Fall detection method and device based on multi-source data fusion, equipment, medium and product

Abstract

The application discloses a fall detection method and device based on multi-source data fusion, equipment, medium and product, relates to the technical field of fall detection, and the method comprises the following steps: acquiring three-axis acceleration of a to-be-detected person to obtain first and acceleration; wherein the three-axis acceleration is obtained according to a three-axis accelerometer, and the three-axis accelerometer is worn by the to-be-detected person; determining whether the to-be-detected person has an impact event at the current time according to the first and acceleration; if it is determined that the impact event occurs, determining whether the to-be-detected person has a free fall event in a first preset time window before the current time according to a second and acceleration sequence in the first preset time window; if it is determined that the free fall event exists, acquiring a tracking heat map sequence in a second preset time window and determining whether the to-be-detected person has a fall event in the second preset time window; if the fall event exists, it is determined that the to-be-detected person has a fall. The application can improve the recognition accuracy of fall detection.

Description

Technical Field

[0001] This application relates to the field of fall detection technology, and in particular to a fall detection method, apparatus, equipment, medium and product based on multi-source data fusion. Background Technology

[0002] Fall detection is an active protection system that uses sensors and artificial intelligence technology to identify accidental falls in real time and automatically trigger an alarm. Its core value lies in transforming traditional post-accident rescue into proactive monitoring and early warning.

[0003] Currently, most fall detection solutions based on wearable sensors use accelerometers or gyroscopes to determine a fall by detecting impact acceleration and subsequent stationary state.

[0004] In existing technologies, sudden movements such as sitting down quickly, bending over to pick something up, or sneezing are easily misjudged as falls, resulting in a high false alarm rate. Summary of the Invention

[0005] The purpose of this application is to provide a fall detection method, device, equipment, medium, and product based on multi-source data fusion, which can improve the accuracy of fall detection.

[0006] To achieve the above objectives, this application provides the following solution: Firstly, this application provides a fall detection method based on multi-source data fusion, including: The triaxial acceleration of the subject to be tested is obtained, and the first acceleration of the subject to be tested is obtained based on the triaxial acceleration; wherein the triaxial acceleration is obtained from a triaxial accelerometer, which is worn by the subject to be tested; Based on the first and acceleration, determine whether the detector has experienced an impact event at the current time; If the impact event is determined to have occurred, a second sum acceleration sequence within a first preset time window prior to the current time is obtained; based on the second sum acceleration sequence, it is determined whether the subject under test experienced a free fall event within the first preset time window. If the free fall event is confirmed, a tracking heatmap sequence of the subject within the second preset time window is obtained; based on the tracking heatmap sequence, it is determined whether the subject falls within the second preset time window; if the subject falls within the second preset time window, it is determined that the subject has fallen.

[0007] Optionally, determining whether the detector has experienced an impact event at the current time based on the first acceleration includes: Determine whether the first acceleration is greater than a first preset value. If the first acceleration is greater than the first preset value, determine that the subject under test has experienced an impact event at the current time.

[0008] Optionally, determining whether the subject under test has a freefall event within the first preset time window based on the second and acceleration sequences includes: Determine whether there is a sum of accelerations lower than a second preset value in the second sum of acceleration sequences; If there is a sum and acceleration lower than the second preset value, it is determined that the subject under test has the free fall event within the first preset time window; Otherwise, if it is determined that the subject does not have the free fall event within the first preset time window, return to the step of obtaining the three-axis acceleration of the subject and obtaining the first acceleration of the subject based on the three-axis acceleration.

[0009] Optionally, the tracking heatmap sequence includes multiple frames of heatmaps of the person to be detected; the step of determining whether the person to be detected has fallen within the second preset time window based on the tracking heatmap sequence includes: Traverse the heatmap, establish a two-dimensional coordinate system with the top of the heatmap as the x-axis, and obtain the ordinate of the human body's centroid coordinate in the heatmap; Determine whether there is a heat map in which the vertical coordinate of the human body's center of mass is greater than a third preset value. If so, determine that the person to be tested experienced the fall event within the second preset time window. Otherwise, it is determined that the person to be tested did not experience the fall event within the second preset time window.

