Bioradar-based fall monitoring method and system, terminal, and storage medium

Abstract

Disclosed in the present invention are a bioradar-based fall monitoring method and system, a terminal, and a storage medium. The method comprises: acquiring point cloud data of the environment where a target user is located within a predetermined time, and clustering the point cloud data to obtain target point cloud data; inputting the target point cloud data into a deep learning algorithm model to obtain human target information; and analyzing the human target information on the basis of different monitoring content to obtain an analysis result, generating a corresponding fall alert event on the basis of the analysis result, and processing the fall alert event and triggering an alert, so as to alert a guardian of the target user. In the present invention, point cloud data acquired by a bioradar is analyzed to monitor the state of a target user in real time, so that a quick alert response can be provided when the user falls, and a guardian of the user can be notified in time, thereby ensuring the safety of the user.

Description

Bioradar-based fall monitoring method, system, terminal and storage medium Technical Field

[0001] The present invention relates to the field of behavior monitoring technology, and in particular to a fall monitoring method, system, terminal and storage medium based on bioradar. Background Art

[0002] With the development of the economy, people's living standards are constantly improving, and they are paying more and more attention to home safety, especially the safety of the elderly at home. Falls have become one of the main factors leading to death and accidental injuries among the elderly. As a public health issue, falls have become a common obstacle to independent living for the elderly. Therefore, in the home-based elderly care scenario, timely and effective medical and health monitoring has become crucial to detect falls in the elderly in a timely manner and provide active and proactive assistance. If falls are not detected in time, treatment will be delayed, and the safety of the elderly cannot be guaranteed.

[0003] The technical solutions currently used in fall detection mainly include accelerometers, camera-based vision technology and millimeter-wave radar. Accelerometers are usually integrated into smart watches or bracelets, and have low recognition rates for slow falls, poor comfort, inconvenient charging, and require users to wear them for detection. Camera-based vision technology requires sufficient light for fall detection, and there must be no objects blocking the camera and the person being detected, and there are privacy leaks. Millimeter-wave radar, as an emerging sensing method, is widely used in scenarios such as traffic flow recognition, vehicle speed detection, and obstacle perception. However, the false alarm rate of devices that support millimeter-wave radar is high and cannot accurately determine the type of fall. The detection of human presence and activity is not accurate enough, making it impossible to issue early warnings. It is also unable to perform calls, gesture recognition, and voice recognition, resulting in an inability to promptly determine whether the user has fallen, and thus cannot effectively ensure user safety.

[0004] Therefore, the existing technology still needs to be improved and developed. Summary of the Invention

[0005] The main purpose of the present invention is to provide a fall monitoring method, system, terminal and storage medium based on bioradar, aiming to solve the problems of existing fall monitoring methods, such as the inability to accurately determine the type of fall, low accuracy in detecting human presence and activity, and inability to accurately identify user instructions.

[0006] To achieve the above objectives, the present invention provides a bioradar-based fall monitoring method, which comprises the following steps:

[0007] Acquire point cloud data of the target user's environment within a predetermined time, and perform clustering processing on the point cloud data to obtain target point cloud data;

[0008] Inputting the target point cloud data into a deep learning algorithm model and outputting human target information, wherein the deep learning algorithm model is trained based on a training set, and the training set includes environmental point cloud data and human target data corresponding to the environmental point cloud data;

[0009] The human target information is analyzed according to different monitoring contents to obtain analysis results, a corresponding fall alarm event is generated according to the analysis results, and an alarm processing is performed on the fall alarm event to remind the guardian of the target user.

[0010] Optionally, in the bioradar-based fall monitoring method, the monitoring content includes fall prediction, gesture recognition and activity monitoring.

[0011] Optionally, in the bioradar-based fall monitoring method, if the monitoring content is fall prediction, analyzing the human target information according to different monitoring contents to obtain analysis results, and generating a corresponding fall alarm event according to the analysis results, specifically includes:

[0012] Calculating and processing the human target information to obtain the three-dimensional coordinates, displacement information, velocity information, acceleration information of the target user and the three-dimensional shape of the point cloud data;

[0013] Inputting the three-dimensional coordinates, the displacement information, the velocity information, the acceleration information, and the three-dimensional shape into a temporal machine learning model or a deep learning model to obtain a fall type, wherein the temporal machine learning model or the deep learning model is trained based on a training set of different fall types;

[0014] A corresponding processing method is matched according to the fall type to obtain a target processing method. If the target processing method is an alarm processing, a fall alarm event is generated according to the processing method.

