Human behavior analysis and early warning method and system based on image technology

Abstract

The present invention provides a human behavior analysis and early warning method and system based on image technology, which are applied to the field of image data processing; the present invention can automatically identify users whose conditions are getting worse by monitoring the behavior data of waiting users in real time, and highlight these users through a camera terminal according to a preset criticality level, so as to ensure that medical staff can quickly locate emergency users who need priority treatment, thereby optimizing the allocation of medical resources, and at the same time, through continuous voice broadcast and behavior data synchronization, the response speed of medical staff to sudden illness is improved, and medical delays caused by late registration sequence are reduced, and this real-time early warning and tracking mechanism enhances the user's sense of security, improves the medical experience, and effectively solves the problem of not being discovered in time due to worsening of the disease.

Description

Technical Field

[0001] The present invention relates to the field of image data processing, and in particular to a human behavior analysis and early warning method and system based on image technology. Background Art

[0002] Medical informatization refers to the digitization, networking and informationization of medical services. It means using computer science, modern network communication technology and database technology to provide collection, storage, processing, extraction and data exchange of patient information and management information between hospitals and between departments of hospitals, and to meet the functional requirements of all authorized users.

[0003] At present, users who go to the hospital outpatient department to register have different severity of their illnesses, which means that the outpatient department can only provide medical services to users according to the priority sequence of registration. If some users with serious illnesses worsen prematurely while waiting for medical treatment because their registration sequence is late, it will be difficult for medical staff to discover these users with sudden illnesses in the first time, and therefore it is difficult to ensure the safety of these users with sudden illnesses. Summary of the invention

[0004] The present invention aims to solve the problem that users with serious illnesses become ill earlier than expected while waiting for medical treatment because their registration sequence is delayed, and medical staff find it difficult to detect these users with sudden illnesses in the first place. The present invention provides a human behavior analysis and early warning method and system based on image technology.

[0005] The present invention adopts the following technical means to solve the technical problem:

[0006] The present invention provides a human behavior analysis and early warning method based on image technology, comprising:

[0007] Based on the camera terminal preset in the outpatient department, obtain the camera data in the preset area;

[0008] Determining whether preset human body data is detected in the camera data;

[0009] If yes, then mark the corresponding waiting user from the preset area according to the human body data, generate the registration sequence information of the waiting user in the camera terminal according to the identity information of the waiting user pre-entered by the outpatient department, and collect the behavior data of the waiting user in the preset time period in real time;

[0010] Determine whether the behavior data meets the preset warning content;

[0011] If it meets the requirements, the waiting user is highlighted in the preset area based on the behavior data, a preset critical level corresponding to the waiting user is generated, emergency users are classified among the waiting users according to the critical level, the behavior data of the emergency user is synchronized to the outpatient department, and the emergency user is continuously broadcast to the outpatient department using a preset voice terminal until the emergency user can no longer be detected in the preset area.

[0012] Furthermore, before the step of highlighting the waiting user in the preset area based on the behavior data, the method further includes:

[0013] Based on the facial features preset by the camera terminal, collecting remote facial data of the waiting user;

[0014] Determining whether the remote facial data detects pre-recorded surface symptoms, wherein the surface symptoms specifically include facial symptoms, skin symptoms, and blood flow symptoms;

[0015] If so, the registration waiting seat corresponding to the waiting user is selected in the preset area, the preset distress call content of the registration waiting seat is activated according to the surface symptoms, and the reserved registration information of the waiting user is adaptively uploaded to the corresponding outpatient department based on the distress call content, wherein the reserved registration information specifically includes the required consultation information, the targeted consultation department and the historical consultation information.

[0016] Furthermore, after the step of generating the registration sequence information of the waiting user in the camera terminal, the method further includes:

[0017] Acquire additional symptom information pre-entered by the waiting user into the outpatient department, and identify the registration waiting seat of the waiting user in the preset area based on the additional symptom information, wherein the additional symptom information specifically includes past symptoms, chronic diseases and symptoms under treatment;

[0018] Determine whether the user sitting on the registered waiting seat is the waiting user;

[0019] If so, then the image data of the waiting user is collected in real time according to the additional symptom information, the symptom characteristics of the waiting user are extracted from the image data, the duration of the symptom characteristics is obtained, and the corresponding trend symptom characteristics are generated according to the duration, and the trend symptom characteristics are input into the pre-trained prediction model for prediction, so as to predict the probability of the waiting user having the symptom corresponding to the additional symptom information, wherein the image data specifically includes facial data, body posture and movement characteristics.

[0020] Furthermore, the step of marking the corresponding waiting user from the preset area according to the human body data includes:

[0021] Based on the virtual coordinates of the outpatient department divided in the preset area, the coordinate information of the waiting user corresponding to the preset area is collected, and a virtual boundary box of the waiting user is generated according to the coordinate information;

[0022] determining whether the virtual bounding box can continuously capture the waiting user;

[0023] If possible, the coordinate information is synchronized in real time in the preset multi-camera perspectives according to the virtual boundary frame, the coordinate information is uniformly converted into the same physical coordinate system, and the multi-perspective behavior data of the waiting user is obtained from the multi-camera perspectives.

[0024] Furthermore, the step of determining whether preset human body data is detected in the camera data further includes:

[0025] Based on the surface features pre-collected by the outpatient department from the waiting user, identifying image features corresponding to the surface features from the preset area;

[0026] Determining whether the image feature can be detected in the preset area;

[0027] If possible, based on the different image features, the waiting users corresponding to the image features are captured in real time in the preset area, and the registration waiting seats corresponding to the waiting users are generated through the outpatient department.

[0028] Furthermore, the step of determining whether the behavior data meets the preset warning content also includes:

[0029] The abnormal behavior of the waiting user is captured by the camera terminal to obtain the stagnation time corresponding to the abnormal behavior, wherein the abnormal behavior specifically includes individual behavior and collective behavior;

[0030] Determining whether the stagnation time exceeds a preset time;

[0031] If so, identify the concurrent behaviors attached to the abnormal behavior, substitute the abnormal behavior and the concurrent behaviors into the symptom map pre-constructed by the outpatient department, and divide the abnormal combination rules of the waiting users from the symptom map.

[0032] Furthermore, the step of acquiring the camera data in the preset area based on the camera terminal preset in the outpatient department also includes:

[0033] Acquire the identity information of the waiting user pre-entered in the outpatient department, and identify the corresponding waiting user in the preset area based on the identity information;

[0034] Determining whether the waiting user exists in the preset area;

[0035] If so, the image acquisition mode of the camera terminal is detected, and the image acquisition mode is converted into a preset continuous frame format, and the image data of the waiting user is synchronized to the outpatient department in the continuous frame format, wherein the image acquisition mode specifically includes single frame acquisition and continuous frame acquisition.

[0036] The present invention also provides a human behavior analysis and early warning system based on image technology, comprising:

[0037] An acquisition module is used to acquire the camera data in a preset area based on the camera terminal preset in the outpatient department;

[0038] A judging module, used for judging whether preset human body data is detected in the camera data;

[0039] an execution module, for marking the corresponding waiting user from the preset area according to the human body data, generating the registration sequence information of the waiting user in the camera terminal according to the identity information of the waiting user pre-entered by the outpatient department, and collecting the behavior data of the waiting user in a preset time period in real time;

[0040] The second judgment module is used to judge whether the behavior data conforms to the preset warning content;

[0041] The second execution module is used for highlighting the waiting user in the preset area based on the behavior data if it meets the requirements, generating a preset criticality level corresponding to the waiting user, classifying emergency users among the waiting users according to the criticality level, synchronizing the behavior data of the emergency user to the outpatient department, and continuously broadcasting the emergency user through the outpatient department using a preset voice terminal until the emergency user can no longer be detected in the preset area.

[0042] Furthermore, it also includes:

[0043] A collection module, used for collecting remote facial data of the waiting user based on the facial features preset by the camera terminal;

[0044] A third determination module is used to determine whether the remote facial data detects pre-recorded surface diseases, wherein the surface diseases specifically include facial diseases, skin diseases and blood flow diseases;

[0045] The third execution module is used for selecting the registration waiting seat corresponding to the waiting user in the preset area, activating the preset distress content of the registration waiting seat according to the surface symptoms, and adaptively uploading the reserved registration information of the waiting user to the corresponding outpatient department according to the distress content, wherein the reserved registration information specifically includes the required consultation information, the targeted consultation department and the historical consultation information.

[0046] Furthermore, it also includes:

[0047] A second acquisition module is used to acquire additional symptom information pre-entered by the waiting user into the outpatient department, and identify the registration waiting seat of the waiting user in the preset area based on the additional symptom information, wherein the additional symptom information specifically includes past symptoms, chronic diseases and symptoms under treatment;

[0048] A fourth judgment module is used to judge whether the user sitting on the registered waiting seat belongs to the waiting user;

[0049] The fourth execution module is used to, if yes, collect the image data of the waiting user in real time according to the additional symptom information, extract the symptom characteristics of the waiting user from the image data, obtain the duration of the symptom characteristics, generate corresponding trend symptom characteristics according to the duration, input the trend symptom characteristics into the pre-trained prediction model for prediction, and predict the probability of the waiting user having the symptom corresponding to the additional symptom information, wherein the image data specifically includes facial data, body posture and movement characteristics.