[0010] Optionally, the ordinate of the human body's center of mass is a temperature-weighted ordinate of the human body's center of mass, and obtaining the ordinate of the human body's center of mass in the thermal image includes: The temperature of each pixel in the heat map is obtained, and all pixels with a temperature greater than a fourth preset value are extracted as human body pixels. The vertical coordinates of the human body pixels are weighted based on the corresponding temperature to obtain the temperature-weighted vertical coordinates of the human body's centroid.

[0011] Optionally, the acquisition of the first acceleration of the subject based on the triaxial accelerometer includes: The x-axis acceleration, y-axis acceleration, and z-axis acceleration of the subject to be tested are acquired in real time using a triaxial accelerometer, and the first acceleration is obtained based on the x-axis acceleration, y-axis acceleration, and z-axis acceleration.

[0012] Secondly, this application provides a fall detection device based on multi-source data fusion, comprising: A dynamic detection module is used to acquire the triaxial acceleration of the subject being tested, and to acquire the first acceleration of the subject being tested based on the triaxial acceleration; wherein, the triaxial acceleration is acquired based on a triaxial accelerometer, which is worn by the subject being tested; Based on the first and acceleration, determine whether the detector has experienced an impact event at the current time; If the impact event is determined to have occurred, a second sum acceleration sequence within a first preset time window prior to the current time is obtained; based on the second sum acceleration sequence, it is determined whether the subject under test experienced a free fall event within the first preset time window. The posture detection module is used to, if the free fall event is determined to exist, acquire the tracking heatmap sequence of the subject within the second preset time window; determine whether the subject has a fall event within the second preset time window based on the tracking heatmap sequence; and determine that the subject has fallen if the fall event occurs within the second preset time window.

[0013] Thirdly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the fall detection method based on multi-source data fusion as described above.

[0014] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the fall detection method based on multi-source data fusion described above.

[0015] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the fall detection method based on multi-source data fusion described above.

[0016] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a fall detection method, device, equipment, medium, and product based on multi-source data fusion. It identifies the free fall phase and impact phase in a fall action by using triaxial acceleration data to identify the motion state of the user during the fall. It also identifies the posture of the user based on the tracking heat map sequence, thereby accurately identifying the fall action and reducing the false judgment rate. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 An application environment diagram of a fall detection method based on multi-source data fusion provided in an embodiment of this application; Figure 2 A flowchart illustrating a fall detection method based on multi-source data fusion, provided as an embodiment of this application; Figure 3 A flowchart illustrating a fall detection method based on multi-source data fusion, provided for another embodiment of this application; Figure 4 A schematic diagram of the functional modules of a fall detection device based on multi-source data fusion provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0019] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0020] It should be noted that the terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented, for example, in orders other than those illustrated or described herein.

[0021] It should be noted that "at the time of..." in the embodiments of this application can be either at the instant when a certain situation occurs, or for a period of time after the occurrence of a certain situation. The embodiments of this application do not make specific limitations on this.

[0022] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0023] This application provides a fall detection method based on multi-source data fusion, which can be applied to, for example... Figure 1 In the application environment shown, the detection terminal 101 communicates with the wearable sensor 102 via a network, and also communicates with the thermal imaging sensor 103 via the network. The wearable sensor 102, a triaxial accelerometer, is worn by the person being tested and is used to acquire the person's triaxial acceleration to determine their motion state. The thermal imaging sensor 103 acquires the person's thermal imaging data to determine their posture. The method of this application is used by the detection terminal 101 to determine whether the person has fallen by acquiring the person's triaxial acceleration and thermal imaging data. Furthermore, if a fall is detected, the detection terminal 101 can send an alarm to the family / medical terminal 104.

