Fall detection method, device, equipment and medium based on crowd gathering

Abstract

The present invention relates to the field of big data processing technology, and specifically to a fall detection method, device, equipment and medium based on crowd gathering. After human target recognition is performed on point cloud data and crowd gathering is determined, the point cloud message is reconstructed in three-dimensional space to obtain a three-dimensional point cloud distribution map, and the highest point cloud at each location of the distribution map is extracted as a key point. The height surface is fitted with the key point, and the concave and convex division of the segmented height row is determined by all the height values ​​of the height surface. When the area of ​​a concave area is within a preset area range and the height change rate between it and the surrounding convex area is also within a preset change rate interval, it is determined that a fall has occurred. The present invention directly uses the original point cloud data to judge a fall, which can avoid the missed detection of a fall due to the defect of radar that is easy to miss the target in a crowded situation when the human target is used for fall judgment, thereby improving the accuracy of judging a fall in a crowd.

Description

Technical Field

[0001] The present invention relates to the field of big data processing technology, and in particular to a fall detection method, device, equipment and medium based on crowd gathering. Background Art

[0002] With the improvement of education level and parents paying more attention to the quality of their children's education, students' safety at school has become more concerned by parents and society. Students are more likely to have adverse events because they study and live in a centralized manner and are more active. At the same time, with the frequent occurrence of campus safety incidents, it is becoming more and more important to promptly discover and prevent student safety situations.

[0003] Currently, most campus perception systems require the use of cameras for image acquisition and image processing to generate warning events, etc., to facilitate timely supervision. However, in private scenes on campus, such as dormitory supervision, images cannot be used for acquisition, so the ability to detect falls is limited.

[0004] Currently, radar and other devices that cannot form intuitive images are generally used for perception in private scenes. However, when dealing with complex on-site situations, radar cannot meet the accuracy requirements for most data processing. Especially when dealing with large amounts of point cloud data and complex information, accurate and rapid recognition and judgment becomes an urgent problem that needs to be solved. Summary of the invention

[0005] In view of this, the embodiments of the present invention provide a fall detection method, device, equipment and medium based on crowd gathering to solve the problem that the correlation between radar data and fall judgment is weak, resulting in low recognition accuracy of fall judgment.

[0006] In a first aspect, an embodiment of the present invention provides a fall detection method based on crowd gathering, the fall detection method comprising:

[0007] Obtaining a point cloud message obtained by the radar scanning a set area, performing human target recognition on the point cloud message, and obtaining the number of human targets;

[0008] If the number of human targets detected reaches a preset condition, it is determined to be a crowd gathering, and the point cloud message is reconstructed in three-dimensional space using the world coordinate system to obtain a three-dimensional point cloud distribution map of the set area;

[0009] Extracting height key points from the three-dimensional point cloud distribution map, and fitting the height key points to obtain a height surface, wherein the height key points are point clouds with the highest height value at the same horizontal position;

[0010] Determine a segmentation height according to all height values ​​in the height surface, and use the segmentation height to perform concave-convex division on the height surface to obtain at least one concave area;

[0011] For any concave area, the area of ​​the concave area and the rate of change between the center point of the concave area and the surrounding convex surface are obtained. When the area is within a preset area range and the rate of change is within a preset rate of change interval, it is determined that the human target corresponding to the concave area has fallen.

[0012] In a second aspect, an embodiment of the present invention provides a fall detection device based on crowd gathering, the fall detection device comprising:

[0013] A radar data processing module is used to obtain a point cloud message obtained by the radar scanning a set area, perform human target recognition on the point cloud message, and obtain the number of human targets;

[0014] A three-dimensional reconstruction module is used to determine that a crowd is gathering if the number of human targets detected reaches a preset condition, and to perform three-dimensional spatial reconstruction of the point cloud message in a world coordinate system to obtain a three-dimensional point cloud distribution map of the set area;

[0015] A height surface determination module is used to extract height key points in the three-dimensional point cloud distribution map, and to fit the height key points to obtain a height surface, wherein the height key points are point clouds with the highest height value at the same horizontal position;

[0016] A concave surface determination module, used to determine a segmentation height according to all height values ​​in the height surface, and use the segmentation height to perform concave-convex division on the height surface to obtain at least one concave surface area;

[0017] The fall determination module is used to obtain the area of ​​any concave area and the rate of change between the center point of the concave area and the surrounding convex surface, and when the area is within a preset area range and the rate of change is within a preset rate of change interval, determine that the human target corresponding to the concave area has fallen.

[0018] In a third aspect, an embodiment of the present invention provides a computer device, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor implements the fall detection method as described in the first aspect when executing the computer program.

[0019] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the fall detection method as described in the first aspect is implemented.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] The present invention directly performs screening processing on the original point cloud data for the representation of a person falling. Specifically, in the processing process, the overall ups and downs of the top point cloud height surface in the three-dimensional space obtained by reconstructing the point cloud message are selected to reflect the person falling, and the performance characteristics of the person falling on the height surface are combined to accurately judge whether the actual concave area of ​​the height surface is a concave surface formed by the fall, thereby completing the fall detection. Compared with the existing method of judging a fall by identifying a human target, the present invention uses the original point cloud data to judge a fall, avoiding the missed detection of a fall due to the performance defect of the radar that is easy to miss the target in a specific scene with dense crowds, and improving the accuracy of judging a fall in a crowd. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0023] Figure 1 1 is a schematic diagram of an application environment of a fall detection method based on crowd gathering provided in Embodiment 1 of the present invention;

[0024] Figure 2 is a flow chart of a fall detection method based on crowd gathering provided in Embodiment 2 of the present invention;

