Pedestrian tracking system based on behavior analysis

Abstract

The invention discloses a pedestrian tracking system based on behavior analysis in the technical field of video monitoring, and the system comprises an image collection module which is used for collecting crowd image information and environment information; the cross-lens tracking module is used for realizing character seamless identification and tracking; the behavior analysis module is used for analyzing to obtain character action data, human body attribute data and people flow tracking data; the judgment module is used for judging the physical state of the person and predicting the injury cause of the person according to the person action data and the human body attribute data, planning a smooth action route in the hospital according to the people flow tracking data and predicting the people flow change; the report module is used for giving an alarm to a hospital first-aid room according to the physical state of the person and prompting the fastest arrival route to a first-aid doctor; and the feedback module is used for optimizing the physical state of the person and judging the injury cause of the person. The system is intelligent and efficient, the survival rate of patients is increased in limited time, meanwhile, the hospital management efficiency is improved, and the safety and service quality of hospitals are improved.

Description

Technical Field

[0001] The invention belongs to the technical field of video monitoring, and in particular is a pedestrian tracking system based on behavior analysis. Background Art

[0002] With the continuous advancement of medical technology and the growing demand for medical care, hospitals are facing increasing management pressure. Traditional hospital management methods often rely on manual recording and monitoring, which is not only inefficient but also prone to errors. In order to improve the management efficiency of hospitals, more intelligent and automated management methods are needed. In addition, as a place to save lives and heal the wounded, the safety of patients is of vital importance. However, due to the complex hospital environment and frequent personnel turnover, patients may encounter safety problems such as getting lost and falling during the treatment process. Since the injured are patients themselves, they may need timely treatment and care after falling, which places extremely high demands on the response and speed of emergency personnel.

[0003] Although the existing monitoring system has a relatively wide coverage, it does not have the function of automatically alerting emergency doctors when the injured are found, and it cannot count the changes in the density of people in the hospital in real time and plan emergency routes. Therefore, it is necessary to propose a pedestrian tracking system based on behavior analysis that can identify and analyze the patient's movements and human attributes, determine whether emergency treatment is needed and alert the emergency doctor, and analyze the density and change trend of people flow, provide emergency doctors and different patients with route planning in the hospital, and improve the flow efficiency of the hospital. Summary of the invention

[0004] In order to solve the above problems, the purpose of the present invention is to provide a pedestrian tracking system based on behavioral analysis, which determines whether emergency treatment is needed and reports to the emergency doctor by analyzing the patient's movements and body attributes. It has the function of planning the fastest arrival route for the emergency doctor, thereby increasing the patient's survival rate in a limited time, while improving hospital management efficiency, safety and service quality of the hospital.

[0005] In order to achieve the above-mentioned object, the technical solution of the present invention is as follows: a pedestrian tracking system based on behavior analysis, comprising an image acquisition module, a cross-lens tracking module, a behavior analysis module, a judgment module, a reporting module and a feedback module;

[0006] Image acquisition module, used to collect crowd image information and environmental information within the hospital area;

[0007] Cross-lens tracking module, which is used to achieve seamless recognition and tracking of people between different cameras based on cross-lens tracking technology;

[0008] A behavior analysis module, used to analyze the crowd image information and environmental information collected by the image acquisition module to obtain character action data, human attribute data and crowd flow tracking data;

[0009] The judgment module is used to judge the physical condition of the person and predict the cause of the person's injury or illness based on the person's motion data and body attribute data, and to plan the smooth movement route in the hospital and predict the change of the flow of people based on the flow of people tracking data;

[0010] The reporting module is used to alert the hospital emergency room based on the person's physical condition and possible injuries, and to inform the emergency doctor of the fastest arrival route based on the unobstructed movement routes in the hospital for the purpose of priority rescue. It also pushes the flow changes of people to various departments to hospital visitors based on the current unobstructed movement routes in the hospital and the prediction of the flow changes;

[0011] The feedback module is used to feed back the medical information of emergency patients to the judgment module, and to refine the character's physical condition and the cause of the character's injury and illness by combining the character's motion data and human attribute data.

[0012] The principle of the basic solution is: collect crowd image information through the image acquisition module and then use cross-lens tracking technology to seamlessly identify and track people; the behavior analysis module analyzes the movements and body attributes of people, as well as crowd tracking data; the judgment module determines whether emergency treatment is needed; and the reporting module reports the rescue of patients to the emergency doctor.

[0013] The beneficial effects of the basic solution are: 1. The image acquisition module and cross-lens tracking module can capture the dynamics of the crowd in the hospital in real time. Once a possible emergency situation is detected, such as a person falling down, suddenly falling to the ground, and other abnormal behaviors, the system can immediately trigger the alarm mechanism and quickly notify the emergency room through the reporting module. The judgment module plans the smooth movement route in the hospital based on the crowd tracking data, and predicts the changes in crowd flow, providing emergency doctors with the fastest route to reach the patient, significantly shortening the emergency response time.

