Fall detection method and device

Abstract

The embodiment of the invention provides a fall detection method and device. The tumble detection method comprises the following steps: in a first stage, performing tumble detection according to radar point cloud information, and entering a second stage under the condition that a tumble event is not detected; in the second stage, tumble detection is carried out according to the radar point cloud information and the radar angle FFT information, and the third stage is entered under the condition that no tumble event is detected; and in the third stage, fall detection is carried out according to the radar angle FFT information. According to the embodiment of the invention, in different stages where tumble may occur, the radar point cloud information and the radar angle FFT information are used for tumble detection to judge whether a tumble event occurs, so that the tumble detection speed and reliability are improved.

Description

Technical Field

[0001] This application relates to the field of life detection, and particularly to a fall detection method and device. Background Art

[0002] Currently, the number and proportion of the elderly population are increasing continuously, and the global population is entering the aging stage. Paying attention to the physical health and quality of life of the elderly and solving the problems faced by the elderly have important social and economic significance. Fall accidents pose great harm to the physical health and life safety of the elderly. According to the statistics of the World Health Organization of the United Nations, falls are the second leading cause of unintentional injury death worldwide. Fall detection technology helps to quickly discover fall accidents, implement rescue in a timely manner, and prevent the injury from worsening.

[0003] Common fall detection solutions include video-based detection technologies and detection methods using wearable devices, etc. Video-based fall detection technology uses intelligent detection methods to analyze video data to detect fall events. However, obtaining video data requires good lighting conditions, so this technology cannot be used in a dark environment at night. In addition, the camera seriously exposes people's privacy and cannot be deployed in private environments such as bedrooms and bathrooms. The detection method using wearable devices is to analyze the motion characteristics of users using sensors such as gyroscopes and accelerometers to determine whether a fall has occurred. This method requires the user to wear the device to work. Due to the comfort problem and frequent charging problem of wearable devices, the acceptance of users is low.

[0004] It should be noted that the above introduction of the technical background is only for the convenience of clearly and completely explaining the technical solution of this application and facilitating the understanding of those skilled in the art. It cannot be considered that the above technical solutions are well-known to those skilled in the art just because these solutions are described in the background art part of this application. Summary of the Invention

[0005] The inventors found that the fall detection technology based on millimeter-wave radar does not require the user to wear a device and does not require light, has the advantage of protecting privacy, can be applied to private places such as bedrooms and bathrooms, and has good market prospects. The fall detection technology based on radar can judge whether a fall has occurred by analyzing the motion characteristics of the human body and using methods such as machine learning or template matching; it can also detect whether a person is lying on the ground through the signal fluctuation of the radar, and then judge whether a fall has occurred. However, the former is prone to false negative problems, and the latter requires a long detection time and is prone to false detection. In addition, the information that the radar can provide is relatively small, and it is also a difficult problem to quickly and reliably detect falls.

[0006] To address at least one of the above problems or other similar problems, an embodiment of the present application provides a fall detection method and device. It uses radar point cloud information and radar angle FFT information to detect falls at different stages where a fall may occur, so as to determine whether a fall event has occurred, thereby improving the speed and reliability of fall detection.

[0007] According to one aspect of the embodiments of the present application, there is provided a fall detection device, the device comprising:

[0008] A detection unit, which in the first stage performs fall detection based on radar point cloud information, and enters the second stage if no fall event is detected; in the second stage, performs fall detection based on radar point cloud information and radar angle FFT information, and enters the third stage if no fall event is detected; and in the third stage, performs fall detection based on radar angle FFT information.

[0009] According to another aspect of the embodiments of the present application, there is provided a fall detection method, the method comprising:

[0010] In the first stage, perform fall detection based on radar point cloud information, and enter the second stage if no fall event is detected;

[0011] In the second stage, perform fall detection based on radar point cloud information and radar angle FFT information, and enter the third stage if no fall event is detected;

[0012] In the third stage, perform fall detection based on radar angle FFT information.

[0013] According to still another aspect of the embodiments of the present application, there is provided a computer device, the computer device comprising a memory and a processor, the memory storing a computer program, and the processor being configured to execute the computer program to implement the method as described above.

[0014] According to yet another aspect of the embodiments of the present application, there is provided a storage medium storing a computer-readable program, the computer-readable program causing a computer to execute the method as described above.

[0015] One of the beneficial effects of the embodiments of the present application is that: according to the embodiments of the present application, at different stages where a fall may occur, radar point cloud information and radar angle FFT information are used to detect falls to determine whether a fall event has occurred, thereby improving the speed and reliability of fall detection.

[0016] Specific embodiments of the present application are disclosed in detail with reference to the following description and the accompanying drawings, indicating the ways in which the principles of the present application can be employed. It should be understood that the embodiments of the present application are not limited in scope thereby. Within the scope of the terms of the appended claims, the embodiments of the present application include many variations, modifications, and equivalents.

[0017] Features described and / or illustrated for one embodiment can be used in the same or similar manner in one or more other embodiments, combined with features in other embodiments, or substituted for features in other embodiments.

[0018] It should be emphasized that the term "comprising / including" as used herein refers to the presence of features, whole things, steps, or components, but does not exclude the presence or addition of one or more other features, whole things, steps, or components. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Elements and features described in one drawing or one embodiment of the embodiments of the present application can be combined with elements and features shown in one or more other drawings or embodiments. In addition, in the drawings, like reference numerals denote corresponding components in several drawings and can be used to indicate corresponding components used in more than one embodiment.

