring data (temperature (T) and humidity (Rh), and feature point motion path data output from the video processing module. These data, after time-axis alignment and preprocessing, are used as input to the model.

[0111] Feature extraction layer:

[0112] Physiological feature extraction: Long Short-Term Memory (LSTM) network is used to process time series data (such as changes in heart rate and blood pressure) to capture time-dependent features and extract dynamic features that reflect the patient's physiological state.

[0113] Motion feature extraction: Convolutional neural networks (CNNs) are used to process motion path images of feature points to capture spatial features and motion patterns, such as the motion patterns of joints.

[0114] Environmental feature extraction: Statistical analysis of environmental data is performed to extract the characteristics of environmental influencing factors related to rehabilitation.

[0115] Feature fusion layer:

[0116] The extracted physiological, motor, and environmental features are weighted and fused using an attention mechanism to form a comprehensive feature vector. This layer aims to integrate multi-source information and enhance the model's ability to understand complex rehabilitation situations.

[0117] Prediction / Classification Layer:

[0118] Rehabilitation stage classification: Support vector machine (SVM) is used to classify the fused features to determine the patient's current rehabilitation stage (early, middle, or late).

[0119] Rehabilitation type determination: Patients are clustered using the K-means clustering algorithm to identify their rehabilitation type (rapid recovery, normal recovery, slow recovery).

[0120] Rehabilitation score prediction: The Gradient Boosting Regression Tree (GBRT) model is applied to predict the patient's rehabilitation score based on the comprehensive feature vector, thereby quantifying the rehabilitation effect.

[0121] Output layer:

[0122] It provides the final rehabilitation stage classification, rehabilitation type and rehabilitation score, providing medical staff with clear rehabilitation assessment results to guide the development and adjustment of personalized rehabilitation care plans.

[0123] In summary, this invention integrates multi-module technology to achieve intelligent monitoring of the patient's recovery process. The monitoring system host is connected to user terminal modules, monitoring cameras, wearable devices, etc., constructing a highly efficient platform for data exchange.

[0124] Wearable devices are worn by patients to collect heart rate, blood pressure, and activity levels in real time, continuously monitoring physiological parameters and responding instantly to changes in health status. Monitoring camera modules are installed in the activity area to capture time-series images and transmit them to the host computer, providing visual feedback and aiding in behavioral and environmental adaptation analysis. Calibration points are deployed in fixed locations as spatial references to improve motion capture accuracy, enhance positioning reliability, and ensure accurate motion analysis. A video processing module processes and optimizes image data before transmitting it to the host computer, using image recognition technology to refine motion analysis and behavioral pattern recognition. Environmental monitoring modules are distributed in corners to detect temperature and humidity in real time, monitoring environmental comfort and ensuring a suitable rehabilitation environment. A data storage module stores all monitoring information for easy review and research, establishing a comprehensive rehabilitation history archive and supporting in-depth data mining. A cloud module provides remote data access and analysis services, enabling cross-regional resource sharing and accelerating rehabilitation strategy decision-making. User terminals display real-time data and alarms, supporting remote management and enhancing the participation and intervention capabilities of medical staff and families. The monitoring system host computer integrates data and sends it to the cloud, using rehabilitation analysis models to assess the rehabilitation stage and type, providing scores, scientifically guiding personalized rehabilitation pathways, and improving rehabilitation effectiveness.

[0125] Thus far, the description of the above embodiments has been provided for illustrative and descriptive purposes. This is not intended to be exhaustive or limiting of the present disclosure. Individual elements or features of particular embodiments are generally not limited to those particular embodiments, but may be interchanged and used in selected embodiments where applicable, even if not specifically shown or described. In many respects, the same elements or features may also be varied. Such variations are not considered a departure from this disclosure, and all such modifications are intended to be included within the scope of this disclosure.

[0126] Example embodiments are provided so that this disclosure will become thorough and will fully convey the scope to those skilled in the art. Numerous details, such as examples of specific parts, apparatus, and methods, are set forth to provide a thorough understanding of embodiments of this disclosure. It will be apparent to those skilled in the art that the specific details are not required, and the example embodiments may be implemented in many different forms, neither of which should be construed as limiting the scope of this disclosure. In some example embodiments, well-known processes, well-known apparatus structures, and well-known techniques are not described in detail.