[0014] 2. The system can predict the trend of changes in hospital traffic by real-time monitoring and analysis of traffic data, and provide decision-making support for hospital managers, such as adjusting registration windows, adding guidance services, etc., to alleviate congestion during peak hours. According to the traffic conditions of each department and patient needs, the system can assist hospital managers in dynamically adjusting the allocation of medical resources, such as increasing doctor shifts, deploying medical equipment, etc., to ensure the effective use of medical resources.

[0015] 3. The reporting module can push the best route suggestions to various departments for visitors based on the current unobstructed routes and traffic changes in the hospital, reducing the waiting time and inconvenience of patients during medical treatment. The behavior analysis module can not only identify abnormal behaviors, but also analyze the patient's physical condition and predict the cause of injury to a certain extent.

[0016] 4. The system can monitor abnormal situations in the hospital in real time, such as gathering of people, abnormal behavior, etc., and issue early warnings in time to help hospitals strengthen safety management and prevent safety accidents. With the data support of the behavior analysis module and the judgment module, hospital managers can more accurately grasp the safety status of the hospital and provide a scientific basis for formulating and adjusting safety management measures.

[0017] 5. The system provides data-driven decision support for hospital managers by collecting and analyzing a large amount of data, and promotes hospital management to develop in a more intelligent and refined direction. As an important part of the development of medical intelligence, the system not only improves the hospital's operating efficiency and service quality, but also provides valuable practical experience and data support for medical technology innovation.

[0018] Furthermore, the image acquisition module includes a number of surveillance cameras and a number of environmental sensors;

[0019] Several surveillance cameras include several indoor cameras installed in the outpatient hall of the hospital and several outdoor cameras installed in the hospital area, which are used to collect image information of people in the entire hospital and in secluded corners;

[0020] Several environmental sensors including infrared cameras installed along with surveillance cameras, temperature, humidity and light sensors are used to collect environmental information at different locations within the hospital.

[0021] The beneficial effects of the basic solution are: 1. By installing indoor and outdoor cameras in the outpatient hall and hospital area, all-round monitoring of the entire hospital area is achieved, ensuring safety without blind spots. Indoor cameras can clearly capture crowd image information in key areas such as the outpatient hall, while outdoor cameras can monitor every corner of the hospital area, including secluded areas, improving the accuracy and meticulousness of monitoring.

[0022] 2. The addition of environmental sensors such as infrared cameras, temperature, humidity and light sensors enables the system to simultaneously collect environmental information from different locations in the hospital, including temperature, humidity, light intensity and human infrared radiation, providing more comprehensive data support for monitoring the hospital environment. These environmental data can be used to intelligently adjust the environmental conditions in the hospital, such as automatically adjusting the temperature and humidity of the air conditioner, turning on or off the lighting equipment, etc., to create a more comfortable and safe treatment environment.

[0023] 3. Combining data from surveillance cameras and environmental sensors, the system can more accurately identify abnormal behaviors, such as people suddenly falling to the ground, abnormally high ambient temperature, etc., and issue early warnings in a timely manner, providing strong support for the hospital's safety management.

[0024] 4. By real-time monitoring of environmental conditions within the hospital, the system can promptly detect and address potential risks that may lead to infection, such as bacterial growth caused by excessive humidity, thereby reducing the risk of hospital-acquired infection.

[0025] Further, the cross-shot tracking module includes an image preprocessing unit, a person recognition unit, and a cross-shot matching unit;

[0026] The image preprocessing unit preprocesses crowd image information, including denoising, enhancement and correction, and spatially aligns images collected by different cameras through image registration technology;

[0027] A person recognition unit is used to extract and recognize the features of people in the crowd image information to establish a unique identification of the target;

[0028] The cross-shot matching unit is used to establish associations between different cameras based on the feature identifiers of the person, so as to achieve cross-shot matching of the target.

[0029] The beneficial effects of the basic scheme are as follows: 1. The image preprocessing unit denoises, enhances and corrects the collected crowd image information, effectively improving the quality of the surveillance video. By removing noise, the interference factors in the video are reduced; by image enhancement, the clarity and contrast of the image are improved, making the characteristics of the people more obvious; by image correction and registration technology, the spatial alignment of images collected by different cameras is achieved, providing a basis for subsequent cross-lens tracking.

[0030] 2. The person recognition unit uses advanced algorithms to extract and identify the features of people in the crowd image and establish a unique identification of the target. These features can be facial features, body contours, gait features, etc. Through the feature matching algorithm, the system can accurately identify the same person in different cameras, improving the accuracy of person recognition.

[0031] 3. The cross-lens matching unit establishes associations between different cameras based on the person’s feature identifiers, thus achieving cross-lens matching of the target. This function enables the system to seamlessly track the target person between different cameras, and no matter how the person moves in the hospital, the system can continue to track and record his or her movement trajectory.

