Indoor human body detection method and 5G electronic device

Abstract

The present invention discloses an indoor human body detection method and 5G electronic device, wherein the 5G electronic device includes a data acquisition unit, a data storage unit, a data processing unit, a 5G communication unit, an alarm unit, and a front-end control unit. The data acquisition unit is communicatively connected to the data storage unit, the data storage unit is communicatively connected to the data processing unit, the data acquisition unit, the data storage unit, the data processing unit, and the alarm unit are respectively communicatively connected to the front-end control unit, and the front-end control unit is communicatively connected to a remote device via the 5G communication unit. The present invention utilizes changes in radar point cloud images of side observations of the human body and radar point cloud images of upward observations of the human body and speed data related to the human body to analyze and judge human body posture, thereby reducing misjudgments in human posture detection.

Description

Technical Field

[0001] The present invention relates to the field of indoor human body monitoring technology, and more specifically to an indoor human body detection method and a 5G electronic device. Background Art

[0002] People of different ages suffer different injuries when they fall, and different types of falls also produce different injuries. For example, elderly people are more likely to suffer fractures when they fall, while young people and adolescents are less likely to suffer such injuries. For another example, when falling face-first or sideways, people subconsciously bend their knees and curl up to reduce the severity of the fall. However, when falling on their back, it is difficult to take other effective protective measures other than trying to keep their back in contact with the ground first.

[0003] Many elderly people and young children fall indoors without being promptly discovered and treated, resulting in permanent injuries or serious consequences. To reduce the workload of caregivers and alleviate the burden on children and parents, indoor fall detection systems have been developed. Existing fall detection systems often analyze falls based on changes in point cloud images, which can lead to misjudgments. For example, if an elderly person squats and then sits on the ground, the changes in the point cloud image indicate a change in their posture, similar to the changes in the point cloud image when a person falls. If this phenomenon is not considered a fall, further misjudgments can occur. For example, if an elderly person's legs are insufficiently strong during a squat, their buttocks may hit the ground, creating the illusion that they are sitting. For those with weak waist strength or poor spinal strength, this difference in posture can still cause harm. Summary of the Invention

[0004] To this end, the technical problem to be solved by the present invention is to provide an indoor human body detection method and a 5G electronic device, which uses the changes in the human body side observation radar point cloud map and the human body upward observation radar point cloud map and the speed data related to the human body to analyze and judge the human body posture, thereby reducing misjudgment in human body posture detection.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] A method for indoor human body detection comprises the following steps:

[0007] (S100) collecting radar data of a human body from above and sides in real time;

[0008] (S200) Generate a radar point cloud image of the side view of the human body and a radar point cloud image of the upward view of the human body using the radar data collected in step (S100), and calculate the velocity data V related to the human body.人 , wherein the speed data includes the horizontal moving speed V1 of the whole body, the horizontal moving speed V2 of the upper body, the horizontal moving speed V3 of the lower body, the angular moving speed ω1 of the upper body, and the angular moving speed ω2 of the lower body;

[0009] (S300) Using the human body side observation radar point cloud map, human body upward observation radar point cloud map and speed data V obtained in step (S200) 人 Analyze and judge the human body posture to determine whether someone has fallen. The specific judgment steps are as follows:

[0010] (S310) first determine the change in the human body posture based on the side observation radar point cloud image and the upward observation radar point cloud image of the human body obtained in step (S200), and when the human body posture changes from standing upright to lying down or lying down, execute step (S320);

[0011] (S320) Counting the time t when the human body posture changes from standing to lying down or lying down 变 When the difference ΔV between the horizontal movement speed V2 of the upper body and the horizontal movement speed V3 of the lower body exceeds the threshold value V0 and the angular movement speed ω1 of the upper body and the angular movement speed ω2 of the lower body both exceed the threshold value ω0, a preliminary judgment result is given;

[0012] (S330) For time t 变 The regularity analysis of the changes in the human body side observation radar point cloud map and the human body upward observation radar point cloud map is performed. When the changes in the human body side observation radar point cloud map and the human body upward observation radar point cloud map are both irregular changes, the final judgment result is given and the judgment time t 变 A fall occurs when the body posture changes from an upright position to a prone or lying position.