[0024] Furthermore, the detection terminal 101 also includes a data storage system, which can process the data that the detection terminal 101 needs to process. The data storage system can be set up separately, integrated into the detection terminal 101, or placed in the cloud or on other servers. The detection terminal 101 can be implemented using a standalone server, a server cluster consisting of multiple servers, or a cloud server.

[0025] Furthermore, wearable sensors can be integrated into portable wearable devices, such as smartwatches, smart bracelets, and head-mounted devices.

[0026] Optionally, the wearable sensor uses an inertial measurement unit of model MPU6050. It integrates a triaxial accelerometer (measuring acceleration in the X, Y, and Z directions).

[0027] Furthermore, the thermal imaging sensor can be a thermal imaging sensor of model MLX90640.

[0028] In one exemplary embodiment, such as Figure 2 As shown, a fall detection method based on multi-source data fusion is provided. This method is executed by a computer device, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is applied to... Figure 1 Taking the detection terminal 101 as an example, the explanation includes the following steps 201 to 204. Wherein: Step 201: Obtain the triaxial acceleration of the subject, and obtain the first acceleration of the subject based on the triaxial acceleration; wherein, the triaxial acceleration is obtained from a triaxial accelerometer, which is worn by the subject. Step 202: Determine whether an impact event has occurred to the detector at the current time based on the first and acceleration. Specifically, this application studies two phases of a fall, in which there is first a brief period of free fall, followed by impact with the ground. This step is used to identify the impact event.

[0029] Step 203: If an impact event is confirmed, obtain the second acceleration sequence within the first preset time window before the current time; determine whether the subject to be tested has a free fall event within the first preset time window based on the second acceleration sequence. Specifically, when it is determined that an impact event occurs at the current time, the time period before the current time is traced back to determine whether the event during that time period was a free fall event; Step 204: If a free fall event is confirmed, obtain the tracking heat map sequence of the subject within the second preset time window; determine whether the subject has fallen within the second preset time window based on the tracking heat map sequence; if the subject has fallen within the second preset time window, determine that the subject has fallen.

[0030] Specifically, after confirming the existence of both free fall and impact based on the above steps, the motion state of the person being tested is determined to conform to the pattern of a fall. The posture of the person being tested is further identified based on the tracking heat map sequence. If the posture determines that the person being tested fell at a certain moment (regardless of whether they stood up), then it is determined that the person being tested has fallen, which can trigger a family alarm.

[0031] By implementing steps 201 to 204 above, this application can obtain triaxial acceleration data from a triaxial accelerometer and a tracking heat map sequence from an image tracking sensor. The two types of sensor data complement each other: by identifying the free fall phase and impact phase in the fall action through the triaxial acceleration data, the motion state of the user to be detected when falling can be identified, and the posture of the user to be detected can be identified according to the tracking heat map sequence, so as to accurately identify the fall action and reduce the misjudgment rate.

[0032] In one exemplary embodiment, such as Figure 3 As shown, another embodiment of the fall detection method based on multi-source data fusion provided in this application specifically includes steps 301 to 305: Step 301: Real-time acquisition of triaxial acceleration based on triaxial accelerometer, including x-axis acceleration, y-axis acceleration and z-axis acceleration, and acquisition of the first sum acceleration at the current time based on x-axis acceleration, y-axis acceleration and z-axis acceleration.

[0033] Specifically, the first sum of accelerations is the sum of accelerations at the current time, and the formula for obtaining the first sum of accelerations is: in, For first and acceleration, The x-axis acceleration at the current time. The y-axis acceleration at the current time. This represents the z-axis acceleration at the current time.

[0034] Specifically, the reason for using the resultant acceleration instead of data from individual axes is that the resultant acceleration is a scalar quantity, independent of the sensor's wearing orientation. There's no need to strictly define the orientation of each axis or the wearing location.

[0035] Specifically, regardless of whether the elderly person (the subject of the test) wears the device on their wrist, waist, or other parts of their body, as long as a fall occurs, the resultant acceleration will exhibit the characteristics of free fall and subsequent stillness after impact. In the stationary state, the resultant acceleration is approximately 1g (gravitational acceleration), the resultant acceleration during the impact phase of the fall usually exceeds 3g, and the resultant acceleration during the free fall phase is close to 0g. The impact phase and the free fall phase are the core judgment criteria of this application.