[0015] Optionally, in the bioradar-based fall monitoring method, if the monitoring content is gesture recognition, analyzing the human target information according to different monitoring contents to obtain analysis results, and generating a corresponding fall alarm event according to the analysis results, specifically includes:

[0016] Performing time-series tracking on the human body target information to obtain tracking results, and performing joint recognition on the tracking results to obtain wrist joint information and elbow joint information;

[0017] Calculating the palm shape and arm movement trajectory of the target user according to the wrist joint information and the elbow joint information to obtain a target palm shape and a target movement trajectory;

[0018] Performing gesture recognition on the target palm shape and the target activity trajectory to obtain a target gesture, and determining whether the target gesture is a preset gesture;

[0019] If the target gesture is the preset gesture, it is determined that the target user has fallen, and a fall alarm event is generated according to the target gesture.

[0020] Optionally, in the bioradar-based fall monitoring method, if the monitoring content is activity monitoring, analyzing the human target information according to different monitoring contents to obtain analysis results, and generating a corresponding fall alarm event according to the analysis results, specifically includes:

[0021] Performing type recognition processing on the human target information to obtain a target activity type and an activity time and activity frequency of the target activity type;

[0022] Analyzing the activity time and the activity frequency to obtain health guidance data of the target activity type, and determining whether the health guidance data of the target user exceeds a predetermined change threshold;

[0023] If the health guidance data exceeds the predetermined change threshold, a target activity corresponding to the target activity type is obtained, and a fall alarm event is generated according to the target activity.

[0024] Optionally, the bioradar-based fall monitoring method, wherein the step of performing alarm processing on the fall alarm event to alert the guardian of the target user, specifically includes:

[0025] Generate a voice call instruction according to the fall alarm event, call the address book of the target user according to the voice call instruction, and obtain the number priority of the address book;

[0026] The target contact information of the guardian of the target user is obtained according to the number priority, and a voice call is made to the guardian according to the target contact information to remind the guardian to handle the fall alarm event.

[0027] Optionally, the bioradar-based fall monitoring method further includes:

[0028] Acquire voice information of the target user, recognize the voice information, obtain a recognition result, and determine whether the recognition result contains preset sensitive words;

[0029] If the recognition result includes the preset sensitive word, generating a fall alarm event according to the recognition result, and generating a voice call instruction according to the fall alarm event;

[0030] The contact information of the emergency contact is retrieved according to the voice call instruction, and a voice call is made to the emergency contact according to the contact information, wherein the emergency contact is a guardian bound to the target user.

[0031] Optionally, in the bioradar-based fall monitoring method, the bioradar-based fall monitoring system includes:

[0032] A data acquisition module is used to acquire point cloud data of the target user's environment within a predetermined time, and perform clustering processing on the point cloud data to obtain target point cloud data;

[0033] a data processing module, configured to input the target point cloud data into a deep learning algorithm model and output human target information, wherein the deep learning algorithm model is trained based on a training set, the training set including environmental point cloud data and human target data corresponding to the environmental point cloud data;

[0034] The fall warning module is used to analyze the human target information according to different monitoring contents, obtain analysis results, generate corresponding fall alarm events according to the analysis results, and perform alarm processing on the fall alarm events to remind the guardian of the target user.

[0035] In addition, to achieve the above-mentioned purpose, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a bioradar-based fall monitoring program stored on the memory and runnable on the processor, and when the bioradar-based fall monitoring program is executed by the processor, the steps of the bioradar-based fall monitoring method as described above are implemented.

[0036] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a bioradar-based fall monitoring program, and when the bioradar-based fall monitoring program is executed by the processor, the steps of the bioradar-based fall monitoring method as described above are implemented.