[0013] In the above-mentioned indoor human body detection method, in step (S330), time t 变 When analyzing the regularity of the changes in the radar point cloud images of the side observation of the human body and the changes in the radar point cloud images of the upward observation of the human body, the time t 变 Split into t 变1 and t 变2 And t 变1 :t 变2 is 10~5:1, then t 变1 and t 变2 Compare and analyze the changes in the radar point cloud map of the human body observed from the side and the radar point cloud map of the human body observed from above. The specific operations are as follows:

[0014] (S331) 变1The time period is divided into m time periods, and the human body side observation radar point cloud image and the human body upward observation radar point cloud image in the m time period are compared. When the similarity difference between the human body side observation radar point cloud images in two consecutive adjacent time periods is greater than the threshold SV1 or the similarity difference between the human body upward observation radar point cloud images in two consecutive adjacent time periods is greater than the threshold SV2, it is considered that time t 变1 The changes in the radar point cloud image of the side observation of the human body and the changes in the radar point cloud image of the upward observation of the human body are both irregular changes, where m is a natural number greater than or equal to 5;

[0015] (S332) 变2 The time segment is divided into n time segments, and the human body side observation radar point cloud image and the human body upward observation radar point cloud image in the n time segments are compared. When the similarity difference between the human body side observation radar point cloud images in two consecutive adjacent time segments is greater than the threshold SV3 or the similarity difference between the human body upward observation radar point cloud images in two consecutive adjacent time segments is greater than the threshold SV4, it is considered that time t 变2 The changes in the radar point cloud image of the lateral observation of the human body and the changes in the radar point cloud image of the upward observation of the human body are both irregular changes, where n is a natural number greater than or equal to 3.

[0016] In the above-mentioned indoor human detection method, in step (S200), when the vertical size change of the side-view radar point cloud image of the human body exceeds 10% or / and the horizontal size change of the upward-view radar point cloud image of the human body exceeds 15% within the unit time T, the velocity data V related to the human body is calculated. 人 Otherwise, the speed data V related to the human body is not calculated. 人 , wherein the unit time T is a preset parameter value greater than or equal to 10ms and less than or equal to 50ms.

[0017] In the above-mentioned indoor human body detection method, when calculating the angular velocity ω1 of the upper body and the angular velocity ω2 of the lower body in step (S200), ω1 and ω2 are corrected according to the following strategy:

[0018] (CL1) When the human body begins to tilt, if the direction of ω1 is the same as the overall horizontal movement direction of the human body, the following formula is used to correct ω1 and ω2:

[0019]

[0020]

[0021] Wherein, σ1 and σ2 are parameters related to the human body forward lean angle threshold α, and σ1 and σ2 are calculated by the following formula:

[0022]

[0023]

[0024] (CL2) When the human body begins to tilt, if the direction of ω1 is opposite to the overall horizontal movement of the human body, the following formula is used to correct ω1 and ω2:

[0025]

[0026]

[0027] Where, σ3 and σ4 are parameters related to the human body backward tilt angle threshold β, and σ3 and σ4 are calculated by the following formula:

[0028]

[0029] .

[0030] In the above-mentioned indoor human body detection method, when ω1 and ω2 are in the same direction, a first-level risk fall is identified in the final judgment result given in step (S330); otherwise, a second-level risk fall is identified, wherein the degree of harm of a first-level risk fall is higher than that of a second-level risk fall.

[0031] In the above-mentioned indoor human body detection method, in step (S300), the human body side observation radar point cloud map, the human body upward observation radar point cloud map and the speed data V obtained in step (S200) are used. 人 When analyzing and judging the human body posture and determining that someone has fallen, the direction of the person's fall is analyzed using the side observation radar point cloud map and the upward observation radar point cloud map to determine the risk of head collision, and mark it in the final judgment result.

[0032] In the above-mentioned indoor human body detection method, in step (S300), the human body side observation radar point cloud map, the human body upward observation radar point cloud map and the speed data V obtained in step (S200) are used. 人 When analyzing and judging the human body posture and determining that someone has fallen, calculate the time t 变 The value of time t 变 The numerical value of the fall risk level is marked in the final judgment result.

