Exercise prescription recommendation method and system for old people, product and medium

Abstract

The invention discloses an exercise prescription recommendation method and system for old people, a product and a medium. The method comprises the following steps: shooting a picture of a sports place, monitoring basic body parameters of a person in the picture of the sports place, capturing a real-time face video of the person, measuring a real-time heart rate, evaluating a cardiovascular condition grade, and formulating a real-time sports scheme. Monitoring real-time vital sign information and real-time limb actions in the movement process of the person after the person in the picture of the sports place is monitored to start to move; when the real-time vital sign information is not in the preset vital sign information threshold range, displaying a real-time vital sign data chart and an instructive animation through a display screen; when it is detected that the real-time limb movement has the preset typical movement characteristics, position information and real-time vital sign information are sent to a preset management organization. By implementing the technical scheme provided by the invention, the safety in the process of recommending the exercise prescriptions of the elderly groups is improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and particularly to a method, system, product, and medium for recommending exercise prescriptions for the elderly population. Background Art

[0002] With the intensification of population aging, elderly health management has become particularly important and a key link concerning the quality of life of the elderly and the overall well-being of society. Especially in the specific area of fitness exercise, as the health awareness of the elderly continues to increase, their demand for maintaining physical functions and enhancing physical fitness through scientific and reasonable exercise shows an increasingly significant and continuous growth trend.

[0003] Currently, there are various common methods for recommending exercise prescriptions for the elderly population. For example, wearing smart fitness devices such as smart bracelets and smart watches, or offering fitness courses specifically for the elderly through online platforms to meet the fitness needs of the elderly population.

[0004] However, whether wearing smart fitness devices or watching online courses, if abnormal behaviors such as inappropriate exercise postures or sudden physical discomfort occur during the exercise of the elderly, it is impossible to detect them in the first time and change their exercise methods, which may lead to safety problems or even cause damage to the body after exercise, lacking safety. Summary of the Invention

[0005] This application provides a method, system, product, and medium for recommending exercise prescriptions for the elderly population, aiming to improve the safety during the process of recommending exercise prescriptions for the elderly population.

[0006] In the first aspect, this application provides a method for recommending exercise prescriptions for the elderly population, specifically including: Shoot the video of the sports venue. After detecting a complete person in the video of the sports venue, determine the number of complete persons. When the number of complete persons is one, detect the basic body parameters of the complete person. The basic body parameters include height data, weight data, and body circumference data. Capture the real-time facial video of the complete person, measure the real-time heart rate by analyzing the facial features in the real-time facial video, and evaluate the cardiovascular condition level. The facial features are the change situation of the facial skin color in the real-time facial video. According to the basic body parameters, real-time heart rate, and cardiovascular condition level, formulate a real-time exercise plan. The real-time exercise plan includes exercise type and exercise duration. After detecting that the complete person in the sports venue video starts to exercise, monitor the real-time vital sign information and real-time limb movements during the exercise of the complete person. The real-time vital sign information includes real-time body temperature change, real-time breathing frequency, and real-time breathing depth. When the real-time vital sign information is not within the preset vital sign information threshold range, calculate the real-time vital sign data chart. According to the real-time vital sign information and real-time limb movements, generate a real-time guiding animation, and display the real-time vital sign data chart and guiding animation through the display screen. When detecting that the real-time limb movement appears with the preset typical movement characteristics, send the first prompt information. The preset typical movement characteristics are the movement characteristics corresponding to the preset twitching, spasm, and falling actions. The first prompt information is the position information and real-time vital sign information of the complete person sent to the preset management organization.

[0007] In the above embodiment, by judging the number of people in the video of the sports venue, when there is only one person in the video, detect their basic body parameters, capture the facial video to measure the heart rate and evaluate the cardiovascular condition, so as to formulate a suitable exercise plan. After the exercise starts, continuously track the real-time vital signs and limb movements. If the vital signs are abnormal, assist in adjustment. Once a typical dangerous action appears, immediately send the position and vital sign information to the preset organization, improving the safety in the process of recommending exercise prescriptions for the elderly group.

[0008] Combined with some embodiments of the first aspect, in some embodiments, when the number of complete persons is one, detecting the basic body parameters of the complete person specifically includes: When the number of complete persons is one, construct the body contour information of the complete person and identify the clothing and equipment features; use the body contour information and clothing and equipment features to calculate the actual body circumference value of the complete person; through the actual body circumference value, calculate the basic body parameters of the complete person.

[0009] In the above embodiment, when there is only one person in the sports venue, first construct the body contour, accurately identify the characteristics of clothing and equipment, and calculate the actual values of each dimension of the body based on this information, accurately understand the physical condition of a single person, and avoid misrecommending exercise plans due to measurement errors.

[0010] In some embodiments in combination with some embodiments of the first aspect, when the real-time vital sign information is not within the preset vital sign information threshold range, a real-time vital sign data chart is calculated, and a real-time guiding animation is generated according to the real-time vital sign information and the real-time body movement. The real-time vital sign data chart and the guiding animation are displayed through a display screen, specifically including: When the real-time vital sign information is not within the preset vital sign information threshold range, the real-time vital sign information of the past preset duration is arranged in chronological order to obtain an organized real-time vital sign data set; according to the real-time vital sign data set, a real-time vital sign data chart is drawn; the real-time body movement is recognized and classified based on a human pose recognition algorithm to obtain real-time body movement data; the human pose recognition algorithm is obtained by training a deep learning model in advance using a body movement data group; according to the real-time body movement data and the real-time vital sign information, a corresponding guiding animation is selected from a preset animation template library; the preset animation template library is a preset library storing guiding animations corresponding to different vital sign information; in combination with the real-time vital sign information, the data in the guiding animation is modified to obtain a real-time guiding animation; the real-time vital sign data chart and the real-time guiding animation are displayed through a display screen.

[0011] In the above embodiments, after it is found that the real-time vital sign information exceeds the preset threshold, the vital sign information of a previous period of time is sorted in sequence and drawn into an intuitive chart. At the same time, the trained human pose recognition algorithm is used to analyze the real-time body movement to obtain corresponding data. Then, based on the two, a guiding animation is selected from the preset library, and the animation is improved in combination with the current vital sign information. Finally, it is displayed on the display screen to avoid potential health hazards caused by incorrect exercise methods for the elderly during exercise.

[0012] In some embodiments in combination with some embodiments of the first aspect, when it is detected that the real-time body movement exhibits a preset typical action feature, a first prompt message is sent, specifically including: Semantic labeling and learning are performed on a large number of different types of movement actions to construct an action semantic knowledge base; in the action semantic knowledge base, keyword labeling is performed on the typical semantics corresponding to the preset typical action features; the real-time semantics corresponding to the real-time body movement is searched for in the action semantic knowledge base; in the case where the real-time semantics cannot be found, the real-time body movement is disassembled frame by frame into real-time basic actions; the real-time basic semantics corresponding to the real-time basic actions is searched for in the action semantic knowledge base; if the real-time semantics or the real-time basic semantics is a typical semantics, a first prompt message is sent; the first prompt message is the position information and the real-time vital sign information of the complete person sent to a preset management organization.