[0037] In the present invention, point cloud data of the target user's environment within a predetermined time period is acquired, clustered, and obtained as target point cloud data. This target point cloud data is then fed into a deep learning algorithm model to obtain human target information. This human target information is then analyzed according to different monitoring contents to obtain analysis results. A corresponding fall alarm event is generated based on the analysis results, and an alarm is processed for the fall alarm event to alert the target user's guardian. The present invention analyzes point cloud data acquired by bioradar to monitor the user's status in real time, enabling a rapid alarm response in the event of a fall, and promptly notifying the user's guardian to ensure the user's safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] FIG1 is a flow chart of a preferred embodiment of a fall monitoring method based on bioradar according to the present invention;

[0039] FIG2 is a schematic diagram showing the principle of a preferred embodiment of a bioradar-based fall monitoring system according to the present invention;

[0040] FIG3 is a schematic diagram of an operating environment of a preferred embodiment of the terminal of the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of the present invention more clear and distinct, the present invention is further described in detail below with reference to the accompanying drawings and examples. 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.

[0042] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative position relationship, movement status, etc. between the various components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indications will also change accordingly.

[0043] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features specified as "first" and "second" may explicitly or implicitly include at least one of such features. In addition, the technical solutions between the various embodiments can be combined with each other, but this must be based on the fact that ordinary technicians in this field can implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0044] The bioradar-based fall monitoring method according to a preferred embodiment of the present invention is shown in FIG1 . The bioradar-based fall monitoring method comprises the following steps:

[0045] Step S10: Acquire point cloud data of the target user's environment within a predetermined time, perform clustering processing on the point cloud data, and obtain target point cloud data.

[0046] Specifically, the technical solutions currently used in fall detection mainly include three types: acceleration sensor solutions, camera-based visual technology solutions and bio-radar solutions. However, the acceleration sensor solution is usually integrated in smart watches or bracelets, and has problems such as low recognition rate of slow falls, poor comfort, forgetting to charge, and requiring users to wear it for detection; the camera-based visual technology solution requires sufficient light for fall detection, and there must be no objects blocking the camera and the person being detected to complete the detection, and the camera has the problem of privacy leakage; the current bio-radar solution not only has a high false alarm rate, but also cannot accurately determine the type of fall; the detection of human presence and activity is not accurate enough, and it cannot perform gesture recognition and voice recognition; and voice calls cannot be made, so when the target user falls, it is impossible to effectively establish a remote call to call the corresponding guardian for processing or communication, resulting in the inability to effectively ensure the safety of the target user; the present invention solves the above-mentioned problems through a fall monitoring method based on bio-radar. In an embodiment of the present invention, a corresponding terminal (for example, a bioradar device) is pre-installed in the environment where the target user is located. The bioradar device supports not only ceiling mounting but also wall mounting, so as to accurately identify different degrees of falls in various environments. The millimeter-wave radar sensor is used to collect point cloud data in the target user's current living environment, and the point cloud data is clustered to obtain target point cloud data, so as to facilitate the subsequent real-time analysis of the target user's status based on the target point cloud data, so as to provide timely alarm reminders.

[0047] Step S20: input the target point cloud data into a deep learning algorithm model and output human target information, wherein the deep learning algorithm model is trained according to a training set, and the training set includes environmental point cloud data and human target data corresponding to the environmental point cloud data.

[0048] Specifically, after clustering the collected point cloud data to obtain target point cloud data, it is necessary to analyze the target point cloud data to obtain human target information; in an embodiment of the present invention, the target point cloud data is processed by a deep learning algorithm model to obtain human target information; then the training process of the deep learning algorithm model is: creating a deep learning algorithm training model and obtaining a training set, wherein the training set includes various environmental point cloud data and each environmental point cloud data is marked to obtain corresponding human target data; inputting the training set into the deep learning algorithm training model to train the deep learning algorithm model, the deep learning algorithm model processes the environmental point cloud data, outputs predicted human target data, corrects the deep learning algorithm training model according to the human target data, and inputs the next set of training sets into the deep learning algorithm training model , until the training times of the deep learning algorithm training model reaches a preset number of times, a deep learning algorithm model is obtained, and the deep learning algorithm model is output; the target point cloud data is input into the deep learning algorithm model, and the deep learning algorithm model processes the target point cloud data and outputs the corresponding human target information; when the target user is stationary for a long time, the existing radar monitoring scheme will mistakenly believe that there is no one in the current environment, thereby failing to detect the abnormal situation of the target user in time, and the present invention can monitor the presence of the human body. By obtaining the human target information, it can be identified that there are people in the current environment when the target user is stationary. If the target user has not moved for a long time, it may indicate other conditions (for example, fainting), and the corresponding guardian can be reminded in time to go to the target user to check the target user, thereby ensuring the safety of the target user.