[0033] 5G electronic devices that use the above-mentioned indoor human detection method for human detection include:

[0034] A data acquisition unit is used to collect radar data of the human body; the data acquisition unit includes an upward 4D millimeter-wave radar installed on the roof and a side 4D millimeter-wave radar installed on the upper half of the side wall;

[0035] A data storage unit, used to store radar data collected by the data acquisition unit;

[0036] The data processing unit is used to process the radar data collected by the data acquisition unit; the data processing unit includes a radar point cloud image generation module, a module for calculating the speed data V related to the human body, and a data processing unit for processing the radar data collected by the data acquisition unit. 人 The speed calculation module and the radar point cloud image generated by the radar point cloud image generation module and the speed data V calculated by the speed calculation module 人 an analysis module to perform the analysis;

[0037] 5G communication unit, used to send alarm information to remote devices, where the remote devices include servers, handheld smart terminals and networked computers;

[0038] An alarm unit, used for issuing an alarm through sound;

[0039] Front-end control unit, used to control the data acquisition unit, data storage unit, data processing unit, 5G communication unit and alarm unit;

[0040] The data acquisition unit is communicatively connected to the data storage unit, the data storage unit is communicatively connected to the data processing unit, the data acquisition unit, the data storage unit, the data processing unit and the alarm unit are communicatively connected to the front-end control unit respectively, and the front-end control unit is communicatively connected to the remote device through the 5G communication unit.

[0041] For the above-mentioned 5G electronic devices, the alarm unit also provides an alarm by flashing indicator lights.

[0042] In the above-mentioned 5G electronic device, the data acquisition unit is communicatively connected to the data storage unit via the UWB communication module.

[0043] The technical solution of the present invention achieves the following beneficial technical effects:

[0044] 1. The present invention utilizes the changes in the human body side observation radar point cloud map and the human body upward observation radar point cloud map, as well as the changes in the human body's overall horizontal movement speed, the human body's upper body horizontal movement speed, the human body's lower body horizontal movement speed, the human body's upper body movement angular velocity, and the human body's lower body movement angular velocity when the above two point cloud maps change, to judge whether a person indoors has fallen, thereby avoiding misjudgment when judging whether a person has fallen solely based on the changes in the radar point cloud map, and effectively detecting low-level falls of the elderly. Among them, low-level falls refer to falls that occur when the human body is in a half-squatting or squatting state. Although this type of fall is less harmful than a high-level fall, it can also cause more serious injuries to some elderly people.

[0045] 2. The present invention divides the time periods in which the human body lateral observation radar point cloud map and the human body upward observation radar point cloud map change, and analyzes the regularity of the changes in the human body lateral observation radar point cloud map and the human body upward observation radar point cloud map in different time periods to ensure that the human body's fall is not intentional. Specifically, a person leans forward, then bends his knees, and then lies on the ground. At this time, based on the changes in the human body lateral observation radar point cloud map and the human body upward observation radar point cloud map, as well as the changes in the overall horizontal movement speed of the human body, the horizontal movement speed of the human body upper body, the horizontal movement speed of the human body lower body, the angular velocity of the human body upper body movement, and the angular velocity of the human body lower body movement, there is also the possibility of judging that the human body has fallen. However, the changes in the human body lateral observation radar point cloud map and the human body upward observation radar point cloud map in this process have a regularity that can be followed.

[0046] 3. Identifying the risk level of a fall in the final judgment result will help people take appropriate measures based on the risk level and avoid worsening injuries caused by untimely treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 Schematic diagram of the working principle of 5G electronic equipment capable of indoor human body detection;

[0048] Figure 2 Flowchart for wireless positioning of 5G terminal devices;

[0049] Figure 3 Schematic diagram of the calculation principle of the relative angular velocity of two points during translation. DETAILED DESCRIPTION

[0050] like Figure 1 As shown, the present invention provides a 5G electronic device that can perform indoor human body detection, which includes a data acquisition unit, a data storage unit, a data processing unit, a 5G communication unit, an alarm unit and a front-end control unit. The data acquisition unit is communicatively connected to the data storage unit through a UWB communication module, and the data storage unit is communicatively connected to the data processing unit. The data acquisition unit, the data storage unit, the data processing unit and the alarm unit are respectively communicatively connected to the front-end control unit, and the front-end control unit is communicatively connected to the remote device through the 5G communication unit, wherein the remote device includes a server, a handheld smart terminal and a networked computer, etc.