[0036] Step 302: Determine if the first acceleration is greater than a first preset value to determine if the current time belongs to an impact event. If it belongs to an impact event, proceed to step 303; otherwise, refresh the current time and return to step 301.

[0037] Specifically, if the first acceleration is greater than the first preset value, it is determined that the subject under test has experienced an impact event at the current time. Step 303 is executed. If the first acceleration is not greater than the first preset value, it is determined that the subject under test has not experienced an impact event at the current time, the current time is refreshed, and the process returns to step 301. Specifically, the combined acceleration during the impact phase of a fall typically exceeds 3g, with the first preset value generally set at 1.5g~2.5g. g is the acceleration due to gravity.

[0038] Optionally, the first preset value can be set to 1.5g, 2g, or 2.5g.

[0039] Preferably, the first preset value can be set to 1.5g. The reason for setting the impact threshold to 1.5g is that the peak resultant acceleration generated by normal activities such as walking and sitting generally does not exceed 1.5g, while the impact acceleration during a fall is often above 3g. Setting it to 1.5g can filter out most daily activities without missing real falls.

[0040] Step 303: Obtain the second sum acceleration sequence within the first preset time window before the current time; determine whether there is a sum acceleration lower than the second preset value in the second sum acceleration sequence to determine whether there is a free fall event; if there is a free fall event, proceed to step 304; if there is no free fall event, refresh the current time and return to step 301. Specifically, after confirming the existence of historical acceleration data (second and acceleration sequences) within the first preset time window before the impact, we determine that a freefall event occurred within the first preset time window before the impact, as long as any acceleration is lower than the second preset value, and proceed to step 304. If any acceleration is not lower than the second preset value, we determine that no freefall event occurred within the first preset time window before the impact, indicating that the impact may have been merely a rapid sitting or jumping landing, not a fall. We then refresh the current time and return to step 301.

[0041] Specifically, the first preset time window is set between 150ms and 250ms; Preferably, the first preset time window is set to 200ms. The reason for choosing 200ms is that in a typical fall, the time from the onset of weightlessness to impact with the ground is approximately between 150ms and 250ms. A 200ms window can cover the vast majority of cases.

[0042] Specifically, the second preset value is set to 0.5g~0.8g.

[0043] Preferably, the second preset value is set to 0.8g. If the combined acceleration of at least one frame of data within this window is less than 0.8g, it indicates that free fall did indeed occur before the impact, which is consistent with the kinematic characteristics of a fall.

[0044] Step 304: Obtain the tracking heat map sequence of the subject within the second preset time window; traverse the heat map, obtain the temperature of each pixel in the heat map, and extract all pixels with temperatures greater than the fourth preset value as human body pixels; obtain the temperature-weighted human body centroid ordinate by weighting the ordinate of the human body pixels based on the corresponding temperature.

[0045] Specifically, even if the previous steps confirm the existence of both free fall and impact motion, acceleration data alone may still produce false alarms. For example, quickly sitting down and then immediately standing up can also cause impact and a brief period of weightlessness. Therefore, further judgment is needed by combining this data with tracking data.

[0046] Specifically, the sequence of tracking heatmaps within a second preset time window from the occurrence of the impact event is obtained.

[0047] Specifically, the second preset time window is set to 3 seconds.

[0048] Specifically, as previously determined, free fall and impact occur within the first 200ms. However, considering that the time from the completion of the falling action to body stabilization is usually less than 1 second, and that there may be a slight delay in heat map acquisition and processing, as well as differences in the speed of movement among different people, the second preset time window is set to 3s.

[0049] Specifically, a sequence of tracking heat maps is acquired at a preset frame rate using a thermal imaging sensor; In one embodiment, the thermal imaging sensor used is the MLX90640, configured with a frame rate of 16fps. The interval between each frame is approximately 62.5ms, which is sufficient for real-time confirmation.