[0049] Step S30: Analyze the human target information according to different monitoring contents to obtain analysis results, generate corresponding fall alarm events according to the analysis results, and perform alarm processing on the fall alarm events to remind the guardian of the target user.

[0050] Specifically, by monitoring the target user in real time in the environment where the target user is currently located, it is possible to predict the target user before he falls, and to promptly issue an early warning to the target user's guardian, so that after the target user falls, the guardian can reach the target user in time, so that the target user's safety is guaranteed; in the present invention, the early warning effect is achieved by analyzing the target user's gait, and the specific process of early warning processing is: extracting data from the human target information to obtain gait data, performing regularity analysis and frequency analysis on the gait data in turn to obtain gait analysis results, and analyzing the gait regularity and frequency model of the target user. The gait data is analyzed, and the gait law and frequency model is obtained by training the created gait law and frequency training; after the gait analysis result is obtained through the gait law and frequency model, it is judged whether the gait analysis result has an abnormal gait; if there is no abnormal gait in the gait analysis result, it means that the status of the target user is normal; if there is an abnormal gait in the gait analysis result, it is determined that the target user may be at risk of falling. For example, the analysis result shows that the target user may not walk flexibly and may be at risk of falling. It is necessary to promptly remind the guardian of the target user to take care of the target user to ensure the safety of the target user.

[0051] After obtaining the human target information based on the deep learning algorithm model, different monitoring contents will process the human target information in different ways. In an embodiment of the present invention, the monitoring content includes fall prediction, gesture recognition and activity monitoring; if the monitoring content is fall prediction, the three-dimensional coordinates, displacement information, speed information and acceleration information of the target user, and the three-dimensional shape of the point cloud data are analyzed in real time to determine the type of the target user's current fall, so as to perform different processing; the human target information is calculated and processed to obtain the three-dimensional coordinates, displacement information, speed information, acceleration information of the target user and the three-dimensional shape of the point cloud data; wherein, the displacement information includes vertical displacement and horizontal displacement, the speed information includes vertical movement speed and horizontal movement speed, and the acceleration information includes vertical acceleration and horizontal acceleration, while tracking the temporal data of the human target, inputting the temporal data, the three-dimensional coordinates, the displacement information, the speed information, the acceleration information and the three-dimensional shape into a temporal machine learning model or a deep learning model to obtain a fall type, wherein the temporal machine learning model or the deep learning model is trained based on a training set of different fall types; the fall types include but are not limited to fast Falling, falling slowly and staying still, falling slowly and struggling slowly, squatting slowly and then falling, falling from the sofa, falling slowly against the wall, supporting yourself with your back after falling, crawling forward after falling (without getting up from the upper body), moving forward after falling, getting up with a mobile phone by slightly moving, and falling slowly (the whole process is in a curled-up state), etc.; matching the corresponding processing method according to the fall type, obtaining the target processing method (for example, if the fall type is falling from the sofa, the corresponding processing method is alarm processing); if the target processing method is alarm processing, generating a fall alarm event according to the processing method; generating a voice call instruction according to the fall alarm event Command, according to the voice call command, call the address book of the target user, wherein the contact information of the guardian bound to the target user (i.e., emergency contact) is pre-stored in the address book, and the number priority of the address book is obtained, because the address book can pre-store the contact information of multiple guardians, i.e., the first contact, the second contact, etc., and the priority of the first contact is higher than the second contact, that is, when a subsequent call is made, the first contact is dialed first; the target contact information of the guardian of the target user is obtained according to the number priority, and a voice call is made to the guardian according to the target contact information;If the target contact information cannot be successfully dialed, that is, the first contact cannot be successfully connected, the contact information of the second contact is obtained and dialed to ensure that the guardian can be reminded in time to handle the fall alarm event, so that the guardian can understand the situation of the target user in the first place and make appropriate arrangements based on the situation to ensure the safety of the target user.