[0032] Further, the behavior analysis module includes a skeleton point analysis unit, a character attribute analysis unit and a crowd flow analysis unit;

[0033] A skeleton point analysis unit, used to detect the time-series change motion data of a single human body and a single human body key point based on skeleton key point analysis technology;

[0034] A person attribute analysis unit, used to detect the person's body surface attribute data and identify the person's action data based on the person ID classification detection technology;

[0035] The crowd analysis unit is used to count and track crowds based on deduplication crowd statistics technology.

[0036] The beneficial effects of the basic solution are: 1. The skeleton point analysis unit can accurately detect the skeleton key points of a single human body based on the skeleton key point analysis technology, and further analyze the time series changes of these key points, so as to accurately identify the human body's motion data. This technology is of great significance for analyzing the patient's movement status, gait characteristics, etc., and helps to assist doctors in disease diagnosis and treatment. By real-time monitoring and analysis of the human body's skeleton point data, the system can promptly detect and warn abnormal behaviors, such as falls, sudden acceleration, etc., providing strong support for the hospital's safety management.

[0037] 2. The person attribute analysis unit can detect and identify human surface attribute data and person action data, such as height, body shape, clothing, expression, etc., based on person ID classification detection technology. These data help hospitals to identify and manage patients in a personalized manner and improve service quality and efficiency. By analyzing the patient's body surface attributes and action data, the system can provide doctors with auxiliary diagnosis information, such as whether the patient's gait is abnormal, whether the expression is painful, etc., thereby helping doctors to more accurately judge the patient's condition and provide strong data support for the judgment of the subsequent judgment module.

[0038] 3. The crowd flow analysis unit uses deduplication crowd flow statistics technology to accurately count the number of people in the hospital and track the trajectory of the people. These data help hospital managers understand the flow of people in the hospital, optimize the allocation of medical resources, and improve service efficiency.

[0039] Further, the judgment module includes a character situation judgment unit and a route planning unit;

[0040] A character situation judgment unit is used to judge whether the character needs emergency treatment and possible injury or illness symptoms based on the character's motion data and body attribute data;

[0041] The route planning unit is used to plan smooth movement routes within the hospital based on the crowd tracking data combined with the preset hospital route map.

[0042] The beneficial effects of the basic scheme are: 1. The person situation judgment unit can quickly judge whether the person needs emergency treatment based on the person's motion data and human attribute data. For example, by analyzing the person's gait, expression, heart rate and other parameters, the system can promptly detect possible abnormal conditions of the patient, such as falling, coma, abnormal heart rate, etc., thereby triggering the emergency response mechanism in time to buy precious treatment time for the patient.

[0043] 2. The unit can also further analyze the possible symptoms of the patient's injuries and illnesses. By comparing historical data and the medical knowledge base, the system can make a preliminary classification and judgment of the patient's injuries and illnesses, and provide doctors with auxiliary preliminary diagnostic information, which helps doctors understand the patient's condition more accurately and develop more effective treatment plans.

[0044] 3. The route planning unit can plan the smooth movement routes in the hospital in real time based on the crowd tracking data and the preset hospital route map. By dynamically adjusting the route, the system can effectively avoid crowd congestion and patient retention, improve the traffic efficiency within the hospital, and provide patients with a better medical experience.

[0045] 4. The route planning unit can also provide personalized navigation services for patients. By combining the patient's location information and destination, the system can plan the best route for them and update the road conditions in real time to help patients quickly find the target department or facility.

[0046] Further, the reporting module includes an alarm unit, a push unit and a terminal application;

[0047] The alarm unit is used to alert the emergency doctor based on the first aid judgment result and possible injury and illness symptoms, and plan the fastest arrival route for the emergency doctor based on the unobstructed movement route in the hospital and give him prompts;

[0048] The push unit is used to push the real-time flow changes on the routes to various departments to hospital visitors based on the unobstructed movement routes and the prediction of flow changes in the hospital;

[0049] Terminal applications are used to prompt and promote smooth action routes and real-time changes in passenger flow along the routes, and to display spatial passenger flow density distribution and temporal passenger flow change trends.

[0050] The beneficial effects of the basic solution are: 1. The alarm unit can immediately send an alarm message to the emergency doctor based on the results of the person's first aid judgment and possible symptoms of injuries and illnesses. This ensures that the emergency doctor can receive notification of the emergency situation at the first time and take prompt action. The unit can also plan the fastest arrival route for the emergency doctor based on the unobstructed route of action in the hospital and provide real-time prompts. This helps the emergency doctor quickly find the best path in an emergency, thereby shortening the response time and improving the efficiency of emergency treatment. Through instant alarms and optimized route planning, the alarm unit can significantly improve the level of medical safety. In an emergency, it can ensure that patients receive timely and effective treatment, reducing medical risks and adverse consequences.