[0013] In the above embodiments, by learning the semantics of numerous motion actions, an action semantics knowledge base is established, and special identifiers are given to the typical semantics corresponding to specific actions. During exercise, the real-time limb action semantics are searched. If not found, the real-time limb action is disassembled frame by frame and searched again. Once the real-time semantics point to the typical semantics, the real-time position and physical signs information are immediately sent to the preset organization for its quick rescue to ensure the safety of the exerciser.

[0014] Combined with some embodiments of the first aspect, in some embodiments, when a preset typical action feature appears in the detected real-time limb action and a prompt message is sent, it further includes: When the number of complete persons is greater than or equal to two, a separate identity identifier is established for each complete person; the identity identifier is used to mark and distinguish different complete persons; the basic body parameters, heart rate of each complete person are monitored, and the cardiovascular condition level is evaluated to obtain a basic body parameter data group, a heart rate data group, and a cardiovascular condition level data group; according to the basic body parameter data group, the heart rate data group, and the cardiovascular condition level data group, a multi-person exercise plan is formulated; when the number of complete persons starting to exercise in the monitored exercise venue scene is greater than or equal to one, the vital signs information and limb actions of each complete person during the exercise process are monitored to obtain a vital signs information data group and a limb action group; when there is abnormal vital signs information in the vital signs information data group, the identity identifier of the complete person corresponding to the abnormal vital signs information is detected, marked as an abnormal identity identifier, and a second prompt message is displayed through the display screen; the abnormal vital signs information is the vital signs information in the vital signs information data group that is not within the preset vital signs information threshold range; the second prompt message is to prompt that the current vital signs information of the complete person corresponding to the abnormal identity identifier is abnormal.

[0015] In the above embodiments, when the number of people in the exercise venue is not less than two, an identity identifier is established for each person, the body parameters, heart rate, etc. are monitored to generate corresponding data groups, and a multi-person exercise plan is formulated accordingly; after the exercise starts, the vital signs and limb actions are continuously monitored. Once someone's vital signs are found to be abnormal, it is locked by the identity identifier and a prompt is displayed on the display screen to accurately ensure the safety of individual exercise and make the multi-person exercise scientific and orderly.

[0016] Combined with some embodiments of the first aspect, in some embodiments, when the number of complete persons is greater than or equal to two and a separate identity identifier is established for each complete person, it further includes: Detect the real-time positions of each complete person to obtain a real-time position data group; the identity identifiers in the real-time position data group correspond to the real-time positions; in the real-time position data group, compare the distances between the real-time positions corresponding to different identity identifiers pairwise; when the distance between the real-time positions is less than a preset minimum distance threshold, switch to an auxiliary perspective to monitor the complete person corresponding to the first smallest distance identity identifier and the complete person corresponding to the second smallest distance identity identifier; after switching to the auxiliary perspective, detect the first real-time position of the complete person corresponding to the first smallest distance identity identifier and the second real-time position of the complete person corresponding to the second smallest distance identity identifier; when the distance between the first real-time position and the second real-time position is less than the preset minimum distance threshold, emit a prompt sound through a speaker; the prompt sound is used to prompt the complete person corresponding to the first smallest distance identity identifier and the complete person corresponding to the second smallest distance identity identifier to increase the distance.

[0017] In the above embodiment, by detecting the real-time positions of each complete person to form a data group, comparing the distances between the real-time positions of different persons, switching to an auxiliary perspective to detect the positions again when the distance is too close, and if it is still less than the threshold, emitting a prompt sound through a speaker to remind the relevant personnel to increase the distance, timely reminding those with too close distances to avoid collisions and other situations, and ensuring the safety of the personnel in the sports venue.

[0018] Combined with some embodiments of the first aspect, in some embodiments, when there is abnormal vital sign information in the vital sign information data group, detect the identity identifier of the complete person corresponding to the abnormal vital sign information, mark it as an abnormal identity identifier, and after displaying a second prompt message through a display screen, further include: Search for the semantics corresponding to each limb movement in the limb movement group in the action semantics knowledge base to obtain a limb movement semantics group; when a limb movement that cannot be searched is detected, disassemble the limb movement that cannot be searched frame by frame to obtain a disassembled limb movement group; the limb movement that cannot be searched is a limb movement in the limb movement group that cannot be found in the action semantics knowledge base; search for the disassembled semantics group corresponding to the disassembled limb movement group in the action semantics knowledge base; if there is a typical semantics in the limb movement semantics group or the disassembled semantics group, detect the identity identifier of the complete person corresponding to the typical semantics, mark it as an abnormal action identity identifier, and send a third prompt message; the third prompt message is sent to a preset management organization, and includes the position information of the complete person corresponding to the abnormal action identity identifier and the vital sign information of the complete person corresponding to the abnormal action identity identifier.

[0019] In the above embodiment, search for the semantics of all limb movements of multiple people in the action semantics knowledge base, and if not found, disassemble frame by frame and search again; once there is a typical semantics in the limb movement group or the disassembled group, lock the identity of the corresponding person and send their position and vital sign information to the preset management organization, which can quickly report dangers and ensure the safety of the exercisers.

[0020] In a second aspect, an exercise prescription recommendation system for the elderly provided by an embodiment of the present application includes one or more processors and a memory. The memory is coupled to the one or more processors and is used to store computer program code. The computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the exercise prescription recommendation system for the elderly to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0021] In a third aspect, an embodiment of the present application provides a computer program product containing instructions. When the computer program product runs on the exercise prescription recommendation system for the elderly, it causes the exercise prescription recommendation system for the elderly to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0022] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium including instructions. When the instructions run on the exercise prescription recommendation system for the elderly, it causes the exercise prescription recommendation system for the elderly to execute the method described in the first aspect and any possible implementation manner in the first aspect.

[0023] It can be understood that the exercise prescription recommendation system for the elderly provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the exercise prescription recommendation method provided by the embodiments of the present application. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method and will not be elaborated here.

[0024] One or more technical solutions provided by the embodiments of the present application have at least the following technical effects or advantages: 1. In the present application, by shooting the picture of the exercise venue, when it is judged that there is only one person, the basic physical parameters are detected, the facial video is captured to measure the heart rate, and the cardiovascular condition is evaluated, so as to formulate a suitable exercise plan. After the exercise starts, the real-time vital signs and limb movements are continuously tracked. If the signs are abnormal, a chart and animation are generated to assist in adjustment. Once typical dangerous actions such as convulsions occur, the location and sign information are immediately sent to a preset organization, improving the safety in the process of recommending exercise prescriptions for the elderly.

[0025] 2. In the present application, by learning the semantics of numerous exercise actions, an action semantics knowledge base is established, and special identifiers are given to the typical semantics corresponding to dangerous actions. During the exercise, the semantics of the real-time limb movements are searched. If not found, the real-time limb movements are disassembled frame by frame and searched again. Once the real-time semantics point to the typical semantics, the real-time location and sign information are immediately sent to a preset organization for its quick rescue to ensure the safety of the exerciser.