[0025] Figure 3 is a flow chart of a fall detection method based on crowd gathering provided in Embodiment 3 of the present invention;

[0026] Figure 4 is a flow chart of a fall detection method based on crowd gathering provided in Embodiment 5 of the present invention;

[0027] Figure 5 is a flow chart of a fall detection method based on crowd gathering provided in Embodiment 6 of the present invention;

[0028] Figure 6 is a structural schematic diagram of a fall detection device based on crowd gathering provided in Embodiment 7 of the present invention;

[0029] Figure 7 It is a structural diagram of a computer device provided in Embodiment 8 of the present invention. DETAILED DESCRIPTION

[0030] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0031] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0032] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0033] As used in the specification and appended claims of this application, the term "if" can be interpreted as "when" or "uponce" or "in response to determining" or "in response to detecting", depending on the context. Similarly, the phrase "if it is determined" or "if [described condition or event] is detected" can be interpreted as meaning "uponce it is determined" or "in response to determining" or "uponce [described condition or event] is detected" or "in response to detecting [described condition or event]", depending on the context.

[0034] In addition, in the description of the present application specification and the appended claims, the terms "first", "second", "third", etc. are only used to distinguish the descriptions and cannot be understood as indicating or implying relative importance.

[0035] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the statements "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. that appear in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0036] It should be understood that the size of the serial numbers of the steps in the following embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0037] In order to illustrate the technical solution of the present invention, specific embodiments are provided below for illustration.

[0038] A fall detection method based on crowd gathering provided in the first embodiment of the present invention can be applied in the following aspects: Figure 1 In the application environment, the sensing device is a device installed in a target scene and used to collect corresponding data in the target scene. For example, the sensing device can be a radar, specifically a millimeter wave radar, a laser radar, etc. The sensing device is connected to the server through a network, etc., so that the server can obtain the corresponding data and process the data. The computer device corresponding to the server can be implemented with an independent server or a server cluster composed of multiple servers. Correspondingly, the server communicates with the client, client A is a mobile client that can be viewed by users on mobile devices such as mobile phones, and client B is a computer for users to view in a fixed place. Among them, the client includes but is not limited to PDAs, desktop computers, laptops, ultra-mobile personal computers (Ultra-Mobile Personal Computer, UMPC), netbooks, cloud computer devices, personal digital assistants (Personal Digital Assistant, PDA) and other devices. The server, all clients, etc. form the entire supervision system, and the server is used to collect data and perform data analysis and processing to achieve supervision. Of course, the crowd gathering-based fall detection method of the present application can also be applied to sensing devices, provided that the sensing device is equipped with a processing chip capable of performing analytical calculations. At this time, the sensing device only needs to send the processing results to the server to achieve supervision.

[0039] See also Figure 2 , is a flow chart of a fall detection method based on crowd gathering provided in Embodiment 2 of the present application. The above-mentioned fall detection method based on crowd gathering is applied to Figure 1 The server or sensing device in Figure 2 As shown, the fall detection method based on crowd gathering may include the following steps:

[0040] Step S201, obtaining a point cloud message obtained by a radar scanning a set area, performing human target recognition on the point cloud message, and obtaining the number of human targets.

[0041] In this embodiment, the set area may refer to the installation scene of the sensing device. For example, the set area may be a student dormitory, a corridor or a staircase in a teaching building or a dormitory building. The sensing device is a device including a radar. The radar may specifically be a millimeter-wave radar or a laser radar. This embodiment is described with a millimeter-wave radar, which can realize the measurement of objects in the installation scene.

[0042] Millimeter-wave radar can perform real-time scanning of a set area, and obtain information such as the distance, angle, etc. of the surface of objects in the set area compared to the radar. Combined with the radar installation height, or by proposing a virtual reference horizontal plane, the distance, angle, etc. of the surface of objects in the set area compared to the radar can be further combined to determine the position and height of each point on the surface of objects in the set area. The position and height information of each point on the surface of objects in the set area determined is the point cloud data at the current moment obtained by the radar scanning. Each point on the surface of objects in the set area can be represented by each point in the point cloud containing the position and height information. The point cloud data will be recorded in the form of a message to obtain the corresponding point cloud message.

[0043] The method for human target recognition of point cloud messages can adopt any one of the existing technologies. This embodiment only introduces one implementation process:

[0044] After using radar such as millimeter-wave radar to obtain ADC data corresponding to the set area, static clutter is removed and the CFAR detection algorithm is used to obtain point cloud data containing target list information. The total data length of the target list and the data length of each target are then compared to determine the number of targets contained in the scanned point cloud data.

[0045] The static clutter removal of the ADC data and the CFAR detection algorithm mentioned above all adopt relevant algorithms in the prior art, so they will not be described in detail here.

[0046] After obtaining the number of targets, considering that the purpose of this embodiment is to judge the fall of people in a crowd, the effective identification objects only include the human body. Therefore, it is optional to use the trained neural network to identify the shape of each cluster after completing the clustering to screen out the clusters belonging to the human body. The number of clusters belonging to the human body is the number of human targets required for this step.

[0047] Step S202: If the number of human targets detected reaches a preset condition, it is determined to be a crowd gathering, and the point cloud message is reconstructed in three-dimensional space using the world coordinate system to obtain a three-dimensional point cloud distribution map of the set area.

[0048] Based on the number of human targets identified, it is possible to determine whether there is a gathering of people in the current set area by judging whether the number of human targets reaches the preset conditions. The simplest and easiest to implement method for judging the gathering of people is to set a threshold for judging the gathering of people according to the size of the set area, and then use the number of targets in the set area exceeding the threshold for judging the gathering of people as a preset condition to complete the judgment. When the number of targets in the set area exceeds the threshold for judging the gathering of people, it can be determined that there is a gathering of people.