[0050] The present invention provides a human behavior analysis and early warning method and system based on image technology, which has the following beneficial effects:

[0051] The present invention can automatically identify users whose conditions are getting worse by real-time monitoring of the behavioral data of waiting users, and highlight these users through the camera terminal according to the preset criticality level, ensuring that medical staff can quickly locate emergency users who need priority treatment, thereby optimizing the allocation of medical resources. At the same time, through continuous voice broadcast and behavioral data synchronization, the response speed of medical staff to sudden medical conditions is improved, and medical delays caused by late registration sequences are reduced. In addition, this real-time early warning and tracking mechanism enhances the user's sense of security, improves the medical experience, and effectively solves the problem of untimely detection of worsening conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 A schematic diagram of a flow chart of an embodiment of a human behavior analysis and early warning method based on image technology of the present invention;

[0053] Figure 2This is a structural block diagram of an embodiment of the human behavior analysis and early warning system based on image technology of the present invention. DETAILED DESCRIPTION

[0054] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. The implementation of the objectives, functional features and advantages of the present invention will be further described in conjunction with the embodiments and with reference to the accompanying drawings.

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] Reference Figure 1 , is a human behavior analysis and early warning method based on image technology in an embodiment of the present invention, comprising:

[0057] S1: Based on the camera terminal preset in the outpatient department, obtain the camera data in the preset area;

[0058] S2: Determine whether preset human body data is detected in the camera data;

[0059] S3: If yes, mark the corresponding waiting user from the preset area according to the human body data, generate the registration sequence information of the waiting user in the camera terminal according to the identity information of the waiting user pre-entered by the outpatient department, and collect the behavior data of the waiting user in the preset time period in real time;

[0060] S4: Determine whether the behavior data meets the preset warning content;

[0061] S5: If it meets the requirements, the waiting user is highlighted in the preset area based on the behavior data, a preset critical level corresponding to the waiting user is generated, emergency users are classified among the waiting users according to the critical level, the behavior data of the emergency user is synchronized to the outpatient department, and the emergency user is continuously reported to through the outpatient department using a preset voice terminal until the emergency user can no longer be detected in the preset area.

[0062] In this embodiment, the system obtains the camera data in each preset waiting area based on the camera terminal pre-set in the waiting area of ​​the outpatient department, and then the system determines whether the camera data detects the pre-set human body surface data, such as body contour or wearing accessories, to execute the corresponding steps; for example, when the system determines that the camera data in each waiting area cannot detect the pre-set human body surface data, the system will consider that there are no waiting users in these waiting areas, and the system will check whether the camera terminal is working properly, including network connection, camera clarity and light conditions, to ensure that there is no fault in the equipment, and re-call the detection function of the camera terminal to scan each waiting area again. To confirm that no user has been detected. If no user has been detected for multiple consecutive times, an alarm will be triggered to notify relevant staff to check the actual situation to prevent technical failures or other abnormalities, and notify medical staff or security to go to the relevant waiting area for on-site inspection to confirm whether there are really no waiting users. The system will adjust the monitoring parameters of the camera terminal, such as changing the camera angle and adjusting the sensitivity, to ensure that users in the area will not be missed. By notifying the front desk or relevant staff of the outpatient department that there is no one waiting in the current area, they will ensure that they understand the situation and make corresponding arrangements to avoid waiting users waiting in the wrong area. For example, when the system determines that the camera data in each waiting area can be detected When the system sees the pre-set human body data, it will think that there are waiting users in the waiting area. The system will mark and identify the corresponding waiting users in the waiting area according to different human body surface data, that is, different body contours or wearing accessories, and generate the registration sequence information of these waiting users in the camera terminal according to the identity information of these waiting users pre-entered by the outpatient department, and collect the behavior data of these waiting users in the pre-set time period in real time; the system accurately marks and identifies each waiting user according to the human body surface data to ensure user identity matching and avoid confusion or misidentification of different types of waiting users. At the same time, the registration sequence information is automatically generated according to the user identity information, making the registration process more efficient. Efficient and seamless connection reduces waiting time. Since the registration sequence information can be updated in real time, it ensures that the information of each waiting user accurately reflects the current status. In addition, the behavioral data of waiting users is collected in real time, which can timely capture the status changes of users during the waiting period. For example, in case of emergency situations such as worsening of the condition, through real-time monitoring, the system can dynamically adjust or notify medical staff to make necessary interventions on users, giving priority to users who need urgent medical attention. Waiting users know that their status is being monitored in real time, and they can get timely attention and treatment in case of emergency, thus enhancing their sense of security. Then the system determines whether the behavioral data of waiting users meets the pre-set warning content to execute the corresponding steps.For example, when the system determines that the behavior data of a waiting user does not meet the pre-set warning content, the system will consider that the waiting user has not detected any symptoms requiring emergency assistance during the waiting stage. The system will continue to monitor the behavior data of the waiting user in real time to ensure that any potential health changes during the waiting period can be captured in time, and regularly analyze the user's behavior data. Even if there are minor changes, they can attract the attention of the system in a short time. At the same time, the system will provide users with regular health tips through voice terminals or display screens, reminding them to inform staff in time if they feel uncomfortable, and synchronize the behavior data and analysis results of waiting users to the outpatient department for reference by medical staff when the user is seeing a doctor, helping them to make a more accurate diagnosis; for example, when the system determines that the behavior data of a waiting user meets the pre-set warning content, the system will consider that the waiting user has a sudden illness while waiting. The system will highlight the waiting user in the waiting area of ​​the camera terminal based on different behavior data, generate a pre-set criticality level corresponding to the waiting user, and divide all waiting users into emergency users according to different criticality levels. These emergency users The user's behavior data is synchronized to the outpatient department, and the pre-set voice terminal is used to continuously broadcast to the emergency user through the outpatient department to call on the corresponding emergency user to enter the corresponding outpatient department for consultation, until the camera terminal cannot detect the emergency user in the waiting area, that is, the continuous broadcast is stopped; the system can quickly identify the sudden symptoms of waiting users by real-time monitoring of the user's behavior data, trigger the early warning mechanism in time to avoid delays in treatment, and highlight the user with sudden symptoms in the camera terminal, so that medical staff can quickly locate and take action. At the same time, according to different behavior data, the system can accurately divide emergency users to ensure that users who need emergency medical intervention most can be given priority. By setting the critical level, the system can manage users in a graded manner according to the severity of the disease and optimize the allocation of medical resources. The behavior data of emergency users can be synchronized to the outpatient department in real time, so that medical staff can immediately obtain the latest health status of the user when he arrives, reducing the diagnosis time, because through the real-time synchronization of data, all relevant personnel can obtain the latest information, improving the overall collaboration efficiency of the medical team. ;

[0063] It should be noted that the outpatient department only collects video data through the camera terminal to observe whether the users waiting in the waiting area have sudden symptoms in advance. If sudden symptoms occur, they can be urgently sent to the corresponding outpatient department or emergency room, rather than diagnosing the specific condition of the waiting user in advance through the camera data.

[0064] It should be added that the camera terminal specifically uses traditional visible light and thermal imaging technology to scan and obtain the user's facial expressions, body movements, clothing features and other surface information from the waiting area. Through visible light images, the system can perform routine analysis such as facial recognition and behavior detection to identify the user's identity and monitor their movements and abnormal behaviors. Infrared thermal imaging technology detects changes in the user's body surface temperature and captures information such as skin fever, inflammation, and abnormal blood flow. This enables the system to identify some health problems that cannot be detected by visible light, such as fever, local inflammation or blood circulation problems.

[0065] In this embodiment, before step S5 of highlighting the waiting user in the preset area based on the behavior data, the method further includes:

[0066] S501: Based on the facial features preset by the camera terminal, remote facial data of the waiting user is collected;

[0067] S502: Determine whether the remote facial data detects pre-recorded surface symptoms, wherein the surface symptoms specifically include facial symptoms, skin symptoms, and blood flow symptoms;

[0068] S503: If so, then select the registration waiting seat corresponding to the waiting user in the preset area, activate the preset distress content of the registration waiting seat according to the surface symptoms, and adaptively upload the reserved registration information of the waiting user to the corresponding outpatient department according to the distress content, wherein the reserved registration information specifically includes the required consultation information, the targeted consultation department and the historical consultation information.

