Life detection method, device and storage medium

Abstract

The application provides a life body detection method and device and a storage medium. The method comprises the following steps: acquiring a first radar echo of a to-be-detected area, and constructing a sampling window to traverse the first radar echo to acquire a second radar echo in the sampling window at each sliding window; acquiring a trained target life body detection model, and acquiring a life body detection result of the second radar echo based on the target life body detection model; and acquiring a target life body detection result of the to-be-detected area corresponding to the first radar echo according to the life body detection result of each second radar echo. In the application, the calculation amount of the life body detection result of the radar echo is reduced, the efficiency of the life body detection is improved, the accuracy of the life body detection in the to-be-detected area is improved, and the life body detection algorithm is optimized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of signal processing, and particularly relates to a living body detection method and device and a storage medium. BACKGROUND

[0002] With the development of technology, people can detect the life information of trapped personnel at a rescue site by using a living body detection technology, so as to realize rescue of the trapped personnel.

[0003] In the related art, the living body detection technology can extract the vital sign signals of trapped personnel in a rescue site by using a related short-time Fourier transform algorithm, a wavelet transform and an empirical mode decomposition processing algorithm. However, due to the limitation of the algorithm performance and the complexity of the environment of the rescue site, the living body detection technology in the related art can be affected by the complex environment of the rescue site, thereby resulting in low accuracy of the vital sign signal detection of the trapped personnel. SUMMARY

[0004] The present application aims to solve one of the problems in the related art to some extent. To this end, one object of the present application is to provide a living body detection method, device and storage medium to solve the problem of low accuracy of vital sign signal detection of trapped personnel due to the influence of the complex environment of the rescue site in the related art to some extent. The technical solution of the present application is as follows:

[0005] The first aspect of the present application provides a living body detection method, comprising: acquiring a first radar echo of a to-be-detected area, and constructing a sampling window to traverse the first radar echo to acquire a second radar echo in the sampling window at each sliding window; acquiring a trained target living body detection model, and acquiring a living body detection result of the second radar echo based on the target living body detection model; and acquiring a target living body detection result of the to-be-detected area corresponding to the first radar echo according to the respective living body detection results of all the second radar echoes.

[0006] In addition, the living body detection method provided by the first aspect of the present application also has the following additional features:

[0007] According to an embodiment of the present application, the acquiring of the target living body detection result of the to-be-detected area corresponding to the first radar echo according to the respective living body detection results of all the second radar echoes comprises: integrating the respective living body detection results of all the second radar echoes to acquire a candidate living body detection result of the first radar echo; determining and deleting a misrecognition result in the candidate living body detection result, and determining the candidate living body detection result after the deletion as the target living body detection result of the to-be-detected area corresponding to the first radar echo.

[0008] According to an embodiment of the present application, the integrating the life body detection results of all the second radar echoes to obtain the candidate life body detection result of the first radar echo comprises: splicing the life body detection results of all the second radar echoes according to positions of the second radar echoes in the first radar echo, and obtaining the candidate life body detection result of the first radar echo according to the spliced life body detection results.

[0009] According to an embodiment of the present application, the determining the misrecognition result in the candidate life body detection result and deleting, and determining the target life body detection result of the to-be-detected region corresponding to the first radar echo as the candidate life body detection result after deletion, comprises: clustering the candidate life body detection result to obtain a maximum Euclidean distance and an average Euclidean distance of each cluster; determining the misrecognition result from the candidate life body detection result according to the maximum Euclidean distance and the average Euclidean distance; and deleting the misrecognition result in the candidate life body detection result to obtain the target life body detection result.

[0010] According to an embodiment of the present application, after the obtaining the target life body detection result, comprising: obtaining a detection state value, a prediction state value and a Kalman gain of a life body in the to-be-detected region at a single slow time from the target life body detection result; determining a target state value of the life body at the single slow time according to the detection state value, the prediction state value and the Kalman gain; and obtaining a moving track of the life body in the to-be-detected region according to target state values corresponding to all the slow times in the first radar echo.

[0011] According to an embodiment of the present application, the obtaining the first radar echo of the to-be-detected region comprises: obtaining an initial radar echo of the to-be-detected region according to an ultra-wideband radar; eliminating a static background wave component and a direct current component in the initial radar echo to obtain a candidate radar echo of the to-be-detected region after elimination; performing gain control on the candidate radar echo, and taking the radar echo obtained after the gain control as the first radar echo of the to-be-detected region.

[0012] According to an embodiment of the present application, the eliminating the static background wave component and the direct current component in the initial radar echo to obtain the candidate radar echo of the to-be-detected region after elimination comprises: eliminating the static background wave component in the initial radar echo based on a channel signal subtraction method; and eliminating the direct current component in the initial radar echo based on the time average subtraction method.

[0013] According to an embodiment of the present application, before the target life body detection model is obtained and the life body detection result of the second radar echo is extracted based on the target life body detection model, the method comprises: obtaining a life body detection model to be trained; obtaining a sample life body signal, a sample environment signal and a sample interference signal to generate a training sample of the life body detection model; and training the life body detection model to be trained based on the training sample until the training is completed to obtain the target life body detection model.

[0014] The second aspect of the present application further provides a life body detection device, comprising: a traversal module configured to obtain a first radar echo of a to-be-detected area, and to construct a sampling window to traverse the first radar echo to obtain a second radar echo in the sampling window at each sliding window; an identification module configured to obtain a target life body detection model trained, and to obtain a life body detection result of the second radar echo based on the target life body detection model; and an obtaining module configured to obtain a target life body detection result of the to-be-detected area corresponding to the first radar echo according to the life body detection result of each of the second radar echoes.

[0015] The third aspect of the present application provides a computer readable storage medium having a computer program stored thereon, wherein the program is executed by a processor to implement the life body detection method according to the first aspect.