[0020] The included drawings are used to provide a further understanding of the embodiments of the present application, which form a part of the specification, illustrate the embodiments of the present application, and, together with the written description, explain the principles of the present application. Obviously, the drawings in the following description are only some embodiments of the present application, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts. In the drawings:

[0021] Figure 1 is a schematic diagram of the fall detection method according to an embodiment of the present application;

[0022] Figure 2 is a schematic diagram of each stage of fall detection according to the method of an embodiment of the present application;

[0023] Figure 3 is a schematic diagram of fall detection based on radar point cloud in the first stage;

[0024] Figure 4 [[ID=3]]is a schematic diagram of fall detection based on radar point cloud information and radar angle FFT information in the second stage;

[0025] Figure 5 is a schematic diagram of fall detection based on radar angle FFT information in the third stage;

[0026] Figure 6It is a schematic diagram of the fall detection process of the method according to the embodiments of the present application;

[0027] Figure 7 It is a schematic diagram of the fall detection device according to the embodiments of the present application;

[0028] Figure 8 It is a schematic diagram of the computer device according to the embodiments of the present application. Detailed implementation manners

[0029] Referring to the accompanying drawings, through the following description, the foregoing and other features of the present application will become apparent. In the description and drawings, specific embodiments of the present application are specifically disclosed, which show some embodiments in which the principles of the present application can be adopted. It should be understood that the present application is not limited to the described embodiments. On the contrary, the present application includes all modifications, variations, and equivalents falling within the scope of the appended claims.

[0030] In the embodiments of the present application, terms such as "first" and "second" are used to distinguish different elements in terms of name, but do not represent the spatial arrangement or time sequence of these elements, and these elements should not be limited by these terms. The term "and / or" includes any one and all combinations of one or more of the related listed terms. Terms such as "comprising", "including", and "having" mean the presence of the stated features, elements, components, or assemblies, but do not exclude the presence or addition of one or more other features, elements, components, or assemblies.

[0031] In the embodiments of the present application, the singular forms "a", "the", etc. include the plural forms and should be broadly understood as "a kind" or "a class" rather than being limited to the meaning of "one"; in addition, the term "the" should be understood to include both the singular form and the plural form unless the context clearly indicates otherwise. In addition, the term "according to" should be understood as "at least partially according to...", and the term "based on" should be understood as "at least partially based on...", unless the context clearly indicates otherwise.

[0032] The following describes various embodiments of the present application with reference to the accompanying drawings.

[0033] Embodiments of the first aspect

[0034] The embodiments of the present application provide a fall detection method. More specifically, a multi-stage fall detection method based on a millimeter-wave radar is provided, which uses the point cloud information of the millimeter-wave radar and the radar angle FFT (Fast Fourier Transform) information to perform detections at different stages of a fall event.

[0035] Figure 1 It is a schematic diagram of the fall detection method according to the embodiments of the present application. As Figure 1 shown, the method includes:

[0036] 110: In the first stage, fall detection is performed based on radar point cloud information. If no fall event is detected, it enters the second stage;

[0037] 120: In the second stage, fall detection is performed based on radar point cloud information and radar angle FFT information. If no fall event is detected, it enters the third stage;

[0038] 130: In the third stage, fall detection is performed based on radar angle FFT information.

[0039] According to the embodiments of the present application, in different stages where a fall may occur, radar point cloud information and radar angle FFT information are used for fall detection to determine whether a fall event has occurred, improving the speed and reliability of fall detection.

[0040] In the embodiments of the present application, the radar periodically emits radar signals into space and receives the reflected signals. By processing the reflected signals, it perceives the objects that reflect the radar signals. The reflected signals received by the radar are the superposition of signals reflected from different objects. After processing such as distance FFT, Doppler FFT, and radar angle FFT, the reflected signals can be distinguished from multiple dimensions such as distance, Doppler, and angle.

[0041] In the embodiments of the present application, the radar angle FFT information refers to the result obtained after performing the above processing on the reflected signals received by the radar, and can be represented by S = {s x,y,z}. Among them, s x,y,z represents the radar signal emitted from the spatial position (x, y, z).

[0042] In the embodiments of the present application, a frame of point cloud of the radar can be represented by P = {(x i , y i , z i , v i ), 1 ≤ i ≤ n}, where (x i , y i , z i ) is the spatial coordinate of a point cloud, z i is the height of the point cloud relative to the ground, v i is the Doppler velocity of the point cloud, and n is the number of point clouds. In the embodiments of the present application, the point cloud reflecting the moving target is called the effective point cloud. Therefore, when the radar cannot generate an effective point cloud, the radar cannot detect the moving target within its coverage range; conversely, when the radar can generate an effective point cloud, the radar can detect the moving target within its coverage range.

[0043] In an embodiment of the present application, taking a millimeter-wave radar as an example, such as an FCMW (Frequency Modulated Continuous Wave) radar, the present application is not limited thereto, and this radar may also be other types of radars, as long as it can generate point cloud information.

[0044] Figure 2 It is a schematic diagram of each stage of fall detection according to the method of the embodiment of the present application.

[0045] In an embodiment of the present application, as Figure 2 shown, the first stage is the stage before time T0. In the first stage, the radar can continuously output radar point clouds. Based on these radar point clouds, moving targets can be detected. The activities of the targets may be normal actions, such as walking, sitting, standing, etc., or a fall may have occurred, or the target may be struggling or crawling on the ground after the fall. In the first stage, since moving targets can be detected, fall judgment can be made according to the radar point cloud information. In addition, when the radar no longer outputs radar point clouds, that is, when the radar point clouds no longer appear, for example, at time T0, the second stage is entered.

[0046] In an embodiment of the present application, as Figure 2 shown, the second stage is the time period from time T0 to time T1. Time T0 is the moment when the output of the radar point cloud disappears, and there is no point cloud output until time T1. In the second stage, the detected target changes from a moving state to a stationary state. In this stage, radar point clouds and angle FFT information can be used for fall detection. The reason is that if the target loses consciousness immediately or is extremely weak and unable to move after falling, the radar can only capture the point clouds before the target hits the ground. Since the duration of the fall event is very short and the point cloud data reflecting the fall action is very little, it is very difficult to detect the fall event only through the radar point clouds. When the target hits the ground, as long as the target has physiological activities (breathing and heartbeat), even if there are no large-scale movements of the body, the radar will generate relatively large signal fluctuations at the position of the target. Therefore, this type of fall event can be detected by combining radar point clouds and radar signal fluctuation information.