[0051] In the present invention, the data acquisition unit is used to collect radar data of the human body; the data acquisition unit includes an upward 4D millimeter wave radar installed on the roof and a side 4D millimeter wave radar installed on the upper half of the side wall; the data storage unit is used to store the radar data collected by the data acquisition unit; the data processing unit is used to process the radar data collected by the data acquisition unit; the data processing unit includes a radar point cloud map generation module, a module for calculating the speed data V related to the human body, and a data storage unit for storing the radar data collected by the data acquisition unit. 人 The speed calculation module and the radar point cloud image generated by the radar point cloud image generation module and the speed data V calculated by the speed calculation module 人 An analysis module for performing analysis; a 5G communication unit for sending alarm information to a remote device; an alarm unit for issuing an alarm through sound and flashing indicator lights; and a front-end control unit for controlling the data acquisition unit, data storage unit, data processing unit, 5G communication unit, and alarm unit.

[0052] When the 5G electronic device in the present invention is applied to indoor human body detection, an upward 4D millimeter-wave radar and a side 4D millimeter-wave radar should be set up in the indoor structure, and these two 4D millimeter-wave radars should be communicatively connected to the data storage unit. At the same time, the data storage unit, data processing unit, alarm unit, 5G communication unit and front-end control unit should be centrally arranged in an equipment box.

[0053] like Figure 2 As shown, in a room where the 5G electronic device of the present invention is installed, human body detection in the room is achieved through the following steps:

[0054] (S100) Radar data of a human body is collected from above and sides in real time.

[0055] When using 4D millimeter-wave radar to collect radar data of the human body, sensors can be set at the entrance and exit of the room. When someone enters the room, the 4D millimeter-wave radar is started to work. When no one enters the room, the 4D millimeter-wave radar can be put into standby mode. This not only reduces the energy consumption of the equipment, but also extends the service life of the equipment.

[0056] (S200) Generate a radar point cloud image of the side view of the human body and a radar point cloud image of the upward view of the human body using the radar data collected in step (S100), and calculate the velocity data V related to the human body. 人 , where the speed data includes the overall horizontal movement speed V1 of the human body, the horizontal movement speed V2 of the upper body of the human body, the horizontal movement speed V3 of the lower body of the human body, the angular movement speed ω1 of the upper body of the human body, and the angular movement speed ω2 of the lower body of the human body.

[0057] The velocity data V related to the human body is calculated using the radar data collected in step (S100). 人The Doppler effect principle is used to calculate. Among them, the angular velocity of the upper body ω1 and the angular velocity of the lower body ω2 can be calculated using the horizontal displacement difference of two vertical points, such as Figure 3 As shown, suppose that two points A and B on an object are in the same vertical direction and are separated by L1. Within time t, point A has a horizontal displacement of L2 and point B has a horizontal displacement of L3. Then the angular velocity of point B relative to point A is ω A-B It can be calculated by the following formula:

[0058]

[0059]

[0060] Figure 3 In the figure, the dotted line AB represents the initial position of the object, the solid line AB represents the position of the object after translation time t, and θ represents the angle of rotation of B relative to A.

[0061] In the present invention, the angular velocity ω1 of the upper body and the angular velocity ω2 of the lower body can be calculated with the human foot as the center of the circle, or the angular velocity ω1 of the upper body and the angular velocity ω2 of the lower body can be calculated with the human waist as the center of the circle, and preferably, the angular velocity ω1 of the upper body and the angular velocity ω2 of the lower body are calculated with the human foot as the center of the circle.

[0062] (S300) Using the human body side observation radar point cloud map, human body upward observation radar point cloud map and speed data V obtained in step (S200) 人 Analyze and judge the human body posture to determine whether someone has fallen. The specific judgment steps are as follows:

[0063] (S310) First, determine the change in the human body posture based on the side observation radar point cloud image of the human body and the upward observation radar point cloud image of the human body obtained in step (S200). When the human body posture changes from upright to lying down or lying down, execute step (S320).

[0064] When a person's posture changes, the shape and area of ​​the point cloud image obtained by the upward-facing 4D millimeter-wave radar and the side-facing 4D millimeter-wave radar scanning the body may change, especially when a person changes from an upright position to a fallen position. However, if the shape and area of ​​the point cloud image obtained by the upward-facing 4D millimeter-wave radar and the side-facing 4D millimeter-wave radar scan the body change, it does not mean that the person's posture has changed from upright to fallen. The overall and local motion parameters of the human body during the posture change must be calculated to confirm the change in posture based on this calculation.