[0052] Furthermore, the function of making an automatic voice call after the target user falls is built into the bioradar device. The bioradar device supports SIM cards. The automatic voice call function can be completed by simply inserting the SIM card of the target user into the bioradar device. There is no need to add a voice gateway to make a voice call. Not only is the cost lower, but it is also more convenient to use.

[0053] If the monitoring content is gesture recognition, it is necessary to perform time-series tracking on the human target information to obtain tracking results, and perform joint recognition on the tracking results to obtain joint information of each joint of the target user, and then extract wrist joint information and elbow joint information from all joint information; calculate the palm shape and arm movement trajectory of the target user according to the wrist joint information and the elbow joint information to obtain the target palm shape and target movement trajectory; perform gesture recognition on the target palm shape and the target movement trajectory to obtain the target gesture, and determine whether the target gesture is a preset gesture, wherein the preset gesture includes but is not limited to shaking the hand; if the target If the gesture is the preset gesture, it is determined that the target user has fallen, and a fall alarm event is generated according to the target gesture; a voice call instruction is generated according to the fall alarm event, the address book of the target user is retrieved according to the voice call instruction, and the number priority of the address book is obtained; the target contact information of the guardian of the target user is obtained according to the number priority, and a voice call is made to the guardian according to the target contact information, and the guardian is reminded to handle the fall alarm event in time, so that the guardian can understand the situation of the target user at the first time and make appropriate arrangements according to the situation to ensure the safety of the target user.

[0054] If the monitoring content is activity monitoring, it is necessary to perform type identification processing on the human target information to obtain the target activity type and the activity time and frequency of the target activity type; analyze the activity time and the activity frequency to obtain the health guidance data of the target activity type, wherein the millimeter wave radar module will collect point cloud data within a period of time (for example, a week or a month) to analyze the activity habits of the target user in various activities in the current time period, including but not limited to the average time of getting up, the average time of resting at night, the average daily walking time, the average daily sitting time, the average daily frequency of using the bathroom, and the average daily frequency of getting up at night in the current environment. Each of the activity habits corresponds to a health instruction data; determine whether the health guidance data of the target user exceeds the predetermined A change threshold (for example, 30%); if the health guidance data exceeds the predetermined change threshold, for example, the target user gets up at night significantly more frequently, then the target activity corresponding to the target activity type is obtained, and a fall alarm event is generated according to the target activity; a voice call instruction is generated according to the fall alarm event, the address book of the target user is retrieved according to the voice call instruction, and the number priority of the address book is obtained; the target contact information of the guardian of the target user is obtained according to the number priority, and a voice call is made to the guardian according to the target contact information, so as to promptly remind the guardian to handle the fall alarm event, so that the guardian can understand the situation of the target user at the first time, and make appropriate arrangements according to the situation to ensure the safety of the target user.

[0055] Furthermore, the present invention can also recognize the voice of the target user in real time so as to call the guardian of the target user in time, so that the guardian can understand the situation of the target user at the first time and make corresponding arrangements; specifically, obtain the voice information of the target user, recognize the voice information, obtain the recognition result, and judge whether the recognition result contains preset sensitive words, wherein the sensitive words include but are not limited to falling, making a phone call and feeling uncomfortable; if the recognition result contains the preset sensitive words, a fall alarm event is generated according to the recognition result, and a voice call instruction is generated according to the fall alarm event; the contact information of the emergency contact is retrieved according to the voice call instruction, and a voice call is made to the emergency contact according to the contact information, wherein the emergency contact is the guardian bound to the target user, and the corresponding guardian is reminded in time to handle the fall alarm event, so that the guardian can understand the situation of the target user at the first time and make appropriate arrangements according to the situation to ensure the safety of the target user.