[0047] In an embodiment of the present application, as Figure 2 shown, the third stage is the time period from time T1 to time T2 after the radar point clouds disappear. If no fall event is detected through the radar point clouds and angle FFT information in the first stage and the second stage, the third stage is entered. In actual scenarios, there are cases of slow falls, such as a human target slowly sliding to the ground against a wall or other objects. Due to the slow movement of the human body, the radar may not be able to output effective point clouds, which will cause the fall detection in the first two stages to fail. Therefore, the embodiment of the present application proposes fall detection in the third stage, that is, only using the radar angle FFT information to judge whether a fall has occurred.

[0048] In an embodiment of the present application, optionally, a fourth stage may also be maintained after the third stage, such as Figure 2 a certain period after time T2 when the radar point cloud shown disappears. Since the radar point cloud disappears, the radar coverage area is an environment without active targets. This stationary state may last for a long time. For example, the target leaves the radar detection area, or the target is in a sleeping state within the radar coverage area. Since no fall event has been detected after the previous three stages, it no longer makes sense to continue using radar data for fall detection at this time. On the other hand, due to the influence of noise, long-term detection may lead to false detections. Therefore, in the fourth stage, if the radar still does not output valid point clouds after time T2 and no fall event is detected, the fall detection is stopped. This operation can avoid long-term ineffective detection, reduce false detections caused by noise, and save computing resources.

[0049] In the above embodiment, in the second to fourth stages, when the radar point cloud reappears, that is, when the radar starts to output valid point clouds, it enters the first stage, that is, fall detection is performed according to the radar point cloud information. When the radar point cloud gradually decreases until it disappears, it re-enters the second stage.

[0050] In the above embodiment, the time lengths of the second, third, and fourth stages are not limited. In a possible implementation, the time length of the third stage is greater than that of the second stage. For example, the time length of the second stage is preset to 2 minutes, and the time length of the third stage is preset to 10 minutes. Since the third stage relies only on the radar angle FFT information for fall detection and longer data accumulation is required to obtain reliable detection results, by setting the time length of the third stage to be greater than that of the second stage, the reliability of the detection results can be improved.

[0051] The fall detection for each stage will be described below.

[0052] In the first stage, the radar continuously outputs point clouds and can detect active targets. Figure 3 is a schematic diagram of fall detection based on the radar point cloud in the first stage, as Figure 3 shown, including the following steps:

[0053] 310: Filter each frame of the radar point cloud;

[0054] 320: Save the filtered radar point cloud to the point cloud information list;

[0055] 330: Perform fall detection based on the point clouds in the point cloud information list.

[0056] It should be noted that the above appendix Figure 3The embodiments of the present application are only illustrated schematically, but the present application is not limited thereto. For example, some other operations can be added or some of the operations can be reduced. Those skilled in the art can make appropriate modifications based on the above content, not limited to the above appendix Figure 3 description.

[0057] In the above embodiment, the point cloud information is collected first, and then the fall detection is performed according to the collected point cloud information.

[0058] When collecting the point cloud information, for each frame of radar data, first filter out the noise in the point cloud, and then save the filtered point cloud to the point cloud information list.

[0059] Generally, if the reflection intensity of the radar signal by a stationary object is very high, point clouds will also be generated. However, the Doppler velocity of the point cloud of a stationary object is zero or very small. In order to eliminate the influence of the point cloud generated by the stationary object on the fall judgment, in operation 310, the point cloud with a Doppler velocity less than V (referred to as the first threshold) can be filtered out, that is, the point cloud with |v i | < V is filtered out.

[0060] In addition, in operation 310, the noise in the radar point cloud can also be filtered. For example, the clustering method is used to cluster the point cloud of each frame, and the point cloud that is successfully clustered and belongs to a certain cluster is retained, and the point cloud that fails to cluster and does not belong to any cluster is removed. The condition for successful clustering is, for example: the number of point clouds in the cluster is greater than n c (referred to as the second threshold), and the minimum distance from each point in the point cloud to other point clouds in the same cluster is less than d c (referred to as the third threshold).

[0061] In the above embodiment, after the clustering operation, the point cloud is divided into two categories, the point cloud that is successfully clustered and belongs to a certain cluster and the point cloud that fails to cluster and does not belong to any cluster. The former is the effective point cloud, and the latter is the noise. Thus, by performing the clustering operation on each frame of radar point cloud, the point cloud can be divided into multiple clusters, the number of point clouds in each cluster is greater than n c , and the minimum distance from each point in the point cloud to other point clouds in the same cluster is less than d c .

[0062] In the above embodiment, a conventional clustering algorithm can be used to implement this operation, such as the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm. The present application is not limited thereto.

[0063] In operation 320, the point cloud information list can be represented as L = {P′ j , 1 ≤ j ≤ N L} indicates that, where P′ j is the valid point cloud in the j-th frame of radar data, and N L is the total number of frames.

[0064] In a possible implementation, the point cloud information list stores at most the most recent F L frames of valid point cloud data, that is, N L ≤F L , and the point cloud data at earlier times is discarded. In addition, the number of valid point clouds in each frame in the point cloud information list is greater than zero. If the number of valid point clouds in a certain frame is zero, the radar data of that frame is ignored in the point cloud information list and does not affect the number of data frames in the point cloud information list. For example, assuming that there are currently 50 frames of radar data, and the number of valid point clouds in 45 frames of the radar data is greater than zero, then the number of data frames stored in the point cloud information list is 45, and N L = 45.

[0065] In the above embodiment, when the number of data frames in the point cloud information list is equal to F L , it is determined whether a fall event has occurred according to the point cloud information in the point cloud information list.