[0016] The life body detection method and device provided by the present application obtain a first radar echo of a to-be-detected area, traverse the first radar echo through a slidingly constructed sampling window to obtain a second radar echo framed in the sampling window. Further, the second radar echo is input into a target life body detection model trained, and a life body detection result of the second radar echo is obtained according to the output result of the model. The life body detection result of the first radar echo is obtained according to the life body detection result of each of the second radar echoes, and is determined as a target life body detection result of the to-be-detected area. In the present application, the first radar echo is traversed and sampled to obtain the second radar echo, and the life body detection result of the first radar echo is obtained through the life body detection result of the second radar echo, which reduces the calculation amount of obtaining the life body detection result of the radar echo, improves the efficiency of life body detection, determines the life body detection result based on the target life body detection model trained, and improves the accuracy of life body detection in the to-be-detected area, and optimizes the life body detection algorithm.

[0017] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0018] The accompanying drawings, which are incorporated herein and form a part of the specification, illustrate embodiments consistent with the present application and, together with the description, further serve to explain the principles of the application and, do not limit its scope.

[0019] Figure 1 A schematic diagram of a living body detection method according to an embodiment of the present application;

[0020] Figure 2 A schematic diagram of a radar echo according to an embodiment of the present application;

[0021] Figure 3 A schematic diagram of a living body detection method according to another embodiment of the present application;

[0022] Figure 4 A schematic diagram of a radar echo according to another embodiment of the present application;

[0023] Figure 5 A schematic diagram of a living body detection method according to another embodiment of the present application;

[0024] Figure 6 A schematic diagram of a radar echo according to another embodiment of the present application;

[0025] Figure 7 A schematic diagram of a living body detection method according to another embodiment of the present application;

[0026] Figure 8 A schematic diagram of a living body detection method according to another embodiment of the present application;

[0027] Figure 9 A schematic diagram of a living body detection model according to an embodiment of the present application;

[0028] Figure 10 A schematic diagram of a radar echo according to another embodiment of the present application;

[0029] Figure 11 A schematic diagram of a loss value change of a living body detection model training according to an embodiment of the present application;

[0030] Figure 12 A schematic diagram of a recognition accuracy of a living body detection model according to an embodiment of the present application;

[0031] Figure 13 A schematic diagram of an output of a living body detection model according to an embodiment of the present application;

[0032] Figure 14 A schematic diagram of a living body detection method according to another embodiment of the present application;

[0033] Figure 15 A schematic diagram of a living body detection device according to an embodiment of the present application;

[0034] Figure 16 This is a block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0035] In order to enable ordinary persons in the art to better understand the technical solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. The same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be understood as limiting the present application.

[0036] Figure 1 This is a schematic diagram of a method for detecting living organisms according to an embodiment of the present application. Figure 1 As shown, the method includes:

[0037] S101, obtaining a first radar echo of the area to be measured, and constructing a sampling window to traverse the first radar echo, and obtaining a second radar echo within the sampling window each time the window is slid.

[0038] In some scenarios, there may be life in an area covered by an entity. In this scenario, it is impossible to directly obtain information about life in the area through observation. However, radar detection of the area can be used to obtain information about life in the area.

[0039] The area where life body detection is required may be identified as the area to be detected.

[0040] Optionally, a pulse signal may be sent to the area to be measured by a radar detection instrument, and a radar echo of the area to be measured may be collected as a first radar echo of the area to be measured, wherein the first radar echo may be a radar echo matrix.

[0041] Furthermore, the first radar echo returned from the area to be tested includes radar echoes returned by life in the area to be tested, as well as radar echoes returned by other non-living objects in the area to be tested. Therefore, the life information in the area to be tested can be determined by analyzing and identifying the information carried by the first radar echo.

[0042] like Figure 2 As shown, the first radar echo of the area to be measured is set to be obtained Figure 2 As shown in the image number 1, the dotted box in the figure is the partial radar echo returned by the living body in the test area obtained based on the analysis of the first radar echo in the test area, wherein the partial radar echo is Figure 2 From the mapping position information in the fast time dimension and the slow time dimension shown in the image number 1, it can be seen that the life form is located 4 meters away from the radar detection instrument and has not moved to other locations.

[0043] For example, the first radar echo of the to-be-detected region is set to be obtained Figure 2 As shown in the image numbered 2, the part of the radar echo returned by the living body in the to-be-detected region is analyzed based on the first radar echo of the to-be-detected region, which is framed by the dashed box in the image. Figure 2 As shown in the mapping position information of the fast time dimension and the slow time dimension of the image numbered 2, the living body moves back and forth between 3 m and 12 m away from the radar detection instrument.

[0044] Further, in order to more accurately detect the information of the living body from the first radar echo, the first radar echo can be divided and sampled, and the vital sign information of the part of the radar echo of the living body after sampling can be analyzed.

[0045] Optionally, a corresponding sampling window can be constructed according to the first radar echo, and the first radar echo is traversed through sliding of the sampling window, wherein the part of the radar echo framed in the sampling window when the sampling window is sliding is determined as the second radar echo.

[0046] S102, obtaining a trained target living body detection model, and obtaining a living body detection result of the second radar echo based on the target living body detection model.

[0047] In the embodiment of the application, the living body in the second radar echo can be identified through the trained target living body detection model, wherein the second radar echo obtained by sampling can be input into the trained target living body detection model, and the living body detection result corresponding to the second radar echo can be obtained through analysis and processing of the second radar echo by the target living body detection model.

[0048] Optionally, the output result of the target living body detection model can be used to determine whether the second radar echo carries vital sign data of the living body in the to-be-detected region, so as to identify and detect the living body in the to-be-detected region.