[0065] (S320) Counting the time t when the human body posture changes from standing to lying down or lying down 变 When the difference ΔV between the horizontal movement speed V2 of the upper body and the horizontal movement speed V3 of the lower body exceeds the threshold value V0 and the angular movement speed ω1 of the upper body and the angular movement speed ω2 of the lower body both exceed the threshold value ω0, a preliminary judgment result is given.

[0066] When human body posture changes from upright state to lying down or lying down, by calculating the motion parameters of human body as a whole and part, such as human upper body horizontal movement speed V2, human lower body horizontal movement speed V3, human upper body movement angular velocity ω1 and human lower body movement angular velocity ω2, then by comparing with a preset threshold value, human body posture change is preliminarily confirmed. In this step, human body posture change is not finally confirmed, the reason is that when conscious falling, human body posture change and when unconscious falling, human body posture change have regularity differences, that is, when people are consciously falling, human body posture changes from upright to lying down or lying down in the process, the motion of human trunk and limbs has certain regularity, reflected on the radar point cloud map, the change of radar point cloud map also has certain regularity, and when unconscious falling, the motion of human trunk and limbs does not have regularity at all, especially when the elderly fall. In the present invention, conscious falling refers to people's active falling, such as curling up and rolling forward and falling, while unconscious falling refers to accidental falling.

[0067] (S330) For time t 变 The regularity analysis of the changes in the human body side observation radar point cloud map and the human body upward observation radar point cloud map is performed. When the changes in the human body side observation radar point cloud map and the human body upward observation radar point cloud map are both irregular changes, the final judgment result is given and the judgment time t 变 A fall occurs when the body posture changes from an upright position to a prone or lying position.

[0068] In this step, the time t 变 When analyzing the regularity of the changes in the radar point cloud images of the side observation of the human body and the changes in the radar point cloud images of the upward observation of the human body, the time t 变 Split into t 变1 and t 变2 And t 变1 :t 变2 is 10~5:1, then t 变1 and t 变2 Compare and analyze the changes in the radar point cloud map of the human body observed from the side and the radar point cloud map of the human body observed from above. The specific operations are as follows:

[0069] (S331) 变1The time period is divided into m time periods, and the human body side observation radar point cloud image and the human body upward observation radar point cloud image in the m time period are compared. When the similarity difference between the human body side observation radar point cloud images in two consecutive adjacent time periods is greater than the threshold SV1 or the similarity difference between the human body upward observation radar point cloud images in two consecutive adjacent time periods is greater than the threshold SV2, it is considered that time t 变1 The changes in the radar point cloud image of the side observation of the human body and the changes in the radar point cloud image of the upper observation of the human body are both irregular changes, wherein m is a natural number greater than or equal to 5, preferably, m is 5;

[0070] (S332) 变2 The time segment is divided into n time segments, and the human body side observation radar point cloud image and the human body upward observation radar point cloud image in the n time segments are compared. When the similarity difference between the human body side observation radar point cloud images in two consecutive adjacent time segments is greater than the threshold SV3 or the similarity difference between the human body upward observation radar point cloud images in two consecutive adjacent time segments is greater than the threshold SV4, it is considered that time t 变2 The changes in the radar point cloud map of the inner human body side observation and the changes in the radar point cloud map of the human body upward observation are both irregular changes, where n is a natural number greater than or equal to 3, and preferably, n is 4.

[0071] During an unconscious fall, it takes less than 1 second for the human body to change from an upright posture to a prone or lying posture. In the early stage of the fall, the human body falls slowly, but the shape changes of the human body's lateral observation radar point cloud map and the human body's upward observation radar point cloud map are more obvious. In the later stage of the fall, the human body falls quickly, but the shape changes of the human body's lateral observation radar point cloud map and the human body's upward observation radar point cloud map are relatively insignificant. The reason is that during an unconscious fall, when a person feels that his center of gravity is unstable, he will choose to grab something. At this time, the upper limbs will swing irregularly. When he is about to fall completely, he will choose to take protective measures. At this time, the upper limbs will not swing irregularly, but will look for the ground to seek support for the body. This will cause the shape changes of the human body's lateral observation radar point cloud map and the human body's upward observation radar point cloud map to be more obvious in the early stage of the fall, and the shape changes of the human body's lateral observation radar point cloud map and the human body's upward observation radar point cloud map to be relatively insignificant in the later stage of the fall. Therefore, the present invention will t 变 The image is divided into two time periods, and the shape changes of the radar point cloud images within the two time periods are analyzed to more accurately determine whether it is a fall.