[0066] For example, for each frame of valid point cloud P′ j in the point cloud information list, calculate its average height z′ j . If the number of frames in which the average height z′ j of the point cloud in the point cloud information list is less than Z0 (referred to as the fourth threshold) is greater than M1 (referred to as the fifth threshold), and at the same time the average height z′ j of all frames in the point cloud information list is less than Z1 (referred to as the sixth threshold), it is considered that a fall event has occurred, otherwise it is considered that no fall has occurred.

[0067] The above is only an example, and there may also be no limit on F L , and only the limit on N L . That is, the point cloud information list stores at most the most recent N L frames of valid point cloud data, and the point cloud data at earlier times is discarded. If the number of valid point clouds in a certain frame among the multi-frame valid point cloud data closest to the current time is zero, then the radar data of that frame is ignored in the point cloud information list and does not affect the number of data frames in the point cloud information list.

[0068] In the above embodiment, if the radar continuously generates point cloud data, when adding new point cloud information to the point cloud information list, the above operations can be repeated, and a fall event can be detected according to the valid point cloud information in the point cloud information list. The detection method in this first stage is mainly used to detect the situation where the target struggles or crawls on the ground after falling.

[0069] In the above embodiments, in the case where a fall event is detected in the first stage, the detection is stopped; in the case where no fall event is detected, the second stage is entered.

[0070] The above embodiments only illustrate the method for fall detection based on radar point cloud information. In the specific implementation process, other methods can be obtained by appropriately modifying the above method, which will not be elaborated here.

[0071] In the second stage, the detection target enters the stationary state from the active state. In this stage, radar point cloud and angle FFT information can be used for fall detection. Figure 4 is a schematic diagram of fall detection based on radar point cloud information and radar angle FFT information in the second stage, as Figure 4 shown, including the following steps:

[0072] 410: Obtain the valid point cloud data of N1 frames from the point cloud information list;

[0073] 420: Perform a clustering operation on the valid point cloud data of the N1 frames;

[0074] 430: Detect the position where the stationary target is located according to the radar angle FFT information;

[0075] 440: Determine whether a fall event has occurred according to the valid point cloud data of the N1 frames and the position where the stationary target is located.

[0076] It should be noted that the above appendix Figure 4 only schematically illustrates the embodiments of the present application, but the present application is not limited thereto. For example, some other operations can be added or some of the operations can be reduced. Those skilled in the art can make appropriate modifications according to the above content, not limited to the records in the above appendix Figure 4 alone.

[0077] In operation 410, the valid point cloud data of N1 frames is obtained from the point cloud information list. If the number of frames of the valid point cloud data in the point cloud information list is less than N1, it is considered that the activity time and action amplitude of the target are very small, and it is impossible to have a fall, and there is no need to enter the subsequent fall detection in the second stage; otherwise, subsequent fall detection needs to be performed. The present application does not limit the value of N1, which can be determined according to experience.

[0078] In operation 410, the information of the N1 frames of valid point clouds can also be statistically analyzed for subsequent fall detection. Among them, the number of all point clouds is n1, and the central position of all point clouds is P c =(x c , y c , z c ), and the number of point clouds with a height less than Z2 (referred to as the sixth threshold) is nz2 , the number of point clouds with a height less than Z3 (referred to as the eighth threshold) is n z3 , the number of point clouds with a height greater than Z4 (referred to as the tenth threshold) is n z4 . The proportion of point clouds with a height less than the sixth threshold Z2 is r z2 = n z2 / n1, the proportion of point clouds with a height less than the eighth threshold Z3 is r z3 = n z3 / n1.

[0079] In operation 420, the N1-frame valid point cloud data is integrated together for clustering operation, and c point cloud clusters are obtained. In the above embodiments, the algorithm for the clustering operation is not limited. For example, the above-mentioned DBSCAN algorithm can be used to implement this clustering operation.

[0080] If the number of point cloud clusters obtained through the above clustering operation is greater than 1, that is, c > 1, there are multiple targets within the radar monitoring range, and the possibility of falling and causing serious injuries is low. There is no need to perform the second-stage fall detection, so it is determined as a safe scenario and the detection ends;

[0081] If the number of point cloud clusters obtained through the above clustering operation is not greater than 1, the position of the stationary target is detected according to the radar angle FFT information.

[0082] In operation 430, since the physiological activities of the human body affect the radar signal, the radar angle FFT information can be used to estimate the position of the stationary target (such as the human body).

[0083] A possible implementation is to estimate the position of the stationary target through the signal fluctuation of the radar angle FFT. Calculate the fluctuation of the angle FFT amplitude at each spatial position within a period of time, and this fluctuation can be characterized by the standard deviation of the angle FFT amplitude. If the fluctuation value of the above radar signal is greater than the threshold, it is considered that there is a target within the radar monitoring range, and the position with the largest radar signal fluctuation value is the position of the stationary target.

[0084] Another possible implementation is to estimate the position of the stationary target by using the signal fluctuation and spectral characteristics of the radar angle FFT. For details, reference can be made to the related technology, which is omitted here.

[0085] In the above embodiments, as long as the monitoring target remains stationary, the radar will continuously output the angle FFT information. Therefore, the position of the stationary target can be detected multiple times.

[0086] In the above embodiments, use P s = {(x s,i , y s,i , z s,i ), 1 ≤ i ≤ ns} represents the position of the stationary target detected based on the angular FFT information after the radar point cloud disappears, where n s is the number of detections.

[0087] First, calculate the distance d s between the position of the stationary target in P c and the center position P i of the valid point cloud, which is represented by the following formula (1):

[0088]

[0089] where d i is the distance between the position (x s,i , y s,i , z s,i ) of the stationary target obtained from the i-th detection and the center position P c of the valid point cloud, and the calculation of this distance does not consider the difference in height. For the position of the stationary target obtained from each detection, determine whether the distance between the position of the stationary target and the center position of the valid point cloud is less than D1 (referred to as the twelfth threshold) and whether the height z s,i where the position of the stationary target is located is less than Z5 (referred to as the thirteenth threshold), that is, d i < D1 and z s,i < Z5. Count the proportion r s of the number n z5 of the detection results that meet the above conditions in P z5 , where r z5 = n z5 / n s .