[0051] 2. The push unit can push real-time crowd change information on the routes to various departments to hospital visitors based on the unobstructed routes and crowd change predictions within the hospital. This helps visitors choose the best route to avoid crowd congestion and waiting time. By providing personalized route planning and real-time crowd change information, the push unit can significantly improve the medical experience of hospital visitors. Visitors can find the target department more conveniently, reducing the time and inconvenience of staying in the hospital. The push of real-time crowd change information helps hospital managers better understand the resource utilization within the hospital. By monitoring and analyzing crowd flow data, hospitals can optimize department layout, adjust medical staff configuration, etc., thereby improving resource utilization efficiency and service quality.

[0052] 3. The terminal application can intuitively display the spatial crowd density distribution and the temporal crowd change trend. This helps hospital managers and visitors to understand the crowd flow situation inside the hospital more clearly, so as to make more informed decisions and choices. Through the terminal application, users can check the smooth movement routes and real-time crowd change information inside the hospital anytime and anywhere. This provides great convenience for users and reduces the inconvenience of finding departments and waiting time inside the hospital. The terminal application also provides user interaction functions, such as navigation guidance, route planning suggestions, etc. This helps to enhance the interaction and communication between users and hospitals and improve user satisfaction.

[0053] Furthermore, the character attribute analysis unit includes human surface color block analysis, facial emotion analysis and character action recognition.

[0054] Human body surface color block analysis, used to analyze the color of clothing, stains and blood on the human body surface;

[0055] Facial emotion analysis, used to analyze emotional tendencies based on facial expressions;

[0056] Human action recognition, used to identify human actions.

[0057] The beneficial effects of the basic solution are: 1. The human body surface color block analysis can accurately identify and track the color of a person's clothing, which is particularly important in a surveillance environment. By identifying the color of clothing, the system can more easily lock on the target person and achieve accurate tracking and positioning. This function can also detect stains and blood on the human body surface, which is of great significance for medical emergency, criminal investigation and other scenarios.

[0058] 2. Human action recognition can accurately identify and classify human actions, such as walking, running, covering wounds, etc. This helps the system analyze the behavior patterns of people, so as to determine whether they have injuries and the general type of injuries. By monitoring human actions in real time, the system can promptly detect and warn of abnormal behaviors, such as falls and fights. This helps reduce safety risks and ensure personnel safety.

[0059] Furthermore, the crowd flow analysis unit includes multi-dimensional crowd flow analysis, crowd flow attribute analysis and crowd flow trajectory tracking.

[0060] Multi-dimensional crowd flow analysis is used to analyze the crowd density distribution in the spatial dimension, record historical crowd flow data, and count the crowd flow change trend in the time dimension based on the preset three-dimensional map of the hospital;

[0061] Flow attribute analysis is used to combine with the character attribute analysis unit to analyze the gender, age, and facial expression of the crowd, and provide the hospital with a refined flow portrait;

[0062] Crowd flow tracking is used to track the flow of people and predict the real-time trend of crowd density changes.

[0063] The beneficial effects of the basic solution are: 1. Multi-dimensional crowd flow analysis can display the crowd density distribution in the spatial dimension in real time according to the preset three-dimensional map of the hospital. This helps hospital managers to intuitively understand the crowd flow conditions in each area, promptly discover congested areas, and take corresponding diversion measures. This function can also record and analyze historical crowd flow data, including peak and trough periods. By comparing crowd flow data in different time periods, hospitals can optimize resource allocation, such as adjusting the medical staff schedule, adding service windows, etc., to improve service efficiency. Multi-dimensional crowd flow analysis can also count the crowd flow change trends in the time dimension, such as daily crowd flow, weekly crowd flow, monthly crowd flow, etc. Through trend prediction, hospitals can make personnel deployment, material reserves and other work in advance to cope with possible peaks in crowd flow.

[0064] 2. The flow of people attribute analysis can be combined with the character attribute analysis unit to conduct a detailed analysis of the gender, age, facial expression and other attributes of the crowd. This helps the hospital build a refined flow of people portrait, understand the needs and preferences of different groups of people, and provide a basis for formulating personalized service strategies. By understanding the needs and preferences of different groups of people, the hospital can provide targeted services, such as setting up children's play areas, elderly rest areas, etc., to improve service quality and patient satisfaction. Refined flow of people portraits can also provide support for the hospital's market research and decision-making. For example, based on the results of the flow of people attribute analysis, the hospital can adjust the department layout, optimize the service process, etc. to improve market competitiveness.

[0065] 3. Crowd flow tracking can track the flow of people in real time and predict the changing trend of crowd density. This helps hospitals to promptly identify potential congested areas and take corresponding diversion measures to ensure smooth traffic within the hospital. By analyzing the flow of people, hospitals can understand patients' medical habits and preferences, thereby optimizing the medical process. For example, hospitals can adjust the layout of departments, add guidance services, etc. to shorten patients' waiting time and medical process. In emergency situations, such as fires and earthquakes, crowd flow tracking can quickly locate personnel and provide key information for emergency response. This helps hospitals to evacuate personnel in a timely manner and reduce casualties.