[0056] Furthermore, as shown in FIG2 , based on the above-mentioned bioradar-based fall monitoring method, the present invention also provides a bioradar-based fall monitoring system, wherein the bioradar-based fall monitoring system includes:

[0057] The data acquisition module 51 is used to acquire point cloud data of the target user's environment within a predetermined time, and perform clustering processing on the point cloud data to obtain target point cloud data;

[0058] a data processing module 52 for inputting the target point cloud data into a deep learning algorithm model and outputting human target information, wherein the deep learning algorithm model is trained based on a training set including environmental point cloud data and human target data corresponding to the environmental point cloud data;

[0059] The fall warning module 53 is used to analyze the human target information according to different monitoring contents, obtain analysis results, generate corresponding fall warning events according to the analysis results, and perform alarm processing on the fall warning events to remind the guardian of the target user.

[0060] Furthermore, as shown in FIG3 , based on the above-mentioned bioradar-based fall monitoring method, the present invention also provides a terminal, which includes a processor 10, a memory 20, and a display 30. FIG3 only shows some components of the terminal, but it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0061] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory of the terminal. In other embodiments, the memory 20 may also be an external storage device of the terminal, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the terminal. Furthermore, the memory 20 may also include both an internal storage unit of the terminal and an external storage device. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code of the installation terminal. The memory 20 may also be used to temporarily store data that has been output or is to be output. In one embodiment, a bioradar-based fall monitoring program 40 is stored on the memory 20, and the bioradar-based fall monitoring program 40 can be executed by the processor 10, thereby realizing the bioradar-based fall monitoring method in the present application.

[0062] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, configured to execute program codes or process data stored in the memory 20, such as executing the bioradar-based fall monitoring method.

[0063] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The components 10-30 of the terminal communicate with each other via a system bus.

[0064] In one embodiment, when the processor 10 executes the bio-radar-based fall detection program 40 in the memory 20, the following steps are implemented:

[0065] Acquire point cloud data of the target user's environment within a predetermined time, and perform clustering processing on the point cloud data to obtain target point cloud data;

[0066] Inputting the target point cloud data into a deep learning algorithm model and outputting human target information, wherein the deep learning algorithm model is trained based on a training set, and the training set includes environmental point cloud data and human target data corresponding to the environmental point cloud data;

[0067] The human target information is analyzed according to different monitoring contents to obtain analysis results, a corresponding fall alarm event is generated according to the analysis results, and an alarm processing is performed on the fall alarm event to remind the guardian of the target user.

[0068] The monitoring content includes fall prediction, gesture recognition and activity monitoring.

[0069] Wherein, if the monitoring content is fall prediction, the human target information is analyzed according to different monitoring contents to obtain analysis results, and a corresponding fall alarm event is generated according to the analysis results, specifically including:

[0070] Calculating and processing the human target information to obtain the three-dimensional coordinates, displacement information, velocity information, acceleration information of the target user and the three-dimensional shape of the point cloud data;

[0071] Inputting the three-dimensional coordinates, the displacement information, the velocity information, the acceleration information, and the three-dimensional shape into a temporal machine learning model or a deep learning model to obtain a fall type, wherein the temporal machine learning model or the deep learning model is trained based on a training set of different fall types;

[0072] A corresponding processing method is matched according to the fall type to obtain a target processing method. If the target processing method is an alarm processing, a fall alarm event is generated according to the processing method.

[0073] Wherein, if the monitoring content is gesture recognition, analyzing the human target information according to different monitoring contents to obtain analysis results, and generating a corresponding fall alarm event according to the analysis results, specifically includes:

[0074] Performing time-series tracking on the human body target information to obtain tracking results, and performing joint recognition on the tracking results to obtain wrist joint information and elbow joint information;

[0075] Calculating the palm shape and arm movement trajectory of the target user according to the wrist joint information and the elbow joint information to obtain a target palm shape and a target movement trajectory;

[0076] Performing gesture recognition on the target palm shape and the target activity trajectory to obtain a target gesture, and determining whether the target gesture is a preset gesture;

[0077] If the target gesture is the preset gesture, it is determined that the target user has fallen, and a fall alarm event is generated according to the target gesture.