[0069] In this embodiment, the system collects remote facial data of different waiting users in the waiting area based on the facial features preset by the camera terminal, and then the system determines whether the remote facial data detects pre-recorded surface symptoms, which specifically include facial symptoms, skin symptoms and blood flow symptoms, so as to execute corresponding steps; for example, when the system determines that the remote facial data of the waiting user does not detect the pre-recorded surface symptoms, the system will believe that the waiting user is in the waiting process, and the camera terminal has not detected the waiting user's sudden symptoms from the waiting user's surface state, and the system will continue to monitor the user's facial data, and combine the facial data with other behavioral data (such as posture, movement behavior, etc.) for comprehensive analysis, so as to determine the waiting user's symptoms. Ensure comprehensive health monitoring, and provide users with regular health advice through voice or display screens, reminding them to pay attention to their own health status, and actively seek help if they feel unwell. The current remote facial data is stored in the user's health record for reference during subsequent visits or long-term health trend analysis. For users without obvious symptoms, the system can arrange subsequent monitoring and analysis to observe whether new symptoms or changes appear; for example, when the system determines that the remote facial data of the waiting user detects a pre-recorded surface symptom, the system will consider that the waiting user may have a sudden symptom during the waiting process. The camera terminal detects from the waiting user's surface state that the waiting user may have a sudden symptom, and the system will select the waiting user in the waiting area. The registration waiting seat corresponding to the waiting user activates the pre-set help content on the registration waiting seat according to different superficial symptoms. The medical staff can observe what kind of superficial symptoms the waiting user has suddenly developed from the registration waiting seat where the waiting user sits. According to different help content, the system will adaptively upload the reserved registration information of the waiting user to the corresponding outpatient department. The reserved registration information specifically includes the required consultation information, the targeted consultation department and the historical consultation information. The system automatically detects the superficial symptoms of the waiting user through the camera terminal and immediately activates the help content on the registration waiting seat to ensure that it can attract the attention of medical staff when the symptoms first appear, achieve rapid response, and shorten the delay time of treatment. The medical staff can observe the registration The call for help content on the waiting seat can quickly understand the user's specific type of illness and take corresponding emergency measures according to the actual situation. At the same time, the system will upload the user's reserved registration information to the corresponding outpatient department according to the detected type of illness, ensuring that the user receives the most suitable medical service and avoiding waste of resources. By uploading the user's detailed information, the system helps the outpatient department to prepare in advance, shorten the consultation time and improve the overall treatment effect. In addition, the preset call for help content on the registration waiting seat can be automatically activated when the patient is unable to seek help on his own, thereby improving the patient's safety in the event of sudden illness. Through automatic uploading and targeted consultation arrangements, patients feel more at ease, knowing that once a problem occurs, the system will deal with it immediately and it will not be ignored.

[0070] In this embodiment, after step S3 of generating the registration sequence information of the waiting user in the camera terminal, the following steps are also included:

[0071] S301: Acquire additional symptom information pre-entered by the waiting user into the outpatient department, and identify the registration waiting seat of the waiting user in the preset area based on the additional symptom information, wherein the additional symptom information specifically includes past symptoms, chronic diseases and symptoms under treatment;

[0072] S302: Determine whether the user sitting on the registered waiting seat is a waiting user;

[0073] S303: If so, then the image data of the waiting user is collected in real time according to the additional symptom information, the symptom characteristics of the waiting user are extracted from the image data, the duration of the symptom characteristics is obtained, and the corresponding trend symptom characteristics are generated according to the duration, and the trend symptom characteristics are input into the pre-trained prediction model for prediction, so as to predict the probability of the waiting user having the symptom corresponding to the additional symptom information, wherein the image data specifically includes facial data, body posture and movement characteristics.

[0074] In this embodiment, the system obtains the additional symptom information pre-entered into the outpatient department by the waiting user, the additional symptom information specifically includes past symptoms, chronic diseases and treatment-related symptoms, and identifies the registration waiting seats of different waiting users in the waiting area based on different additional symptom information, and then the system determines whether the users sitting on these registration waiting seats belong to the corresponding waiting users to perform corresponding steps; for example, when the system determines that the user sitting on the registration waiting seat does not belong to the corresponding waiting user, the system will think that the waiting user may have accidentally sat in the wrong seat, resulting in the mismatch between the user detected by the system and the pre-entered information. The system will remind the user in the current seat to check whether their seat is correct through voice prompts or display notifications on the display screen, and guide them to the correct seat, and re-identify the user's identity from multiple camera angles to ensure the accuracy of identification, eliminate misidentification caused by light or angle, and if the identity of the user in the seat cannot be confirmed after multiple attempts to identify, the medical staff or the guide can be notified to perform manual verification to help the user find the correct seat; for example, when the system determines that the user sitting on the registration waiting seat belongs to the corresponding waiting user, the system will think that the waiting user is already in the waiting area Corresponding to the patient waiting for consultation on the registration waiting seat, the system will collect the image data of the waiting user in real time according to the additional symptom information. The image data specifically includes facial data, body posture and movement features, extract the symptom features that may appear in the waiting user from these image data, obtain the duration of these symptom features, generate corresponding trend symptom features according to different durations, input the trend symptom features into a pre-trained prediction model for prediction, and obtain the probability of the waiting user having the symptom corresponding to these additional symptom information; the system collects relevant image data based on the additional symptom information, so that health monitoring is more personalized and targeted, and can better reflect the actual health status of the user, provide personalized health predictions according to the symptom features of different users, and meet the medical needs of different users. At the same time, by obtaining the duration of symptom features and generating trend symptom features, the system can track the development trend of the disease. This dynamic tracking can reveal the changing pattern of the disease and help medical staff better understand the evolution of the disease. The trend symptom features are input into a pre-trained prediction model for prediction, and specific disease probabilities can be obtained, which provides data support for medical decision-making and helps doctors determine whether further examination or emergency treatment is needed.

[0075] It should be noted that the trend disease feature is input into the pre-trained prediction model for prediction, and the probability of the disease corresponding to the additional disease information occurring in the waiting user is predicted. The specific example is as follows:

[0076] Assume that in the waiting area of ​​the outpatient department, the system collects image data of waiting users in real time through a preset high-definition camera terminal. These data include facial features, body postures and movement features. For example, the system detects that a user's facial image shows typical symptoms of cardiovascular disease, such as obvious facial swelling and abnormal skin color (such as pale or cyanotic), and the user's body movements show abnormal behavior, such as continuous stroking of the chest showing obvious discomfort. The system will convert these image data into possible trend disease features, including the start time, duration, and change pattern of the symptoms. These trend disease features will be input into a pre-trained deep learning prediction model, and the model will compare these features with its learned disease data to calculate the probability that the user may have cardiovascular disease. Assume that the model predicts that the probability of cardiovascular disease in the user is 75%. At this time, the system will highlight the user's location in the waiting area, and the medical staff of the outpatient department will receive an alarm, indicating that the user's health status needs to be treated first, and the medical staff can quickly take further measures based on the alarm, such as conducting an emergency consultation or arranging further examinations.

[0077] It should be added that the specific example of the training process of the prediction model is as follows:

[0078] Suppose the system collects an image dataset containing various diseases (such as cardiovascular disease, diabetes, etc.), and these images should be annotated with detailed disease information, including facial and limb features; the dataset should include a large number of normal and abnormal images, covering different disease manifestations and health states, to ensure that the model can learn the characteristics of various diseases; the collected image data is preprocessed, including image cropping, scaling, normalization and enhancement (such as rotation, flipping, color adjustment), etc., to improve data quality and model robustness; and the label data also needs to be preprocessed, such as converting disease classification labels into numerical labels to ensure that the model can understand and process these labels; after selecting a deep learning model (convolutional neural network) suitable for image data processing, by extracting spatial features in the image and designing the model architecture, including convolutional layers, pooling layers, activation functions, batch normalization layers and full The connection layer is used to adapt to the complexity of the data and the requirements of the task; the training set is then used for model training, and the loss function (cross entropy loss function) and optimization algorithm (Adam) are set to minimize the prediction error. The model parameters are updated through the back propagation algorithm for multiple rounds of iterations to gradually improve the accuracy of the model on the training data; then early stopping technology and cross-validation are used to prevent overfitting and ensure the generalization ability of the model. The model performance is evaluated using the test set, and indicators including accuracy, recall rate and F1 score are calculated to test the performance of the model on new data; finally, based on the verification results, the model hyperparameters (including learning rate and batch size) are adjusted to further optimize the model performance, and the trained and optimized model is deployed to the actual system to realize the prediction of real-time data. After deployment, the model is regularly updated to adapt to new data and changes in symptoms to maintain the accuracy and reliability of the prediction.

[0079] In this embodiment, the step S5 of marking the corresponding waiting user from the preset area according to the human body data includes:

[0080] S51: based on the virtual coordinates of the outpatient department divided in the preset area, collecting the coordinate information of the waiting user corresponding to the preset area, and generating a virtual boundary box of the waiting user according to the coordinate information;

[0081] S52: Determine whether the virtual boundary box can continuously capture the waiting user;

[0082] S53: If yes, synchronizing the coordinate information in real time in the preset multi-camera perspectives according to the virtual boundary frame, converting the coordinate information into the same physical coordinate system, and acquiring the multi-perspective behavior data of the waiting user from the multi-camera perspectives.