[0066] Furthermore, the character situation judgment unit includes first aid judgment and injury and illness refinement.

[0067] First aid judgment, which uses deep learning algorithms to train models to determine whether a person has symptoms of injury or illness and whether first aid is needed;

[0068] Injury refinement is used to determine the possible types of injuries a character may have based on the character's motion data and body attribute data.

[0069] The beneficial effects of the basic solution are: 1. Using deep learning algorithms, the first aid judgment module can quickly analyze the status of the person, determine whether there are symptoms of injury or illness, and whether immediate first aid is needed. This rapid response capability is crucial to saving lives, especially in emergency situations. By training the model, the first aid judgment module can learn and identify a variety of symptoms of injury and illness, thereby improving the accuracy of judgment. This helps to reduce misdiagnosis and missed diagnosis, and ensure that patients receive timely and correct treatment. The first aid judgment module can reasonably allocate medical resources based on the judgment results. For patients who need first aid, the system can prioritize rescue personnel and medical equipment to ensure that patients receive timely treatment. At the same time, for patients who do not need first aid, the system can guide them to perform other appropriate treatments to avoid wasting medical resources.

[0070] 2. The injury refinement module can determine the possible types of injuries and illnesses of a person based on the person's motion data and body attribute data. This precise diagnosis capability helps doctors develop personalized first aid plans and improve treatment outcomes.

[0071] Furthermore, terminal applications include personal terminals and hospital terminals.

[0072] Personal terminals are used to push smooth routes or the fastest routes to reach;

[0073] Hospital terminals are used to display multi-dimensional crowd flow analysis data.

[0074] The beneficial effects of the basic solution are: 1. The personal terminal can intelligently push the smooth action route or the fastest arrival route based on the user's real-time location and destination, combined with the background crowd analysis data. This saves time and energy for users, especially in crowded areas, which can effectively avoid congestion and improve admission efficiency. By updating route information in real time, the personal terminal can ensure that users always have the latest road conditions and crowd conditions. This dynamic information service not only improves the user's travel experience, but also enhances the user's trust and satisfaction with the system.

[0075] 2. Hospital terminals can display multi-dimensional crowd flow analysis data in real time, including crowd density, crowd speed, crowd direction, etc. These data are presented in visual forms such as charts and maps, so that hospital managers can intuitively understand the crowd flow situation inside the hospital and provide support for decision-making. By analyzing crowd flow data, hospital managers can understand the distribution of crowd flow in various departments and regions, so as to reasonably allocate medical resources, such as adding service windows and adjusting the shifts of medical staff. This helps to improve the service efficiency of the hospital, reduce the waiting time of patients, and improve patient satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 Schematic diagram of a pedestrian tracking system based on behavior analysis in an embodiment of the present invention.

[0077] Figure 2 Schematic diagram of an image acquisition module of a pedestrian tracking system based on behavior analysis in an embodiment of the present invention.

[0078] Figure 3 Schematic diagram of a cross-shot tracking module of a pedestrian tracking system based on behavior analysis in an embodiment of the present invention.

[0079] Figure 4 Schematic diagram of a behavior analysis module and a judgment module of a pedestrian tracking system based on behavior analysis in an embodiment of the present invention.

[0080] Figure 5 Schematic diagram of a reporting module of a pedestrian tracking system based on behavior analysis in an embodiment of the present invention.

[0081] Figure 6 Schematic diagram of a person attribute analysis unit of a pedestrian tracking system based on behavior analysis in an embodiment of the present invention

[0082] Figure 7 Schematic diagram of a pedestrian flow analysis unit of a pedestrian tracking system based on behavior analysis in an embodiment of the present invention

[0083] Figure 8 Schematic diagram of a person situation determination unit of a pedestrian tracking system based on behavior analysis in an embodiment of the present invention DETAILED DESCRIPTION

[0084] The following is further described in detail through specific implementation methods:

[0085] Example 1

[0086] Basically as attached Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 , Figure 6 , Figure 7 and Figure 8 As shown: A pedestrian tracking system based on behavior analysis, characterized by: including an image acquisition module, a cross-lens tracking module, a behavior analysis module, a judgment module, a reporting module and a feedback module.

[0087] The image acquisition module is used to collect image information of the crowd and environmental information within the hospital campus; the image acquisition module includes a number of surveillance cameras and a number of environmental sensors; the several surveillance cameras include a number of indoor cameras installed in the outpatient hall inside the hospital and a number of outdoor cameras installed in the hospital campus, which are used to collect image information of the crowd in the entire hospital and image information of the crowd in secluded corners; the several environmental sensors include infrared cameras, temperature, humidity and light sensors installed together with the surveillance cameras, which are used to collect environmental information at different locations in the hospital.