[0078] Among them, if the monitoring content is activity monitoring, analyzing the human target information according to different monitoring contents to obtain analysis results, and generating a corresponding fall alarm event according to the analysis results specifically includes:

[0079] Performing type recognition processing on the human target information to obtain a target activity type and an activity time and activity frequency of the target activity type;

[0080] Analyzing the activity time and the activity frequency to obtain health guidance data of the target activity type, and determining whether the health guidance data of the target user exceeds a predetermined change threshold;

[0081] If the health guidance data exceeds the predetermined change threshold, a target activity corresponding to the target activity type is obtained, and a fall alarm event is generated according to the target activity.

[0082] The step of performing alarm processing on the fall alarm event to alert the guardian of the target user specifically includes:

[0083] Generate a voice call instruction according to the fall alarm event, call the address book of the target user according to the voice call instruction, and obtain the number priority of the address book;

[0084] The target contact information of the guardian of the target user is obtained according to the number priority, and a voice call is made to the guardian according to the target contact information to remind the guardian to handle the fall alarm event.

[0085] The bioradar-based fall monitoring method further includes:

[0086] Acquire voice information of the target user, recognize the voice information, obtain a recognition result, and determine whether the recognition result contains preset sensitive words;

[0087] If the recognition result includes the preset sensitive word, generating a fall alarm event according to the recognition result, and generating a voice call instruction according to the fall alarm event;

[0088] The contact information of the emergency contact is retrieved according to the voice call instruction, and a voice call is made to the emergency contact according to the contact information, wherein the emergency contact is a guardian bound to the target user.

[0089] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a bioradar-based fall monitoring program, and when the bioradar-based fall monitoring program is executed by a processor, the steps of the bioradar-based fall monitoring method as described above are implemented.

[0090] In summary, the present invention provides a fall monitoring method, system, terminal, and storage medium based on bioradar, the method comprising: obtaining point cloud data of the target user's environment within a predetermined time, clustering the point cloud data to obtain target point cloud data; inputting the target point cloud data into a deep learning algorithm model to obtain human target information, wherein the deep learning algorithm model is trained based on a training set, the training set including environmental point cloud data and human target data corresponding to the environmental point cloud data; analyzing the human target information according to different monitoring contents to obtain analysis results, generating corresponding fall alarm events based on the analysis results, and performing alarm processing on the fall alarm events to alert the guardian of the target user. The present invention analyzes the point cloud data obtained by the bioradar to monitor the user's status in real time, so as to quickly respond to the alarm in the event of a fall, and promptly notify the user's guardian to ensure the user's safety.

[0091] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or terminal comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or terminal. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or terminal comprising the element.

[0092] Of course, those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware (such as a processor, controller, etc.) through a computer program. The program can be stored in a computer-readable storage medium that can be read by a computer. When the program is executed, it can include the processes in the above-described method embodiments. The computer-readable storage medium can be a memory, a magnetic disk, an optical disk, etc.

[0093] It should be understood that the application of the present invention is not limited to the above examples. For those skilled in the art, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A fall monitoring method based on a bio-radar, characterized in that, The fall monitoring method based on bio-radar includes: Obtain the point cloud data of the environment where the target user is located within a predetermined time, perform clustering processing on the point cloud data to obtain target point cloud data; Input the target point cloud data into a deep learning algorithm model to output human target information, where the deep learning algorithm model is trained according to a training set, and the training set includes environmental point cloud data and human target data corresponding to the environmental point cloud data; Analyze the human target information according to different monitoring contents to obtain an analysis result, generate a corresponding fall warning event according to the analysis result, and perform an alarm process on the fall warning event to remind the guardian of the target user.

2. The fall monitoring method based on a bio-radar according to claim 1, characterized in that, The monitoring contents include fall prediction, gesture recognition, and activity monitoring.