[0026] 3. When the number of people exercising is not less than two, this application detects the real-time positions of each complete person to form a data set, compares the distances between the real-time positions of different people, switches to an auxiliary view for position detection when the distance is too close, and if it is still less than the threshold, a prompt sound is emitted through the speaker to remind the relevant personnel to increase the distance, timely reminding those who are too close to avoid collisions and other situations, ensuring the safety of people in the sports venue. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a schematic structural diagram of an applicable system framework for the exercise prescription recommendation method for the elderly group in an embodiment of this application; Figure 2 is a schematic flowchart of an exercise prescription recommendation method for the elderly group in an embodiment of this application; Figure 3 is another schematic flowchart of an exercise prescription recommendation method for the elderly group in an embodiment of this application; Figure 4 is another schematic flowchart of an exercise prescription recommendation method for the elderly group in an embodiment of this application; Figure 5 is an exemplary hardware structural diagram of an exercise prescription recommendation system for the elderly group in an embodiment of this application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] The terms used in the following embodiments of this application are only for the purpose of describing specific embodiments and are not intended to limit this application. As used in the specification and appended claims of this application, the singular forms "a", "an", "", "above-mentioned", "the", and "this" are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0029] Hereinafter, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as implying or suggesting relative importance or implicitly indicating the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of this application, unless otherwise stated, the meaning of "a plurality" is two or more.

[0030] Figure 1 is a schematic structural diagram of an applicable system framework for the exercise prescription recommendation method for the elderly group in an embodiment of this application.

[0031] Please refer to Figure 1 , this system includes a collection device 110, a data processing and storage device 120, a display screen 130, and a communication device 140.

[0032] The acquisition device 110 is used to capture images of the sports venue and detect the body parameters and vital signs of the people in the images, including a camera 111, a body parameter monitoring device 112, and a vital sign monitoring device 113. Among them, the camera 111 can be one or more, which is used to capture images of the sports venue, can cover the entire sports venue, monitor whether there are complete people and judge the number of complete people, and at the same time capture the real-time facial video of the complete people; the body parameter monitoring device 112 can include a laser rangefinder, an ultrasonic height measuring instrument, a non-contact body circumference measuring device based on millimeter-wave radar, etc., and obtain the circumference data of multiple body parts through non-contact measurement; the vital sign monitoring device 113 can include an infrared thermal imager, a non-contact breathing monitoring device based on radar, etc., and can monitor the vital signs of complete people in real time without contacting the body.

[0033] The data processing and storage device 120, as the core data processing and storage part of the system, includes a server 121 and a data storage device 122. Among them, the server 121 receives data from various acquisition devices and performs complex calculations and analyses; the data storage device 122 can be a hard disk array or a cloud storage service, which is used to store various data collected by the system, such as basic body parameter data, real-time vital sign data, facial video data, exercise plan data, etc.

[0034] The display screen 130 is used to display real-time vital sign data charts and guiding animations. It can be a liquid crystal display screen, a light-emitting diode display screen, etc. The size is selected according to the size of the sports venue and the viewing distance, and it can clearly display the data charts and animation content.

[0035] The communication device 140 can be a network switch or a router, which is used to build an internal network, connect each acquisition device, processing device, and display device together, realize data transmission and sharing, and ensure the stability and high speed of the network to meet the requirements of real-time data transmission.

[0036] In some embodiments of the present application, an audio device 150 is further included, which is used to emit a prompt sound when needed.

[0037] The present application provides a method for recommending exercise prescriptions for the elderly population. It uses video to monitor the people in the exercise venue, and judges the number of people through algorithms such as human contour detection. For the single-person situation, it obtains basic physical parameters, uses the change in skin color in the facial video, combines with a physiological model to analyze the heart rate and cardiovascular level, and formulates an exercise plan based on the above data. During exercise, it continuously monitors vital signs and limb movements. When abnormal, it generates charts and animations for guidance, and sends a prompt message when specific movement characteristics appear. This enables the elderly population to exercise reasonably according to a scientific exercise plan, reduces the exercise risk, and improves the safety in the process of recommending exercise prescriptions for the elderly population.

[0038] Please refer to Figure 2 , which is a schematic flowchart of using the method for recommending exercise prescriptions for the elderly population in the embodiments of the present application, including the following steps: S201. Shoot the picture of the exercise venue. When a complete person appears in the picture of the exercise venue, judge the number of complete people; Use high-definition cameras arranged at various key positions in the exercise venue to continuously collect the picture information in the venue. These cameras capture images at a certain frame rate per second, and then, with the help of image recognition and analysis technology, process each frame of the picture. By identifying the human contour features, head and limb shapes, etc. in the picture, it is judged whether it is a complete person. Use the algorithm for detecting human targets and judging the number in the picture of the exercise venue to compare and count multiple consecutive frames of images, and determine the number of complete people that appear.

[0039] The algorithm for detecting human targets and judging the number in the picture of the exercise venue can be a target detection algorithm, a human pose estimation auxiliary judgment algorithm, a data fusion and post-processing algorithm, etc. From the image data of the exercise venue, locate and identify human targets, extract, analyze and fuse multi-faceted information such as the appearance features (such as the target detection algorithm) and pose features (such as the human pose estimation auxiliary judgment algorithm) of the people, and use a series of rules and post-processing means to exclude interference factors, and finally achieve an accurate judgment of the number of complete people.

[0040] S202. When the number of complete people is one, detect the basic physical parameters of the complete person; Specifically, the basic physical parameters include height data, weight data and body circumference data.

[0041] When the system determines that there is only one old person in the picture of the exercise venue, start the process of detecting basic physical parameters. Use a laser rangefinder installed at a specific position to emit a laser beam and receive the reflected light, measure the distance from the top of the old person's head to the ground, and obtain the height data; use a millimeter wave radar to penetrate the clothes, construct a body contour based on the reflected wave, calculate the waist circumference, hip circumference and other circumference values, and combine with the height data to calculate the weight data.

[0042] In some embodiments of the present application, the laser ranging method is used to obtain height data; in some other embodiments of the present application, methods such as the ultrasonic ranging method or the method based on machine vision can also be used to obtain height data, which is not limited herein.

[0043] In some embodiments of the present application, a millimeter-wave radar is used to construct a body contour to obtain body girth data; in some other embodiments of the present application, methods such as the three-dimensional human body scanning method can also be used to obtain body girth data, which is not limited herein.

[0044] S203. Capture the real-time facial video of a complete person, and measure the real-time heart rate and evaluate the cardiovascular condition level by analyzing the facial features in the real-time facial video; Specifically, the facial feature is the change in the facial skin color in the real-time facial video.

[0045] When the camera continuously captures the video of a complete person, lock the facial skin area to obtain the real-time facial video of the complete person, and monitor the subtle changes in the skin color. Based on the fact that blood flow affects the light transmittance and reflected light characteristics of the facial skin, for example, when the heart rate accelerates, the blood supply to specific parts of the face changes, resulting in regular fluctuations in the reflection of skin color in different bands such as red light and blue light. According to this, extract the characteristic data, convert it into the real-time heart rate, and then combine the big data model to evaluate the cardiovascular condition level.