[0049] S103, obtaining a target living body detection result of the to-be-detected region corresponding to the first radar echo according to the living body detection result of each of the second radar echoes.

[0050] In the embodiment of the application, the second radar echo is the part of the radar echo in the first radar obtained by traversing the first radar echo, so that the second radar echoes can cover the first radar echo.

[0051] In this scenario, the living body detection result corresponding to the first radar echo can be obtained according to the living body detection result of each of the second radar echoes.

[0052] Optionally, the pulse signal sent by the radar detection instrument has a corresponding transmission time, the first radar echo corresponding to the pulse signal has a corresponding return time, and the second radar echo obtained by traversing sampling of the first radar echo has a corresponding return time. Based on the return time corresponding to each of the second radar echoes, the life body detection results of all the second radar echoes can be spliced in the time sequence dimension, and the life body detection result in the first radar echo can be obtained as the target life body detection result of the to-be-detected region according to the life body detection result obtained after splicing.

[0053] The life body detection method provided in the present application obtains the first radar echo of the to-be-detected region, traverses the first radar echo through a sliding constructed sampling window, and obtains the second radar echo framed by the sampling window. Further, the second radar echo is input into the trained target life body detection model, and the life body detection result of the second radar echo is obtained according to the output result of the model. The life body detection result of the first radar echo is obtained according to the life body detection result of each of the second radar echoes, and is determined as the target life body detection result of the to-be-detected region. In the present application, the second radar echo is obtained by traversing sampling of the first radar echo, and the life body detection result of the first radar echo is obtained through the life body detection result of the second radar echo, which reduces the calculation amount of obtaining the life body detection result of the radar echo, improves the efficiency of life body detection, and improves the accuracy of life body detection in the to-be-detected region based on the trained target life body detection model, thereby optimizing the life body detection algorithm.

[0054] In the above embodiments, the target life body detection result of the to-be-detected region can be obtained in combination with Figure 3 It is further understood that Figure 3 The flowchart of the life body detection method of another embodiment of the present application is shown in Figure 3 As shown in the figure, the method comprises the following steps:

[0055] S301, integrating the life body detection result of each of the second radar echoes to obtain the candidate life body detection result of the first radar echo.

[0056] In the embodiments of the present application, the life body detection result of each of the second radar echoes can be integrated based on the data integration method in the related art, and the integrated life body detection result can be determined as the candidate life body detection result of the first radar echo.

[0057] Optionally, the second radar echo has a corresponding position in the first radar echo, and therefore, the life body detection results of all the second radar echoes are spliced according to the positions of the second radar echoes in the first radar echo, and the candidate life body detection result of the first radar echo is obtained according to the spliced life body detection result.

[0058] In implementations, the first radar echo is a radar echo signal generated by a radar detection instrument sending a pulse signal sequence with a set duration in a to-be-detected region, and in a scenario of traversing the first radar echo by a sliding sampling window, the second radar echo sampled has a corresponding position in the first radar echo.

[0059] Therefore, the position relationship between the life body detection results of all the second radar echoes can be determined based on the positions of the second radar echoes in the first radar echo, and further, the life body detection results of all the second radar echoes are spliced based on the position relationship, and the candidate life body detection result of the first radar echo is obtained according to the spliced life body detection result.

[0060] In some implementations, there may be an overlap of part of the radar echo between the second radar echoes with adjacent positions in the first radar echo, and therefore, the spliced life body detection result can be de-duplicated to obtain the candidate life body detection result of the first radar echo.

[0061] S302, determine the misrecognition result in the candidate life body detection result and delete it, and determine the candidate life body detection result after deletion as the target life body detection result of the to-be-detected region corresponding to the first radar echo.

[0062] In implementations, due to the complexity of the environment in the to-be-detected region, there may be misrecognition results in the candidate life body detection result of the first radar echo, and therefore, the misrecognition results in the candidate life body detection result need to be determined and deleted from the candidate life body detection result.

[0063] Optionally, the candidate life body detection result can be clustered to obtain the maximum Euclidean distance and the average Euclidean distance of each cluster.

[0064] In the embodiments of the present application, each life body detection result in the candidate life body detection result can be clustered, wherein the candidate life body detection result can be clustered according to a K-means clustering algorithm, and the misrecognition result in the candidate life body detection result is identified according to the cluster obtained after clustering.

[0065] The candidate life form detection results may be clustered according to the Euclidean distances between all life form detection results included in the candidate life form detection results.

[0066] Set the candidate life detection result x i The total number of life detection results included is x i ={x1,x2,x3,…,x m}, then we can get x according to the calculation formula of Euclidean distance i The corresponding k nearest neighbor distance matrices, where the Euclidean distance d xy The calculation formula is as follows:

[0067]

[0068] In the above formula, n represents the data space dimension of each life form detection result in the candidate life form detection results, and k is x i The number of corresponding close distance matrices is calculated according to the above formula in {x1,x2,x3,…,x m}, and calculate the k cluster sample points based on the k cluster sample points. i Clustering is performed, where the set of sample points consisting of k cluster sample points can be identified as N k (x).

[0069] Furthermore, based on the clustering rules in the related art, k Determine x in (x) i The cluster y i , the formula is as follows:

[0070]

[0071] In the above formula, K is the cluster y i The total number of candidate life detection results x i The total number of life detection results included is x i ={x1,x2,x3,…,x m} Construct the corresponding cluster structure, such as K-MST structure, to achieve the detection result x of candidate life forms i Clustering and obtaining cluster y after clustering i , c j It is the set of life detection results within each cluster after clustering.

[0072] In the implementation, the coverage of the cluster in the candidate life form detection results can be obtained through the maximum Euclidean distance and average Euclidean distance of the cluster. Therefore, the maximum Euclidean distance and average Euclidean distance of all clusters obtained after clustering can be calculated, and the false recognition results can be determined from the candidate life form detection results based on the maximum Euclidean distance and average Euclidean distance.