[0083] In this embodiment, the system collects coordinate information corresponding to different waiting users in the waiting area based on the virtual coordinates divided by the outpatient department in the waiting area, generates virtual bounding boxes of different waiting users in the camera terminal according to these coordinate information, and then the system determines whether these virtual bounding boxes can continuously capture the corresponding waiting users to execute the corresponding steps; for example, when the system determines that the virtual bounding boxes of different waiting users cannot continuously capture the corresponding waiting users, the system will consider that the waiting users may have left the monitoring range of the camera terminal or the waiting area, resulting in the virtual bounding box being unable to continuously track the user, and the system will check whether the camera terminal has a fault, such as whether the lens is blocked, whether the connection is normal, etc. If a problem is found, the medical staff will be notified to make necessary repairs or adjustments, and the bounding box will be recalibrated to ensure that it accurately covers the position of the waiting user, and at the same time, the data update frequency of the camera terminal will be increased to capture the user's movement faster, reduce the tracking loss caused by the user's rapid movement, and if the user cannot be captured by the virtual bounding box for a long time, the system will trigger an alarm to notify the medical staff to check whether there is any abnormality in the waiting area; for example, when the system determines that the virtual bounding boxes of different waiting users can continuously capture the corresponding waiting users, the system will consider that the waiting users may have left the monitoring range of the camera terminal or the waiting area, and ... Because the camera terminal is always collecting the behavior data of the waiting user, the system will synchronize the coordinate information of the waiting user in real time in the pre-set multi-camera perspectives according to the virtual boundary box, convert the coordinate information in each camera perspective into the same physical coordinate system, and obtain the multi-perspective behavior data of the waiting user from the multi-camera perspectives; the system can comprehensively observe the behavior of the waiting user from different angles by synchronizing the coordinate information of the multi-camera perspectives. This multi-perspective data fusion makes the monitoring of user behavior more accurate and comprehensive. For example, the frontal perspective may capture the changes in the user's facial expression, while the side perspective can observe the user's body movements or posture changes. By integrating the information from these perspectives, it is helpful to have a more comprehensive understanding of the user's health status and behavior patterns. At the same time, multi-perspective data can help the system more accurately identify and analyze abnormal behaviors. For example, when the user shows abnormal body movements or discomfort, different camera perspectives can capture different aspects of these abnormal behaviors. The system can integrate this information to improve the accuracy of abnormal behavior recognition, and continuously obtain multi-perspective behavior data to ensure that medical staff have real-time understanding of the status of waiting users. If the system detects an abnormal situation, the system will promptly notify the medical staff to intervene and optimize the user's waiting experience.

[0084] In this embodiment, the step S2 of determining whether preset human body data is detected in the camera data further includes:

[0085] S21: Based on the surface features pre-collected by the outpatient department from the waiting user, identifying image features corresponding to the surface features in the preset area;

[0086] S22: Determine whether the image feature can be detected in the preset area;

[0087] S23: If possible, based on the different image features, the waiting users corresponding to the image features are captured in real time in the preset area, and the registration waiting seats corresponding to the waiting users are generated through the outpatient department.

[0088] In this embodiment, the system identifies image features corresponding to the surface features in the waiting area based on the surface features of the waiting users collected in advance by the outpatient department, and then the system determines whether these image features can be detected in the waiting area to execute the corresponding steps; for example, when the system determines that the image features corresponding to a certain surface feature cannot be detected in the waiting area, the system will believe that the waiting user may be blocked by other objects, or that the user is moving too fast, resulting in the camera terminal being unable to capture and identify the image features in real time, and the system will check whether the camera terminal is working properly. Ensure that the lens is clear and free of obstructions, verify whether the device connection is stable, check whether the lighting conditions in the waiting area are suitable for video recording, adjust the lighting equipment to provide a good image acquisition environment, and activate the angle or position of the multi-view camera to cover possible blind spots or obstructed areas to ensure comprehensive monitoring of the waiting area. If the image features cannot be detected for a long time, the system will notify medical staff to conduct manual inspections on site to confirm the waiting status of the waiting users. For example, when the system determines that the image features corresponding to a certain surface feature can be detected in the waiting area, the system will consider that there are waiting users waiting for consultation in the waiting area. The system will capture the waiting users corresponding to the image features in real time in the waiting area based on different image features, and generate registration waiting seats corresponding to the waiting users through the outpatient department to avoid waiting users sitting on the waiting seats of other waiting users. The system can Image features accurately identify and locate waiting users, ensuring that the position of each user in the waiting area is accurately recorded, which helps to avoid scheduling problems caused by position errors or data loss. By capturing image features in real time, the position and status information of waiting users can be continuously updated, so that medical staff can grasp the waiting situation at any time and optimize waiting management. At the same time, according to the identified image features, corresponding registration waiting seats can be generated for each waiting user, reducing manual operations and errors, and improving the efficiency of the registration and queuing process. By capturing the user's position and status in real time, the system can adjust the waiting seat allocation according to real-time data to deal with emergencies, such as temporary changes in users or increased waiting needs, and monitor the user's waiting status in real time, and deal with possible problems in a timely manner, such as user discomfort due to long waiting, to improve the overall service quality and user satisfaction.

[0089] In this embodiment, the step S4 of determining whether the behavior data meets the preset warning content further includes:

[0090] S41: capturing abnormal behaviors of the waiting users through the camera terminal, and obtaining the stagnation time corresponding to the abnormal behaviors, wherein the abnormal behaviors specifically include individual behaviors and collective behaviors;

[0091] S42: Determine whether the stagnation time exceeds a preset time;

[0092] S43: If yes, identify the concurrent behaviors attached to the abnormal behavior, substitute the abnormal behavior and the concurrent behaviors into the symptom map pre-constructed by the outpatient department, and divide the abnormal combination rules of the waiting users from the symptom map.

[0093] In this embodiment, the system collects abnormal behaviors of different waiting users through the camera terminal, and the abnormal behaviors specifically include individual behaviors and collective behaviors, obtains the stagnation time corresponding to these abnormal behaviors, and then the system determines whether these stagnation times exceed the preset time to execute the corresponding steps; for example, when the system determines that the abnormal behavior of a waiting user does not exceed the preset time, the system will consider that the abnormal behavior of the waiting user may be short-term, belonging to an occasional and non-continuous problem, and the system will continue to monitor the user's behavior to ensure that any changes can be captured in time, and record the abnormal behavior and its stagnation time for subsequent analysis and comparison. At the same time, the behavior pattern of the waiting user is compared in the database to check whether there is repeated abnormal behavior or potential previous medical history, and the system will arrange for medical staff to actively ask the user whether he needs help to ensure that the user's comfort and needs are met; for example, when the system determines that the abnormal behavior of a waiting user exceeds the preset time, the system will consider that the waiting user If the abnormal behavior has lasted for a long time, the system will identify the concurrent behaviors attached to the abnormal behavior, substitute the abnormal behavior and concurrent behaviors into the symptom map pre-built by the outpatient department, and divide the abnormal combination rules of the waiting user from the symptom map; the system accurately identifies the specific symptoms or health problems related to the user's abnormal behavior by substituting the abnormal behavior and its concurrent behaviors into the symptom map, and identifying the combination of abnormal behavior and its concurrent behaviors helps to fully understand the user's health status. At the same time, concurrent behaviors reveal more disease information, such as accompanying symptoms or disease progression, thereby improving the accuracy of later medical staff's diagnosis of the disease. The system can give priority to users who may have serious health problems based on the duration of abnormal behavior and abnormal combination rules, helping medical staff to concentrate on patients who really need urgent attention. According to the identified abnormal combination rules, the system can optimize the allocation of medical resources. For example, if a certain abnormal behavior combination is strongly correlated with a specific disease, the medical institution can allocate relevant medical resources and experts to handle these cases.