[0046] In some embodiments of the present application, the cardiovascular condition level can be divided into a good level, a medium-risk level, and a high-risk level.

[0047] Good level: The heart rate is stable within the normal range. Generally, for the elderly, the resting heart rate is about 60-80 beats per minute, and it rises appropriately during exercise, such as rising to 100-120 beats per minute, and can quickly return to near the resting heart rate level after exercise; the change in the facial skin color is regular and small. For example, during exercise, due to the normal acceleration of blood circulation, the face may show slight rosiness, but this rosiness is evenly distributed in areas such as the cheeks, and there are no local abnormal color changes (such as suddenly turning white or red). This indicates that the heart can effectively pump blood throughout the body, and the cardiovascular system functions well.

[0048] Medium risk level: The heart rate often exceeds the normal range during exercise, and may reach 80-90 beats per minute at rest. During exercise, it will rapidly rise to 120-140 beats per minute, and it takes a relatively long time to return to the resting level, or the heart rate fluctuates greatly. For example, there are obvious fluctuations up and down within a short period of time (such as fluctuations of 10-20 beats per minute). The color change of the facial skin is more obvious and uneven. Local flushing or paleness may occur. For example, the forehead, the tip of the nose and other parts are overly red, while the color of other parts is normal or relatively pale. This may imply a certain degree of disorder in blood circulation, that the heart may need to work harder to maintain blood supply, or there are some problems with the regulatory function of blood vessels.

[0049] High risk level: The resting heart rate is often higher than 90 beats per minute. During exercise, the heart rate rises sharply, exceeding 140 beats per minute and is difficult to control, or there are arrhythmia phenomena, such as obvious irregularities in the heartbeat intervals. After exercise, the heart rate cannot return to the normal range for a long time, and even shows a continuous upward trend. There are serious abnormal changes in the color of the facial skin. For example, the face suddenly turns pale or shows a bluish-purple color. This color change may be due to the heart's inability to effectively pump blood out, resulting in insufficient blood oxygenation or serious obstruction of blood circulation. In addition, the color change of the skin may be accompanied by other symptoms, such as facial swelling, etc.

[0050] S204. Develop a real-time exercise plan based on basic body parameters, real-time heart rate and cardiovascular condition level; Specifically, the real-time exercise plan includes exercise type and exercise duration.

[0051] Based on the obtained basic body parameters such as the height, weight and body circumference of the elderly, initially judge their basic physical condition. For example, judge whether they are overweight through the body mass index. Combine the current heart load reflected by the real-time heart rate and the evaluated cardiovascular condition level, and comprehensively analyze to develop an exercise plan. If the elderly are relatively strong and have good cardiovascular health, medium-intensity exercises such as brisk walking and aerobics can be recommended, and the exercise duration is set at 30-40 minutes; if they are weak and have poor cardiovascular health, arrange slow walking for relaxation, with a duration of 15-20 minutes.

[0052] S205. After monitoring that a complete person in the exercise venue starts to exercise, monitor the real-time vital signs information and real-time limb movements of the complete person during the exercise process; Specifically, the real-time vital signs information includes real-time body temperature changes, real-time respiratory rate and real-time respiratory depth.

[0053] After the system starts to monitor a complete person's movement, the millimeter-wave radar emits and receives electromagnetic waves. By modulating the echo through the chest wall, phase data containing respiration and heartbeat information is obtained. After processing, the respiration rate and heartbeat rate are calculated, and then converted into the real-time heart rate. At the same time, the infrared thermal imager captures the body's thermal radiation and monitors the real-time body temperature based on the heat change. The camera takes pictures and uses image recognition technology to analyze the amplitude and speed of limb movements, etc.

[0054] S206. When the real-time vital sign information is not within the preset vital sign information threshold range, calculate the real-time vital sign data chart, generate a real-time guiding animation according to the real-time vital sign information and real-time limb movements, and display the real-time vital sign data chart and the guiding animation on the display screen. When the system detects that the real-time vital signs exceed or are lower than the normal threshold, collect and organize multiple sets of data such as recent body temperature and respiration. Through the built-in program, these data are plotted into intuitive line charts, bar charts and other charts. At the same time, use image recognition to analyze limb movements, combine abnormal sign information, match and generate corresponding guiding animations from the preset animation templates, and finally present them on the display screen.

[0055] S207. When it is detected that the real-time limb movement shows preset typical movement characteristics, send a first prompt message.

[0056] Specifically, the preset typical movement characteristics are the movement characteristics corresponding to preset twitching, spasm, falling and other movements; the first prompt message is the position information and real-time vital sign information of the complete person sent to the preset management organization.

[0057] Through the high-definition camera, the limb movement pictures of the elderly are captured in real time. Through the built-in image recognition algorithm, each frame of the image is analyzed. When abnormal states such as twitching, spasm, and falling are recognized, preset typical movement characteristics such as the rapid shaking of the limbs, the tense state of the muscles during spasm, and the sharp change of the body posture at the moment of falling are identified. After successful matching, the location information of the elderly and the monitored real-time vital sign information are packaged and sent to the preset management organization through the network. The preset management organization can be a community management center or an emergency center, etc.

[0058] In the above embodiments, through the all-round intelligent monitoring, the movement of the elderly is made safer and more scientific. The number of people exercising is judged, and the exercise plan is customized according to the individual situation by measuring physical parameters, heart rate, and evaluating the cardiovascular level, so as to ensure that the exercise is suitable for the individual's physical fitness; during the exercise, the vital signs and limb movements are monitored in real time. When there are abnormalities, charts and animations are used to assist in adjustment. When dangerous movements are detected, information is sent to the management organization in time, which improves the safety during the recommendation of exercise prescriptions for the elderly group and the process of the elderly exercising according to the exercise prescriptions.

[0059] In some embodiments, some special situations may be encountered during the process of recommending exercise prescriptions for the elderly population. For example, if the elderly wear relatively thick clothes or large ornaments on the day, it will interfere with the detection of basic body data, resulting in incorrect detection of basic body data and thus incorrect recommendation of exercise prescriptions. The exercise prescription recommendation system for the elderly population can exclude the interference of clothing by constructing the body contour information of the complete person and identifying the clothing and equipment features. As Figure 3 shown, it is another flowchart of the exercise prescription recommendation method for the elderly population provided by the embodiment of the present application. This method can be used in Figure 1 the system architecture shown, and includes the following steps: S301. Shoot the picture of the exercise venue. When a complete person appears in the picture of the exercise venue, judge the number of complete persons; S302. In the case where the number of complete persons is one, construct the body contour information of the complete person and identify the clothing and equipment features; When it is confirmed that there is only one elderly person in the exercise venue, use the camera to capture the images of the elderly person from multiple angles. With the help of image recognition technology and deep neural network algorithms, process the collected images, construct the body contour by identifying the boundaries and shapes of various body parts, and at the same time distinguish the clothing texture, color, style and equipment features such as glasses, hats, and protective gear worn.