[0049] It is easy to understand that the above-mentioned threshold for judging the gathering of people is preferably an integer value corresponding to the number of people. The specific value can be adaptively set by technical personnel in this field in combination with the actual scenario and the level of warning sensitivity, and this embodiment does not specifically limit it.

[0050] Because student safety incidents, such as stampedes and school bullying incidents, are most likely to occur in crowds, it is necessary to perform fall detection only after a crowd is determined to have gathered. Therefore, in this embodiment, after a crowd is determined to have gathered, the point cloud message is reconstructed in three-dimensional space using the world coordinate system to obtain a three-dimensional point cloud distribution map corresponding to the set area, which is used as the basis for subsequent correlation reaction analysis of the point cloud data to the fall situation.

[0051] Step S203, extracting height key points from the three-dimensional point cloud distribution map, and fitting the height key points to obtain a height surface, wherein the height key points are point clouds with the highest height value at the same horizontal position.

[0052] In order to realize the association of data information in point cloud messages with people's falls and thus reflect the fall situation, the idea of ​​the present invention is to fit the topmost point cloud of all point cloud data in a set area into a surface to fit the ups and downs of the obtained surface, specifically the area of ​​the concave area and the degree of decline compared to other non-concave areas, whether it is consistent with the performance of a person's fall, to complete the association and response of point cloud data to the fall situation of the person.

[0053] As described in step S201, since the millimeter-wave radar obtains point cloud messages while scanning the set area, each point cloud contains distance and angle information compared to the millimeter-wave radar, and the angle information specifically includes vertical angle and horizontal angle. Therefore, by combining the radar installation height or setting a virtual reference horizontal plane, the specific position and height value of each point cloud can be determined.

[0054] Then, this application first determines that the point cloud with the highest height value at each horizontal position is the height key point, where the horizontal position can be generally understood to represent the horizontal plane, and each horizontal position can be generally understood to be a different position on the horizontal plane. Then, the so-called height key point refers to the point cloud with the highest height value at a certain position on the horizontal plane.

[0055] Regarding the point cloud with the highest height value at the same horizontal position, it is easy to understand that the same horizontal position can be either a horizontal position, i.e., a point on the horizontal plane, or an area with an area. Since the present embodiment is to complete the posture response of the human target through the specific point cloud data in the point cloud message, the present embodiment preferably selects the same horizontal position, specifically, a horizontal position, i.e., an area on the horizontal plane with an area, and the area of ​​the area can be determined based on the size of the ground occupied by the person when standing.

[0056] After determining each height key point in the point cloud data, the height key point can be subjected to surface fitting to obtain a height surface. The specific method of surface fitting can use any feasible method in the art, such as the least squares method, the spline interpolation method, etc., which is not limited in this embodiment.

[0057] Step S204: determining a segmentation height according to all height values ​​in the height surface, and using the segmentation height to perform concave-convex segmentation on the height surface to obtain at least one concave region.

[0058] Since there is a possibility of people falling, and there may be non-personnel point cloud data in the set area, there is a high possibility of unevenness in the height surface obtained by selecting height key points and fitting them from the point cloud data.

[0059] To this end, a segmentation height for distinguishing concave areas from convex areas can be determined through the height values ​​of all points on the height surface, and the height surface is divided into concave and convex areas using the segmentation height to determine at least one concave area.

[0060] The determination of the segmentation height can be achieved in any appropriate way, such as directly calculating the average value of the height values ​​of each point on the height surface, and taking the obtained average value as the segmentation height; for example, first selecting points from the height surface that belong to the horizontal area without depressions or protrusions, and then taking the average value of the height values ​​of these selected points as the segmentation height, and so on.

[0061] Step S205, for any concave area, obtain the area of ​​the concave area and the rate of change between the center point of the concave area and the surrounding convex surface. When the area is within a preset area range and the rate of change is within a preset rate of change interval, determine that the human target corresponding to the concave area is a fall.

[0062] For any one of the determined concave areas, since the present invention is to complete the characterization of whether a person has fallen or not through the point cloud data corresponding to the concave area, for further judgment of the concave area, it is necessary to set the screening conditions according to the characteristics of the concave area that may be formed when a person falls.

[0063] Since the size of human bodies is roughly similar, there is a rough range for the size of the falling area formed after a fall, and the present embodiment thus sets a preset area range for screening the area of ​​the concave area; at the same time, considering that the height difference between a person and the non-fallen people around him after a fall can also be estimated, the present embodiment thus sets a preset change rate interval for screening the change rate between the center point of the concave area and the surrounding convex surface.

[0064] That is, when the area of ​​the concave area is too small, it may be caused by the height difference between people or the fitting error. When the area of ​​the concave area is too large, it is likely that the radar scans large non-human objects such as seats, boxes, etc.; at the same time, when the change rate between the center point of the concave area and the surrounding convex surface is too small, it may also be caused by the height difference between people. When the change rate between the center point of the concave area and the surrounding convex surface is too large, the point cloud data at this time is likely to be a noise error signal that needs to be eliminated.

[0065] Combining the above two aspects, this embodiment identifies the concave area whose area is within a preset area range and whose change rate between the center point of the concave area and the surrounding convex surface is within a preset change rate interval as the concave area corresponding to the human body fall.

[0066] Regarding the rate of change between the center point of the concave area and the surrounding convex surfaces, technical personnel in this field can use any existing implementation method to determine it, such as calculating the height difference between the center point of the concave area and the surrounding convex surfaces, and then comparing the obtained difference with the height of the convex surface to determine the rate of change, and so on.