[0088] The cross-lens tracking module is used to realize seamless recognition and tracking of people between different cameras based on the cross-lens tracking technology; the cross-lens tracking module includes an image preprocessing unit, a person recognition unit and a cross-lens matching unit; the image preprocessing unit preprocesses the crowd image information, including denoising, enhancement and correction, and spatially aligns the images collected by different cameras through image registration technology; the person recognition unit is used to extract and recognize the features of people in the crowd image information to establish a unique identification of the target; the cross-lens matching unit is used to establish associations between different cameras based on the feature identification of the person to realize cross-lens matching of the target.

[0089] The behavior analysis module is used to analyze the crowd image information and environmental information collected by the image acquisition module to obtain character motion data, human attribute data and crowd tracking data; the behavior analysis module includes a skeleton point analysis unit, a character attribute analysis unit and a crowd analysis unit; the skeleton point analysis unit is used to detect the time-series change motion data of a single human body and a single human key point based on the skeleton key point analysis technology; the character attribute analysis unit is used to detect the character body surface attribute data and identify the character motion data based on the character ID classification detection technology. The character attribute analysis unit includes the body surface color block analysis, face emotion analysis and character motion recognition. The body surface color block analysis is used to analyze the color of clothing on the body surface, dirt, etc. stains and bloodstains, facial emotion analysis is used to analyze emotional tendencies based on character expressions, and character action recognition is used to identify character actions; the crowd flow analysis unit is used to count and track crowds based on deduplication crowd flow statistics technology. The crowd flow analysis unit includes multi-dimensional crowd flow analysis, crowd flow attribute analysis and crowd flow trajectory tracking. The multi-dimensional crowd flow analysis is used to analyze the crowd density distribution in the spatial dimension according to the preset three-dimensional map of the hospital, record historical crowd flow data, and count the crowd flow change trend in the time dimension. The crowd flow attribute analysis is used to combine with the character attribute analysis unit to analyze the gender, age, and facial expressions of the crowd, and provide the hospital with a refined crowd flow portrait. The crowd flow trajectory tracking is used to track the flow trajectory of the crowd and predict the real-time crowd density change trend.

[0090] The judgment module is used to judge the physical condition of the person and predict the cause of the person's injury or illness based on the person's motion data and human attribute data, plan an unobstructed route in the hospital based on the crowd tracking data and predict changes in crowd flow; the judgment module includes a person situation judgment unit and a route planning unit; the person situation judgment unit is used to judge whether the person needs first aid and possible symptoms of injury or illness based on the person's motion data and human attribute data, the person situation judgment unit includes first aid judgment and injury refinement, the first aid judgment is used to use a deep learning algorithm to train a model to judge whether the person has symptoms of injury or illness and whether first aid is needed, the injury refinement is used to judge the possible type of injury or illness of the person based on the person's motion data and human attribute data; the route planning unit is used to plan an unobstructed route in the hospital based on the crowd tracking data combined with a preset hospital route map.

[0091] The reporting module is used to alert the hospital emergency room based on the person's physical condition and possible injuries and illnesses, and based on the purpose of priority rescue, it prompts the emergency doctor the fastest arrival route based on the unobstructed movement routes in the hospital, and pushes the flow of people changes to the routes of various departments to hospital visitors based on the current unobstructed movement routes in the hospital and the prediction of crowd changes; the reporting module includes an alarm unit and a terminal application; the alarm unit is used to alert the emergency doctor based on the person's first aid judgment results and possible injuries and illness symptoms, and plans the fastest arrival route for the emergency doctor based on the unobstructed movement routes in the hospital and prompts him; the terminal application is used to prompt and promote unobstructed movement routes and real-time crowd changes on routes, and display the spatial crowd density distribution and the temporal crowd change trend. The terminal application includes a hospital terminal, and the hospital terminal is used to display multi-dimensional crowd analysis data.

[0092] The feedback module is used to feed back the medical information of emergency patients to the judgment module, and to refine the character's physical condition and the cause of the character's injury and illness by combining the character's motion data and human attribute data.

[0093] The specific implementation process is as follows: In actual medical situations, in addition to patients who are unable to move and are sent directly to the hospital by ambulance for emergency treatment, even patients with trauma or acute illnesses may go to the hospital on their own. On the way to the hospital, these patients' walking posture, walking speed, facial expressions, clothing surface marks and hand movements are different from those of ordinary people due to their injuries. In addition, these patients may ultimately find it difficult to reach the clinic on their own due to their injuries, and even if they arrive, the best time for treatment may be delayed.

[0094] Under this premise, the system can determine whether the patient needs emergency treatment by collecting images and tracking people across lenses, and then using the behavior analysis module to analyze the patient's movements and character attributes. After determining that emergency treatment is needed, the reporting module will alert the emergency doctor and plan the fastest route for the emergency doctor. For example, a single patient with swollen feet due to gout is several times slower than ordinary people in going to the department in the hospital due to pain, and accidentally falls on the way. At this time, through the collected patient images and the patient's fall analyzed by the bone point analysis unit, the judgment module determines that the patient needs emergency treatment and initially refines the injury into a lower limb disease. Then the reporting module alerts the emergency doctor through the hospital terminal with the emergency information, the crowd analysis module analyzes the crowd density distribution in the spatial dimension, and the route planning unit plans a route that can reach the patient's location as quickly as possible. The emergency doctor leads the medical staff and carries the corresponding medical materials for the injury to quickly reach the patient's location for emergency treatment and send patients with difficulty in moving to the emergency room, thereby increasing the efficiency of emergency treatment for patients.