[0060] S303. Use the body contour information and clothing and equipment features to calculate the actual body circumference value of the complete person; Based on the constructed body contour information, determine the boundary ranges of various body parts. For example, outline the general areas of waist circumference, hip circumference, etc. through the contour. Then, combined with the clothing and equipment features, refer to the known conventional size ratio relationship for auxiliary calibration, and use the image analysis algorithm to convert the corresponding actual length according to the pixel distance, etc., so as to obtain the actual body circumference values such as chest circumference and limb circumference.

[0061] When the elderly wear loose clothes or multiple layers of clothes, the body contour will be blurred, increasing the difficulty of accurately calculating the body dimensions. At this time, a depth camera or laser scanning technology can be used to obtain more accurate body shape information. The depth camera can obtain the depth information of the object. Even in the case of loose clothes, it can more accurately locate the body boundary by analyzing the depth difference between the human body and the clothes. For multiple layers of clothing, according to the heat conduction characteristics of clothes made of different materials, combined with infrared thermal imaging technology, identify the contour of the layer of clothing closest to the body, and calculate the body dimensions based on this.

[0062] For the case of carrying large equipment or assistive devices, if the elderly carry large sports equipment (such as carrying a large sports backpack) or assistive devices (such as wheelchairs, crutches), it will interfere with the recognition of the body contour. In this case, first use the image recognition algorithm to identify the shapes of the equipment and devices and segment them from the overall image. Then, based on the relative positions of the human body and these objects and the normal proportional relationship of the human body, infer the body contour of the occluded part. For example, given the size of the wheelchair and the position of the elderly sitting on the wheelchair, the approximate contours of the legs and hips occluded by the wheelchair can be estimated, and then the body dimensions can be calculated.

[0063] For the case of irregular body postures, the elderly may have irregular postures such as bending over or leaning sideways. This will cause the body contour to deform at certain angles, affecting the accuracy of dimension calculation. Images from multiple cameras can be obtained at different angles, and the body contour information from multiple angles can be comprehensively analyzed. At the same time, through the motion tracking algorithm, predict the positions of various parts of the body in the normal standing posture, and then calculate the body dimensions according to the proportional relationship of the normal posture. For example, when the elderly bends over, by analyzing the changes in the back curve and the information of other non-deformed parts, calculate the body dimensions in the normal upright state.

[0064] S304. Calculate the basic body parameters of the complete person through the actual body circumference values; Based on the calculated actual body circumference values, such as waist circumference, hip circumference, chest circumference and other data, combined with the existing proportional relationship of the human body structure and related mathematical models, use the algorithm to calculate the height. For example, estimate the height through a certain proportion between the hip circumference and the height. Then, referring to the body fat distribution corresponding to the circumference and combining the general conversion formula of weight and circumference, roughly calculate the weight, so as to integrate and obtain the complete basic body parameters.

[0065] S305. Capture the real-time face video of the complete person, measure the real-time heart rate and evaluate the cardiovascular condition level by analyzing the facial features in the real-time face video; S306. Develop a real-time exercise plan according to the basic body parameters, real-time heart rate and cardiovascular condition level; S307. After monitoring that the complete person in the sports venue starts to exercise, monitor the real-time vital signs information and real-time limb movements during the exercise process of the complete person; S308. When the real-time vital signs information is not within the preset vital signs information threshold range, arrange the real-time vital signs information of the past preset duration in chronological order to obtain the sorted real-time vital signs data set; When it is detected that the real-time vital sign information exceeds or is lower than the preset threshold range, the system will automatically retrieve the vital sign data collected in the past preset duration, such as body temperature, respiratory rate, etc., and then arrange them in the order of collection time. These ordered data are integrated to obtain the sorted real-time vital sign data set.

[0066] S309. Draw a real-time vital sign data chart according to the real-time vital sign data set; After the system obtains the real-time vital sign data set, it will, according to the data type. For example, body temperature is a set of values that change over time, and respiratory rate is another set of data, etc. Using a drawing software or the built-in drawing module, with time as the horizontal axis and the corresponding vital sign values as the vertical axis, mark and connect each data point. For example, a line chart for body temperature and a bar chart for respiratory rate are used to draw a chart that can show the changes in vital signs.

[0067] S310. Identify and classify the real-time limb movements based on the human body posture recognition algorithm to obtain real-time limb movement data; It can be understood that the human body posture recognition algorithm is obtained by pre-training a deep learning model with a set of limb movement data in advance.

[0068] First, collect a large number of sets of limb movement data containing different types and angles, and use them to train the deep learning model, enabling the model to learn the characteristic laws of each movement and form the human body posture recognition algorithm. Input the real-time limb movements into the human body posture recognition algorithm, analyze key contents such as the changes in movement joint points and movement trajectories, accurately identify and classify these movements, and output the corresponding real-time limb movement data.

[0069] S311. Select the corresponding guiding animation from the preset animation template library according to the real-time limb movement data and real-time vital sign information; It can be understood that the preset animation template library is a preset library that stores guiding animations corresponding to different vital sign information.

[0070] After the system obtains the real-time limb movement data and real-time vital sign information, it matches and compares these data with the different vital signs and limb movement characteristics corresponding to each animation in the preset animation template library, and selects the guiding animation that fits the current physical condition and movement performance of the elderly. For example, when the vital signs are abnormal and the movements are unsteady, select the corresponding animation that prompts to adjust breathing and stabilize movements.

[0071] S312. Modify the data in the guiding animation in combination with the real-time vital sign information to obtain the real-time guiding animation; When a guiding animation in the preset animation template library is selected, the system will, based on the real-time monitored vital sign information, such as specific values like too high body temperature and rapid breathing, make targeted adjustments to the data such as the prompt information and the rhythm of action demonstrations involved in the animation. For example, if the body temperature is on the high side, the frequency of the cooling action will be highlighted in the animation; when the breathing is too fast, the demonstration speed of guiding the breathing adjustment will be slowed down, so as to generate a real-time guiding animation that fits the current situation.

[0072] When an elderly person has multiple abnormal vital signs simultaneously, such as too high body temperature, rapid breathing, and abnormal heartbeat, it is necessary to comprehensively consider these factors to modify the animation. According to the mutual relationship between these abnormal vital signs and the degree of impact on the overall body, the animation content is adjusted. For example, in such a complex situation, the animation may first show the action of asking the elderly person to sit down or lie down to rest, and at the same time give adjustment suggestions in the order of priority of heartbeat, breathing, and body temperature, such as first guiding the adjustment of the breathing rhythm, and then prompting to take physical cooling measures, etc.

[0073] S313. Display the real-time vital sign data chart and the real-time guiding animation through the display screen; S314. Perform semantic marking and learning on a large number of different types of movement actions to build an action semantic knowledge base; Collect a large number of movement action videos and image materials covering different fields such as fitness, daily exercise, and rehabilitation training. Using image recognition and text annotation technologies, annotate semantic information such as the name, action essentials, force application parts, and applicable scenarios for each action. Then input these marked data into a preset machine learning model, allowing the model to deeply analyze the logical relationships and similar features between actions, and thus build a rich and comprehensive action semantic knowledge base.