[0073] Among them, the detection results that are not covered by the cluster in the candidate life form detection results can be determined by the maximum Euclidean distance and average Euclidean distance of the cluster. These detection results can be understood as outliers that do not belong to the cluster. Therefore, the detection results corresponding to the outliers can be determined as misidentification results in the candidate life form detection results.

[0074] Furthermore, the misidentification results in the candidate life form detection results are deleted to obtain the target life form detection results.

[0075] The determined misidentification result may be deleted from the candidate life form detection result, and the deleted candidate life form detection result is the target life form detection result corresponding to the to-be-detected area carried in the first radar echo.

[0076] like Figure 4 As shown, in Figure 4 In the image numbered 1, points 1 and 2 are outliers in the image. The life form detection results corresponding to points 1 and 2 can be deleted from the candidate life form detection results, thereby obtaining the target life form detection result of the test area to which the first radar echo corresponding to the image numbered 1 belongs.

[0077] exist Figure 4 In the image numbered 2, point 3 is an outlier in the image. The life form detection result corresponding to point 3 can be deleted from the candidate life form detection results, and the target life form detection result of the test area to which the first radar echo corresponding to the image numbered 2 belongs can be obtained.

[0078] exist Figure 4 In the image numbered 3, points 4, 5, and 6 are outliers in the image. The life form detection results corresponding to points 4, 5, and 6 can be deleted from the candidate life form detection results, and the target life form detection result of the test area to which the first radar echo corresponding to the image numbered 3 belongs can be obtained.

[0079] exist Figure 4 In the image numbered 4, points 7 and 8 are outliers in the image. Therefore, the life body detection results corresponding to points 7 and 8 can be deleted from the candidate life body detection results, and the target life body detection result of the test area to which the first radar echo corresponding to the image numbered 4 belongs can be obtained.

[0080] The life body detection method provided in the application integrates respective life body detection results of all second radar echoes to obtain candidate life body detection results of the first radar echoes, further identifies misrecognition results in the candidate life body detection results and deletes the misrecognition results, so as to obtain target life body recognition results of the to-be-detected region carried by the first radar echoes. In the application, the accuracy of the target life body detection results is improved and the life body detection method is optimized by determining and deleting the misrecognition results.

[0081] Further, the state trajectory of the life body in the to-be-detected region can be obtained according to the target life body detection result, and the state trajectory can be combined with the target life body detection result to obtain the target life body detection result in the to-be-detected region. Figure 5 It is further understood that Figure 5 The flowchart of the life body detection method of another embodiment of the application is shown in FIG. 2, which includes the following steps. Figure 5

[0082] S501, obtaining, from the target life body detection result, a detection state value, a prediction state value and a Kalman gain of the life body in the to-be-detected region at a single slow time.

[0083] In the embodiment of the application, the Kalman filtering algorithm can be used to extract the trajectory of the target life body detection result, so as to obtain the moving trajectory of the life body in the to-be-detected region.

[0084] In the implementation, the first radar echo has multiple corresponding slow times, wherein the moving trajectory of the life body in the to-be-detected region can be determined according to the position, moving speed and other related information of the life body in the to-be-detected region at each slow time.

[0085] Optionally, the detection state value, the prediction state value and the corresponding Kalman gain of the life body at each single slow time can be obtained from the target life body detection result.

[0086] The detection state value can include the position and moving speed of the life body in the to-be-detected region obtained by the first radar echo, and the prediction state value can include the position and moving speed of the life body in the to-be-detected region predicted according to the Kalman filtering algorithm.

[0087] Optionally, the prediction state value X k of the life body in the to-be-detected region at the slow time k can be calculated according to the following formula:

[0088] X k = A k X k + B k U k + w k

[0089] In the above formula, U k ​is the control quantity applied to the radar detection system at slow time k, A k is the state transfer matrix acting on slow time k-1, B k For the effect on U k The control matrix on k It can be used to describe the influence of the control quantity on the state of the radar detection system, w k is the process noise of radar detection, where w k Subject to mean 0 and covariance Q k independent multivariate normal distributions.

[0090] Optionally, the detection state value Z of the life form in the detection area at the slow time k k The calculation formula is as follows:

[0091] Z k =H k X k +v k

[0092] In the above formula, X k is the predicted state value of the life in the test area at slow time k, H k is the detection matrix, where H k It can be used to describe the mapping from the actual state space to the detection space, v k is the detection noise, where v k Subject to mean 0 and covariance R k independent multivariate normal distributions.

[0093] Among them, the detection state value Z k The position of the life form in the detection area is included in the radar detection instrument. In the implementation, the radar detection instrument has a corresponding set coordinate system, and the position of the life form in the detection area is mapped in the set coordinate system.

[0094] Furthermore, the horizontal and vertical coordinates corresponding to the mapping position can be obtained by the following formula:

[0095]

[0096] In the above formula, The target detection result detected by the radar detection instrument at the slow time k includes the average value of all horizontal coordinates of the position of the living body in the detection area mapped to the set coordinate system, It is the mean value of all vertical coordinates of the position of the living body in the detection area mapped to the set coordinate system, included in the target detection result detected by the radar detection instrument at slow time k, and t is time.

[0097] In some implementations, the predicted state value of the living body in the area to be tested at the current slow time can be determined by the predicted state value of the living body in the area to be tested at the previous slow time of the current slow time. Furthermore, the target state value of the living body in the area to be tested at the current slow time can be obtained based on the detected state value and the predicted state value of the living body in the area to be tested at the current slow time.