[0095] It is also possible that a victim with abdominal trauma needs to go to emergency treatment, but the wound is difficult to stop bleeding and is in great pain. The human body surface color block analysis in the character attribute analysis unit can analyze that the victim has suffered trauma based on the blood on the victim's body surface, and the facial emotion analysis can show that the victim is in great pain. The judgment module calculates and judges that the victim needs emergency treatment based on the victim's speed and the spread of blood. The reporting module then alerts the emergency doctor, who quickly reaches the victim's location according to the fastest arrival route planned by the route planning unit, and initiates emergency treatment to send the victim to the emergency room to save the victim's life.

[0096] In addition, patients receiving inpatient care may walk in open areas of the hospital, where they may fall or have other accidents. Since some places may be secluded and deserted corners, by the time others discover the patient, the best time for treatment may have been lost. Therefore, cameras covering the entire hospital can avoid such blind spots for rescue. When a patient falls, the image acquisition module collects the image, the behavior analysis module analyzes the image, and the judgment module determines that the fallen patient needs first aid. The reporting module alerts the emergency doctor and the attending physician of the inpatient, and the route planning unit plans the fastest route to the patient. The emergency doctor and the attending physician can quickly reach the patient's location for first aid and send the patient back to the ward.

[0097] After first aid and diagnosis, the patient's actual injury information is fed back to the behavior analysis module and the judgment module through the feedback module. The accuracy of the character attribute analysis and character situation judgment is calibrated through the deep learning algorithm, and the patient's injury data is refined to provide data support for the next patient's injury judgment, improve the accuracy of the system, and optimize the allocation of hospital emergency resources.

[0098] Example 2

[0099] The difference from the above embodiment is that, as shown in the attached Figure 5 As shown: the reporting module also includes a push unit, which is used to push real-time crowd flow changes on routes to various departments to hospital visitors based on unobstructed movement routes and crowd flow change predictions within the hospital.

[0100] Terminal applications also include personal terminals, which are used to push smooth routes or fastest arrival routes.

[0101] The specific implementation process is as follows: Due to the complex staff in hospitals, congestion is prone to occur in buildings. The crowd analysis unit can analyze the crowd density distribution and attributes based on the crowd image information collected by the image acquisition module. Combined with the preset hospital three-dimensional map model, it can display the crowd distribution pattern for the hospital and optimize resource allocation. At the same time, it can also predict the trend of crowd changes through trajectory tracking, providing data support for the hospital's response.

[0102] After analyzing the distribution of crowd density, the route planning unit can plan a route with less crowd distribution. On the one hand, it can plan the fastest route for emergency doctors to improve emergency efficiency and grasp the first aid timing of patients. On the other hand, ordinary patients or visitors can use personal terminals, such as mobile apps and WeChat applets, to input the starting point and the department they want to reach. The route planning unit can plan a route that is not congested and display the changes in crowd flow on the route in real time, thereby realizing automatic diversion of medical personnel, improving medical efficiency and experience, and reducing the possibility of accidents in hospitals due to large crowds.

[0103] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0104] The above is only an embodiment of the present invention. The common sense such as the known specific structure and characteristics in the scheme is not described in detail here. The ordinary technicians in the relevant field know all the common technical knowledge in the technical field of the invention before the application date or priority date, can obtain all the existing technologies in the field, and have the ability to apply the conventional experimental means before that date. The ordinary technicians in the relevant field can improve and implement this scheme in combination with their own abilities under the enlightenment given by this application. Some typical known structures or known methods should not become obstacles for ordinary technicians in the relevant field to implement this application. It should be pointed out that for those skilled in the art, without departing from the structure of the present invention, several deformations and improvements can be made, which should also be regarded as the protection scope of the present invention, which will not affect the effect of the implementation of the present invention and the practicality of the patent. The protection scope required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A pedestrian tracking system based on behavior analysis, characterized in that: It includes image acquisition module, cross-lens tracking module, behavior analysis module, judgment module, report module and feedback module; Image acquisition module, used to collect crowd image information and environmental information within the hospital area; Cross-lens tracking module, which is used to achieve seamless recognition and tracking of people between different cameras based on cross-lens tracking technology; A behavior analysis module, used to analyze the crowd image information and environmental information collected by the image acquisition module to obtain character action data, human attribute data and crowd flow tracking data; The judgment module is used to judge the physical condition of the person and predict the cause of the person's injury or illness based on the person's motion data and body attribute data, and to plan the smooth movement route in the hospital and predict the change of the flow of people based on the flow of people tracking data; The reporting module is used to alert the hospital emergency room based on the person's physical condition and possible injuries, and to inform the emergency doctor of the fastest arrival route based on the unobstructed movement routes in the hospital for the purpose of priority rescue. It also pushes the flow changes of people to various departments to hospital visitors based on the current unobstructed movement routes in the hospital and the prediction of the flow changes; The feedback module is used to feed back the medical information of emergency patients to the judgment module, and to refine the character's physical condition and the cause of the character's injury and illness by combining the character's motion data and human attribute data.