[0067] After identifying the presence of a crowd gathering in a set area, this embodiment uses the topmost point cloud of the original point cloud data as a key point to perform height surface fitting, and selects the concave areas on the height surface. For each concave area, the area size and the height change rate of the center point compared to the surrounding convex area are observed to see whether they meet the characteristics of the concave area formed by a person falling, thereby completing the association and characterization of the point cloud data and the person falling. Since this embodiment directly uses the original point cloud data to make a fall judgment, it can directly avoid the missed detection of falls caused by the defect of radar that is easy to miss targets in crowded situations when making a fall judgment based on the identified human target, thereby improving the accuracy of the judgment of falls in crowds.

[0068] See also Figure 3 , which is the third embodiment of the present application, in step S203, after extracting the height key points in the three-dimensional point cloud distribution map, it also includes:

[0069] Step S301, eliminating N height key points with the smallest values ​​among the height key points, and determining the remaining height key points.

[0070] Considering that the heights of the people in the set area are likely to be different, the height differences between the people will cause the height key points extracted by the millimeter wave radar scanning to obtain the point cloud data to fluctuate, thereby causing the height surface obtained by fitting to fluctuate. Therefore, in order to avoid false alarms caused by the deviation caused by the height differences between the people, this embodiment selects the discrete degree of the height values ​​between the height key points to correct the preset area pre-range threshold used to determine the area of ​​the concave area, and obtain the preset area range described in step S205.

[0071] Furthermore, considering that among the height key points obtained from the three-dimensional point cloud distribution map, there are likely to be height key points corresponding to non-human objects, and these key points cannot actually be used to characterize the differences between people's own heights, this embodiment also first removes the N height key points with the smallest values ​​among the height key points, and determines the correction weights based on the discreteness of the remaining height key points to obtain the preset area range described in step S205. Among them, the value of N can be set by those skilled in the art based on experience, or based on the proportion of non-human targets identified in historical point cloud data, and this embodiment does not specifically limit its value.

[0072] Step S302: obtaining a first correction weight according to the remaining height key points.

[0073] After removing N height key points with the smallest values ​​from all height key points, the first correction weight can be obtained according to the degree of discreteness of the height values ​​of the remaining height key points.

[0074] Since the height difference between people leads to an increase in the degree of fluctuation on the height surface, that is, the area of ​​the concave region on the height surface increases, the role of the first correction weight should be used to expand the preset area range threshold, that is, to correct the boundary value of the preset area range. Therefore, it is easy to understand that the first correction weight should not be less than 1, and the first correction weight is positively correlated with the degree of discreteness of the height values ​​of the remaining height key points.

[0075] There are various methods for constructing the first correction weight that is positively correlated with the degree of discreteness of the height values ​​of the remaining height key points and whose value is not less than 1, such as summing the variance of the height values ​​of the remaining height key points with 1 as the first correction weight, or summing the standard deviation of the height values ​​of the remaining height key points with 1 as the first correction weight, or calculating the average value of the difference between each value of the height value of the remaining height key points and the minimum height value in turn, and then summing the obtained average value with 1 as the first correction weight, etc. This embodiment does not specifically limit the calculation method of the first correction weight.

[0076] Step S303: Correct the preset area range threshold according to the first correction weight to obtain the preset area range.

[0077] After obtaining the first correction weight, the preset area range threshold can be corrected according to the obtained first correction weight to obtain the preset area range.

[0078] The so-called correction is to multiply the first correction weight by the preset area range threshold, that is, the boundary value, to redefine the boundary.

[0079] The first cultivation weight should at least perform an enlargement correction on the smaller boundary value of the preset area range threshold, or simultaneously perform an enlargement correction on both boundary values ​​of the preset area range threshold to obtain the preset area range.

[0080] The detection process of the present application takes into account the degree of height difference among individuals in a set area, and according to the degree of height difference among individuals in the set area, the area threshold used for judging a fall is expanded and corrected to a corresponding degree, thereby avoiding errors in the fall judgment due to the differences in the heights of the individuals themselves, and the degree of error correction in the judgment process is adaptively determined according to the size of the error, thereby further improving the accuracy of judging the fall situation of people in a crowd.

[0081] In the fourth embodiment of the present application, in the above step S202, a first correction weight is obtained according to the remaining height key points, including:

[0082] According to the discrete degree of the remaining height key points, the first correction weight is obtained:

[0083]

[0084] Among them, L1 is the first correction weight, is the variance of the remaining height key points, and e represents a natural constant.

[0085] This embodiment clarifies a specific method for determining the first correction weight based on the degree of discreteness of the height values ​​of the remaining height key points. This embodiment uses the variance corresponding to the height values ​​of the remaining height key points as the exponent of an exponential function with the natural constant e as the base, and provides a calculation method in which the first correction weight increases faster as the degree of discreteness of the remaining height key points increases. This can more effectively characterize the impact of differences in people's own heights and improve the accuracy of fall judgment.

[0086] Optionally, in other embodiments, considering the set area scanned by the radar, there will inevitably be scenes where the ground itself is uneven, such as physical training scenes suitable for young students or outdoor exploration scenes. In these scenes, although the ground as a whole can be regarded as horizontal, there are uneven ups and downs everywhere. This ups and downs are just like the inconsistent heights of the various people in the scene introduced in Example 2, which will also lead to errors in fall detection. For this reason, this embodiment additionally considers correcting the preset area threshold for fall judgment by constructing a correction weight that is positively correlated with the degree of ground unevenness.

[0087] Specifically, considering that the unevenness of the ground itself causes an increase in the degree of ups and downs on the height surface, that is, the area of ​​the concave region on the height surface increases, the point cloud with the lowest height value at the same horizontal position in the three-dimensional point cloud distribution map can be selected as the lowest key point, and then the second correction weight is constructed according to the discrete degree of the height values ​​of all the lowest key points.