[0098] Optionally, the current slow time is set as slow time k, and the formula for the predicted state value of the living body in the test area at the previous slow time of the current slow time can be as follows:

[0099]

[0100]

[0101] In the above formula, is the prior prediction value at slow time k obtained based on the predicted state value at slow time k-1, is the target state value of the life form in the test area at slow time k-1, is the estimated error between the predicted state value and the detected state value of the radar detection system at slow time k, Q k is the detection process noise w of the radar detection system k The covariance matrix, P k-1 is the posterior error covariance matrix at slow time k-1, U k is the control quantity applied to the radar detection system at slow time k, A k is the state transfer matrix acting on slow time k-1, A k The transposed matrix, B k For the effect on U k The control matrix on k It can be used to describe the influence of the control quantity on the state of the radar detection system, w k is the process noise of radar detection, where w k Subject to mean 0 and covariance Q k independent multivariate normal distributions.

[0102] Further, obtain the Kalman gain K at slow time k k , the formula can be shown as follows:

[0103]

[0104] In the above formula, R k is the detection noise v of the radar detection system k The covariance matrix of Hk is an estimation error value between the predicted state value and the detected state value of the radar detection system at the slow time k k Hk is a detection matrix, wherein, H k Hk can be used to describe the mapping from the actual state space to the detection space, Hk is the transpose matrix of H k .

[0105] Further, the association between the predicted state value and the detected state value at the slow time k is determined according to the Kalman gain, so as to obtain the target state value of the living body in the to-be-detected region at the slow time k.

[0106] S502, the target state value of the living body at a single slow time is determined according to the detected state value, the predicted state value and the Kalman gain.

[0107] In the embodiments of the present application, the detected state value, the predicted state value and the Kalman gain can be calculated based on a set algorithm, so as to obtain the target state value of the living body at a single slow time.

[0108] Wherein, the current slow time is set as the slow time k, and the target state value The calculation formula of is as follows:

[0109]

[0110] In the above formula, Hk-1 is an a priori prediction value at the slow time k obtained according to the predicted state value at the slow time k-1, K k Hk is the Kalman gain at the slow time k, Z k Hk is the detected state value of the living body in the to-be-detected region at the slow time k, H k Hk is a detection matrix, wherein, H k Hk can be used to describe the mapping from the actual state space to the detection space.

[0111] It should be noted that the target state value of the living body at the slow time k includes the target position and the target moving speed of the living body in the to-be-detected region at the slow time k, wherein, the target position and the target moving speed can be understood as the position parameter and the moving speed parameter which are the most possible to approach the actual position and the actual moving speed of the living body in the to-be-detected region.

[0112] S503, the moving track of the living body in the to-be-detected region is obtained according to the target state values corresponding to all slow times in the first radar echo.

[0113] In the embodiments of the present application, according to the target position and target speed included in the target state value of each of the living beings at all individual slow times, the moving position and moving speed of the living beings in the to-be-detected region can be obtained, so as to generate the moving track of the living beings in the to-be-detected region.

[0114] For example, it is known from the target state value corresponding to each of all slow times included in the first radar echo by the living beings that there are three living beings in the to-be-detected region, wherein the first living being is located at a position 5 m away from the radar detection instrument and does not move, the second living being is located at a position 8 m away from the radar detection instrument and does not move, and the third living being is located at a position 11 m away from the radar detection instrument and does not move, and then the moving track of the three living beings in the to-be-detected region in this scenario can be as shown in the image numbered 1 in FIG. 1. Figure 6

[0115] For another example, it is known from the target state value corresponding to each of all slow times included in the first radar echo by the living beings that there is one living being in the to-be-detected region, wherein the living being reciprocates between a position 3 m away from the radar detection instrument and a position 7.5 m away from the radar detection instrument based on a set speed, and then the moving track of the living being in the to-be-detected region in this scenario can be as shown in the image numbered 2 in FIG. 1. Figure 6

[0116] For another example, it is known from the target state value corresponding to each of all slow times included in the first radar echo by the living beings that there are three living beings in the to-be-detected region, wherein the three living beings reciprocate between a position 4 m away from the radar detection instrument and a position 12 m away from the radar detection instrument based on a set speed, and then the moving track of the three living beings in the to-be-detected region in this scenario can be as shown in the image numbered 3 in FIG. 1. Figure 6

[0117] For another example, it is known from the target state value corresponding to each of all slow times included in the first radar echo by the living beings that there is one living being in the to-be-detected region, wherein the living being randomly moves in the to-be-detected region, and then the moving track of the living being in the to-be-detected region in this scenario can be as shown in the image numbered 4 in FIG. 1. Figure 6

[0118] It should be noted that after the target state value of the living being at the current slow time is obtained, the posterior error covariance matrix corresponding to the current slow time can be updated based on the target state value at the current slow time, so that the target state value of the living being in the to-be-detected region at the next slow time can be obtained according to the updated posterior error covariance matrix.

[0119] wherein it is assumed that the current slow time is k, and the posterior error covariance matrix P k at the slow time k​​​​k The calculation formula is:

[0120]

[0121] In the above formula, K k is the Kalman gain on the slow time k, H k is a detection matrix, wherein H k can be used to describe the mapping from the actual state space to the detection space, is the estimated error value between the predicted state value and the detection state value of the radar detection system on the slow time k, and I is a unit matrix.

[0122] The life body detection method provided in the present application obtains the target state value of the life body on a single slow time according to the target life body detection result, and further obtains the moving track of the life body in the to-be-detected region according to the respective target state values on all slow times. In the present application, the moving track of the life body in the to-be-detected region is obtained through the target life body detection result, which improves the accuracy of the moving track and optimizes the precision of the moving track.