[0090] In the above embodiment, if r z2 (the proportion of the point cloud with a height less than the sixth threshold Z2) is greater than R2 (referred to as the seventh threshold), r z3 (the proportion of the point cloud with a height less than the eighth threshold Z3) is greater than the threshold R3 (referred to as the ninth threshold), n z4 (the number of point clouds with a height greater than the tenth threshold Z4) is less than N z4 (the eleventh threshold), and at the same time r z5 (the number of detection results that meet the above conditions among the multiple detection results of the position where the stationary target is located) is greater than R4 (referred to as the fourteenth threshold), then it is determined that a fall event has occurred; otherwise, it is considered that no fall has occurred.

[0091] In the above embodiment, in the case where a fall event is detected in the second stage, the detection is stopped, and in the case where no fall event is detected, the third stage is entered.

[0092] In the above embodiments, when radar point clouds appear in the second stage, it enters the first stage, and the method for fall detection based on radar point cloud information. The specific detection method has been described above and will not be elaborated here.

[0093] The above embodiments only illustrate the fall detection based on radar point cloud information and radar angle FFT information. In the specific implementation process, other methods can be obtained by appropriately modifying according to the above method, which will not be elaborated here.

[0094] In the third stage, only the radar angle FFT information is used to determine whether a fall has occurred. Figure 5 It is a schematic diagram of fall detection based on radar angle FFT information in the third stage, as Figure 5 shown, including the following steps:

[0095] 510: Detect the position where the stationary target is located according to the radar angle FFT information;

[0096] 520: Determine whether a fall event has occurred according to the position where the stationary target is located.

[0097] In the above embodiments, as described above, the position of the stationary target can be estimated using the radar angle FFT information.

[0098] In the above embodiments, use P s ={(x s,i ,y s,i ,z s,i ), 1≤i≤n s} to represent the position where the stationary target is detected according to the angle FFT information after the radar point cloud disappears, and n s is the number of detections.

[0099] In a possible implementation, before time T2, if n s is greater than N s (referred to as the fifteenth threshold), and at the same time, the proportion of z s (the height of the position where the stationary target is located) in P s,i (the set of positions where the stationary target is detected through multiple detections) that is less than Z5 (the thirteenth threshold) is greater than R5 (referred to as the sixteenth threshold), then it is determined that a fall event has occurred; otherwise, it is determined that no fall has occurred.

[0100] In the above embodiments, similar to the detection in the second stage, when radar point clouds appear in the third stage, it enters the first stage, and the method for fall detection based on radar point cloud information. The specific detection method has been described above and will not be elaborated here.

[0101] The above embodiments only illustrate the method for fall detection based on radar angle FFT information. In the specific implementation process, other methods can be obtained by appropriately modifying the above method, which will not be elaborated here.

[0102] In the embodiments of the present application, optionally, as Figure 2 shown, a fourth stage can be added. As described above, in the fourth stage, if the radar still does not output valid point clouds after time T2 and no fall event is detected, the fall detection is stopped. This operation can avoid long-term invalid detection, reduce false detections caused by noise interference, and save computing resources.

[0103] The above embodiments only exemplarily illustrate the method of the embodiments of the present application, but the present application is not limited thereto, and appropriate modifications can be made on the basis of the above embodiments. For example, the above embodiments can be used alone, or one or more of the above embodiments can be combined.

[0104] Figure 6 is a schematic diagram of the fall detection process of the method according to the embodiments of the present application. As Figure 6 shown, the process includes:

[0105] 610: Filter the point clouds from static targets, |v i |<V;

[0106] 620: Filter the noise in the point clouds through clustering to obtain valid point clouds;

[0107] 630: Save the valid point clouds to the point cloud information list L = {P′ j , 1 ≤ j ≤ N L};

[0108] 640: Calculate the average height z′ of each frame of point cloud in L j ;

[0109] 650: If N L = F L , and the number of frames with z j ′ < Z0 is greater than M1 and for all frames, z j ′ < Z1, a fall is detected, and the detection ends; otherwise, operation 660 is executed;

[0110] 660: Between time T0 and time T1 (the second stage), obtain N1 frames of point clouds from L;

[0111] 670: Cluster all the point clouds of N1 frames to obtain c clusters;

[0112] 680: If c > 1, a safe scenario is detected, and the detection ends; otherwise, operation 690 is executed;

[0113] 690: Calculate r z2 = n z2 / n1 and r z3 = n z3 / n1;

[0114] 6100: Use the angular FFT information to detect stationary targets and obtain P s = {(x s,i , y s,i , z s,i ), 1 ≤ i ≤ n s};

[0115] 6110: Calculate the ratio r of the stationary target position result z5 = n z5 / n s ;

[0116] 6120: If r z2 > R2, r z3 > R3, n z4 < N z4 , and r z5 > R4, a fall is detected, otherwise perform operation 6130;

[0117] 6130: Between time T1 and time T2 (the third stage), if n s > N s and P s the proportion of z s,i less than Z5 is greater than R5, a fall is detected, end the detection, otherwise perform operation 6140;

[0118] 6140: After time T2 (the fourth stage), stop fall detection.

[0119] It should be noted that the above appendix Figure 6 only schematically illustrates the embodiments of the present application, but the present application is not limited thereto. For example, some other operations can be added or some of the operations can be reduced. Those skilled in the art can make appropriate modifications according to the above content and the foregoing embodiments, not limited to the records of the above appendix Figure 6 .

[0120] According to the embodiments of the present application, fall detection is performed using radar point cloud information and angular FFT information at different stages where a fall may occur, improving the speed and reliability of fall detection.