[0088] Similar to the fact that the difference in height between people will cause the area of ​​the concave region on the height surface to increase, the role of the second correction weight is the same as the first correction weight. It should be used to expand the preset area range threshold, that is, the boundary value of the preset area range. Therefore, it is easy to understand that the second correction weight should also be no less than 1, and the second correction weight is positively correlated with the discreteness of the height value of the lowest key point.

[0089] The specific method for constructing the second revised weight is the same as the method for constructing the first revised weight, and will not be repeated here.

[0090] See also Figure 4 In the fifth embodiment of the present application, step S204 specifically includes:

[0091] Step S501: sampling the height surface to determine the height value of each sampling point.

[0092] Taking into account that the distribution of height key points corresponding to the original point cloud data may be relatively sparse, and it is impossible to accurately characterize the details of the height surface during the calculation process, this embodiment preferably re-samples the height surface according to the set sampling interval and determines the height value of each sampling point after the height surface has been fitted.

[0093] Among them, the sampling interval can be set based on the requirements for the accuracy of the fall warning and the speed of the fall warning. It is easy to understand that the smaller the sampling interval is, the higher the sampling precision of the height surface, the more accurate the low-lying area determined accordingly, and the higher the accuracy of the fall warning. On the contrary, the smaller the sampling interval is, the lower the sampling precision of the height surface, the fewer the corresponding sampling points obtained, the smaller the calculation amount in the process of determining the low-lying area, and the faster the speed of the fall warning.

[0094] Step S502, based on the height value, DBSCAN clustering is performed on all sampling points to obtain a clustering result. If there are at least two cluster sets in the clustering result, the average value of all height key points in the height surface is used as the segmentation height.

[0095] After sampling, the sampling points can be used to more precisely characterize the height surface. At this time, DBSCAN clustering without specifying the number of clusters is performed on all sampling points based on the height value. When there are multiple clusters in the clustering result, it can be shown that there are observable ups and downs on the height surface. At this time, there may be depressions and convexities, indicating that there may be a possibility of a person falling. In order to finally determine whether a person has fallen, it is necessary to distinguish between depressions and convexities. In order to distinguish between depressions and convexities, it is necessary to first determine a segmentation threshold for distinguishing them.

[0096] This embodiment selects the average value of all height key points in the height surface to be directly calculated as the segmentation height, as the segmentation threshold for distinguishing the height key points. In other embodiments, after performing DBSCAN clustering without specifying the number of clusters based on the height value for all sampling points, the cluster with the largest number of sampling points in the obtained clustering results is selected, and the height average of each sampling point in the cluster is used as the segmentation height. Since the sampling points represent the height surface more finely than the height key points, and the cluster with the largest number of sampling points is closer to the area without fluctuations on the height surface compared to all the height key points, this method of determining the segmentation height as the segmentation threshold using the cluster with the largest number of sampling points in the clustering results can further improve the accuracy of determining the segmentation threshold.

[0097] Step S503: taking an area in the height surface whose height value is less than the segmentation height as a to-be-determined area; if the to-be-determined area matches a sampling point in any clustering set in the clustering result, determining the to-be-determined area as a concave area.

[0098] After obtaining the segmentation height as the segmentation threshold, the height surface is segmented according to the height value to determine a number of areas to be determined. Further, if the area to be determined matches a sampling point in a cluster obtained by clustering, it can be determined that the area to be determined is a concave area with obvious depression.

[0099] The so-called matching mentioned above specifically refers to whether the sampling points in a certain cluster are included in the area to be determined. It is easy to understand that due to the existence of errors, this inclusion can be that the sampling points in a certain cluster are completely included in the area to be determined, or that more than a set proportion of the sampling points in a certain cluster are included in the area to be determined. The specific value of the set proportion is directly related to the allowable error. The larger the allowable error, the smaller the value of the set proportion, such as setting the set proportion to 95%.

[0100] The reason why it is necessary to judge whether the area to be determined is matched with each cluster obtained by clustering is that the clustering process is to classify the sampling points that can obviously be attributed to the same category into the same cluster, but the areas to be determined determined by segmenting the height surface with the segmented height may also contain a number of very small concave areas that are segmented due to slight ups and downs, and do not belong to the concave areas formed by real falls. Therefore, by using the cluster obtained by clustering to verify the area to be determined, an effective screening of the concave areas can be carried out to improve the accuracy of subsequent fall detection.

[0101] The method for determining the area of ​​the concave region in step S503 is:

[0102] Step S5031: for any concave area, project the concave area on a horizontal plane to obtain a projection area.

[0103] Since the concave area is a non-planar area with curvature, its area size is not accurate, so the concave area is first projected onto a horizontal plane to determine the projection area.

[0104] Step S5032: Calculate the actual area of ​​the projection area in the world coordinate system, and use the actual area as the area of ​​the concave area.

[0105] Then, after the projection area is obtained, the actual area of ​​the projection area can be calculated according to the world coordinate system used when the point cloud message is reconstructed in three-dimensional space, and used as the area of ​​the concave area.

[0106] See also Figure 5 In the sixth embodiment of the present application, before step S201, the following steps are also included:

[0107] Step S601, collecting the environmental sound of the set area, performing speech recognition on the environmental sound, and obtaining a speech recognition result.

[0108] Considering that falls usually occur in gathering situations, there will be loud voices in the set area before or when a fall occurs. In addition, when a fall occurs, the crowd will usually shout specific keywords, such as "help", "fight", "save", etc.

[0109] To this end, this embodiment chooses to collect the ambient sound of the set area and perform voice recognition before executing step S201 to obtain the voice recognition result to provide a basis for subsequent sound decibel detection and keyword detection.