[0127] Technical terms are used herein for the purpose of describing specific exemplary embodiments only and are not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms “a” and “the” as used herein may also refer to the plural forms. The terms “comprising” and “having” are inclusive and therefore specify the presence of the stated features, integrals, steps, operations, elements, and / or components, but do not exclude the presence or additional having of one or more other features, integrals, steps, operations, elements, components, and / or combinations thereof. Unless expressly indicated in order of execution, the method steps, processes, and operations described herein are not to be construed as necessarily requiring performance in the specific order discussed and shown. It should also be understood that additional or optional steps may be employed.

Claims

1. A real-time monitoring system for rehabilitation nursing of patients with mobility impairments, comprising a monitoring system host, a user terminal module, a monitoring camera, a wearable device, a video processing module, calibration points, a cloud module, an environmental monitoring module, and a data storage module; the monitoring system host and the user terminal module, monitoring camera, wearable device, video processing module, cloud module, environmental monitoring module, and data storage module are connected; Wearable devices are worn by patients to collect their heart rate, blood pressure, and activity levels in real time; Monitoring cameras are installed in the patient's activity area to capture time-stream images, including calibration points, and send them to the monitoring system host. The calibration point is fixed at a key geometric location in the patient's activity area. The surface of the calibration point is provided with a QR code and an edge marker. The QR code corresponds to a unique code of the calibration point, and the edge marker is used to help identify the edge of the calibration point. The monitoring system host sends the time-stream images to the video processing module, which processes the time-stream images captured by the monitoring cameras and sends the processed images back to the monitoring system host. The video processing module identifies the patient's joint movement trajectory based on time-stream images. The specific steps include: Foot localization: By analyzing multiple frames of images of the patient's feet in contact with the ground when standing or walking, contour recognition and shape matching algorithms are used to determine the position of each foot in the image. Then, the precise coordinates of the patient's feet in three-dimensional space are calculated and determined through the spatial coordinate system of the calibration points. Centroid coordinate determination: Given the spatial coordinates of both feet, image processing and geometric analysis methods are used to identify the patient's body contour, calculate the centroid position, and define the projection point of the centroid on the horizontal plane as the center point of the line connecting the projections of the coordinates of both feet on the horizontal plane, thereby determining the three-dimensional coordinates of the patient's centroid. Feature point spatial coordinate determination: Based on foot coordinates and centroid coordinates, the OpenPose algorithm is used to track and predict other feature points of the patient in continuous time-stream images: ankle joint, knee joint, hip joint, shoulder joint, elbow joint and wrist joint. Through the spatial relationship and kinematic model of these points, their coordinates in three-dimensional space are calculated and determined. The environmental monitoring module is installed in different corners of the patient's activity area to detect the ambient temperature T and humidity Rh in real time; The data storage module is installed on a local server to store all monitoring data and analysis results; The cloud module is deployed in a remote data center to provide cloud data storage and cloud data analysis functions for the monitoring system host; The user terminal module is installed on the workstations of medical staff, wards, or mobile devices of family members to display real-time monitoring data, receive alarm information, and enable remote operation and management. The monitoring system host acquires data from wearable devices, monitoring cameras, video processing modules, and environmental monitoring modules, and sends the data to the cloud module. The cloud module runs a rehabilitation analysis model, which calculates and determines the patient's rehabilitation stage and type, and provides the patient's rehabilitation score. The rehabilitation stage, type, and score are then sent to the user terminal module.

2. The real-time monitoring system for rehabilitation nursing of patients with mobility impairments according to claim 1, characterized in that: The wearable device includes: The heart rate detection unit is configured to continuously collect and digitize the patient's electrocardiogram signals in real time, accurately calculate and upload the patient's heart rate parameter HR to the monitoring system host. The blood pressure monitoring unit uses non-invasive or minimally invasive techniques to measure the patient's blood pressure parameters in real time, including systolic blood pressure (SBP) and diastolic blood pressure (DBP), and sends the measurement results to the monitoring system host. The activity level detection unit integrates an accelerometer and a gyroscope to record and quantify the patient's activity status information, including steps S, activity intensity I, and movement trajectory D. The wireless communication module, using Bluetooth or Wi-Fi, encrypts the processed patient heart rate parameters (HR), blood pressure parameters (SBP, DBP), and activity parameters (S, I, D) and transmits them to the monitoring system host in real time.