[0123] In the above embodiment, regarding the acquisition of the first radar echo, the following can be combined Figure 7 for further understanding Figure 7 is a flowchart of the life body detection method of another embodiment of the present application, as shown in Figure 7 , the method comprises the following steps.

[0124] S701, obtaining an initial radar echo of a to-be-detected region according to an ultra-wideband radar.

[0125] In the embodiment of the present application, the radar detection device configured with the ultra-wideband radar can be used to send a continuous pulse signal sequence to the to-be-detected region and receive the echo signal returned by the to-be-detected region.

[0126] Further, the received echo signal returned by the to-be-detected region is used as the initial radar echo of the to-be-detected region.

[0127] S702, eliminating the static background wave component and the direct current component in the initial radar echo to obtain a candidate radar echo of the to-be-detected region after elimination.

[0128] In the embodiment of the present application, due to the complexity of the environment of the to-be-detected region, part of the echo in the initial radar echo may have an impact on the life body detection result, and therefore, the part of the echo needs to be eliminated.

[0129] In some implementations, the initial radar echo can include the following components as shown in the following formula:

[0130] R[m, n] = r[m, n] + c[n] + w[m, n] + d[m] + l[m, n] + t[m, n]

[0131] In the above formula, R[m, n] is the initial radar echo, r[m, n] is the vital sign signal of the living body in the to-be-detected region, c[n] is the static background wave, w[m, n] is the additive white noise, d[m] is the direct current component in the unstable fast time, l[m, n] is the linear trend generated in the slow time due to the amplitude instability of the radar detection system in the detection process, and t[m, n] is the harmonic and signal distortion caused by the moving object in the to-be-detected region.

[0132] In the initial radar echo, the static background wave and the direct current component in the unstable fast time may affect the detection result of the living body, and therefore, it is necessary to eliminate them from the initial radar echo.

[0133] Optionally, the static background wave component in the initial radar echo can be eliminated based on a trace signal subtraction method, where the calculation formula of the trace signal subtraction method is as follows:

[0134] R'[m, 1] = R[m, 1]

[0135] R'[m, n] = R[m, n] - R[m, n-1] (m = 1, 2,..., M) (n = 2, 3,..., N)

[0136] In the above formula, R'[m, n] is the initial radar echo after eliminating the static background wave component from the initial radar echo.

[0137] Optionally, the direct current component in the initial radar echo can be eliminated based on a time average subtraction method, where the calculation formula of the time average subtraction method is as follows:

[0138]

[0139] In the above formula, R''[m, n] is the initial radar echo after eliminating the direct current component from the initial radar echo after eliminating the static background wave component.

[0140] Further, the initial radar echo after eliminating the static background wave component and the direct current component in the unstable fast time can be determined as the candidate radar echo of the to-be-detected region.

[0141] S703, gain control is performed on the candidate radar echo, and the radar echo obtained after the gain control is taken as the first radar echo of the to-be-detected region.

[0142] In the embodiment of the present application, in order to improve the signal-to-noise ratio of the vital sign signal carried in the radar echo, the candidate radar echo can be gain controlled to enhance the vital sign signal in the fast time direction.

[0143] Optionally, the algorithm for gain control of the candidate radar echo may be as shown in the following formula:

[0144]

[0145]

[0146] In the above formula, d is the window length of the time window, g max is the maximum gain threshold, g mask is the gain coefficient.

[0147] Furthermore, the candidate radar echo obtained after gain control is used as the first radar echo of the area to be measured.

[0148] The method for detecting life forms proposed in this application eliminates static background clutter and DC components from the initial radar echo of the test area to obtain a candidate radar echo after the elimination. Furthermore, the gain of the candidate radar echo is controlled to obtain a first radar echo from the test area. By preprocessing the signal of the initial radar echo, interference factors in the first radar echo are eliminated, while the vital sign signal in the first radar echo is strengthened, thereby improving the accuracy of life form detection and optimizing the method for detecting life forms.

[0149] In the above embodiment, the acquisition of the target life body detection model can be combined with Figure 8 Further understanding, Figure 8 This is a flow chart of a method for detecting living organisms according to another embodiment of the present application. Figure 8 As shown, the method includes:

[0150] S801: Obtain a life form detection model to be trained.

[0151] In the embodiment of the present application, a life form detection model to be trained can be constructed based on the CNN model structure, such as Figure 9 As shown, the life form detection model to be trained may include Figure 9 The 5-layer neural network shown includes 4 feature extraction layers and 1 fully connected layer.

[0152] like Figure 9 As shown, the four feature extraction layers include feature extraction layer 1, feature extraction layer 2, feature extraction layer 3, and feature extraction layer 4, wherein feature extraction layer 1, feature extraction layer 2, feature extraction layer 3, and feature extraction layer 4 respectively include 1 convolution layer, 1 batch normalization layer, and 1 nonlinear function activation layer (relu).

[0153] Further, as shown in Figure 9 , the life body detection model to be trained further includes one max-pooling layer and one residual connection, wherein the residual connection includes one convolutional layer and one batch normalization layer, and the feature map output by the feature extraction layer 2 can be transmitted to the nonlinear function activation layer (relu) of the feature extraction layer 4 through the residual connection.

[0154] In the implementation, the neural network layer of the life body detection model has a set of model parameters. Optionally, the model parameter settings of the life body detection model in the embodiment of the present application can be as shown in Figure 9 , wherein the number of convolution kernels N C in the convolutional layer of the feature extraction layer 1 is 8, the convolution kernel size FH×FW is 3×3, the step length SL is 1, and the padding amount P n is 1. The number of convolution kernels N C in the convolutional layer of the feature extraction layer 2 is 16, the convolution kernel size FH×FW is 3×3, the step length SL is 1, and the padding amount P n is 1. The number of convolution kernels N C in the convolutional layer of the feature extraction layer 3 is 8, the convolution kernel size FH×FW is 3×3, the step length SL is 1, and the padding amount P n is 1. The number of convolution kernels N C in the convolutional layer of the feature extraction layer 4 is 32, the convolution kernel size FH×FW is 3×3, the step length SL is 1, and the padding amount P n is 1. The number of convolution kernels N C in the max-pooling layer is 32, the convolution kernel size FH×FW is 2×2, the step length SL is 2, and the padding amount P n is 0. The number of convolution kernels N C in the convolutional layer of the residual connection is 32, the convolution kernel size FH×FW is 1×1, the step length SL is 1, and the padding amount P n is 0.

