Fall behavior recognition and determination method

Abstract

The application provides a fall behavior recognition and determination method, and belongs to the technical field of fall determination, and comprises the following steps: collecting contour information of people and a supporting surface in a certain range, and generating corresponding person models and a supporting surface model in an upper computer according to the contour information; setting strength parameters and hardness parameters of the supporting surface model according to actual conditions; determining the gravity center positions of the person models, and the size and direction of the supporting force of the supporting surface model; and when the time interval of imbalance of the person models and the force when the person models collide with the supporting surface both reach preset requirements, it is determined that a fall has occurred. The fall behavior recognition and determination method provided by the application optimizes the fall determination conditions by the time interval of imbalance and the force when inflation is fitted, so that the method has more guiding significance, avoids misjudgment and improves the accuracy.

Description

Technical Field

[0001] This invention belongs to the field of fall detection technology, and more specifically, relates to a method for fall behavior recognition and determination. Background Technology

[0002] With the global aging population intensifying, falls have become one of the leading health threats to the elderly. More and more seniors are living alone, without anyone to care for them, and accidents may not be detected in time, posing significant safety risks to their lives.

[0003] With the ongoing development of safe city and intelligent transportation projects in my country, integrating machine vision technology into video surveillance systems has become a hot research topic. Currently, most existing methods rely on traditional machine learning for fall detection, resulting in low accuracy and delays in providing timely assistance to elderly individuals. Therefore, how to efficiently, accurately, and in real-time detect falls in the elderly is an urgent problem that needs to be solved.

[0004] Currently, there are three main types of methods for fall detection: wearable device-based methods, environmental sensor-based methods, and computer vision-based methods. However, all of these methods suffer from over-judgment or under-judgment. Even if a person loses their balance, if there is support below, it should not be considered a fall. For these reasons, the current detection methods are not very practical and have low accuracy. Summary of the Invention

[0005] The purpose of this invention is to provide a method for recognizing and determining fall behavior, aiming to solve the problems of poor practicality and low accuracy in determining falls.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is: to provide a method for recognizing and determining fall behavior, comprising:

[0007] Collect the contour information of a person and a supporting surface within a certain range, and generate the corresponding person model and supporting surface model in the host computer using the contour information;

[0008] Set the strength and hardness parameters of the support surface model according to the actual situation;

[0009] The position of the center of gravity of each character model and the magnitude and direction of the supporting force exerted by the supporting surface model on the center of gravity are determined.

[0010] When the time interval of the character model becoming unbalanced and the force exerted when the character model collides with the supporting surface both reach the preset requirements, it is determined that a fall has occurred.

[0011] In one possible implementation, the acquisition of the contour information of the person and the supporting surface within a certain range includes:

[0012] The data acquisition device scans all people and supporting surfaces within a certain range; the supporting surfaces include the ground and the outer perimeter of each object.

[0013] In one possible implementation, generating the corresponding character model and support surface model using the contour information within the host computer includes:

[0014] The contour information transmitted back in real time by the acquisition device enables the character model and the support surface model in the host computer to be synchronized with reality.

[0015] In one possible implementation, generating the corresponding character model and support surface model using the contour information within the host computer includes:

[0016] The difference between two outline information in adjacent time periods is determined, and the host computer adjusts the character model and the support surface model according to the difference.

[0017] In one possible implementation, setting the strength and hardness parameters of the support surface model according to the actual situation includes:

[0018] The hardness and strength information of the support surface are determined, and the same and corresponding hardness and strength parameters are set for the support surface model in the host computer according to the hardness and strength information.

[0019] In one possible implementation, generating the corresponding character model and support surface model using the contour information within the host computer includes:

[0020] Set a reference object and determine the proportional relationship between the actual size of the reference object and the size of the model generated in the host computer;

[0021] The required actual size is determined based on the proportional relationship using the character model and the supporting surface model.