3. The real-time monitoring system for rehabilitation nursing of patients with mobility impairments according to claim 1, characterized in that: The video processing module utilizes image recognition technology and computer vision algorithms to perform calibration and correction processing on the received time-stream images; the specific process of the correction processing is as follows: 1) Perform real-time analysis on time-stream image sequences and extract the position information of each calibration point in the image through template matching or feature point detection techniques; 2) Based on the principle of perspective transformation, the position of the calibration point in the two-dimensional image coordinate system is transformed to the pre-set three-dimensional world coordinate system, so as to realize the accurate mapping of the image to the real space; 3) By comparing the observed position of the calibration point in the image with its known position in actual space, the least squares method is used to correct the internal and external parameters of the camera, thereby achieving effective correction of the camera's field of view and lens distortion. Internal parameters include focal length and principal point coordinates, while external parameters include rotation matrix and translation matrix.

4. The real-time monitoring system for rehabilitation nursing of patients with mobility impairments according to claim 3, characterized in that: The determination of the spatial coordinates of the feature points specifically includes: Initial feature point localization: First, based on the determined foot coordinates and centroid coordinates, the initial spatial coordinates of the ankle, knee, hip, shoulder, elbow and wrist joints are estimated using a pre-trained human key point detection model. Spatiotemporal continuity optimization: For each feature point, optical flow or continuous inter-frame matching techniques are applied in a temporally continuous image sequence to track the positional changes of these feature points at different time points; Motion trajectory generation: Based on the dynamic tracking results of feature point positions in continuous frames, the three-dimensional spatial motion trajectory of each joint of the patient is constructed; for each feature point, the position change in the time dimension is smoothed by interpolation or fitting methods to obtain a continuous and smooth motion path.

5. The real-time monitoring system for rehabilitation nursing of patients with mobility impairments according to claim 4, characterized in that: The cloud-based module runs the rehabilitation analysis model as follows: Time axis alignment: Unify the timestamps of heart rate parameters (HR), blood pressure parameters (SBP and DBP), activity parameters (S, I, D) acquired in real time by wearable devices; ambient temperature (T) and humidity (Rh) acquired by the environmental monitoring module; and motion path data of feature points acquired by the video processing module; ensure that the data is standardized according to the unified timestamps and that the time axis is aligned. Data preprocessing: missing value imputation, outlier detection and handling, and data standardization are performed on the aligned time series data; Feature extraction: LSTM network is used to process time series physiological index data to obtain physiological features, CNN is used to process motion paths to extract motion features, and statistical feature extraction is performed on environmental data to extract environmental features. feature Fusion: The extracted physiological features, motion features, and environmental features are fused to form a comprehensive feature vector, and an attention mechanism is used for weighted fusion; Rehabilitation analysis and assessment: Support vector machine was used to classify rehabilitation stages, K-means clustering algorithm was used to determine rehabilitation type, and gradient boosting regression tree was used to calculate rehabilitation score.

6. The real-time monitoring system for rehabilitation nursing of patients with mobility impairments according to claim 5, characterized in that: The rehabilitation analysis model is constructed as follows: Collect monitoring parameters from patients with movement disorders of different ages, genders, and etiologies to ensure that the dataset includes patients at different stages of rehabilitation, different types of rehabilitation, and different rehabilitation scores; The preprocessed dataset is divided into a training set, a validation set, and a test set; Use monitoring data as input; The rehabilitation analysis model was trained using different rehabilitation stages, different rehabilitation types, and different rehabilitation scores as outputs. The monitoring data includes: heart rate parameters (HR), blood pressure parameters (SBP and DBP), activity parameters (S, I, D) acquired in real time by the wearable device; ambient temperature (T) and humidity (Rh) acquired by the environmental monitoring module; and motion paths of feature points acquired by the video processing module. The rehabilitation stages include initial, intermediate, and final stages; the rehabilitation types include rapid recovery, normal recovery, and slow recovery. The rehabilitation score uses a scale of 1-100, where 1 indicates the weakest rehabilitation effect and requires adjustment of the rehabilitation care plan; 100 indicates the best rehabilitation effect and requires no adjustment of the rehabilitation care plan.

Images