[0121] Embodiments of the second aspect

[0122] An embodiment of the present application provides a fall detection device. Since the principle of the device for solving problems is similar to that of the method in the embodiment of the first aspect, the specific implementation thereof may refer to the implementation of the method in the embodiment of the first aspect, and the same content will not be repeated.

[0123] Figure 7 is a schematic diagram of the fall detection device according to the embodiment of the present application. As Figure 7 shown, the fall detection device 700 according to the embodiment of the present application includes:

[0124] A detection unit 710 that performs fall detection based on radar point cloud information in the first stage and enters the second stage when no fall event is detected; in the second stage, performs fall detection based on radar point cloud information and radar angle FFT information and enters the third stage when no fall event is detected; in the third stage, performs fall detection based on radar angle FFT information.

[0125] In some embodiments, in the first stage, when a fall event is detected, the detection is stopped (such as the operation 650→Y as Figure 6 shown); in the second stage, when a fall event is detected (such as the operation 6120→Y as Figure 6 shown) or when a safe scenario is detected (such as the operation 680→Y as Figure 6 shown), the detection is stopped; in the third stage, when a fall event is detected (such as the operation 6130→Y as Figure 6 shown), the detection is stopped, and when no fall event is detected (such as the operation 6130→N as Figure 6 shown), it is determined that no fall has occurred.

[0126] In some embodiments, the first stage is the stage where there is radar point cloud. When the radar point cloud disappears, the second stage is entered; the second stage and the third stage are the stages where there is no radar point cloud. The time length of the second stage can be less than that of the third stage.

[0127] In some embodiments, in the second stage and the third stage, when radar point cloud appears, the first stage is re-entered.

[0128] In some embodiments, the detection unit 710 performs fall detection based on radar point cloud information in the first stage, including:

[0129] Filter each frame of radar point cloud;

[0130] Save the filtered radar point cloud to the point cloud information list;

[0131] Perform fall detection based on the point cloud in the point cloud information list.

[0132] In the above embodiment, filtering the radar point cloud of each frame may include:

[0133] Filtering out the point cloud with Doppler velocity less than the first threshold V.

[0134] In the above embodiment, filtering the radar point cloud of each frame may also include:

[0135] Using a clustering method to cluster the point cloud of each frame, retaining the point cloud that is successfully clustered and belongs to a certain cluster, and removing the point cloud that fails to cluster and does not belong to any cluster.

[0136] In the above embodiment, the condition for successful clustering may be: the number of point clouds in the cluster is greater than the second threshold n c , and the minimum distance from each point in the point cloud to other point clouds in the same cluster is less than the third threshold d c .

[0137] In the above embodiment, the point cloud information list is represented by L = {P′ j , 1 ≤ j ≤ N L}, where P′ j is the valid point cloud in the j-th frame of radar data, and N L is the total number of frames.

[0138] In the above embodiment, if the number of frames in the point cloud information list where the average height z′ j of the point cloud is less than the fourth threshold Z0 is greater than the fifth threshold M1, and at the same time the average height z′ j of the point cloud in all frames of the point cloud information list is less than the sixth threshold Z1, it is considered that a fall event has occurred, otherwise it is considered that no fall has occurred.

[0139] In some embodiments, the detection unit 710 performs fall detection according to the radar point cloud information and the radar angle FFT information in the second stage, including:

[0140] Obtaining the valid point cloud data of N1 frames from the point cloud information list;

[0141] Performing a clustering operation on the valid point cloud data of the N1 frames;

[0142] If the number of point cloud clusters obtained through the clustering operation is not greater than 1, then detect the position of the stationary target according to the radar angle FFT data;

[0143] Judge whether a fall event has occurred according to the valid point cloud data of the N1 frames and the position of the stationary target.

[0144] In the above embodiment, if the number of frames of the valid point cloud data in the point cloud information list is less than N1, it is determined that no fall has occurred (detected as not fallen), and the detection ends.

[0145] In the above embodiment, if the number of point cloud clusters is greater than 1, it is determined as a safe scenario and the detection ends.

[0146] In the above embodiment, if the following conditions are met, it is determined that a fall event has occurred; otherwise, it is considered that no fall has occurred. These conditions include:

[0147] The proportion r of point clouds with a height less than the sixth threshold Z2 z2 is greater than the seventh threshold R2;

[0148] The proportion r of point clouds with a height less than the eighth threshold Z3 z3 is greater than the ninth threshold R3;

[0149] The number n of point clouds with a height greater than the tenth threshold Z4 z4 is less than the eleventh threshold N z4 ; and,

[0150] According to the set P of positions where the stationary targets are located obtained from multiple (at least one) detections s among which, the proportion r of the number n of positions that satisfy "the distance d between the position where the stationary target is located and the center of the valid point clouds of the N1 frames i is less than the twelfth threshold D1 and the height z of the position where the stationary target is located s,i is less than the thirteenth threshold Z5" z5 is greater than the fourteenth threshold R4. z5 is greater than the fourteenth threshold R4.

[0151] In some embodiments, the detection unit 710 performs fall detection according to the radar angle FFT information in the third stage, including:

[0152] Detecting the position where the stationary target is located according to the radar angle FFT data;

[0153] Judging whether a fall event has occurred according to the position where the stationary target is located.

[0154] In the above embodiment, if the number of detections n s is greater than the fifteenth threshold N s , and, according to the set P of positions where the stationary targets are located obtained from multiple (at least one) detections s among which, the proportion of the height z of the position where the stationary target is located s,i less than the thirteenth threshold Z5 is greater than the sixteenth threshold R5, it is determined that a fall event has occurred; otherwise, it is considered that no fall has occurred.

[0155] In some embodiments, if the detection unit 710 still does not detect a fall event after the third time period, the fall detection is stopped. Alternatively, if a radar point cloud appears within a period of time (the fourth stage) after the fall detection is stopped, the process may return to the first stage.