[0022] In one possible implementation, determining the center of gravity position of each character model and the magnitude and direction of the supporting force exerted by the supporting surface model on the center of gravity includes:

[0023] Based on a person's gender and the type of clothing they are currently wearing, the system can fit the corresponding person's body shape.

[0024] Based on the mass parameters of each part of the body, the person's weight is determined by the described body shape.

[0025] In one possible implementation, determining the center of gravity position of each character model and the magnitude and direction of the supporting force exerted by the supporting surface model on the center of gravity includes:

[0026] Based on the speed and acceleration of the character model relative to the supporting surface model, and in conjunction with the position of the center of gravity, the magnitude and direction of the supporting force exerted by the supporting surface on the center of gravity are analyzed.

[0027] In one possible implementation, determining that a fall has occurred when both the time interval of the character model becoming unbalanced and the force exerted when the character model collides with the supporting surface meet preset requirements includes:

[0028] Within the host computer, the time interval is determined from the moment the character model becomes unbalanced until the character model regains its balance.

[0029] In one possible implementation, determining that a fall has occurred when both the time interval of the character model becoming unbalanced and the force exerted when the character model collides with the supporting surface meet preset requirements includes:

[0030] Within the host computer, the force exerted when the character model expands the supporting surface is simulated;

[0031] Different judgment intervals are set according to different ages. When the time interval is within the judgment interval and the force is greater than the corresponding standard, it is judged as a fall; otherwise, no warning is issued.

[0032] The beneficial effects of the fall behavior recognition and determination method provided by this invention are as follows: Compared with the prior art, this invention's fall behavior recognition and determination method first collects the contour information of a person and a supporting surface within a certain range. Once the contour information is determined, the corresponding person model and supporting surface model can be generated in the host computer using the contour information. To be more realistic, the strength and hardness parameters of the supporting surface model need to be set according to the actual situation. Since the person model and supporting surface model correspond to reality, the center of gravity position of the person model and the magnitude and direction of the supporting force of the supporting surface model on the person model can be determined.

[0033] Within the host computer, a fall is determined to have occurred when both the time interval of the character model's imbalance and the force exerted when the character model collides with the supporting surface meet preset requirements. In this application, not only is a corresponding model generated based on the actual situation, but the fall determination conditions are also optimized by using the time interval of imbalance and the force exerted during the fitted expansion, thus providing more guidance, avoiding misjudgments, and improving accuracy. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 A flowchart of a fall behavior recognition and determination method provided in an embodiment of the present invention. Detailed Implementation

[0036] To make the technical problems to be solved, the technical solutions, and the beneficial effects of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0037] Please see Figure 1 The present invention will now describe a method for recognizing and determining fall behavior. A method for recognizing and determining fall behavior includes:

[0038] Collect the contour information of people and supporting surfaces within a certain range, and generate corresponding character models and supporting surface models in the host computer based on the contour information.

[0039] Set the strength and hardness parameters of the support surface model according to the actual situation.

[0040] Determine the center of gravity position of each character model, as well as the magnitude and direction of the supporting force exerted by the supporting surface model on the center of gravity.

[0041] When the time interval between the character model's imbalance and the force exerted when the character model collides with the supporting surface both reach the preset requirements, it is determined that a fall has occurred.

[0042] The beneficial effects of the fall behavior recognition and determination method provided by this invention are as follows: Compared with the prior art, this invention's fall behavior recognition and determination method first collects the contour information of a person and a supporting surface within a certain range. Once the contour information is determined, the corresponding person model and supporting surface model can be generated in the host computer using the contour information. To be more realistic, the strength and hardness parameters of the supporting surface model need to be set according to the actual situation. Since the person model and supporting surface model correspond to reality, the center of gravity position of the person model and the magnitude and direction of the supporting force of the supporting surface model on the person model can be determined.