3. The fall monitoring method based on bio-radar according to claim 2, characterized in that, If the monitoring content is fall prediction, the analyzing the human target information according to different monitoring contents to obtain an analysis result, and generating a corresponding fall warning event according to the analysis result specifically includes: Perform calculation processing on the human target information to obtain the three-dimensional coordinates, displacement information, velocity information, acceleration information of the target user, and the three-dimensional shape of the point cloud data; Input the three-dimensional coordinates, the displacement information, the velocity information, the acceleration information, and the three-dimensional shape into a temporal machine learning model or a deep learning model to obtain a fall type, where the temporal machine learning model or the deep learning model is trained according to training sets of different fall types; Match a corresponding processing method according to the fall type to obtain a target processing method. If the target processing method is an alarm process, generate a fall warning event according to the processing method.

4. The fall monitoring method based on a bio-radar according to claim 2, characterized in that If the monitoring content is gesture recognition, the analyzing the human target information according to different monitoring contents to obtain an analysis result, and generating a corresponding fall warning event according to the analysis result specifically includes: Perform temporal tracking on the human target information to obtain a tracking result, and perform joint recognition on the tracking result to obtain wrist joint information and elbow joint information; Calculate the palm shape and the movement trajectory of the arm of the target user according to the wrist joint information and the elbow joint information to obtain a target palm shape and a target movement trajectory; Perform gesture recognition on the target palm shape and the target movement trajectory to obtain a target gesture, and determine whether the target gesture is a preset gesture; If the target gesture is the preset gesture, determine that the target user has a fall behavior, and generate a fall warning event according to the target gesture.

5. The fall monitoring method based on bio-radar according to claim 2, wherein, If the monitoring content is activity monitoring, the analyzing the human target information according to different monitoring contents to obtain an analysis result, and generating a corresponding fall warning event according to the analysis result specifically includes: Perform type recognition processing on the human target information to obtain a target activity type, the activity time, and the activity frequency of the target activity type; Analyze the activity time and the activity frequency to obtain the health guidance data of the target activity type, and determine whether the health guidance data of the target user exceeds a predetermined change threshold; If the health guidance data exceeds the predetermined change threshold, obtain the target activity corresponding to the target activity type, and generate a fall warning event according to the target activity.

6. The fall monitoring method based on bio-radar according to claim 1, wherein The alarm processing of the fall warning event to remind the guardian of the target user specifically includes: Generate a voice call instruction according to the fall warning event, retrieve the address book of the target user according to the voice call instruction, and obtain the number priority of the address book; Obtain the target contact information of the guardian of the target user according to the number priority, and make a voice call to the guardian according to the target contact information to remind the guardian to handle the fall warning event.

7. The fall monitoring method based on a bio-radar according to claim 1, characterized in that The fall monitoring method based on bio-radar further includes: Obtain the voice information of the target user, identify the voice information to obtain an identification result, and determine whether the identification result contains a preset sensitive word; If the identification result contains the preset sensitive word, generate a fall warning event according to the identification result, and generate a voice call instruction according to the fall warning event; Retrieve the contact information of the emergency contact according to the voice call instruction, and make a voice call to the emergency contact according to the contact information, where the emergency contact is the guardian bound to the target user.

8. A fall monitoring system based on a bio-radar, characterized in that, The fall monitoring system based on bio-radar includes: A data acquisition module for acquiring point cloud data of the environment where the target user is located within a predetermined time, and performing clustering processing on the point cloud data to obtain target point cloud data; A data processing module for inputting the target point cloud data into a deep learning algorithm model and outputting human target information, where the deep learning algorithm model is trained according to a training set, and the training set includes environmental point cloud data and human target data corresponding to the environmental point cloud data; A fall warning module for analyzing the human target information according to different monitoring contents to obtain an analysis result, generating a corresponding fall warning event according to the analysis result, and performing alarm processing on the fall warning event to remind the guardian of the target user.

9. A terminal, characterized in that, The terminal includes: a memory, a processor, and a bio-radar-based fall monitoring program stored on the memory and executable on the processor. When the bio-radar-based fall monitoring program is executed by the processor, it implements the steps of the bio-radar-based fall monitoring method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a bio-radar-based fall monitoring program. When the bio-radar-based fall monitoring program is executed by a processor, it implements the steps of the bio-radar-based fall monitoring method according to any one of claims 1-7.

Images