[0094] It should be noted that the concurrent behavior attached to the abnormal behavior is identified, the abnormal behavior and the concurrent behavior are substituted into the symptom map pre-built by the outpatient department, and the abnormal combination rules of the waiting users are divided from the symptom map. The specific examples are as follows:

[0095] Assume that in the waiting area of ​​an outpatient department, a waiting user exhibits the following abnormal and concurrent behaviors during the waiting period:

[0096] Abnormal behavior:

[0097] Long periods of sitting still: Users sit in a chair in the waiting area, barely moving, and maintaining a constant posture;

[0098] Facial expression of pain: The user's facial expression is distorted, showing obvious pain;

[0099] Concurrent Behavior:

[0100] Shortness of breath: The user's breathing becomes noticeably faster and heavier;

[0101] Facial Pallor: The user's face is paler than normal;

[0102] Severe abdominal pain: Users occasionally hold their abdomen, expressing a sense of severe pain;

[0103] The system captures the user's abnormal behavior, such as sitting still for a long time and facial expressions of pain, through the camera terminals installed in the outpatient department. This information is transmitted to the system in real time for analysis. Since the system also detects the user's concurrent behaviors, such as shortness of breath, pale face and severe abdominal pain, these behaviors are also recorded and transmitted to the system for further analysis. By substituting them into the symptom map, the system substitutes these abnormal behaviors and concurrent behaviors into the symptom map pre-built by the outpatient department. The symptom map contains the behavioral characteristics of various symptoms and their combination rules. For example, the map records that "sitting still for a long time" and "severe abdominal pain" are usually related to acute stomach pain, while "pale face and shortness of breath" are related to acute gastritis or gastric ulcer. Therefore

[0104] When dividing abnormal combination rules, according to the combination rules in the symptom map, the system can identify that the user's behavior data is highly matched with the characteristics of acute stomach pain, such as:

[0105] “Sitting still for a long time” + “severe abdominal pain” → may be acute stomach pain;

[0106] “Pale face” + “shortness of breath” → may be related to acute gastritis or gastric ulcer;

[0107] Finally, the system will summarize these matching results and generate warning information, indicating that the waiting user may have acute stomach pain or related symptoms; the system will automatically mark the user as "high risk" in the waiting area and arrange medical examinations on a priority basis; based on the generated warning, medical personnel will give priority to the user's condition and provide necessary medical intervention or emergency treatment, and the system can notify the corresponding outpatient department to prepare for reception and prepare relevant medical resources so that the user can receive diagnosis and treatment as soon as possible.

[0108] In this embodiment, the step S1 of acquiring the camera data in the preset area based on the camera terminal preset in the outpatient department also includes:

[0109] S11: Acquire the identity information of the waiting user pre-entered in the outpatient department, and identify the corresponding waiting user in the preset area based on the identity information;

[0110] S12: Determine whether the waiting user exists in the preset area;

[0111] S13: If yes, detect the image acquisition mode of the camera terminal, convert the image acquisition mode into a preset continuous frame format, and synchronize the image data of the waiting user to the outpatient department in the continuous frame format, wherein the image acquisition mode specifically includes single frame acquisition and continuous frame acquisition.

[0112] In this embodiment, the system obtains the identity information pre-entered by the waiting user in the outpatient department, identifies the corresponding waiting user in the waiting area based on different identity information, and then determines whether these waiting users exist in the waiting area to execute the corresponding steps; for example, when the system determines that a waiting user whose identity information has been entered does not exist in the waiting area, the system will consider that the user may not have arrived in the waiting area, has left the waiting area, or has not been detected at the system preset location due to other reasons, and the system will search for the user's arrival record in the registration system, and ensure that the identity information and location records in the system are up to date by continuously refreshing the system data and rescanning the waiting area. At the same time, if the user does not appear in the waiting area at the scheduled time, the system will contact the user by SMS, phone or push notification to confirm whether he still intends to receive medical services, update the user's status in the system to "not arrived" or "absent", process his registration or appointment status according to the hospital rules, and optimize the layout of the waiting area or the real-time monitoring function of the system according to the actual situation to ensure that the user can arrive smoothly and register; for example, when the system determines that a waiting user whose identity information has been entered The system will detect the image acquisition mode of the camera terminal, which specifically includes single-frame acquisition and continuous-frame acquisition. The system will convert the original single-frame acquisition mode into the pre-set continuous-frame acquisition mode, and synchronize the image data of the waiting user to the outpatient department in the form of continuous frames. The system can capture the movements and behavior changes of the waiting user in real time through the continuous-frame acquisition mode, so that the outpatient department can understand the dynamic status of the user in a timely manner, which is particularly important for timely discovering the abnormal behavior or health status changes of the user. By continuously updating the user's status and providing the latest image data, it is helpful to quickly respond to any potential health problems or emergencies. At the same time, through continuous frame acquisition, the outpatient department can observe the user's facial expressions, body movements and other behavioral characteristics in detail, which is helpful to discover hidden diseases or health problems, including facial changes or abnormal postures, and continuous frame acquisition provides more data samples, so that the system can more accurately detect early signs of abnormal behavior or diseases. Continuous data collection helps to analyze the trend of user behavior, identify potential health risks, and enable medical staff to formulate corresponding intervention measures.

[0113] Reference Figure 2 , is a human behavior analysis and early warning system based on image technology in one embodiment of the present invention, comprising:

[0114] The acquisition module 10 is used to acquire the camera data in the preset area based on the camera terminal preset in the outpatient department;

[0115] A determination module 20, used for determining whether preset human body data is detected in the camera data;

[0116] The execution module 30 is used for marking the corresponding waiting user from the preset area according to the human body data, generating the registration sequence information of the waiting user in the camera terminal according to the identity information of the waiting user pre-entered by the outpatient department, and collecting the behavior data of the waiting user in the preset time period in real time;

[0117] The second judgment module 40 is used to judge whether the behavior data meets the preset warning content;

[0118] The second execution module 50 is used to highlight the waiting user in the preset area based on the behavior data if it meets the requirements, generate a preset criticality level corresponding to the waiting user, classify emergency users among the waiting users according to the criticality level, synchronize the behavior data of the emergency user to the outpatient department, and use a preset voice terminal to continuously broadcast the emergency user through the outpatient department until the emergency user can no longer be detected in the preset area.

[0119] In this embodiment, the acquisition module 10 acquires the camera data in each preset waiting area based on the camera terminal pre-set in the waiting area by the outpatient department, and then the judgment module 20 judges whether the camera data detects the pre-set human body surface data, such as body contour or wearing accessories, to execute the corresponding steps; for example, when the system determines that the camera data in each waiting area cannot detect the pre-set human body surface data, the system will consider that there are no waiting users in these waiting areas, and the system will check whether the camera terminal is working properly, including network connection, camera clarity and light conditions, to ensure that there is no fault in the device, and re-call the detection function of the camera terminal to scan each waiting area again. The system will then check the waiting area to confirm that no user has been detected. If no user has been detected for multiple consecutive times, an alarm will be triggered to notify relevant staff to check the actual situation to prevent technical failures or other abnormalities, and notify medical staff or security guards to go to the relevant waiting area for on-site inspection to confirm whether there are really no waiting users. The system will also adjust the monitoring parameters of the camera terminal, such as changing the camera angle and adjusting the sensitivity, to ensure that users in the area are not missed. By notifying the front desk or relevant staff of the outpatient department that there is no one waiting in the current area, they will be ensured to understand the situation and make corresponding arrangements to avoid waiting users waiting in the wrong area. For example, when the system determines that the camera data in each waiting area can be detected, The execution module 30 will consider that there are waiting users in the waiting area according to the pre-set human body data. The system will mark and identify the corresponding waiting users in the waiting area according to different human body surface data, that is, different body contours or wearing accessories. According to the identity information of these waiting users pre-entered by the outpatient department, the registration sequence information of these waiting users is generated in the camera terminal, and the behavior data of these waiting users in the pre-set time period is collected in real time. The system accurately marks and identifies each waiting user according to the human body surface data to ensure user identity matching and avoid confusion or misidentification of different types of waiting users. At the same time, the registration sequence information is automatically generated according to the user identity information, making the registration process more efficient. and seamless connection, reducing waiting time. Since the registration sequence information can be updated in real time, it is ensured that the information of each waiting user accurately reflects the current status, and the behavior data of the waiting users is collected in real time, so that the status changes of the users during the waiting period can be captured in time. For example, in case of emergency situations such as worsening of the condition, through real-time monitoring, the system can dynamically adjust or notify medical staff to carry out necessary intervention on the users, give priority to users who need urgent medical attention, and the waiting users know that their status is being monitored in real time, and can get timely attention and treatment in case of emergency, thereby enhancing their sense of security. Then the second judgment module 40 judges whether the behavior data of the waiting users meets the pre-set warning content, so as to execute the corresponding steps.For example, when the system determines that the behavior data of a waiting user does not meet the pre-set warning content, the system will consider that the waiting user has not detected any symptoms requiring emergency assistance during the waiting stage, and the system will continue to monitor the behavior data of the waiting user in real time to ensure that any potential health changes during the waiting period can be captured in time, and regularly analyze the user's behavior data. Even if there are minor changes, they can attract the attention of the system in a short time. At the same time, the system will provide regular health tips to the user through the voice terminal or display screen, reminding them to inform the staff in time if they feel uncomfortable, and synchronize the behavior data and analysis results of the waiting user to the outpatient department for reference by medical staff when the user is seeing a doctor, helping them to make a more accurate diagnosis; for example, when the system determines that the behavior data of a waiting user meets the pre-set warning content, the second execution module 50 will consider that the waiting user has a sudden symptom while waiting, and the system will highlight the waiting user in the waiting area of ​​the camera terminal based on different behavior data, generate a pre-set criticality level corresponding to the waiting user, and divide all waiting users into emergency users according to different criticality levels, and divide them into emergency users. The behavior data of emergency users is synchronized to the outpatient department. The pre-set voice terminal is used to continuously broadcast emergency users through the outpatient department to call on the corresponding emergency users to enter the corresponding outpatient department for consultation until the camera terminal cannot detect the emergency user in the waiting area, that is, the continuous broadcast is stopped; the system can quickly identify the sudden symptoms of waiting users by real-time monitoring of user behavior data, trigger the early warning mechanism in time to avoid delays in treatment, and highlight the users with sudden symptoms in the camera terminal, so that medical staff can quickly locate and take action. At the same time, according to different behavior data, the system can accurately divide emergency users to ensure that users who need emergency medical intervention most can be given priority. By setting the critical level, the system can manage users in a graded manner according to the severity of the symptoms and optimize the allocation of medical resources. The behavior data of emergency users can be synchronized to the outpatient department in real time, so that medical staff can immediately obtain the latest health status of the user when he arrives, reducing the diagnosis time, because through the real-time synchronization of data, all relevant personnel can obtain the latest information, improving the overall collaboration efficiency of the medical team. ;