[0043] Within the host computer, a fall is determined to have occurred when both the time interval of the character model's imbalance and the force exerted when the character model collides with the supporting surface meet preset requirements. In this application, not only is a corresponding model generated based on the actual situation, but the fall determination conditions are also optimized by using the time interval of imbalance and the force exerted during the fitted expansion, thus providing more guidance, avoiding misjudgments, and improving accuracy.

[0044] A fall is defined as a sudden, involuntary, and unintentional change of body position resulting in falling to the ground or a lower surface. Data from the 2006 National Disease Surveillance System's cause-of-death surveillance showed that the fall mortality rate among people aged 65 and over in my country was 49.56 per 100,000 for men and 52.80 per 100,000 for women. Falls are the fourth leading cause of injury-related death in my country, and the leading cause among people aged 65 and over. Therefore, timely monitoring of falls to provide rapid assistance and prevent secondary injuries is of paramount importance.

[0045] Existing fall detection technologies mainly employ sensor monitoring systems based on accelerometers and gyroscopes. However, these fall detection technologies suffer from low accuracy and high false alarm rates.

[0046] Currently, there is much discussion both domestically and internationally regarding fall detection methods based on computer vision. Based on the different algorithms and implementation methods used, they can be specifically categorized as follows:

[0047] Body shape analysis is a method that extracts the human body contour from an image using a background removal modeling algorithm. The body is then defined as a region of interest (ROI) using a rectangle, and the aspect ratio is used to determine if a fall has occurred. However, this method is susceptible to changes in lighting and moving objects in the background, resulting in a high false positive rate, and it cannot predict falls.

[0048] This method, which relies on the characteristic of elderly people not recovering for a long time after a fall, distinguishes falls from similar actions such as bending over or squatting, and improves the fall recognition rate by defining specific areas. However, this method is a typical post-fall detection method, unable to predict the outcome, and cannot correctly distinguish between prolonged lying down and a fall.

[0049] This head motion tracking method uses particle filtering to track the human head and detects falls based on the distance from the head to the ground and the rate of descent. While this method achieves a certain degree of fall prediction through stable head tracking, the lack of a fall model library leads to a high false positive rate. Furthermore, it lacks scene understanding algorithms, resulting in poor robustness to complex environments, and the particle filtering algorithm is time-consuming, resulting in a slow detection rate and hindering real-time detection.

[0050] This method, focusing on behavior recognition, trains various behaviors such as walking, squatting, sitting, lying down, and falling using a Convolutional Neural Network (CNN) to generate a fall detection model library. These models are then classified and identified to achieve fall detection. This method generates its own model library, significantly improving the accuracy of fall detection. However, training the CNN model involves a large computational load, resulting in low algorithm efficiency, and it does not include fall prediction functionality.

[0051] Falls are a leading cause of accidental injuries among the elderly. Statistics show that 60% of head injuries and 90% of hip and wrist injuries in the elderly are caused by falls, and 30% of elderly people living alone and 50% of those in long-term care facilities experience at least one fall per year. Therefore, timely fall detection is crucial in the long-term care of the elderly. On the other hand, with the increasingly serious trend of global population aging, the cost of long-term care for the elderly is also rising, especially in care facilities such as hospitals and nursing homes. Therefore, the demand for real-time fall detection systems for the elderly is very high.

[0052] Currently, there are three main types of methods for fall detection: wearable device-based methods, environmental sensor-based methods, and computer vision-based methods.

[0053] Wearable device-based methods detect falls by using sensors to detect acceleration. Their advantages include low computational cost and ease of use, but they require constant wear, disrupting daily life. Environmental sensor-based methods, such as pressure and sound sensors, also have low computational cost, but are greatly affected by changes in environmental pressure and sound, resulting in a high false alarm rate. Computer vision-based methods primarily utilize surveillance video images. These methods do not require wearable devices, but are easily affected by lighting conditions, cannot be used at night, have poor privacy protection, and their accuracy does not meet high standards.