[0072] The threshold values ​​SV1, SV2, SV3 and SV4 can be obtained through fall simulation analysis, and then adjusted based on the threshold values ​​obtained through the fall simulation analysis according to actual application requirements.

[0073] In addition, since the elderly sway their bodies significantly when walking, the radar point cloud image formed by the radar scan of the elderly will change to a certain extent. Therefore, in step (S200), when the vertical size change of the radar point cloud image of the human body observed from the side exceeds 10% or / and the horizontal size change of the radar point cloud image of the human body observed from the top exceeds 15% within the unit time T, the velocity data V related to the human body is calculated. 人 Otherwise, the speed data V related to the human body is not calculated. 人 , where the unit time T is a preset parameter value greater than or equal to 10ms and less than or equal to 50ms. This reduces the computational burden on the data processing unit. Furthermore, the vertical size change threshold for the side-view radar point cloud image of the human body and the horizontal size change threshold for the upward-view radar point cloud image of the human body can be set as needed to meet actual needs.

[0074] When a person is about to fall, the critical values ​​for the forward and lateral tilt angles are both greater than the backward tilt angle. This means that when a person is tilted, they are more likely to fall when leaning backward. Therefore, when determining whether a person has fallen, the common sense that the critical values ​​for the forward and lateral tilt angles are both greater than the backward tilt angle should be considered. Specifically, when calculating the angular velocity ω1 of the upper body and the angular velocity ω2 of the lower body in step (S200), ω1 and ω2 should be modified according to the following strategy:

[0075] (CL1) When the human body begins to tilt, if the direction of ω1 is the same as the overall horizontal movement direction of the human body, the following formula is used to correct ω1 and ω2:

[0076]

[0077]

[0078] Where, σ1 and σ2 are parameters related to the human body's forward lean angle threshold α, and σ1 and σ2 are calculated by the following formula:

[0079]

[0080]

[0081] (CL2) When the human body begins to tilt, if the direction of ω1 is opposite to the overall horizontal movement of the human body, the following formula is used to correct ω1 and ω2:

[0082]

[0083]

[0084] Where, σ3 and σ4 are parameters related to the human body backward tilt angle threshold β, and σ3 and σ4 are calculated by the following formula:

[0085]

[0086] .

[0087] By revising ω1 and ω2, it is possible to avoid falling down and being judged as quickly lying down or lying down. The reason is that, when falling forward, when the human body's center of gravity has shifted forward, ω1 and ω2 will all increase rapidly. On the contrary, when falling backward, the human body's center of gravity can only shift slightly backward, and ω1 and ω2 will then appear to slowly increase first and then rapidly increase. Moreover, for taller people, it is not necessary to revise ω1 and ω2. It is still possible to obtain an ω1 and ω2 that exceeds a threshold value when falling down. For shorter people, especially hunched old men, when falling down, there is a radar that cannot measure an ω1 and ω2 that exceeds a threshold value. Therefore, the present invention utilizes the human body forward tilt angle threshold value and the human body backward tilt angle threshold value to revise the ω1 and ω2 when falling forward and falling backward.

[0088] Since there are significant differences in the degree of harm that may be caused by falling forward, falling backward, and falling sideways (generally, falling backward causes greater harm), when ω1 and ω2 are in the same direction, a level one risk fall is identified in the final judgment result given in step (S330); otherwise, a level two risk fall is identified, wherein the level one risk fall has a higher degree of harm than the level two risk fall.

[0089] In order to remind the user to take timely measures according to the actual fall situation, in the present invention, in step (S300), the human body side observation radar point cloud map, the human body upward observation radar point cloud map and the speed data V obtained in step (S200) are used. 人 When analyzing and judging the human body posture and determining that someone has fallen, the direction of the person's fall is analyzed using the side observation radar point cloud map and the upward observation radar point cloud map to determine the risk of head collision, and mark it in the final judgment result.