[0120] In this embodiment, it also includes:

[0121] A collection module, used for collecting remote facial data of the waiting user based on the facial features preset by the camera terminal;

[0122] A third determination module is used to determine whether the remote facial data detects pre-recorded surface diseases, wherein the surface diseases specifically include facial diseases, skin diseases and blood flow diseases;

[0123] The third execution module is used for selecting the registration waiting seat corresponding to the waiting user in the preset area, activating the preset distress content of the registration waiting seat according to the surface symptoms, and adaptively uploading the reserved registration information of the waiting user to the corresponding outpatient department according to the distress content, wherein the reserved registration information specifically includes the required consultation information, the targeted consultation department and the historical consultation information.

[0124] In this embodiment, the system collects remote facial data of different waiting users in the waiting area based on the facial features preset by the camera terminal, and then the system determines whether the remote facial data detects pre-recorded surface symptoms, which specifically include facial symptoms, skin symptoms and blood flow symptoms, so as to execute corresponding steps; for example, when the system determines that the remote facial data of the waiting user does not detect the pre-recorded surface symptoms, the system will believe that the waiting user is in the waiting process, and the camera terminal has not detected the waiting user's sudden symptoms from the waiting user's surface state, and the system will continue to monitor the user's facial data, and combine the facial data with other behavioral data (such as posture, movement behavior, etc.) for comprehensive analysis, so as to determine the waiting user's symptoms. Ensure comprehensive health monitoring, and provide users with regular health advice through voice or display screens, reminding them to pay attention to their own health status, and actively seek help if they feel unwell. The current remote facial data is stored in the user's health record for reference during subsequent visits or long-term health trend analysis. For users without obvious symptoms, the system can arrange subsequent monitoring and analysis to observe whether new symptoms or changes appear; for example, when the system determines that the remote facial data of the waiting user detects a pre-recorded surface symptom, the system will consider that the waiting user may have a sudden symptom during the waiting process. The camera terminal detects from the waiting user's surface state that the waiting user may have a sudden symptom, and the system will select the waiting user in the waiting area. The registration waiting seat corresponding to the waiting user activates the pre-set help content on the registration waiting seat according to different superficial symptoms. The medical staff can observe what kind of superficial symptoms the waiting user has suddenly developed from the registration waiting seat where the waiting user sits. According to different help content, the system will adaptively upload the reserved registration information of the waiting user to the corresponding outpatient department. The reserved registration information specifically includes the required consultation information, the targeted consultation department and the historical consultation information. The system automatically detects the superficial symptoms of the waiting user through the camera terminal and immediately activates the help content on the registration waiting seat to ensure that it can attract the attention of medical staff when the symptoms first appear, achieve rapid response, and shorten the delay time of treatment. The medical staff can observe the registration The call for help content on the waiting seat can quickly understand the user's specific type of illness and take corresponding emergency measures according to the actual situation. At the same time, the system will upload the user's reserved registration information to the corresponding outpatient department according to the detected type of illness, ensuring that the user receives the most suitable medical service and avoiding waste of resources. By uploading the user's detailed information, the system helps the outpatient department to prepare in advance, shorten the consultation time and improve the overall treatment effect. In addition, the preset call for help content on the registration waiting seat can be automatically activated when the patient is unable to seek help on his own, thereby improving the patient's safety in the event of sudden illness. Through automatic uploading and targeted consultation arrangements, patients feel more at ease, knowing that once a problem occurs, the system will deal with it immediately and it will not be ignored.

[0125] In this embodiment, it also includes:

[0126] A second acquisition module is used to acquire additional symptom information pre-entered by the waiting user into the outpatient department, and identify the registration waiting seat of the waiting user in the preset area based on the additional symptom information, wherein the additional symptom information specifically includes past symptoms, chronic diseases and symptoms under treatment;

[0127] The fourth judgment module is used to judge whether the user sitting on the registered waiting seat belongs to the waiting user;

[0128] The fourth execution module is used to, if yes, collect the image data of the waiting user in real time according to the additional symptom information, extract the symptom characteristics of the waiting user from the image data, obtain the duration of the symptom characteristics, generate corresponding trend symptom characteristics according to the duration, input the trend symptom characteristics into the pre-trained prediction model for prediction, and predict the probability of the waiting user having the symptom corresponding to the additional symptom information, wherein the image data specifically includes facial data, body posture and movement characteristics.

[0129] In this embodiment, the system obtains the additional symptom information pre-entered into the outpatient department by the waiting user, the additional symptom information specifically includes past symptoms, chronic diseases and treatment-related symptoms, and identifies the registration waiting seats of different waiting users in the waiting area based on different additional symptom information, and then the system determines whether the users sitting on these registration waiting seats belong to the corresponding waiting users to perform corresponding steps; for example, when the system determines that the user sitting on the registration waiting seat does not belong to the corresponding waiting user, the system will think that the waiting user may have accidentally sat in the wrong seat, resulting in the mismatch between the user detected by the system and the pre-entered information. The system will remind the user in the current seat to check whether their seat is correct through voice prompts or display notifications on the display screen, and guide them to the correct seat, and re-identify the user's identity from multiple camera angles to ensure the accuracy of identification, eliminate misidentification caused by light or angle, and if the identity of the user in the seat cannot be confirmed after multiple attempts to identify, the medical staff or the guide can be notified to perform manual verification to help the user find the correct seat; for example, when the system determines that the user sitting on the registration waiting seat belongs to the corresponding waiting user, the system will think that the waiting user is already in the waiting area Corresponding to the patient waiting for consultation on the registration waiting seat, the system will collect the image data of the waiting user in real time according to the additional symptom information. The image data specifically includes facial data, body posture and movement features, extract the symptom features that may appear in the waiting user from these image data, obtain the duration of these symptom features, generate corresponding trend symptom features according to different durations, input the trend symptom features into a pre-trained prediction model for prediction, and obtain the probability of the waiting user having the symptom corresponding to these additional symptom information; the system collects relevant image data based on the additional symptom information, so that health monitoring is more personalized and targeted, and can better reflect the actual health status of the user, provide personalized health predictions according to the symptom features of different users, and meet the medical needs of different users. At the same time, by obtaining the duration of symptom features and generating trend symptom features, the system can track the development trend of the disease. This dynamic tracking can reveal the changing pattern of the disease and help medical staff better understand the evolution of the disease. The trend symptom features are input into a pre-trained prediction model for prediction, and specific disease probabilities can be obtained, which provides data support for medical decision-making and helps doctors determine whether further examination or emergency treatment is needed.

[0130] In this embodiment, the second execution module further includes:

[0131] A generating unit, configured to collect coordinate information corresponding to the waiting user in the preset area based on the virtual coordinates of the outpatient department divided in the preset area, and generate a virtual boundary box of the waiting user according to the coordinate information;

[0132] a determination unit, configured to determine whether the virtual boundary box can continuously capture the waiting user;

[0133] The execution unit is used to synchronize the coordinate information in real time in the preset multi-camera perspectives according to the virtual boundary box, if possible, convert the coordinate information into the same physical coordinate system, and obtain the multi-perspective behavior data of the waiting user from the multi-camera perspectives.