[0054] In some embodiments of the fall behavior recognition and determination method provided in this application, collecting contour information of a person and a supporting surface within a certain range includes:

[0055] The data acquisition device scans all people and supporting surfaces within a certain range; the supporting surfaces include the ground and the outer perimeter of each object.

[0056] To identify human bodies, current methods primarily rely on optical data acquisition devices such as cameras for real-time data collection. There are also relatively mature technologies that can accurately identify body contours and determine the identity of the target person based on facial features.

[0057] Based on the aforementioned technological foundation, and through other methods such as image acquisition and sensor detection, it is possible to determine the joints of the human body, thereby determining the relative motion between the joints.

[0058] However, it should be noted that different people have different weights, and people of the same height may have different body shapes, that is, different weight distributions. This leads to a certain difference in the position of each person's center of gravity. Due to the above differences, the degree of change in the center of gravity when a fall occurs is different, so each case needs to be analyzed specifically.

[0059] More importantly, the human body does not always fall while standing; it may also fall in other postures such as squatting. In addition, if there is a support under the body, then even if there is a loss of balance, the timely support from the support should not be considered a fall. Therefore, judging a fall solely based on the movement patterns of the joints is somewhat narrow.

[0060] For the reasons mentioned above, it is necessary to analyze the human body's posture and the surrounding environment, that is, to determine the environmental factors of the current scene and the status of the supporting platforms. When the human body is in different positions in the current scene, the judgment of fall behavior needs to be adjusted accordingly, so as to more accurately and directly determine whether a fall has occurred and the severity of the fall.

[0061] In some embodiments of the fall behavior recognition and determination method provided in this application, generating corresponding human figure and support surface model using contour information in the host computer includes:

[0062] By collecting the contour information transmitted back in real time by the acquisition device, the character model and support surface model in the host computer can be synchronized with reality.

[0063] If a person falls onto a relatively soft support, even if a fall is determined to have occurred using current technology, it is not considered a truly harmful fall in this application because it is not harmful.

[0064] The ultimate goal of determining fall behavior is to provide early warning of potential harm. If a fall occurs and is harmful, then the method of determining fall behavior is valuable. However, if the body loses its balance but there is a relatively soft support and the body's free fall time is very short, or if the loss of balance is spontaneous and the body performs certain actions, then since the above behaviors are not harmful and have a certain purpose, they cannot be determined as a fall.

[0065] However, existing methods mostly analyze the structure of the human body in isolation, neglecting the environment in which the body is situated and its relative position to surrounding objects. This results in the current methods for determining fall behavior being impractical and prone to significant misjudgments.

[0066] This application requires the use of devices such as cameras to identify human bodies and to determine the volume and relative position of objects within the captured area. When a person enters the target area, the host computer can generate a corresponding model based on the data acquired by the camera, and then determine the person's state based on changes in position and posture.

[0067] In some embodiments of the fall behavior recognition and determination method provided in this application, generating corresponding human figure and support surface model using contour information in the host computer includes:

[0068] The differences between two contour information in adjacent time periods are determined, and the host computer adjusts the character model and the support surface model based on the differences.

[0069] In this application, the camera and other data acquisition devices will scan each person and object within the target range in real time. The data acquisition devices will transmit the acquired data to the host computer, and then the host computer will generate a model in real time. Through the model, the fall behavior can be judged in the most intuitive and accurate way.

[0070] However, in order to perform comprehensive and accurate modeling, at least two acquisition devices are typically required, and multiple acquisition devices need to be able to capture images of the object from different angles. For these reasons, this places higher demands on the processing power of the host computer.

[0071] For the reasons mentioned above, in this application, after the data acquisition device transmits data to the host computer, the host computer filters the data. Specifically, the data determined by the acquisition device is divided. This is because the data acquired by the acquisition device includes dynamic and static components; the dynamic component mainly consists of people, while the static component mainly consists of objects in the surrounding environment.