[0090] At the same time, the human body side observation radar point cloud map, human body upward observation radar point cloud map and speed data V obtained in step (S200) are also used. 人 When analyzing and judging the human body posture and determining that someone has fallen, calculate the time t 变 The value of time t 变 The numerical value of the fall risk level is marked in the final judgment result.

[0091] By accurately identifying falls, the present invention can effectively reduce the extra burden on caregivers or guardians due to system misjudgment. At the same time, by predicting and marking the risks of falls, it is beneficial for caregivers or guardians to take corresponding measures according to the risk level, such as rushing to the scene, calling emergency services, or selecting medical assistance tools. It also makes it easier for people to estimate and predict the possible dangers of falling people based on experience.

[0092] Obviously, the above embodiments are merely examples for clarity of explanation and are not intended to limit the implementation methods. Those skilled in the art will readily appreciate that other variations or modifications based on the above descriptions are possible. It is not necessary and impossible to enumerate all implementation methods here. Obvious variations or modifications arising therefrom remain within the scope of protection of the claims of this patent application.

Claims

1. A method for detecting human body indoors, characterized in that: The steps include: The S100 collects radar data of the human body from above and to the side in real time; S200 uses the radar data collected in step S100 to generate a radar point cloud image of the side view of the human body and a radar point cloud image of the upward view of the human body, and calculates the speed data V related to the human body. 人 , wherein the speed data includes the horizontal moving speed V1 of the whole body, the horizontal moving speed V2 of the upper body, the horizontal moving speed V3 of the lower body, the angular moving speed ω1 of the upper body, and the angular moving speed ω2 of the lower body; S300 uses the human body side observation radar point cloud map, human body upward observation radar point cloud map and speed data V obtained in step S200 人 Analyze and judge the human body posture to determine whether someone has fallen. The specific judgment steps are as follows: S310 first determines the change of the human body posture based on the side observation radar point cloud image and the upward observation radar point cloud image of the human body obtained in step S200. When the human body posture changes from standing upright to lying down or lying down, step S320 is executed; S320 counts the time t when the human body posture changes from standing to lying down or lying down 变 When the difference ΔV between the horizontal movement speed V2 of the upper body and the horizontal movement speed V3 of the lower body exceeds the threshold value V0 and the angular movement speed ω1 of the upper body and the angular movement speed ω2 of the lower body both exceed the threshold value ω0, a preliminary judgment result is given; S330 versus time t 变 The regularity analysis of the changes in the human body side observation radar point cloud map and the human body upward observation radar point cloud map is performed. When the changes in the human body side observation radar point cloud map and the human body upward observation radar point cloud map are both irregular changes, the final judgment result is given and the judgment time t 变 A fall occurs when the body posture changes from an upright position to a prone or lying position.

2. The indoor human body detection method according to claim 1, characterized in that: In step S330, the time t 变 When analyzing the regularity of the changes in the radar point cloud images of the side observation of the human body and the changes in the radar point cloud images of the upward observation of the human body, the time t 变 Split into t 变1 and t 变2 And t 变1 :t 变2 is 10~5:1, then t 变1 and t 变2 Compare and analyze the changes in the radar point cloud map of the human body observed from the side and the radar point cloud map of the human body observed from above. The specific operations are as follows: S331 will be 变1 The time period is divided into m time periods, and the human body side observation radar point cloud image and the human body upward observation radar point cloud image in the m time period are compared. When the similarity difference between the human body side observation radar point cloud images in two consecutive adjacent time periods is greater than the threshold SV1 or the similarity difference between the human body upward observation radar point cloud images in two consecutive adjacent time periods is greater than the threshold SV2, it is considered that time t 变1 The changes in the radar point cloud image of the side observation of the human body and the changes in the radar point cloud image of the upward observation of the human body are both irregular changes, where m is a natural number greater than or equal to 5; S332 will be t 变2 The time segment is divided into n time segments, and the human body side observation radar point cloud image and the human body upward observation radar point cloud image in the n time segments are compared. When the similarity difference between the human body side observation radar point cloud images in two consecutive adjacent time segments is greater than the threshold SV3 or the similarity difference between the human body upward observation radar point cloud images in two consecutive adjacent time segments is greater than the threshold SV4, it is considered that time t 变2 The changes in the radar point cloud image of the lateral observation of the human body and the changes in the radar point cloud image of the upward observation of the human body are both irregular changes, where n is a natural number greater than or equal to 3.