[0134] In this embodiment, the system collects coordinate information corresponding to different waiting users in the waiting area based on the virtual coordinates divided by the outpatient department in the waiting area, generates virtual bounding boxes of different waiting users in the camera terminal according to these coordinate information, and then the system determines whether these virtual bounding boxes can continuously capture the corresponding waiting users to execute the corresponding steps; for example, when the system determines that the virtual bounding boxes of different waiting users cannot continuously capture the corresponding waiting users, the system will consider that the waiting users may have left the monitoring range of the camera terminal or the waiting area, resulting in the virtual bounding box being unable to continuously track the user, and the system will check whether the camera terminal has a fault, such as whether the lens is blocked, whether the connection is normal, etc. If a problem is found, the medical staff will be notified to make necessary repairs or adjustments, and the bounding box will be recalibrated to ensure that it accurately covers the position of the waiting user, and at the same time, the data update frequency of the camera terminal will be increased to capture the user's movement faster, reduce the tracking loss caused by the user's rapid movement, and if the user cannot be captured by the virtual bounding box for a long time, the system will trigger an alarm to notify the medical staff to check whether there is any abnormality in the waiting area; for example, when the system determines that the virtual bounding boxes of different waiting users can continuously capture the corresponding waiting users, the system will consider that the waiting users may have left the monitoring range of the camera terminal or the waiting area, and ... Because the camera terminal is always collecting the behavior data of the waiting user, the system will synchronize the coordinate information of the waiting user in real time in the pre-set multi-camera perspectives according to the virtual boundary box, convert the coordinate information in each camera perspective into the same physical coordinate system, and obtain the multi-perspective behavior data of the waiting user from the multi-camera perspectives; the system can comprehensively observe the behavior of the waiting user from different angles by synchronizing the coordinate information of the multi-camera perspectives. This multi-perspective data fusion makes the monitoring of user behavior more accurate and comprehensive. For example, the frontal perspective may capture the changes in the user's facial expression, while the side perspective can observe the user's body movements or posture changes. By integrating the information from these perspectives, it is helpful to have a more comprehensive understanding of the user's health status and behavior patterns. At the same time, multi-perspective data can help the system more accurately identify and analyze abnormal behaviors. For example, when the user shows abnormal body movements or discomfort, different camera perspectives can capture different aspects of these abnormal behaviors. The system can integrate this information to improve the accuracy of abnormal behavior recognition, and continuously obtain multi-perspective behavior data to ensure that medical staff have real-time understanding of the status of waiting users. If the system detects an abnormal situation, the system will promptly notify the medical staff to intervene and optimize the user's waiting experience.

[0135] In this embodiment, the judgment module further includes:

[0136] A recognition unit, configured to recognize image features corresponding to the surface features pre-collected by the outpatient department from the waiting user in the preset area;

[0137] A second judging unit, used to judge whether the image feature can be detected in the preset area;

[0138] The second execution unit is used for, if possible, capturing waiting users corresponding to the image features in real time in the preset area according to different image features, and generating registration waiting seats corresponding to the waiting users through the outpatient department.

[0139] In this embodiment, the system identifies image features corresponding to the surface features in the waiting area based on the surface features of the waiting users collected in advance by the outpatient department, and then the system determines whether these image features can be detected in the waiting area to execute the corresponding steps; for example, when the system determines that the image features corresponding to a certain surface feature cannot be detected in the waiting area, the system will believe that the waiting user may be blocked by other objects, or that the user is moving too fast, resulting in the camera terminal being unable to capture and identify the image features in real time, and the system will check whether the camera terminal is working properly. Ensure that the lens is clear and free of obstructions, verify whether the device connection is stable, check whether the lighting conditions in the waiting area are suitable for video recording, adjust the lighting equipment to provide a good image acquisition environment, and activate the angle or position of the multi-view camera to cover possible blind spots or obstructed areas to ensure comprehensive monitoring of the waiting area. If the image features cannot be detected for a long time, the system will notify medical staff to conduct manual inspections on site to confirm the waiting status of the waiting users. For example, when the system determines that the image features corresponding to a certain surface feature can be detected in the waiting area, the system will consider that there are waiting users waiting for consultation in the waiting area. The system will capture the waiting users corresponding to the image features in real time in the waiting area based on different image features, and generate registration waiting seats corresponding to the waiting users through the outpatient department to avoid waiting users sitting on the waiting seats of other waiting users. The system can Image features accurately identify and locate waiting users, ensuring that the position of each user in the waiting area is accurately recorded, which helps to avoid scheduling problems caused by position errors or data loss. By capturing image features in real time, the position and status information of waiting users can be continuously updated, so that medical staff can grasp the waiting situation at any time and optimize waiting management. At the same time, according to the identified image features, corresponding registration waiting seats can be generated for each waiting user, reducing manual operations and errors, and improving the efficiency of the registration and queuing process. By capturing the user's position and status in real time, the system can adjust the waiting seat allocation according to real-time data to deal with emergencies, such as temporary changes in users or increased waiting needs, and monitor the user's waiting status in real time, and deal with possible problems in a timely manner, such as user discomfort due to long waiting, to improve the overall service quality and user satisfaction.

[0140] In this embodiment, the second determination module further includes:

[0141] an acquisition unit, configured to collect abnormal behaviors of the waiting users through the camera terminal, and acquire the stagnation duration corresponding to the abnormal behaviors, wherein the abnormal behaviors specifically include individual behaviors and collective behaviors;

[0142] A third judging unit, configured to judge whether the stagnation time exceeds a preset time;

[0143] The third execution unit is used for identifying the concurrent behavior attached to the abnormal behavior, substituting the abnormal behavior and the concurrent behavior into the symptom map pre-constructed by the outpatient department, and dividing the abnormal combination rules of the waiting users from the symptom map.

[0144] In this embodiment, the system collects abnormal behaviors of different waiting users through the camera terminal, and the abnormal behaviors specifically include individual behaviors and collective behaviors, obtains the stagnation time corresponding to these abnormal behaviors, and then the system determines whether these stagnation times exceed the preset time to execute the corresponding steps; for example, when the system determines that the abnormal behavior of a waiting user does not exceed the preset time, the system will consider that the abnormal behavior of the waiting user may be short-term, belonging to an occasional and non-continuous problem, and the system will continue to monitor the user's behavior to ensure that any changes can be captured in time, and record the abnormal behavior and its stagnation time for subsequent analysis and comparison. At the same time, the behavior pattern of the waiting user is compared in the database to check whether there is repeated abnormal behavior or potential previous medical history, and the system will arrange for medical staff to actively ask the user whether he needs help to ensure that the user's comfort and needs are met; for example, when the system determines that the abnormal behavior of a waiting user exceeds the preset time, the system will consider that the waiting user If the abnormal behavior has lasted for a long time, the system will identify the concurrent behaviors attached to the abnormal behavior, substitute the abnormal behavior and concurrent behaviors into the symptom map pre-built by the outpatient department, and divide the abnormal combination rules of the waiting user from the symptom map; the system accurately identifies the specific symptoms or health problems related to the user's abnormal behavior by substituting the abnormal behavior and its concurrent behaviors into the symptom map, and identifying the combination of abnormal behavior and its concurrent behaviors helps to fully understand the user's health status. At the same time, concurrent behaviors reveal more disease information, such as accompanying symptoms or disease progression, thereby improving the accuracy of later medical staff's diagnosis of the disease. The system can give priority to users who may have serious health problems based on the duration of abnormal behavior and abnormal combination rules, helping medical staff to concentrate on patients who really need urgent attention. According to the identified abnormal combination rules, the system can optimize the allocation of medical resources. For example, if a certain abnormal behavior combination is strongly correlated with a specific disease, the medical institution can allocate relevant medical resources and experts to handle these cases.

[0145] In this embodiment, the acquisition module further includes:

[0146] A second acquisition unit is used to acquire the identity information of the waiting user pre-entered in the outpatient department, and identify the corresponding waiting user in the preset area based on the identity information;

[0147] A fourth determination unit, used for determining whether the waiting user exists in the preset area;

[0148] The fourth execution unit is used to detect the image acquisition mode of the camera terminal, convert the image acquisition mode into a preset continuous frame format, and synchronize the image data of the waiting user to the outpatient department in the continuous frame format, wherein the image acquisition mode specifically includes single frame acquisition and continuous frame acquisition.

[0149] In this embodiment, the system obtains the identity information pre-entered by the waiting user in the outpatient department, identifies the corresponding waiting user in the waiting area based on different identity information, and then determines whether these waiting users exist in the waiting area to execute the corresponding steps; for example, when the system determines that a waiting user whose identity information has been entered does not exist in the waiting area, the system will consider that the user may not have arrived in the waiting area, has left the waiting area, or has not been detected at the system preset location due to other reasons, and the system will search for the user's arrival record in the registration system, and ensure that the identity information and location records in the system are up to date by continuously refreshing the system data and rescanning the waiting area. At the same time, if the user does not appear in the waiting area at the scheduled time, the system will contact the user by SMS, phone or push notification to confirm whether he still intends to receive medical services, update the user's status in the system to "not arrived" or "absent", process his registration or appointment status according to the hospital rules, and optimize the layout of the waiting area or the real-time monitoring function of the system according to the actual situation to ensure that the user can arrive smoothly and register; for example, when the system determines that a waiting user whose identity information has been entered The system will detect the image acquisition mode of the camera terminal, which specifically includes single-frame acquisition and continuous-frame acquisition. The system will convert the original single-frame acquisition mode into the pre-set continuous-frame acquisition mode, and synchronize the image data of the waiting user to the outpatient department in the form of continuous frames. The system can capture the movements and behavior changes of the waiting user in real time through the continuous-frame acquisition mode, so that the outpatient department can understand the dynamic status of the user in a timely manner, which is particularly important for timely discovering the abnormal behavior or health status changes of the user. By continuously updating the user's status and providing the latest image data, it is helpful to quickly respond to any potential health problems or emergencies. At the same time, through continuous frame acquisition, the outpatient department can observe the user's facial expressions, body movements and other behavioral characteristics in detail, which is helpful to discover hidden diseases or health problems, including facial changes or abnormal postures, and continuous frame acquisition provides more data samples, so that the system can more accurately detect early signs of abnormal behavior or diseases. Continuous data collection helps to analyze the trend of user behavior, identify potential health risks, and enable medical staff to formulate corresponding intervention measures.