2. The pedestrian tracking system based on behavior analysis according to claim 1, characterized in that: The image acquisition module includes several monitoring cameras and several environmental sensors; Several surveillance cameras include several indoor cameras installed in the outpatient hall of the hospital and several outdoor cameras installed in the hospital area, which are used to collect image information of people in the entire hospital and in secluded corners; Several environmental sensors including infrared cameras installed along with surveillance cameras, temperature, humidity and light sensors are used to collect environmental information at different locations within the hospital.

3. The pedestrian tracking system based on behavior analysis according to claim 1, characterized in that: The cross-shot tracking module includes an image preprocessing unit, a person recognition unit, and a cross-shot matching unit; The image preprocessing unit preprocesses crowd image information, including denoising, enhancement and correction, and spatially aligns images collected by different cameras through image registration technology; A person recognition unit is used to extract and recognize the features of people in the crowd image information to establish a unique identification of the target; The cross-shot matching unit is used to establish associations between different cameras based on the feature identifiers of the person, so as to achieve cross-shot matching of the target.

4. The pedestrian tracking system based on behavior analysis according to claim 1, characterized in that: The behavior analysis module includes a skeleton point analysis unit, a character attribute analysis unit, and a crowd flow analysis unit; A skeleton point analysis unit, used to detect the time-series change motion data of a single human body and a single human body key point based on skeleton key point analysis technology; A person attribute analysis unit, used to detect the person's body surface attribute data and identify the person's action data based on the person ID classification detection technology; The crowd analysis unit is used to count and track crowds based on deduplication crowd statistics technology.

5. The pedestrian tracking system based on behavior analysis according to claim 1, characterized in that: The judgment module includes a character situation judgment unit and a route planning unit; A character situation judgment unit is used to judge whether the character needs emergency treatment and possible injury or illness symptoms based on the character's motion data and body attribute data; The route planning unit is used to plan smooth movement routes within the hospital based on the crowd tracking data combined with the preset hospital route map.

6. The pedestrian tracking system based on behavior analysis according to claim 1, characterized in that: The reporting module includes an alarm unit, a push unit, and a terminal application; The alarm unit is used to alert the emergency doctor based on the first aid judgment result and possible injury and illness symptoms, and plan the fastest arrival route for the emergency doctor based on the unobstructed movement route in the hospital and give him prompts; The push unit is used to push the real-time flow changes on the routes to various departments to hospital visitors based on the unobstructed movement routes and the prediction of flow changes in the hospital; Terminal applications are used to prompt and promote smooth action routes and real-time changes in passenger flow along the routes, and to display spatial passenger flow density distribution and temporal passenger flow change trends.

7. The pedestrian tracking system based on behavior analysis according to claim 4, characterized in that: The character attribute analysis unit includes human surface color block analysis, facial emotion analysis and character action recognition. Human body surface color block analysis, used to analyze the color of clothing, stains and blood on the human body surface; Facial emotion analysis, used to analyze emotional tendencies based on facial expressions; Human action recognition, used to identify human actions.

8. The pedestrian tracking system based on behavior analysis according to claim 4, characterized in that: The crowd flow analysis unit includes multi-dimensional crowd flow analysis, crowd flow attribute analysis and crowd flow trajectory tracking. Multi-dimensional crowd flow analysis is used to analyze the crowd density distribution in the spatial dimension, record historical crowd flow data, and count the crowd flow change trend in the time dimension based on the preset three-dimensional map of the hospital; Flow attribute analysis is used to combine with the character attribute analysis unit to analyze the gender, age, and facial expression of the crowd, and provide the hospital with a refined flow portrait; Crowd flow tracking is used to track the flow of people and predict the real-time trend of crowd density changes.

9. The pedestrian tracking system based on behavior analysis according to claim 5, characterized in that: The character situation judgment unit includes first aid judgment and injury and illness refinement. First aid judgment, which uses deep learning algorithms to train models to determine whether a person has symptoms of injury or illness and whether first aid is needed; Injury refinement is used to determine the possible types of injuries a character may have based on the character's motion data and body attribute data.

10. The pedestrian tracking system based on behavior analysis according to claim 6, characterized in that: Terminal applications include personal terminals and hospital terminals. Personal terminals are used to push smooth routes or the fastest routes to reach; Hospital terminal, used to display multi-dimensional crowd flow analysis data.

Images