[0156] It is worth noting that the above description only describes the components or modules related to the present application, but the present application is not limited thereto. The fall detection device 700 may also include other components or modules. For details of these components or modules, reference may be made to the relevant art.

[0157] To keep it simple, Figure 7 The connection relationships and signal paths between various components or modules are shown for illustrative purposes only. However, those skilled in the art will appreciate that various related technologies, such as bus connections, can be employed. The aforementioned components or modules can be implemented using hardware such as processors and memory; this is not a limitation of the present invention.

[0158] The above embodiments are merely exemplary of the present invention, but the present invention is not limited thereto. Appropriate modifications may be made based on the above embodiments. For example, the above embodiments may be used alone, or one or more of the above embodiments may be combined.

[0159] According to the embodiment of the present application, radar point cloud information and angle FFT information are used to perform fall detection at different stages where a fall may occur, thereby improving the speed and reliability of fall detection.

[0160] Embodiments of the third aspect

[0161] This embodiment of the present application provides a computer device including the fall detection apparatus 700 described in the second aspect of the present application, the contents of which are incorporated herein. The computer device may be, for example, a computer, a server, a workstation, a laptop computer, a smartphone, etc., but this embodiment of the present application is not limited thereto.

[0162] Figure 8 Schematic diagram of a computer device according to an embodiment of the present application. Figure 8 As shown, the computer device 800 may include: a processor (e.g., a central processing unit (CPU)) 810 and a memory 820; the memory 820 is coupled to the central processing unit 810. The memory 820 can store various data; in addition, it can store an information processing program 821, and the program 821 is executed under the control of the processor 810.

[0163] In some embodiments, the functions of the fall detection device 700 are integrated into the processor 810. The processor 810 is configured to implement the fall detection method as described in the embodiment of the first aspect.

[0164] In some embodiments, the fall detection device 700 is configured separately from the processor 810. For example, the fall detection device 700 can be configured as a chip connected to the processor 810, and the functions of the fall detection device 700 are realized through the control of the processor 810.

[0165] In addition, as Figure 8 shown, the computer device 800 may further include: input / output (I / O) device 830, display 840, etc.; among them, the functions of the above components are similar to those in the prior art and will not be elaborated here. It should be noted that the computer device 800 does not necessarily have to include Figure 8 all the components shown in Figure 8 ; in addition, the computer device 800 may further include components not shown in

[0166] The embodiments of the present application also provide a computer-readable program, wherein when the program is executed in the fall detection device, the program causes the fall detection device to execute the method described in the embodiments of the first aspect.

[0167] The embodiments of the present application provide a storage medium storing a computer-readable program, wherein the computer-readable program causes the fall detection device to execute the method described in the embodiments of the first aspect.

[0168] The above devices and methods of the present application can be implemented by hardware or by a combination of hardware and software. The present application relates to such a computer-readable program that when the program is executed by a logic component, it can cause the logic component to implement the above-mentioned device or component, or cause the logic component to implement the above-mentioned various methods or steps. The present application also relates to a storage medium for storing the above program, such as a hard disk, a magnetic disk, an optical disk, a DVD, a flash memory, etc.

[0169] The method / devices described in combination with the embodiments of the present application can be directly embodied as hardware, a software module executed by a processor, or a combination of the two. For example, one or more of the functional block diagrams shown in the figure and / or a combination of one or more of the functional block diagrams can correspond to each software module in the computer program flow, and can also correspond to each hardware module. These software modules can respectively correspond to the respective steps shown in the figure. These hardware modules can be realized, for example, by using a field-programmable gate array (FPGA) to solidify these software modules.

[0170] The software module may be located in a RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, register, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art. A storage medium may be coupled to the processor such that the processor can read information from the storage medium and write information to the storage medium; or the storage medium may be an integral part of the processor. The processor and the storage medium may be located in an ASIC. The software module may be stored in the memory of the mobile terminal or in a memory card insertable into the mobile terminal. For example, if the device (such as a mobile terminal) uses a larger-capacity MEGA-SIM card or a large-capacity flash memory device, the software module may be stored in the MEGA-SIM card or the large-capacity flash memory device.

[0171] One or more of the functional blocks described in the accompanying drawings and / or one or more combinations of functional blocks can be implemented as a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, or any suitable combination thereof for performing the functions described in this application. One or more of the functional blocks described in the accompanying drawings and / or one or more combinations of functional blocks can also be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors in communication combination with a DSP, or any other such configuration.

[0172] The present application has been described in conjunction with specific embodiments, but those skilled in the art should understand that these descriptions are exemplary and not a limitation on the protection scope of the present application. Those skilled in the art can make various variations and modifications to the present application according to the spirit and principle of the present application, and these variations and modifications are also within the scope of the present application.

[0173] Regarding the above-described embodiments disclosed in this embodiment, the following additional notes are also disclosed:

[0174] 1. A fall detection method, comprising:

[0175] In a first stage, fall detection is performed based on radar point cloud information, and in the case where no fall event is detected, the second stage is entered;

[0176] In a second stage, fall detection is performed based on radar point cloud information and radar angle FFT information, and in the case where no fall event is detected, the third stage is entered;

[0177] In a third stage, fall detection is performed based on radar angle FFT information.

[0178] 2. The method according to Note 1, wherein

[0179] the first stage is the stage where radar point cloud exists, and when the radar point cloud disappears, it enters the second stage;

[0180] the second stage and the third stage are the stages where there is no radar point cloud.

[0181] 3. The method according to Note 2, wherein

[0182] the time length of the second stage is less than the time length of the third stage.

[0183] 4. The method according to any one of Notes 1-3, wherein in the first stage, fall detection is performed based on radar point cloud information, including:

[0184] filtering each frame of radar point cloud;

[0185] saving the filtered radar point cloud into the point cloud information list;

[0186] performing fall detection based on the point cloud in the point cloud information list.