[0074] S315. In the action semantic knowledge base, perform keyword marking on the typical semantics corresponding to the preset typical action features; Find the content related to the preset typical action features from the already constructed action semantic knowledge base, such as the action descriptions corresponding to twitching, spasm, falling, etc. Then use special marking tools or software functions to add keyword markings to the action semantics associated with these features, such as the semantics of abnormal muscle twitching and limb incoordination involved in the twitching action, so that they can be distinguished from ordinary action semantics.

[0075] S316. Search for the real-time semantics corresponding to the real-time limb actions in the action semantic knowledge base; Obtain the data of the real-time limb actions, analyze key elements such as their action forms and joint activities, and then input these feature information into the action semantic knowledge base. Through the preset classification and retrieval mechanisms in the knowledge base, compare and match them one by one with various stored action semantics, and find the set of semantic descriptions that best fits the current real-time limb actions, and find the corresponding real-time semantics.

[0076] S317. In the case where real-time semantics cannot be found, disassemble the real-time body movement frame by frame into real-time basic movements; When the corresponding real-time body movement cannot be found in the action semantics knowledge base, use image analysis technology to extract the video recording the body movement frame by frame in chronological order, analyze the positions, angles, state changes, etc. of each part of the body in each frame, disassemble it into relatively simple real-time basic movements such as raising the arm and bending the leg, and then further analyze the overall body movement situation based on the basic movements.

[0077] S318. Search for the real-time basic semantics corresponding to the real-time basic movement in the action semantics knowledge base; Step S318 is similar to step S316 above. For the description, please refer to step S316 and will not be repeated here.

[0078] S319. If the real-time semantics or the real-time basic semantics is a typical semantics, send a first prompt message.

[0079] Specifically, the first prompt message is the position information and real-time vital sign information of the complete person sent to the preset management organization.

[0080] When it is determined that the real-time semantics or the real-time basic semantics belongs to the pre-marked typical semantics, the system will immediately obtain the position information of the complete person at this moment, which can be determined by positioning technology, and at the same time capture the real-time monitored vital sign information, and then use the network communication function to package and send this information to the preset management organization.

[0081] Steps S301, S305 - S307, S313 are Figure 2 similar to steps S201, S203 - S206 in the embodiment shown. For the description, please refer to steps S201, S203 - S206 and will not be repeated here.

[0082] In the above embodiments, first, by shooting the video of the sports venue, accurately judge the number of complete persons therein. When there is only one person, use image technology to construct the body contour, identify the clothing and equipment, calculate the actual body dimensions, obtain the basic body parameters, and at the same time capture the facial video, measure the real-time heart rate and evaluate the cardiovascular condition according to the change of facial skin color, so as to formulate a real-time exercise plan. After the exercise starts, monitor the vital signs and limb movements. Once the sign data deviates from the threshold, organize the data to draw a graph, identify the limb movements, match the guiding animation and optimize the animation in combination with the real-time signs, and guide the elderly to adjust in a visual way; build a semantic knowledge base of movements, mark the semantic of special situation movements with keywords, compare the limb movements in real time, and send the location and sign information of the elderly to the management organization when detecting typical dangerous movements, effectively avoiding the situation of recommending wrong exercise prescriptions, and improving the safety in the process of recommending exercise prescriptions for the elderly group and the elderly exercising according to the exercise prescriptions.

[0083] In some embodiments, some special situations may be encountered in the process of recommending exercise prescriptions for the elderly group. For example, when there are multiple elderly people exercising in the same venue, the exercise prescription recommendation system for the elderly group can combine the actual situations of multiple elderly people, recommend exercise prescriptions on average, and monitor all the elderly people in real time during the exercise process. As Figure 4 shown, it is another flow schematic diagram of the exercise prescription recommendation method for the elderly group provided by the embodiment of the present application. This method can be used in Figure 1 the system architecture shown, and includes the following steps: S401. Shoot the video of the sports venue, and when a complete person appears in the monitored video of the sports venue, judge the number of complete persons; S402. In the case where the number of complete persons is greater than or equal to two, establish a separate identity identifier for each complete person; When it is monitored that the number of complete persons in the video of the sports venue is greater than or equal to two, use face recognition technology to extract the unique features of each person's face, such as key information such as the layout of facial features and the characteristics of facial contours, and at the same time combine features such as body shape and clothing to comprehensively distinguish different individuals, and then create a separate identity identifier for each complete person for subsequent correspondence of their relevant information.

[0084] S403. Detect the real-time position of each complete person to obtain a real-time position data group; It can be understood that the identity identifier in the real-time position data group corresponds to the real-time position.

[0085] With the help of positioning devices, such as high-precision cameras, infrared sensors, etc., lock the complete people with different identity identifiers, and accurately judge the specific locations of each complete person in the venue according to the coordinate information or relative position reference feedback by these devices. Then, match and associate the identity identifiers with the corresponding real-time positions one by one, so as to form a real-time position data group.

[0086] S404. In the real-time position data group, compare the distances between the real-time positions corresponding to different identity identifiers pairwise; First, extract the real-time position information corresponding to each different identity identifier from the real-time position data group. This information generally includes key data such as coordinates. Then, use the distance calculation algorithm, substitute the position coordinate data corresponding to every two different identity identifiers, and calculate the distance value between each pair through the mathematical formula for calculating the distance between two points.

[0087] S405. When the distance between the real-time positions is less than the preset minimum distance threshold, switch to the auxiliary perspective to monitor the complete person corresponding to the first smallest distance identity identifier and the complete person corresponding to the second smallest distance identity identifier; Compare the real-time position distances. When it is found that the distance between two complete people is less than the preset minimum distance threshold, find the identity identifier person corresponding to the first smallest distance with the smallest value and the identity identifier person corresponding to the second smallest distance from these distance data. Control the monitoring device, such as adjusting the camera angle, focal length, etc., to focus on these two people, and switch to the auxiliary perspective that can clearly observe them to carry out monitoring.

[0088] S406. After switching to the auxiliary perspective, detect the first real-time position of the complete person corresponding to the first smallest distance identity identifier and the second real-time position of the complete person corresponding to the second smallest distance identity identifier; When switching to the auxiliary perspective, rely on the positioning devices within the perspective range, such as high-precision cameras, infrared positioning sensors, etc., to capture the complete person with the first smallest distance identity identifier, accurately determine the specific coordinate and other position information where it is located as the first real-time position. Similarly, also use these devices to capture the relevant information of the complete person corresponding to the second smallest distance identity identifier, so as to determine its second real-time position, and calculate the distance between the first real-time position and the second real-time position.

[0089] S407. When the distance between the first real-time position and the second real-time position is less than the preset minimum distance threshold, emit a prompt sound through the speaker; Specifically, the prompt sound is used to prompt the complete person corresponding to the first smallest distance identity identifier and the complete person corresponding to the second smallest distance identity identifier to increase the distance.

[0090] If the calculated distance between the two is less than the preset minimum distance threshold, the instruction module connected to the speaker is triggered to retrieve the audio file specifically used to remind of maintaining distance from the pre-stored prompt sound library, and play it to the outside world through the speaker, clearly informing the people corresponding to the first minimum distance identification and the second minimum distance identification, so that they know that they need to increase the distance between each other.