[0155] Correspondingly, the input size and output size of the neural network layer of the life body detection model have a set of size parameters, as shown in Figure 9As shown, the input size in the feature extraction layer 1 is (1, 10, 20), the output size is (8, 10, 20), the input size in the feature extraction layer 2 is (8, 10, 20), the output size is (16, 10, 20), the input size in the feature extraction layer 3 is (16, 10, 20), the output size is (16, 10, 20), the input size in the feature extraction layer 4 is (16, 10, 20), the output size is (32, 10, 20), the input size of the max pooling layer is (32, 20, 20), the output size is (32, 5, 10), the input size of the full connection layer is 32x5x10 = 1600, and the output size is 3.

[0156] It should be noted that the batch normalization layer and the nonlinear function activation layer relu in each feature extraction layer enable the living body detection model to realize the normalization of weight parameters.

[0157] S802, obtaining a sample vital sign signal, a sample environment signal, and a sample interference signal to generate a training sample of the living body detection model.

[0158] In order to effectively train the living body detection model to be trained, the corresponding training sample can be constructed based on the actual rescue site environment.

[0159] In some implementations, the radar echo signal collected at the rescue site can include the vital sign signal of the trapped person, the environment signal of the rescue site, and the interference signal in the rescue site environment caused by the movement of the trapped person. Therefore, the corresponding sample vital sign signal, sample environment signal, and sample interference signal can be obtained, and the obtained sample vital sign signal, sample environment signal, and sample interference signal can be mixed to obtain the training sample required for training the living body detection model.

[0160] As shown in FIG. 8, the image numbered 1 is the corresponding image of the sample vital sign signal, the image numbered 2 is the corresponding image of the sample environment signal, and the image numbered 3 is the corresponding image of the sample interference signal. Figure 10 Figure 10 As shown, the image numbered 1 is the corresponding image of the sample vital sign signal, the image numbered 2 is the corresponding image of the sample environment signal, and the image numbered 3 is the corresponding image of the sample interference signal.

[0161] S803, training the living body detection model to be trained based on the training sample until the training is completed to obtain a trained target living body detection model.

[0162] ​In the embodiments of the present application, the mixed signal of the acquired sample vital sign signal, sample environment signal and sample interference signal can be batch-divided, the training samples are divided into a plurality of batch sub-samples, and the vital body detection model is trained by each sub-sample.

[0163] Optionally, in the training process of the vital body detection model, the training output result of the model can include the belonging label of the sub-sample input into the vital body detection model in the current round and the confidence of each label.

[0164] Further, the output result of the vital body detection model can be a K-dimensional vector, which includes at least one belonging label of the sub-sample input into the vital body detection model for training and recognized by the model, and the confidence of the belonging label, wherein the confidence of the at least one belonging label of the sub-sample output by the model can be identified as The value range of [q i,0 ,q i,1 ...,q i,K-1 ].

[0165] Wherein, the dimension of the vector output by the model can be determined according to the number of labels of the training sample, for example, assuming that the training sample input into the vital body detection model for training includes 3 labels, which are the corresponding labels of the vital sign signal, the corresponding labels of the environment signal and the corresponding labels of the interference signal, in this scenario, the dimension K of the vector output by the model training can be 3.

[0166] Optionally, the model parameters of the vital body detection model to be trained can be adjusted and optimized according to the training output result of each training round of the vital body detection model in the model training process and the loss value between the training sample input into the model corresponding to the training round, until the condition of the end of the model training is met.

[0167] Wherein, the training output result of each training round of the vital body detection model in the model training process and the loss value between the training sample input into the model corresponding to the training round can be obtained based on the calculation formula of the loss function loss as follows:

[0168]

[0169] In the above formula, is the cross-entropy corresponding to the model output result, is a regularization penalty term in the model training process, N is the number of samples included in the sub-sample input into the model for training in the current round, K is the number of classification labels, q i,k represents the confidence of the i-th sample being recognized as the k-th label, and λ is a regularization parameter, The weight parameters of each layer of the living body detection model, l represents the number of layers of the living body detection model, and n is the number of weights of the corresponding layer.

[0170] Optionally, the change of the loss value in the training process of the living body detection model can be as shown in the figure. Figure 11 As the number of iterations increases and the model parameters are optimized, the loss value of the living body detection model presents a downward trend.

[0171] In the embodiments of the present application, the mixed signal of the sample vital sign signal, the sample environment signal and the sample interference signal can be proportionally divided to obtain the divided training samples and test samples. For example, 80% of the mixed signal can be used as the training sample of the living body detection model, and the remaining 20% of the mixed signal can be used as the test sample of the trained living body detection model.

[0172] Optionally, the number of iterations of the living body detection model can be set, for example, the number of iterations of each training round of the living body detection model is set to 80, and the number of iterations in the model training process can be monitored and recorded. If the model parameters are iteratively optimized for the 80th time after the current model training is completed, it can be determined that the current round of model training is completed.

[0173] Further, in the process of model training, the model parameters can be adjusted and optimized based on the set model optimization algorithm.