[0150] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A human behavior analysis and early warning method based on image technology, characterized in that: The following steps are involved: Based on the camera terminal preset in the outpatient department, obtain the camera data in the preset area; Determining whether preset human body data is detected in the camera data; If yes, then mark the corresponding waiting user from the preset area according to the human body data, generate the registration sequence information of the waiting user in the camera terminal according to the identity information of the waiting user pre-entered by the outpatient department, and collect the behavior data of the waiting user in the preset time period in real time; Determine whether the behavior data meets the preset warning content; If it is met, the waiting user is highlighted in the preset area based on the behavior data, a preset critical level corresponding to the waiting user is generated, emergency users are classified from the waiting users according to the critical level, the behavior data of the emergency user is synchronized to the outpatient department, and the emergency user is continuously broadcasted through the outpatient department using a preset voice terminal until the emergency user can no longer be detected in the preset area; Wherein, before the step of highlighting the waiting user in the preset area based on the behavior data, the method further includes: Based on the facial features preset by the camera terminal, collecting remote facial data of the waiting user; Determining whether the remote facial data detects pre-recorded surface symptoms, wherein the surface symptoms specifically include facial symptoms, skin symptoms, and blood flow symptoms; If so, the registration waiting seat corresponding to the waiting user is selected in the preset area, the preset call for help content of the registration waiting seat is activated according to the surface symptoms, and the reserved registration information of the waiting user is adaptively uploaded to the corresponding outpatient room according to the call for help content, wherein the reserved registration information specifically includes the required consultation information, the targeted consultation department and the historical consultation information; Wherein, after the step of generating the registration sequence information of the waiting user in the camera terminal, the method further includes: Acquire additional symptom information pre-entered by the waiting user into the outpatient department, and identify the registration waiting seat of the waiting user in the preset area based on the additional symptom information, wherein the additional symptom information specifically includes past symptoms, chronic diseases and symptoms under treatment; Determine whether the user sitting on the registered waiting seat is the waiting user; If so, the image data of the waiting user is collected in real time according to the additional symptom information, the symptom characteristics of the waiting user are extracted from the image data, the duration of the symptom characteristics is obtained, the corresponding trend symptom characteristics are generated according to the duration, the trend symptom characteristics are input into the pre-trained prediction model for prediction, and the probability of the waiting user having the symptom corresponding to the additional symptom information is predicted, wherein the image data specifically includes facial data, body posture and action characteristics; The step of marking the corresponding waiting user from the preset area according to the human body data includes: Based on the virtual coordinates of the outpatient department divided in the preset area, the coordinate information of the waiting user corresponding to the preset area is collected, and a virtual boundary box of the waiting user is generated according to the coordinate information; determining whether the virtual bounding box can continuously capture the waiting user; If yes, synchronizing the coordinate information in real time in the preset multi-camera perspectives according to the virtual boundary box, converting the coordinate information into a same physical coordinate system, and acquiring the multi-perspective behavior data of the waiting user from the multi-camera perspectives; Wherein, the step of determining whether preset human body data is detected in the camera data further includes: Based on the surface features pre-collected by the outpatient department from the waiting user, identifying image features corresponding to the surface features from the preset area; Determining whether the image feature can be detected in the preset area; If possible, based on the different image features, the waiting users corresponding to the image features are captured in real time in the preset area, and the registration waiting seats corresponding to the waiting users are generated through the outpatient department.

2. The human behavior analysis and early warning method based on image technology according to claim 1 is characterized in that: The step of determining whether the behavior data meets the preset warning content also includes: The abnormal behavior of the waiting user is captured by the camera terminal to obtain the stagnation time corresponding to the abnormal behavior, wherein the abnormal behavior specifically includes individual behavior and collective behavior; Determining whether the stagnation time exceeds a preset time; If so, identify the concurrent behaviors attached to the abnormal behavior, substitute the abnormal behavior and the concurrent behaviors into the symptom map pre-constructed by the outpatient department, and divide the abnormal combination rules of the waiting users from the symptom map.

3. The human behavior analysis and early warning method based on image technology according to claim 1 is characterized in that: The step of acquiring the imaging data in the preset area based on the imaging terminal preset in the outpatient department also includes: Acquire the identity information of the waiting user pre-entered in the outpatient department, and identify the corresponding waiting user in the preset area based on the identity information; Determining whether the waiting user exists in the preset area; If so, the image acquisition mode of the camera terminal is detected, and the image acquisition mode is converted into a preset continuous frame format, and the image data of the waiting user is synchronized to the outpatient department in the continuous frame format, wherein the image acquisition mode specifically includes single frame acquisition and continuous frame acquisition.

4. The human behavior analysis and early warning system based on image technology is characterized by: include: An acquisition module is used to acquire the camera data in a preset area based on the camera terminal preset in the outpatient department; A judging module, used for judging whether preset human body data is detected in the camera data; an execution module, for marking the corresponding waiting user from the preset area according to the human body data, generating the registration sequence information of the waiting user in the camera terminal according to the identity information of the waiting user pre-entered by the outpatient department, and collecting the behavior data of the waiting user in a preset time period in real time; The second judgment module is used to judge whether the behavior data conforms to the preset warning content; The second execution module is used for highlighting the waiting user in the preset area based on the behavior data if it meets the requirements, generating a preset critical level corresponding to the waiting user, classifying the emergency users among the waiting users according to the critical level, synchronizing the behavior data of the emergency users to the outpatient department, and continuously broadcasting the emergency users through the outpatient department using a preset voice terminal until the emergency users cannot be detected in the preset area; Among them, it also includes: A collection module, used for collecting remote facial data of the waiting user based on the facial features preset by the camera terminal; A third determination module is used to determine whether the remote facial data detects pre-recorded surface diseases, wherein the surface diseases specifically include facial diseases, skin diseases and blood flow diseases; The third execution module is used for selecting the registration waiting seat corresponding to the waiting user in the preset area, activating the preset call for help content of the registration waiting seat according to the surface symptoms, and adaptively uploading the reserved registration information of the waiting user to the corresponding outpatient room according to the call for help content, wherein the reserved registration information specifically includes the required consultation information, the targeted consultation department and the historical consultation information; Among them, it also includes: A second acquisition module is used to acquire additional symptom information pre-entered by the waiting user into the outpatient department, and identify the registration waiting seat of the waiting user in the preset area based on the additional symptom information, wherein the additional symptom information specifically includes past symptoms, chronic diseases and symptoms under treatment; A fourth judgment module is used to judge whether the user sitting on the registered waiting seat belongs to the waiting user; a fourth execution module, for, if yes, collecting image data of the waiting user in real time according to the additional symptom information, extracting symptom features of the waiting user from the image data, obtaining duration of the symptom features, generating corresponding trend symptom features according to the duration, inputting the trend symptom features into a pre-trained prediction model for prediction, and predicting the probability of the waiting user having the symptom corresponding to the additional symptom information, wherein the image data specifically includes facial data, body posture and action features; The second execution module further includes: A generating unit, configured to collect coordinate information corresponding to the waiting user in the preset area based on the virtual coordinates of the outpatient department divided in the preset area, and generate a virtual boundary box of the waiting user according to the coordinate information; a determination unit, configured to determine whether the virtual boundary box can continuously capture the waiting user; an execution unit, configured to synchronize the coordinate information in real time in a preset multi-camera perspective according to the virtual bounding box, if possible, convert the coordinate information into a same physical coordinate system, and obtain multi-view behavior data of the waiting user from the multi-camera perspective; The judgment module also includes: A recognition unit, configured to recognize image features corresponding to the surface features pre-collected by the outpatient department from the waiting user in the preset area; A second judging unit, used to judge whether the image feature can be detected in the preset area; The second execution unit is used for, if possible, capturing waiting users corresponding to the image features in real time in the preset area according to different image features, and generating registration waiting seats corresponding to the waiting users through the outpatient department.

Images