[0050] Specifically, a preset coordinate system is established and pixels with a value greater than the fourth preset value are used as human body pixels; Specifically, a two-dimensional coordinate system can be established with the top of the heatmap as the x-axis or with the bottom of the heatmap as the x-axis. Specifically, since the ordinate needs to be determined later, the origin of the coordinate system can be set at any position at the top or bottom of the heatmap. For ease of calculation, the top left or top right corner of the heatmap is generally used as the origin of the coordinate system. Furthermore, the coordinate system of the thermal imaging sensor (the origin of the coordinate system is generally at the lower left or upper right corner of the heat map due to different placement orientations) can be directly used as the preset coordinate system.

[0051] Specifically, the fourth preset value is the set human body temperature threshold, which is generally set between room temperature (about 20°C) and human body surface temperature (about 32°C).

[0052] Preferably, a fourth preset value is set to 26°C. All pixels with a temperature higher than this preset value are considered human body pixels to distinguish the human body area from the background area. In a real home environment, the surface temperature of background objects such as walls, floors, and furniture is usually below 26°C, while the temperature of exposed skin or clothing-covered areas is generally above 26°C. This threshold allows for a simple and effective separation of human body pixels from the background.

[0053] Furthermore, for the human body, the change in the center of mass in the vertical direction is most significant when the body is standing or lying down. The human torso has a higher temperature and the limbs have a lower temperature. Therefore, the torso's contribution to the position of the center of mass should be greater than that of the limbs. The vertical coordinate of the center of mass obtained by using temperature value as a weight better reflects the position of the main body.

[0054] Specifically, the vertical coordinate of the human body's center of mass can be obtained using the following formula: ; in, It is the vertical coordinate of the human body's center of mass. It is the ordinate of the i-th human pixel. It is the temperature value corresponding to the i-th human pixel.

[0055] Step 305: Determine whether the person being tested has fallen within the second preset time window from the current time based on the vertical coordinate of the human body's center of mass. If a fall occurs, determine that the person being tested has fallen and send an alarm to the family / medical terminal.

[0056] Specifically, a two-dimensional coordinate system is established with the top of the heat map as the x-axis; it is determined whether there is a heat map in which the vertical coordinate of the human body's center of mass is greater than the third preset value. If so, it is determined that the person to be tested has fallen to the ground within the second preset time window. As another embodiment, a two-dimensional coordinate system is established with the bottom of the heat map as the x-axis; the ordinate of the human body's centroid coordinate in the heat map is obtained, and it is determined whether the ordinate of the human body's centroid coordinate is less than a fifth preset value. If the ordinate of the human body's centroid coordinate is less than the fifth preset value, it is determined that the person to be tested has fallen to the ground within a second preset time window.

[0057] Specifically, if the person being tested does not fall, then the person being tested has not fallen.

[0058] As one embodiment, the third preset value is set to 56%n~65%n, where n is the height of the heatmap (the number of pixels).

[0059] In another embodiment, the fifth preset value is set to 35%n~44%n.

[0060] Specifically, the wearable device also integrates a heart rate monitoring device, which sends the patient's timestamp and current heart rate when an alarm is triggered.

[0061] In summary, this application leverages the sensitivity of an accelerometer to dynamic events and the ability of a thermal imaging sensor to confirm static posture, acquiring triaxial acceleration data from a triaxial accelerometer and a tracking heatmap sequence from an image tracking sensor. The two types of sensor data complement each other. By identifying the free-fall and impact phases in a fall using triaxial acceleration data, the motion state of the user being detected during the fall can be identified. Based on the tracking heatmap sequence, the posture of the person being detected can be identified, enabling accurate identification of the fall action, reducing the false positive rate, and achieving fast and reliable fall detection.

[0062] This application also provides an application scenario in which the aforementioned fall detection method based on multi-source data fusion is applied. Specifically, the fall detection method based on multi-source data fusion provided in this embodiment can be applied in scenarios such as elderly care institutions and medical monitoring, to detect whether the person to be detected has fallen.