[0072] Therefore, when dividing the data, the outline of the static part is first defined. Typically, the static part is an area where the position reflected by the data does not change within a certain time interval. Once the static part is determined, multiple points are picked up within this area to display the outline of the static object. If multiple points at the same location in adjacent data determined by the acquisition device do not change, then no further modeling is performed; instead, the previous model is used instead. In other words, the host computer only models objects that have changed. Furthermore, within the same time interval, based on the extraction of the object's trajectory and outline, objects whose shape does not change are not modeled. Instead, the model is moved and its angle changed according to the outline, ultimately reducing the amount of data that the host computer needs to process.

[0073] In some embodiments of the fall behavior recognition and determination method provided in this application, the strength parameters and hardness parameters of the support surface model are set according to the actual situation, including:

[0074] The hardness and strength information of the support surface are determined, and the same and corresponding hardness and strength parameters are set on the support surface model in the host computer based on the hardness and strength information.

[0075] Once the model is built, a fall cannot be immediately determined, as the severity of injury varies depending on the material. If the supporting surface is relatively soft, even if someone falls, it cannot be considered a true fall because it poses no harm, such as falling onto a spring mattress. However, if the supporting surface is relatively hard, the situation is different.

[0076] Therefore, in practical applications, relevant personnel need to set corresponding parameters such as hardness and strength in the model according to the actual situation, and also need to set corresponding mass parameters according to the size of the human body model.

[0077] Once the above parameters are set, the force when the models collide can be roughly calculated, and thus the injury situation when a person falls can be predicted.

[0078] In some embodiments of the fall behavior recognition and determination method provided in this application, generating corresponding human figure and support surface model using contour information in the host computer includes:

[0079] Set a reference object and determine the proportional relationship between the actual size of the reference object and the size of the model generated in the host computer.

[0080] The required actual size is determined based on the proportions using the figure model and the supporting surface model.

[0081] In this application, it is necessary to model the human body and objects in the environment in the host computer. In order to perform real-time scanning and finally generate the model, at least two acquisition devices are needed to scan the human body and objects within a certain range from different angles.

[0082] However, in reality, the farther the acquisition device is from the object being measured, the smaller the final model size will be; if the acquisition device is close to the object, the larger the final model size will be.

[0083] In order to ensure that the size of the generated model is the same as the actual size, a reference object is set within the coverage area of ​​the acquisition device in this application. The reference object is scanned along with the acquisition device during the scanning process. Since the actual size of the reference object is stored in the host computer, the host computer marks the generated model according to the actual size of the reference object. That is, the host computer determines the ratio between the unit model length and the actual length. Based on this ratio, the actual size of the object and the height of the person can be determined from the model.

[0084] In some embodiments of the fall behavior recognition and determination method provided in this application, determining the center of gravity position of each character model and the magnitude and direction of the supporting force of the supporting surface model on the center of gravity includes:

[0085] Based on a person's gender and the type of clothing they are currently wearing, the system can fit the corresponding body shape.

[0086] A person's weight is determined by analyzing the mass parameters of each part of their body and their posture.

[0087] After the models of the dynamic and static parts are established through the data acquisition device, the main object of the dynamic part is usually a person. Therefore, in practical applications, it is necessary to determine the time when the person loses balance and the time from when the person loses balance until they come into contact with the support.

[0088] For the reasons mentioned above, it is first necessary to determine the center of gravity of a person. The human model can be determined by the data acquisition device. Once the human model is determined, the volume of the model is the total volume of the person and their clothes. Since the thickness of the clothes worn by people varies in different seasons, if the model is directly regarded as the human body itself, it is impossible to effectively estimate the total weight, and the determination of the center of gravity will be inaccurate.