[0110] Step S602, if it is detected that the sound intensity in the speech recognition result is greater than the preset decibel value, and the speech recognition result contains preset keywords, then execute the acquisition of the point cloud message obtained by the radar scanning the set area, perform human target recognition on the point cloud message, and obtain the number of human targets.

[0111] If the sound intensity in the speech recognition result is greater than the preset decibel value, and the speech recognition result contains preset keywords, it means that there is a high possibility of a gathering of people. It is necessary to start the radar to scan the set area to obtain the corresponding point cloud message, and perform human target recognition on the point cloud message to determine the number of human targets, thereby providing a basis for whether a gathering of people has occurred.

[0112] This embodiment uses the detection of human decibels and keywords in a set area as a prerequisite for radar activation and startup, so that the radar works in a triggered manner. Only when the necessity of radar activation is detected, the radar is controlled to obtain point cloud messages, thereby reducing the working intensity of the radar.

[0113] In other embodiments, after determining that a crowd is gathering in step S202, the method further includes:

[0114] Performing timbre recognition on the environmental sound to obtain a timbre recognition result;

[0115] If it is detected that the difference between the number of timbre in the timbre recognition result and the number of human targets is less than a difference threshold, the three-dimensional spatial reconstruction of the point cloud message using the world coordinate system is performed to obtain a three-dimensional point cloud distribution map of the set area.

[0116] This setting uses the number of voice recognitions to additionally verify the number of human targets identified through point cloud messages, thereby improving the accuracy of crowd gathering recognition and avoiding unnecessary three-dimensional space reconstruction calculations, thereby effectively avoiding the waste of computing resources.

[0117] In other embodiments, after step S601, the method further includes:

[0118] If it is detected that the sound intensity in the speech recognition result is greater than a preset decibel value, and the speech recognition result contains a preset keyword, then obtaining the current sampling frequency and the current filtering threshold of the radar;

[0119] When the current sampling frequency is not the maximum sampling frequency, adjusting the current sampling frequency of the radar to the maximum sampling frequency;

[0120] When the current filtering threshold is not the minimum filtering threshold, the current filtering threshold of the radar is adjusted to the minimum filtering threshold.

[0121] This setting ensures that the radar obtains point cloud messages with maximum performance and improves the accuracy of fall detection.

[0122] In other embodiments, regarding step S601, before executing the step of acquiring the point cloud message obtained by scanning the set area by the radar, performing human target recognition on the point cloud message, and obtaining the number of human targets, the step further includes:

[0123] Get the current time;

[0124] If the interval between the current moment and the time when a fall was last determined is greater than a preset time interval, the step of acquiring the point cloud message obtained by scanning the set area by the radar is executed, and human target recognition is performed on the point cloud message to obtain the number of human targets.

[0125] This setting can prevent the next fall warning from being launched too frequently or too early after the previous fall judgment and alarm have been completed, saving computing resources and reducing the radar's workload.

[0126] Corresponding to the fall detection method based on crowd gathering in the above embodiment, Figure 6 The structure block diagram of the fall detection device based on crowd gathering provided in the seventh embodiment of the present application is shown. The fall detection device is applied to Figure 1 For ease of explanation, only the parts related to the embodiments of the present application are shown.

[0127] See also Figure 6 , the fall detection device comprises:

[0128] The radar data processing module 71 is used to obtain the point cloud message obtained by the radar scanning the set area, perform human target recognition on the point cloud message, and obtain the number of human targets;

[0129] The three-dimensional reconstruction module 72 is used to determine that a crowd is gathering if the number of human targets detected reaches a preset condition, and to perform three-dimensional spatial reconstruction on the point cloud message in a world coordinate system to obtain a three-dimensional point cloud distribution map of the set area;

[0130] A height surface determination module 73 is used to extract height key points in the three-dimensional point cloud distribution map, and to fit the height key points to obtain a height surface, wherein the height key points are point clouds with the highest height values ​​at the same horizontal position;

[0131] A concave surface determination module 74, configured to determine a segmentation height according to all height values ​​in the height surface, and use the segmentation height to perform concave-convex division on the height surface to obtain at least one concave surface area;

[0132] The fall determination module 75 is used to obtain the area of ​​any concave area and the rate of change between the center point of the concave area and the surrounding convex surface, and when the area is within a preset area range and the rate of change is within a preset rate of change interval, determine that the human target corresponding to the concave area has fallen.

[0133] Optionally, the height surface determination module 73 includes:

[0134] A height key point screening unit is used to remove N height key points with the smallest values ​​among the height key points to determine the remaining height key points;

[0135] A first correction weight obtaining unit, used for obtaining a first correction weight according to the remaining height key points;

[0136] The preset area range determining unit is used to correct the preset area range threshold according to the first correction weight to obtain the preset area range.

[0137] Optionally, the first modified weight obtaining unit is specifically configured to:

[0138] According to the discrete degree of the remaining height key points, the first correction weight is obtained:

[0139] The first modified weight calculation formula is as follows:

[0140]

[0141] Among them, L1 is the first correction weight, is the variance of the remaining height key points, and e represents a natural constant.

[0142] Optionally, the concave surface determination module 74 includes:

[0143] A height surface sampling unit, used to sample the height surface and determine the height value of each sampling point;

[0144] A segmentation height determination unit is used to perform DBSCAN clustering on all sampling points based on the height value to obtain a clustering result. If there are at least two clustering sets in the clustering result, the average value of all height key points in the height surface is used as the segmentation height;

[0145] The concave area determination unit is used to take an area in the height surface whose height value is less than the segmentation height as an area to be determined, and if the area to be determined matches a sampling point in any clustering set in the clustering result, the area to be determined is determined to be a concave area.