[0063] Based on the same inventive concept, this application also provides a fall detection device based on multi-source data fusion for implementing the fall detection method based on multi-source data fusion described above. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more fall detection device embodiments based on multi-source data fusion provided below can be found in the limitations of the fall detection method based on multi-source data fusion described above, and will not be repeated here.

[0064] In one exemplary embodiment, such as Figure 4 As shown, a fall detection device based on multi-source data fusion is provided, comprising: The dynamic detection module is used to acquire the triaxial acceleration of the subject and to acquire the first acceleration of the subject based on the triaxial acceleration; wherein, the triaxial acceleration is acquired from a triaxial accelerometer, which is worn by the subject. Based on the first and acceleration, determine whether the detector has experienced an impact event at the current time; If an impact event is confirmed, obtain the second acceleration sequence within the first preset time window before the current time; determine whether the subject under test has a free fall event within the first preset time window based on the second acceleration sequence. The posture detection module is used to obtain the tracking heat map sequence of the subject within a second preset time window if a free fall event is determined; to determine whether the subject has fallen within the second preset time window based on the tracking heat map sequence; and to determine if the subject has fallen within the second preset time window if a fall event has occurred.

[0065] In one embodiment, the dynamic detection module is further configured to: Determine whether the first acceleration is greater than the first preset value. If the first acceleration is greater than the first preset value, determine that the subject under test has experienced an impact event at the current time.

[0066] In one embodiment, the dynamic detection module is further configured to: Determine whether there is a sum and acceleration value lower than the second preset value in the second sum and acceleration sequence; If there is a sum and acceleration lower than the second preset value, it is determined that the subject under test has a free fall event within the first preset time window; Otherwise, if it is determined that there is no free fall event of the subject within the first preset time window, return to the step of obtaining the three-axis acceleration of the subject and obtaining the first acceleration of the subject based on the three-axis acceleration.

[0067] In one embodiment, the attitude detection module is further configured to: Traverse the heatmap, establish a two-dimensional coordinate system with the top of the heatmap as the x-axis, and obtain the ordinate of the human centroid coordinate in the heatmap; Determine if there is a heatmap where the vertical coordinate of the human body's center of mass is greater than the third preset value. If so, determine if the person being tested has fallen within the second preset time window. Otherwise, it is determined that the person to be tested did not fall to the ground within the second preset time window.

[0068] In one embodiment, the attitude detection module is further configured to: Get the temperature of each pixel in the heat map, and extract all pixels with a temperature greater than the fourth preset value as human body pixels; The vertical coordinates of human body pixels are weighted based on the corresponding temperature to obtain the temperature-weighted vertical coordinates of the human body's centroid.

[0069] In one embodiment, the dynamic detection module is further configured to: The x-axis acceleration, y-axis acceleration, and z-axis acceleration of the subject are acquired in real time using a triaxial accelerometer, and the first acceleration is obtained based on the x-axis acceleration, y-axis acceleration, and z-axis acceleration.

[0070] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 5 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When executed by the processor, the computer program implements a fall detection method based on multi-source data fusion.

[0071] Those skilled in the art will understand that Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0072] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0073] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0074] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0075] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0076] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0077] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchain. The processors involved in the embodiments provided in this application may be, but are not limited to, general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc.

[0078] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0079] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be construed as a limitation of this application.

Claims

1. A fall detection method based on multi-source data fusion, characterized in that, The fall detection method based on multi-source data fusion includes: The triaxial acceleration of the subject to be tested is obtained, and the first acceleration of the subject to be tested is obtained based on the triaxial acceleration; wherein the triaxial acceleration is obtained according to a triaxial accelerometer, which is worn by the subject to be tested; Based on the first and acceleration, determine whether the detector has experienced an impact event at the current time; If the impact event is determined to have occurred, a second sum acceleration sequence within a first preset time window prior to the current time is obtained; based on the second sum acceleration sequence, it is determined whether the subject under test experienced a free fall event within the first preset time window. If the free fall event is confirmed, a tracking heatmap sequence of the subject within the second preset time window is obtained; based on the tracking heatmap sequence, it is determined whether the subject falls within the second preset time window; if the subject falls within the second preset time window, it is determined that the subject has fallen.