[0089] For the reasons mentioned above, the host computer needs to make a preliminary judgment on the person's clothing based on the current temperature and weather conditions, and determine the gender of the target person by collecting data and combining it with facial recognition. Based on the above information, the host computer can make a preliminary judgment on the person's body shape based on the model and other information such as gender. After the body shape is determined, the person's shape is determined by combining the approximate model of the clothing with the established model. The person's weight is determined by the shape. Since the model is fixed, the person's center of gravity can be determined within the host computer.

[0090] In some embodiments of the fall behavior recognition and determination method provided in this application, determining the center of gravity position of each character model and the magnitude and direction of the supporting force of the supporting surface model on the center of gravity includes:

[0091] Based on the speed and acceleration of the character model relative to the supporting surface model, and combined with the position of the center of gravity, the magnitude and direction of the supporting force exerted by the supporting surface on the center of gravity are analyzed.

[0092] The key to this application is how to determine the imbalance of the human body's center of gravity. First, the human body posture is determined in the host computer based on the model, the person's gender, and the external environment. Then, by combining the posture with the corresponding density of the human body in various parts, the center of gravity of the human body is finally determined through the model.

[0093] The model within the host computer is constantly changing, and the position of the center of gravity is also constantly changing. To determine whether a person is out of balance, it is necessary to determine the direction and magnitude of the force exerted on the center of gravity by the supporting surface, and to consider the relative position of the person to other objects.

[0094] Before a fall, the center of gravity is inside the body. If there are no objects around, the legs or other parts of the body provide support below the center of gravity. When imbalance occurs, there are no supporting parts of the body below the center of gravity, and surrounding objects cannot provide support. The supporting force from the supporting surfaces is less than the weight of the body, which is considered an imbalance.

[0095] For the reasons mentioned above, it is necessary to use the host computer to determine the center of gravity and the support below the center of gravity. To explain in detail, it is necessary to use the host computer to determine the direction of the supporting force of the support surface on the person. If the magnitude of the supporting force of the support surface on the person is less than the person's weight, then it is determined to be unbalanced.

[0096] In some embodiments of the fall behavior recognition and determination method provided in this application, a fall is determined to have occurred when both the time interval of the human figure becoming unbalanced and the force exerted when the human figure collides with the supporting surface meet preset requirements:

[0097] Within the host computer, the time interval is determined from when the character model becomes unbalanced until it is rebalanced.

[0098] The model in the host computer is synchronized with the actual situation. When not falling, the legs support the body. During a fall, due to the lack of leg support or the supporting force of the support surface being less than gravity, the body is in an unbalanced state. When the body falls onto the ground, i.e., the support surface, the support surface provides support, which allows the body to return to a balanced state.

[0099] Therefore, in actual judgment, it is necessary to analyze the support between the human body's center of gravity and the support surface in real time through the model. Since the model is synchronized with reality and the model's material is the same as reality, the magnitude and direction of the force between the support surface and the human body can be determined in real time in the host computer, so it is possible to determine whether the human body is in an unbalanced state.

[0100] In some embodiments of the fall behavior recognition and determination method provided in this application, a fall is determined to have occurred when both the time interval of the human figure becoming unbalanced and the force exerted when the human figure collides with the supporting surface meet preset requirements:

[0101] The system simulates the forces acting on the expanding support surface of the character model within the host computer.

[0102] Different judgment intervals are set according to different ages. If the time interval is within the judgment interval and the force is greater than the corresponding standard, it is judged as a fall; otherwise, no warning is issued.

[0103] In this application, if the distance between the human body and the supporting surface is low, even if a fall occurs, the injury is not considered significant, and therefore it is not considered a fall. However, for some elderly people, due to their reduced bone strength, even a minor fall can lead to fractures and other problems. For the above reasons, it is necessary to identify the basic information of the person through a data acquisition device, and to establish different judgment intervals according to different age groups. Each judgment interval includes the time of gravity imbalance when a fall is determined. The data acquisition device and the host computer can analyze the time of gravity imbalance, and then, by combining this time with the person's age and the corresponding judgment interval, it can be determined whether a fall has occurred.