[0091] By emitting targeted prompt sounds, people who are too close can receive warnings in a timely manner. This can prevent continuous approaching due to negligence or unawareness, thus avoiding accidents such as collisions and interference with movement, ensuring the safety of people in the sports venue, and also preventing large errors in real-time sports monitoring data caused by people being too close.

[0092] S408. Monitor the basic physical parameters, heart rate of each complete person and evaluate the cardiovascular condition level to obtain a basic physical parameter data group, a heart rate data group and a cardiovascular condition level data group; Obtain the basic physical parameter data, heart rate data and cardiovascular condition level data of each complete person, and mark the data with the identity identification of the corresponding complete person, establish a basic physical parameter data group, a heart rate data group and a cardiovascular condition level data group, and store the corresponding same-type data into the corresponding data groups.

[0093] S409. According to the basic physical parameter data group, the heart rate data group and the cardiovascular condition level data group, formulate a multi-person exercise plan; Analyze the collected basic physical parameter data group to understand the basic physical conditions of each person, such as endurance and strength foundation. Then refer to the heart rate data group to know the cardiorespiratory function response during exercise, and then combine the cardiovascular condition level data group to judge the body's tolerance. Integrate this information, comprehensively judge all the data, and formulate a multi-person exercise plan applicable to everyone.

[0094] Starting from the basic physical parameter data group, calculate the physical fitness baseline in the group. For example, if the endurance indicators shown by the basic physical parameter data of most people are average, high-intensity and long-duration exercises are excluded; then, based on the heart rate data group, determine the overall appropriate exercise intensity to ensure that the heart rate of most people does not soar too fast and exceed the safe range during exercise. Combining the cardiovascular condition level data group, if there are a few members with relatively fragile cardiovascular systems, design exercises such as aerobic exercises with a gentle rhythm and easy team relay walks with their tolerable intensity as the upper limit. By comprehensively weighing these three groups of data and taking into account individual differences, formulate a multi-person exercise plan suitable for all members.

[0095] S410. When the number of complete people starting to exercise in the sports venue screen is greater than or equal to one, monitor the vital sign information and limb movements of each complete person during the exercise process to obtain a vital sign information data group and a limb movement group; S411. When there is abnormal vital sign information in the vital sign information data group, detect the identity identifier of the complete person corresponding to the abnormal vital sign information, mark it as an abnormal identity identifier, and display a second prompt message through the display screen; Specifically, the second prompt message is to prompt the complete person corresponding to the abnormal identity identifier that the current vital sign information is abnormal.

[0096] When abnormal vital sign information appears in the vital sign information data group, the system will lock the identity identifier of the person corresponding to the abnormal information based on the identity identifiers previously matched with the vital sign information of each complete person, and determine it as an abnormal identity identifier. Then, retrieve the preset prompt template, fill in the relevant abnormal situations, and present them through the display screen to prompt the person with the abnormal situation.

[0097] S412. Search for the semantics corresponding to each limb movement in the limb movement group in the action semantics knowledge base to obtain a limb movement semantics group; S413. When a limb movement that cannot be searched is detected, disassemble the limb movement that cannot be searched frame by frame to obtain a disassembled limb movement group; S414. Search for the disassembled semantics group corresponding to the disassembled limb movement group in the action semantics knowledge base; S415. If there is a typical semantics in the limb movement semantics group or the disassembled semantics group, detect the identity identifier of the complete person corresponding to the typical semantics, mark it as an abnormal action identity identifier, and send a third prompt message.

[0098] Specifically, the third prompt message is the location information of the complete person corresponding to the abnormal action identity identifier and the vital sign information of the complete person corresponding to the abnormal action identity identifier sent to the preset management organization.

[0099] Perform semantic analysis on the collected limb movement group, or disassemble and analyze the movement when the semantics cannot be directly matched. Once typical semantics such as falling or convulsing appear, quickly locate the identity of the complete person who made the movement based on the correspondence between the movement and the identity identifier, and mark it as an abnormal action identity identifier. Subsequently, immediately retrieve the current location information and vital sign information of this person and send them to the preset management organization.

[0100] Steps S401, S410 are Figure 2 similar to steps S201, S205 in the embodiments shown; steps S412 - S414 are Figure 3 similar to steps S316 - S318 in the embodiments shown. For the descriptions, reference can be made to the descriptions in steps S201, S205, S316 - S318, and details are not elaborated here.

[0101] In the above embodiments, the number of complete persons is determined by capturing images of the sports venue. When there are many people, an identity identifier is established for each person, and their real-time positions are monitored and the distances are compared. If the distance is too close, the perspective is switched to give a reminder, so as to maintain the sports order and avoid the risk of collision. Then, various physical indicators are monitored to formulate a sports plan suitable for multiple people, ensuring that the sports are scientific and reasonable. During the sports process, the vital signs and limb movements are monitored. If the signs are abnormal, the person himself / herself is reminded. When dangerous limb movements occur, the position and signs of the corresponding person can be quickly sent to a preset management organization for convenient and timely rescue, maximizing the health and safety of the sports participants and improving the safety during the process of recommending sports prescriptions for the elderly group in the case of multiple people.

[0102] The following introduces the exemplary sports prescription recommendation system 500 for the elderly group provided by the embodiments of the present application. Figure 5 It is an exemplary hardware structure diagram of the sports prescription recommendation system 500 for the elderly group provided by the embodiments of the present application.

[0103] In some embodiments, the sports prescription recommendation system 500 for the elderly group includes a computer device. The computer device includes a processor, a memory, and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data. The network interface of the computer device is used to communicate with other external terminals or servers through a network connection. In some embodiments, the network interface can be a wired network interface, and in some embodiments, the network interface can also be a wireless network interface. When the computer program is executed by the processor, it realizes the method in the embodiments of the present application.

[0104] Those skilled in the art can understand that Figure 5 the structure shown in is only a block diagram of some structures related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0105] As mentioned above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present application.

[0106] As used in the foregoing embodiments, depending on the context, the term "when" may be construed to mean "if" or "after" or "in response to determining" or "in response to detecting". Similarly, depending on the context, the phrase "when determining" or "if detecting (the stated condition or event)" may be construed to mean "if determining" or "in response to determining" or "when detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)".

[0107] In the foregoing embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented by software, it may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wire (such as coaxial cable, optical fiber, digital subscriber line) or wirelessly (such as infrared, wireless, microwave, etc.). The computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more integrated available media. The available medium may be a magnetic medium (such as a floppy disk, hard disk, magnetic tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid-state drive), etc.

[0108] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the foregoing embodiments can be completed by hardware instructed by a computer program. The program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the foregoing method embodiments. The foregoing storage medium includes: various media that can store program codes such as ROM or random access memory RAM, magnetic disks, or optical discs.