[0187] 5. The method according to Note 4, wherein

[0188] the point cloud information list is represented by L = {P′ j , 1 ≤ j ≤ N L}, where P′ j is the valid point cloud in the j-th frame of radar data, and N L is the total number of frames.

[0189] 6. The method according to Note 4, wherein

[0190] if the number of frames in which the average height z′ j of the point cloud in the point cloud information list is less than the fourth threshold Z0 is greater than the fifth threshold M1, and at the same time, the average height z′ j of the point cloud in all frames of the point cloud information list is less than the sixth threshold Z1, it is considered that a fall event has occurred, otherwise it is considered that no fall has occurred.

[0191] 7. The method according to any one of Notes 1-3, wherein in the second stage, fall detection is performed based on radar point cloud information and radar angle FFT information, including:

[0192] obtaining valid point cloud data of N1 frames from the point cloud information list;

[0193] performing clustering operations on the valid point cloud data of the N1 frames;

[0194] If the number of point cloud clusters obtained through the clustering operation is not greater than 1, the position where the stationary target is located is detected according to the radar angle FFT information;

[0195] According to the valid point cloud data of the N1 frames and the position where the stationary target is located, it is determined whether a fall event has occurred.

[0196] 8. The method according to item 7 of the attached note, wherein,

[0197] If the number of frames of valid point cloud data in the point cloud information list is less than the N1, it is determined that no fall has occurred, and the detection ends; and / or,

[0198] If the number of point cloud clusters is greater than 1, it is determined as a safe scenario, and the detection ends.

[0199] 9. The method according to any one of items 1 - 3 of the attached note, wherein, in the third stage, fall detection is performed according to the radar angle FFT information, including:

[0200] Detect the position where the stationary target is located according to the radar angle FFT information;

[0201] According to the position where the stationary target is located, it is determined whether a fall event has occurred;

[0202] Wherein,

[0203] If the number of detection times n s is greater than the fifteenth threshold N s , and, according to the set P of positions where the stationary target is located obtained from multiple detections s , the proportion of the height z of the position where the stationary target is located s,i less than the thirteenth threshold Z5 is greater than the sixteenth threshold R5, it is determined that a fall event has occurred; otherwise, it is considered that no fall has occurred.

[0204] 10. The method according to any one of items 1 - 9 of the attached note, wherein, the method further includes:

[0205] If no fall event is still detected after the third stage, the fall detection is stopped.

Claims

1. A fall detection device, characterized in that, The device includes: A detection unit that, in the first stage, performs fall detection based on radar point cloud information and enters the second stage if no fall event is detected; in the second stage, performs fall detection based on radar point cloud information and radar angle FFT information and enters the third stage if no fall event is detected; and in the third stage, performs fall detection based on radar angle FFT information.

2. The device according to claim 1, wherein, In the first stage, when the detection unit detects a fall event, it stops detecting; In the second stage, when the detection unit detects a fall event, it stops detecting; In the third stage, when the detection unit detects a fall event, it stops detecting, and if no fall event is detected, it is determined that no fall has occurred.

3. The device according to claim 1, wherein, In the second stage, when new radar point cloud appears, it enters the first stage; In the third stage, when new radar point cloud appears, it enters the first stage.

4. The apparatus according to claim 1, wherein, The detection unit performs fall detection based on radar point cloud information in the first stage, including: Filtering each frame of radar point cloud to filter out point clouds with Doppler velocity less than the first threshold V and / or noise in the point cloud; Saving the filtered radar point cloud to a point cloud information list; Performing fall detection based on the point clouds in the point cloud information list.

5. The apparatus according to claim 4, wherein, Filtering out noise in the point cloud includes: Using a clustering method to cluster each frame of point cloud, retaining the point clouds that are successfully clustered and belong to a certain cluster, and removing the point clouds that are not successfully clustered and do not belong to any cluster; The condition for successful clustering is that the number of point clouds in a cluster is greater than the second threshold n c , and the minimum distance from each point in the point cloud to other point clouds in the same cluster is less than the third threshold d c .

6. The apparatus according to claim 1, wherein, The detection unit performs fall detection based on radar point cloud information and radar angle FFT information in the second stage, including: Obtaining N1 frames of valid point cloud data from the point cloud information list; Performing a clustering operation on the N1 frames of valid point cloud data; If the number of point cloud clusters obtained through the clustering operation is not greater than 1, detecting the position of the stationary target based on radar angle FFT information; Judging whether a fall event has occurred based on the N1 frames of valid point cloud data and the position of the stationary target.

7. The device according to claim 6, wherein, If the number of frames of valid point cloud data in the point cloud information list is less than the N1, it is determined that no fall has occurred, and the detection unit ends the detection.

8. The device according to claim 6, wherein, If the number of point cloud clusters is greater than 1, it is determined to be a safe scenario, and the detection unit ends the detection.

9. The device according to claim 6, wherein If the following conditions are met, it is determined that a fall event has occurred; otherwise, it is considered that no fall has occurred. These conditions include: The proportion r of the point cloud with a height less than the sixth threshold Z2 z2 is greater than the seventh threshold R2; The point cloud ratio r with a height less than the eighth threshold Z3 z3 is greater than the ninth threshold R3; The number of point clouds n with a height greater than the tenth threshold Z4 z4 Less than the eleventh threshold N z4 ; And, In the set P of positions of static targets obtained through multiple detections s where the distance d between the position of the static target and the center of the valid point cloud of the N1 frames i is less than the twelfth threshold D1 and the height z of the position of the static target s,i is less than the thirteenth threshold Z5, the number n z5 occupies a proportion r z5 greater than the fourteenth threshold R4.

10. The apparatus according to claim 1, wherein, The detection unit performs fall detection based on radar angle FFT information in the third stage, including: Detecting the position of the stationary target based on radar angle FFT information; Judging whether a fall event has occurred based on the position of the stationary target.

Images