[0146] Optionally, the concave surface determination module 74 further includes:

[0147] A concave area projection unit, used for projecting any concave area onto a horizontal plane to obtain a projection area;

[0148] The concave region area determination unit is used to calculate the actual area of ​​the projection area in the world coordinate system and use the actual area as the area of ​​the concave region.

[0149] Optionally, the fall detection device further includes:

[0150] A speech recognition module, used to collect the ambient sound of the set area, perform speech recognition on the ambient sound, and obtain a speech recognition result;

[0151] The radar activation module based on voice recognition is used to execute the acquisition of point cloud messages obtained by the radar scanning a set area, perform human target recognition on the point cloud messages, and obtain the number of human targets if it is detected that the sound intensity in the voice recognition result is greater than a preset decibel value and the voice recognition result contains preset keywords.

[0152] Optionally, the three-dimensional reconstruction module 72 includes:

[0153] A timbre recognition unit, used to perform timbre recognition on the environmental sound to obtain a timbre recognition result;

[0154] The reconstruction start unit based on timbre recognition is used to execute the three-dimensional spatial reconstruction of the point cloud message in the world coordinate system to obtain a three-dimensional point cloud distribution map of the set area if it is detected that the difference between the number of timbre in the timbre recognition result and the number of human targets is less than a difference threshold.

[0155] Optionally, the speech recognition module includes:

[0156] A radar parameter acquisition unit, configured to acquire a current sampling frequency and a current filtering threshold of the radar if it is detected that the sound intensity in the speech recognition result is greater than a preset decibel value and the speech recognition result contains a preset keyword;

[0157] A sampling frequency adjustment unit, configured to adjust the current sampling frequency of the radar to a maximum sampling frequency when the current sampling frequency is not the maximum sampling frequency;

[0158] A filtering threshold adjustment unit is used to adjust the current filtering threshold of the radar to the minimum filtering threshold when the current filtering threshold is not the minimum filtering threshold.

[0159] Optionally, a radar start module based on voice recognition includes:

[0160] A current time acquisition unit, used to acquire the current time;

[0161] The fall alarm interval determination unit is used to execute the acquisition of the point cloud message obtained by the radar scanning the set area if the interval between the current moment and the time when the fall was last determined is greater than the preset time interval, and to perform human target recognition on the point cloud message to obtain the number of human targets.

[0162] Figure 7 This is a schematic diagram of the structure of a computer device provided in Example 8 of the present application. Figure 7 As shown, the computer device of this embodiment includes: at least one processor ( Figure 7Only one is shown in the figure), a memory, and a computer program stored in the memory and executable on at least one processor, and when the processor executes the computer program, the steps in any of the above-mentioned embodiments of the fall detection method based on crowd gathering are implemented.

[0163] The computer device may include, but is not limited to, a processor and a memory. Those skilled in the art will appreciate that Figure 5 This is merely an example of a computer device and does not constitute a limitation on the computer device. The computer device may include more or fewer components than those shown in the figure, or a combination of certain components, or different components. For example, it may also include a network interface, a display screen, and an input device.

[0164] The processor may be a CPU, or other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0165] The memory includes a readable storage medium, an internal memory, etc., wherein the internal memory may be the memory of a computer device, and the internal memory provides an environment for the operation of an operating system and computer-readable instructions in the readable storage medium. The readable storage medium may be a hard disk of a computer device, and in other embodiments may also be an external storage device of a computer device, for example, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on a computer device. Further, the memory may also include both an internal storage unit of a computer device and an external storage device. The memory is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as program codes of computer programs, etc. The memory may also be used to temporarily store data that has been output or is to be output.

[0166] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned device can refer to the corresponding process in the above-mentioned method embodiment, which will not be repeated here. If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned method embodiment when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device capable of carrying computer program code, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium. For example, USB flash drive, mobile hard disk, magnetic disk or optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.

[0167] The present application implements all or part of the processes in the above-mentioned embodiment method, and may also be completed through a computer program product. When the computer program product runs on a computer device, the computer device can implement the steps in the above-mentioned method embodiment when executing the computer program product.

[0168] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0169] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0170] In the embodiments provided in the present application, it should be understood that the disclosed devices / computer equipment and methods can be implemented in other ways. For example, the device / computer equipment embodiments described above are only schematic, for example, the division of modules or units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0171] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0172] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A fall detection method based on crowd gathering, characterized in that: The fall detection method comprises: Obtaining a point cloud message obtained by the radar scanning a set area, performing human target recognition on the point cloud message, and obtaining the number of human targets; If the number of human targets detected reaches a preset condition, it is determined to be a crowd gathering, and the point cloud message is reconstructed in three-dimensional space using the world coordinate system to obtain a three-dimensional point cloud distribution map of the set area; Extracting height key points from the three-dimensional point cloud distribution map, and fitting the height key points to obtain a height surface, wherein the height key points are point clouds with the highest height value at the same horizontal position; Determine a segmentation height according to all height values ​​in the height surface, and use the segmentation height to perform concave-convex division on the height surface to obtain at least one concave area; For any concave area, the area of ​​the concave area and the change rate between the center point of the concave area and the surrounding convex surface are obtained, and when the area is within a preset area range and the change rate is within a preset change rate interval, it is determined that the human target corresponding to the concave area is a fall; The calculation method of the change rate is: Calculating the height difference between the center point of the concave area and the surrounding convex surface, and then comparing the obtained difference with the height of the surrounding convex surface to determine the change rate; The step of determining a segmentation height according to all height values ​​in the height surface, and using the segmentation height to perform concave-convex division on the height surface to obtain at least one concave area comprises: Sampling the height surface to determine the height value of each sampling point; Based on the height value, DBSCAN clustering is performed on all sampling points to obtain a clustering result. If there are at least two clustering sets in the clustering result, the average value of all height key points in the height surface is used as the segmentation height; Taking an area in the height surface whose height value is less than the segmentation height as a to-be-determined area, if the to-be-determined area matches a sampling point in any clustering set in the clustering result, determining the to-be-determined area as a concave area; After obtaining at least one concave area, the method further includes: For any concave area, project the concave area on a horizontal plane to obtain a projection area; The actual area of ​​the projection area in the world coordinate system is calculated, and the actual area is used as the area of ​​the concave area.