Claims

1. A method for recommending exercise prescriptions for the elderly, characterized in that: include: Shooting a sports scene, and when detecting complete persons appearing in the sports scene, determining the number of complete persons; When the number of complete characters is one, detecting basic body parameters of the complete character; the basic body parameters include height data, weight data and body circumference data; Capturing a real-time facial video of a complete person, measuring the real-time heart rate and assessing the cardiovascular condition level by analyzing the facial features in the real-time facial video; the facial features are changes in facial skin color in the real-time facial video; Formulate a real-time exercise plan according to the basic body parameters, the real-time heart rate and the cardiovascular condition level; the real-time exercise plan includes exercise type and exercise duration; When the complete person in the sports venue screen starts to exercise, real-time vital sign information and real-time body movements of the complete person during the exercise are monitored; the real-time vital sign information includes real-time body temperature changes, real-time breathing frequency and real-time breathing depth; When the real-time vital sign information is not within a preset vital sign information threshold range, a real-time vital sign data chart is calculated, a real-time guiding animation is generated according to the real-time vital sign information and the real-time limb movement, and the real-time vital sign data chart and the guiding animation are displayed on a display screen; When it is detected that the real-time body movement has a preset typical movement feature, sending a first prompt message; The preset typical actions are action features corresponding to preset twitching, spasm and falling actions; the first prompt information is the location information and the real-time vital sign information of the complete person sent to the preset management organization.

2. The method according to claim 1, wherein when the number of complete characters is one, detecting basic body parameters of the complete characters comprises: When the number of complete characters is one, construct the body outline information of the complete character and identify the characteristics of clothing and equipment; Calculating the actual body circumference of the complete person using the body contour information and the clothing and equipment features; The basic body parameters of the complete person are calculated using the actual body circumference values.

3. The method according to claim 1, wherein when the real-time vital sign information is not within a preset vital sign information threshold range, a real-time vital sign data chart is calculated, a real-time guiding animation is generated according to the real-time vital sign information and the real-time limb movement, and the real-time vital sign data chart and the guiding animation are displayed on a display screen, specifically comprising: When the real-time vital sign information is not within a preset vital sign information threshold range, the real-time vital sign information of the preset time period in the past is arranged in chronological order to obtain a sorted real-time vital sign data set; Drawing a real-time vital sign data chart according to the real-time vital sign data set; Identify and classify the real-time body movements based on a human posture recognition algorithm to obtain real-time body movement data; The human body posture recognition algorithm is obtained by training the deep learning model in advance using the limb movement data set; Selecting a corresponding guiding animation from a preset animation template library according to the real-time limb motion data and the real-time vital sign information; The preset animation template library is a preset library storing guiding animations corresponding to different vital signs information; In combination with the real-time vital sign information, the data in the guiding animation is modified to obtain a real-time guiding animation; The real-time vital sign data chart and the real-time guiding animation are displayed on a display screen.

4. The method according to claim 1, wherein when the real-time body movement is detected to have a preset typical movement feature, sending the first prompt information specifically comprises: Semantically label and learn a large number of different types of motion actions and build a motion semantic knowledge base; In the action semantics knowledge base, keyword tagging is performed on typical semantics corresponding to preset typical action features; Searching the action semantics knowledge base for the real-time semantics corresponding to the real-time body action; In the case where the real-time semantics cannot be found, the real-time body movements are disassembled frame by frame into real-time basic movements; Searching the action semantics knowledge base for the real-time basic semantics corresponding to the real-time basic action; If the real-time semantics or the real-time basic semantics is the typical semantics, a first prompt message is sent.

5. The method according to claim 1, after sending the first prompt information when the real-time body movement is detected to have a preset typical movement feature, further comprises: When the number of complete characters is greater than or equal to two, a separate identity is established for each complete character; The identity mark is used to mark and distinguish different complete persons; Monitor the basic physical parameters and heart rate of each complete person and evaluate the cardiovascular condition level to obtain a basic physical parameter data group, a heart rate data group and a cardiovascular condition level data group; Formulate a multi-person exercise plan based on the basic body parameter data group, the heart rate data group and the cardiovascular condition level data group; When it is detected that the number of complete characters starting to exercise in the sports venue screen is greater than or equal to one, the vital sign information and body movements of each complete character during the exercise process are monitored to obtain a vital sign information data group and a body movement data group; When abnormal vital signs information exists in the vital signs information data group, the identity identifier of the complete person corresponding to the abnormal vital signs information is detected, marked as an abnormal identity identifier, and a second prompt information is displayed on the display screen; the abnormal vital signs information is the vital signs information in the vital signs information data group that is not within the preset vital signs information threshold range; the second prompt information is to prompt the complete person corresponding to the abnormal identity identifier that the current vital signs information is abnormal.

6. The method according to claim 5, after establishing a separate identity for each complete character when the number of complete characters is greater than or equal to two, further comprising: Detecting the real-time position of each complete person to obtain a real-time position data group; The identity identifier in the real-time location data group corresponds to the real-time location; In the real-time location data group, the distances between the real-time locations corresponding to different identity tags are compared in pairs; When the distance between the real-time positions is less than a preset minimum distance threshold, the complete person corresponding to the first short-distance identity identifier and the complete person corresponding to the second short-distance identity identifier are switched to the auxiliary perspective for monitoring; After switching to the auxiliary viewing angle, detecting a first real-time position of the complete person corresponding to the first short-distance identity identifier and a second real-time position of the complete person corresponding to the second short-distance identity identifier; When the distance between the first real-time position and the second real-time position is less than a preset minimum distance threshold, a prompt sound is emitted through the speaker; the prompt sound is used to prompt the complete person corresponding to the first short-distance identity identifier and the complete person corresponding to the second short-distance identity identifier to increase the distance.

7. The method according to claim 5, when there is abnormal vital sign information in the vital sign information data group, detecting the identity of the complete person corresponding to the abnormal vital sign information, marking it as an abnormal identity, and displaying the second prompt information on the display screen, further comprising: Searching the semantics corresponding to each limb action in the limb action group in the action semantics knowledge base to obtain a limb action semantic group; When it is detected that the limb movement cannot be found, the limb movement that cannot be found is disassembled frame by frame to obtain a disassembled limb movement group; the limb movement that cannot be found is a limb movement in the limb movement group that cannot be found in the action semantic knowledge base; Searching the disassembly semantic group corresponding to the disassembly body action group in the action semantic knowledge base; If the typical semantics exists in the body movement semantic group or the disassembled semantic group, the identity of the complete person corresponding to the typical semantics is detected, marked as an abnormal movement identity, and a third prompt message is sent; the third prompt message is sent to a preset management organization, including the location information of the complete person corresponding to the abnormal movement identity and the vital signs information of the complete person corresponding to the abnormal movement identity.

8. An exercise prescription recommendation system for the elderly, characterized in that: The elderly group motion control system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the elderly group motion control system to execute the method described in any one of claims 1-7.

9. A computer program product comprising instructions, characterized in that When the computer program product runs on the elderly group motion control system, the elderly group motion control system is enabled to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium comprising instructions, characterized in that: When the instruction is executed on the elderly group motion control system, the elderly group motion control system executes the method as described in any one of claims 1-7.

Images