2. The fall detection method based on multi-source data fusion according to claim 1, characterized in that, The step of determining whether the detector has experienced an impact event at the current time based on the first acceleration includes: Determine whether the first acceleration is greater than a first preset value. If the first acceleration is greater than the first preset value, determine that the subject under test has experienced an impact event at the current time.

3. The fall detection method based on multi-source data fusion according to claim 1, characterized in that, The method involves determining whether the subject under test experiences a free fall event within the first preset time window based on the second and acceleration sequences; include: Determine whether there is a sum of accelerations lower than a second preset value in the second sum of acceleration sequences; If there is a sum and acceleration lower than the second preset value, it is determined that the subject under test has the free fall event within the first preset time window; Otherwise, if it is determined that the subject does not have the free fall event within the first preset time window, return to the step of obtaining the three-axis acceleration of the subject and obtaining the first acceleration of the subject based on the three-axis acceleration.

4. The fall detection method based on multi-source data fusion according to claim 1, characterized in that, The tracking heatmap sequence includes multiple frames of heatmaps of the subject being detected; The step of determining whether the subject of the test has fallen within the second preset time window based on the tracking heatmap sequence includes: Traverse the heatmap, establish a two-dimensional coordinate system with the top of the heatmap as the x-axis, and obtain the ordinate of the human body's centroid coordinate in the heatmap; Determine whether there is a heat map in which the vertical coordinate of the human body's center of mass is greater than a third preset value. If so, determine that the person to be tested experienced the fall event within the second preset time window. Otherwise, it is determined that the person to be tested did not experience the fall event within the second preset time window.

5. The fall detection method based on multi-source data fusion according to claim 4, characterized in that, The vertical coordinate of the human body's center of mass is a temperature-weighted vertical coordinate of the human body's center of mass. Obtaining the vertical coordinate of the human body's center of mass in the thermal image includes: The temperature of each pixel in the heat map is obtained, and all pixels with a temperature greater than a fourth preset value are extracted as human body pixels. The vertical coordinates of the human body pixels are weighted based on the corresponding temperature to obtain the temperature-weighted vertical coordinates of the human body's centroid.

6. The fall detection method based on multi-source data fusion according to claim 1, characterized in that, The acquisition of the first acceleration of the subject based on a triaxial accelerometer includes: The x-axis acceleration, y-axis acceleration, and z-axis acceleration of the subject to be tested are acquired in real time using a triaxial accelerometer, and the first acceleration is obtained based on the x-axis acceleration, y-axis acceleration, and z-axis acceleration.

7. A fall detection device based on multi-source data fusion, characterized in that, The fall detection device based on multi-source data fusion includes: A dynamic detection module is used to acquire the triaxial acceleration of the subject being tested, and to acquire the first acceleration of the subject being tested based on the triaxial acceleration; wherein, the triaxial acceleration is acquired based on a triaxial accelerometer, which is worn by the subject being tested; Based on the first and acceleration, determine whether the detector has experienced an impact event at the current time; If the impact event is determined to have occurred, a second sum acceleration sequence within a first preset time window prior to the current time is obtained; based on the second sum acceleration sequence, it is determined whether the subject under test experienced a free fall event within the first preset time window. The posture detection module is used to, if the free fall event is determined to exist, acquire the tracking heatmap sequence of the subject within the second preset time window; determine whether the subject has a fall event within the second preset time window based on the tracking heatmap sequence; and determine that the subject has fallen if the fall event occurs within the second preset time window.

8. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the steps of the fall detection method based on multi-source data fusion as described in any one of claims 1-6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the steps of the fall detection method based on multi-source data fusion as described in any one of claims 1-6.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the fall detection method based on multi-source data fusion as described in any one of claims 1-6.

Images