2. The fall detection method according to claim 1, characterized in that: After extracting the height key points in the three-dimensional point cloud distribution map, the method further includes: Eliminate N height key points with the smallest values ​​among the height key points to determine the remaining height key points; Obtaining a first correction weight according to the remaining height key points; The preset area range threshold is corrected according to the first correction weight to obtain the preset area range.

3. The fall detection method according to claim 2, characterized in that: The step of obtaining a first correction weight according to the remaining height key points includes: According to the discrete degree of the remaining height key points, the first correction weight is obtained: The first modified weight calculation formula is as follows: ;in, is the first modified weight, is the variance of the remaining height keypoints, Represents a natural constant.

4. The fall detection method according to claim 1, characterized in that: Before acquiring the point cloud message obtained by scanning the set area by the radar and performing human target recognition on the point cloud message to obtain the number of human targets, the method further includes: Collecting ambient sound in the set area, performing speech recognition on the ambient sound, and obtaining a speech recognition result; If it is detected that the sound intensity in the speech recognition result is greater than a preset decibel value, and the speech recognition result contains a preset keyword, the step of acquiring the point cloud message obtained by scanning the set area by the radar, performing human target recognition on the point cloud message, and obtaining the number of human targets; After the crowd gathering is determined, it also includes: Performing timbre recognition on the environmental sound to obtain a timbre recognition result; If it is detected that the difference between the number of timbre in the timbre recognition result and the number of human targets is less than a difference threshold, the three-dimensional spatial reconstruction of the point cloud message using the world coordinate system is performed to obtain a three-dimensional point cloud distribution map of the set area.

5. The fall detection method according to claim 4, characterized in that: After collecting the ambient sound of the set area, performing voice recognition on the ambient sound, and obtaining the voice recognition result, the method further includes: If it is detected that the sound intensity in the speech recognition result is greater than a preset decibel value, and the speech recognition result contains a preset keyword, then obtaining the current sampling frequency and the current filtering threshold of the radar; When the current sampling frequency is not the maximum sampling frequency, adjusting the current sampling frequency of the radar to the maximum sampling frequency; When the current filtering threshold is not the minimum filtering threshold, the current filtering threshold of the radar is adjusted to the minimum filtering threshold.

6. The fall detection method according to claim 4, characterized in that: Before the step of acquiring the point cloud message obtained by scanning the set area by the radar and performing human target recognition on the point cloud message to obtain the number of human targets, the method further includes: Get the current time; If the interval between the current moment and the time when a fall was last determined is greater than a preset time interval, the step of acquiring the point cloud message obtained by scanning the set area by the radar is executed, and human target recognition is performed on the point cloud message to obtain the number of human targets.

7. A fall detection device based on crowd gathering, characterized in that: The fall detection device comprises: A radar data processing module is used to obtain a point cloud message obtained by the radar scanning a set area, perform human target recognition on the point cloud message, and obtain the number of human targets; A three-dimensional reconstruction module is used to determine that a crowd is gathering if the number of human targets detected reaches a preset condition, and to perform three-dimensional spatial reconstruction of the point cloud message in a world coordinate system to obtain a three-dimensional point cloud distribution map of the set area; A height surface determination module is used to extract height key points in the three-dimensional point cloud distribution map, and to fit the height key points to obtain a height surface, wherein the height key points are point clouds with the highest height value at the same horizontal position; A concave surface determination module, used to determine a segmentation height according to all height values ​​in the height surface, and use the segmentation height to perform concave-convex division on the height surface to obtain at least one concave surface area; A fall determination module is used to obtain, for any concave area, the area of ​​the concave area and the rate of change between the center point of the concave area and the surrounding convex surface, and when the area is within a preset area range and the rate of change is within a preset rate of change interval, determine that the human target corresponding to the concave area has fallen; The calculation method of the change rate is: Calculating the height difference between the center point of the concave area and the surrounding convex surface, and then comparing the obtained difference with the height of the surrounding convex surface to determine the change rate; The step of determining a segmentation height according to all height values ​​in the height surface, and using the segmentation height to perform concave-convex division on the height surface to obtain at least one concave area comprises: Sampling the height surface to determine the height value of each sampling point; Based on the height value, DBSCAN clustering is performed on all sampling points to obtain a clustering result. If there are at least two clustering sets in the clustering result, the average value of all height key points in the height surface is used as the segmentation height; Taking an area in the height surface whose height value is less than the segmentation height as a to-be-determined area, if the to-be-determined area matches a sampling point in any clustering set in the clustering result, determining the to-be-determined area as a concave area; After obtaining at least one concave area, the method further includes: For any concave area, project the concave area on a horizontal plane to obtain a projection area; The actual area of ​​the projection area in the world coordinate system is calculated, and the actual area is used as the area of ​​the concave area.

8. A computer device, characterized in that: The computer device includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the fall detection method based on crowd gathering as described in any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the fall detection method based on crowd gathering is implemented as described in any one of claims 1 to 6.

Images