3. The indoor human body detection method according to claim 1, characterized in that: In step S200, when the vertical size change of the human body side observation radar point cloud image exceeds 10% or / and the horizontal size change of the human body upward observation radar point cloud image exceeds 15% within the unit time T, the velocity data V related to the human body is calculated. 人 Otherwise, the speed data V related to the human body is not calculated. 人 , wherein the unit time T is a preset parameter value greater than or equal to 10ms and less than or equal to 50ms.

4. The indoor human body detection method according to claim 1, characterized in that: When calculating the angular velocity ω1 of the upper body and the angular velocity ω2 of the lower body in step S200, ω1 and ω2 are corrected according to the following strategy: When the CL1 human body begins to tilt, the direction of ω1 is the same as the overall horizontal movement direction of the human body. Then use the following formula to correct ω1 and ω2: Where, σ1 and σ2 are parameters related to the human body's forward lean angle threshold α, and σ1 and σ2 are calculated by the following formula: When the CL2 human body begins to tilt, the direction of ω1 is opposite to the overall horizontal movement direction of the human body. Then use the following formula to correct ω1 and ω2: Where, σ3 and σ4 are parameters related to the human body backward tilt angle threshold β, and σ3 and σ4 are calculated by the following formula: 。 5. The indoor human body detection method according to claim 1, characterized in that: When ω1 and ω2 are in the same direction, a first-level risk fall is identified in the final judgment result given in step S330. Otherwise, a second-level risk fall is identified. The degree of harm caused by a first-level risk fall is higher than that by a second-level risk fall.

6. The indoor human body detection method according to claim 1, characterized in that: In step S300, the human body side observation radar point cloud image, human body upward observation radar point cloud image and speed data V obtained in step S200 are used to calculate the human body side observation radar point cloud image and human body upward observation radar point cloud image. 人 When analyzing and judging the human body posture and determining that someone has fallen, the direction of the person's fall is analyzed using the side observation radar point cloud map and the upward observation radar point cloud map to determine the risk of head collision, and mark it in the final judgment result.

7. The indoor human body detection method according to claim 1, characterized in that: In step S300, the human body side observation radar point cloud image, human body upward observation radar point cloud image and speed data V obtained in step S200 are used to calculate the human body side observation radar point cloud image and human body upward observation radar point cloud image. 人 When analyzing and judging the human body posture and determining that someone has fallen, calculate the time t 变 The value of time t 变 The numerical value of the fall risk level is marked in the final judgment result.

8. A 5G electronic device that performs human body detection using the indoor human body detection method according to any one of claims 1 to 7, characterized in that: include: A data acquisition unit, used to collect radar data of the human body; The data acquisition unit includes an upward-facing 4D millimeter-wave radar installed on the roof and a side-facing 4D millimeter-wave radar installed on the upper half of the side wall; A data storage unit, used to store radar data collected by the data acquisition unit; The data processing unit is used to process the radar data collected by the data acquisition unit; the data processing unit includes a radar point cloud image generation module, a module for calculating the speed data V related to the human body, and a data processing unit for processing the radar data collected by the data acquisition unit. 人 The speed calculation module and the radar point cloud image generated by the radar point cloud image generation module and the speed data V calculated by the speed calculation module 人 an analysis module to perform the analysis; 5G communication unit, used to send alarm information to remote devices, where the remote devices include servers, handheld smart terminals and networked computers; An alarm unit, used for issuing an alarm through sound; Front-end control unit, used to control the data acquisition unit, data storage unit, data processing unit, 5G communication unit and alarm unit; The data acquisition unit is communicatively connected to the data storage unit, the data storage unit is communicatively connected to the data processing unit, the data acquisition unit, the data storage unit, the data processing unit and the alarm unit are communicatively connected to the front-end control unit respectively, and the front-end control unit is communicatively connected to the remote device through the 5G communication unit.

9. The 5G electronic device according to claim 8, characterized in that The alarm unit also provides warnings by flashing indicator lights.

10. The 5G electronic device according to claim 8, wherein: The data acquisition unit is connected to the data storage unit through the UWB communication module.

Images