[0104] In addition to determining the time of imbalance, it is also necessary to analyze and judge the support surface of the final fall. If the support surface is relatively soft, the damage caused by the fall will be relatively small. If it is judged as a fall in this case, it will undoubtedly be a misjudgment.

[0105] For the reasons mentioned above, in addition to calculating the time of imbalance, this application also needs to calculate the force when the human body hits the support surface through the host computer. Since the host computer has models of the human body and the object and sets the corresponding material parameters, the force when the human body hits the support surface can be determined by analyzing the data with software such as finite element method. If the force is greater than a certain value and the time of imbalance is within the corresponding judgment interval, then it is determined to be a fall.

[0106] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for recognizing and determining fall behavior, characterized in that, include: Collect the contour information of a person and a supporting surface within a certain range, and generate the corresponding person model and supporting surface model in the host computer using the contour information; Set the strength and hardness parameters of the support surface model according to the actual situation; The position of the center of gravity of each character model and the magnitude and direction of the supporting force of the supporting surface model on the center of gravity are determined; When the time interval of the character model becoming unbalanced and the force exerted when the character model collides with the supporting surface both reach the preset requirements, it is determined that a fall has occurred. The collection of contour information of the person and the supporting surface within a certain range includes: The data acquisition device scans all people and supporting surfaces within a certain range; the supporting surfaces include the ground and the outer peripheral surfaces of each object. The process of generating the corresponding character model and support surface model using the contour information within the host computer includes: The contour information transmitted back in real time by the acquisition device enables the character model and the support surface model in the host computer to be synchronized with reality. The step of setting the strength and hardness parameters of the support surface model according to the actual situation includes: The hardness and strength information of the support surface are determined, and the same and corresponding hardness and strength parameters are set for the support surface model in the host computer according to the hardness and strength information. The process of generating the corresponding character model and support surface model using the contour information within the host computer includes: Set a reference object and determine the proportional relationship between the actual size of the reference object and the size of the model generated in the host computer; The required actual size is determined based on the proportional relationship using the character model and the supporting surface model.

2. The method for recognizing and determining fall behavior as described in claim 1, characterized in that, The process of generating the corresponding character model and support surface model using the contour information within the host computer includes: The difference between two outline information in adjacent time periods is determined, and the host computer adjusts the character model and the support surface model according to the difference.

3. The method for recognizing and determining fall behavior as described in claim 1, characterized in that, Determining the center of gravity position of each character model and the magnitude and direction of the supporting force exerted by the supporting surface model on the center of gravity includes: Based on a person's gender and the type of clothing they are currently wearing, the system can fit the corresponding person's body shape. Based on the mass parameters of each part of the body, the person's weight is determined by the described body shape.

4. The method for recognizing and determining fall behavior as described in claim 3, characterized in that, Determining the center of gravity position of each character model and the magnitude and direction of the supporting force exerted by the supporting surface model on the center of gravity includes: Based on the speed and acceleration of the character model relative to the supporting surface model, and in conjunction with the position of the center of gravity, the magnitude and direction of the supporting force exerted by the supporting surface on the center of gravity are analyzed.

5. The method for recognizing and determining fall behavior as described in claim 4, characterized in that, The determination that a fall has occurred when both the time interval of the character model becoming unbalanced and the force exerted when the character model collides with the supporting surface meet preset requirements includes: Within the host computer, the time interval is determined from the moment the character model becomes unbalanced until the character model regains its balance.

6. The method for recognizing and determining fall behavior as described in claim 5, characterized in that, The determination that a fall has occurred when both the time interval of the character model becoming unbalanced and the force exerted when the character model collides with the supporting surface meet preset requirements includes: Within the host computer, the force exerted when the character model expands the supporting surface is simulated; Different judgment intervals are set according to different ages. When the time interval is within the judgment interval and the force is greater than the corresponding standard, it is judged as a fall; otherwise, no warning is issued.

Images