[0174] For example, the Adam algorithm can be selected to adjust and optimize the model parameters, the initial learning rate of the model is set to 0.003, and further, the learning rate decay period is set to 10 training rounds, and the learning rate decay factor is 0.8. After 10 rounds of model training and model weight parameter adjustment and optimization, the initial learning rate can be screened based on the decay factor of 0.8 to slow down the update gradient of the weight parameters, thereby realizing the fine-tuning of the living body detection model in the training process.

[0175] Further, after the end of each round of model training, the recognition accuracy of the trained living body detection model can be obtained by testing the sample, and the end condition of the model training can be set based on the recognition accuracy of the test sample by the living body detection model.

[0176] The termination condition for model training is set as the recognition accuracy of the test sample being greater than or equal to the accuracy threshold. After each round of model training, the test sample can be input into the living organism detection model that has completed training in the current round. If the recognition success rate of the living organism detection model for the test sample after the current round of training is greater than or equal to the accuracy threshold, it can be determined that the living organism detection model that has completed training in the current round of training meets the termination condition for model training. Model training for the living organism detection model can be terminated, and the living organism detection model obtained after the current round of training is used as the trained target living organism detection model.

[0177] Optionally, the recognition accuracy of the life form detection model for the training samples and the recognition accuracy of the test samples after each round of training can be as follows: Figure 12 As shown by Figure 12 It can be seen that with the increase in the number of iterations, the recognition accuracy of the life detection model for training samples and test samples shows an upward trend.

[0178] Furthermore, the radar echo to be identified can be input into a trained target life body detection model for identification, wherein the target life body detection model can output the recognition result of the radar echo to be identified based on the form of a confusion matrix. Optionally, the confusion matrix output by the target life body detection model can be as follows: Figure 13 shown.

[0179] Depend on Figure 13 It can be seen that the radar echo to be identified input into the target life detection model includes vital sign signals, environmental signals, and interference signals. Among them, row 1 is the recognition result of the target life detection model for the vital sign signal component in the radar echo to be identified, row 2 is the recognition result of the target life detection model for the environmental signal component in the radar echo to be identified, and row 3 is the recognition result of the target life detection model for the interference signal component in the radar echo to be identified.

[0180] Among them, it can be seen from the content in row 1 that the recognition result of the target vital sign signal for the vital sign signal component in the radar echo to be identified is the vital sign signal with a confidence level of 0.95, the environmental signal with a confidence level of 0.02, and the interference signal with a confidence level of 0.03. It can be seen that the recognition result of the target vital sign signal for the vital sign signal in the radar echo to be identified is accurate.

[0181] Correspondingly, from the content in row 2, it can be seen that the recognition result of the target vital sign signal for the environmental signal component in the radar echo to be identified is the vital sign signal with a confidence level of 0.01, the environmental signal with a confidence level of 0.92, and the interference signal with a confidence level of 0.07. It can be seen that the recognition result of the target vital sign signal for the environmental signal in the radar echo to be identified is accurate.

[0182] Furthermore, from the content in row 3, it can be seen that the recognition results of the target vital signs signal for the interference signal component in the radar echo to be identified are vital signs signal with a confidence level of 0.06, environmental signal with a confidence level of 0.07, and interference signal with a confidence level of 0.87. From this, it can be seen that the recognition results of the target vital signs signal for the environmental signal in the radar echo to be identified are accurate.

[0183] The life form detection method proposed in this application obtains a life form detection model to be trained and trains the life form detection model based on training samples consisting of a mixture of sample vital sign signals, sample environmental signals, and sample interference signals, thereby obtaining a trained target life form detection model. During the life form detection process, the trained target life form detection model can be used to identify and extract vital sign signals from the first radar echo, thereby improving the accuracy of life form detection within the target area.

[0184] To better understand the above embodiments, Figure 14 , Figure 14 This is a flow chart of a method for detecting living organisms according to another embodiment of the present application. Figure 14 As shown:

[0185] After obtaining the initial radar echo from the target area, it is preprocessed by removing interference components from the initial radar echo through channel signal subtraction and time averaging. Gain control is then applied to the interference-removed initial radar echo to obtain the first radar echo from the target area. A sampling window is constructed and the first radar echo is traversed through the sliding sampling window to obtain the second radar echo. This second radar echo is then input into the trained target life detection model to obtain the life detection results of the second radar echo.

[0186] Based on the life form detection results of all the second radar echoes, the candidate life form detection results of the test area are obtained, and the candidate life form detection results are clustered to obtain the false recognition results. The false recognition results are deleted from the candidate life form detection results to obtain the target life form detection results of the test area carried by the first radar echo.

[0187] Furthermore, the movement trajectory of the life form in the area to be detected is obtained according to the target life form detection result, thereby realizing the detection of the life form in the area to be detected.

[0188] The life body detection method provided in the application obtains the second radar echo by traversing sampling the first radar echo, and obtains the life body detection result of the first radar echo through the life body detection result of the second radar echo, thereby reducing the calculation amount of obtaining the life body detection result of the radar echo, improving the efficiency of life body detection, improving the accuracy of the target life body detection result through determination and deletion of the misrecognition result, optimizing the life body detection algorithm based on the trained target life body detection model, improving the accuracy of the life body detection in the to-be-detected region, improving the accuracy of the moving track of the life body in the to-be-detected region through the target life body detection result, and optimizing the precision of the moving track.

[0189] Corresponding to the life body detection method provided in the above several embodiments, one embodiment of the application further provides a life body detection device. Since the life body detection device provided in the embodiment of the application corresponds to the life body detection method provided in the above several embodiments, the implementation manners of the life body detection method are also applicable to the life body detection device provided in the embodiment of the application, which will not be described in detail in the following embodiments.

[0190] Figure 15 The structure diagram of the life body detection device of one embodiment of the application is shown in FIG. 15, which comprises a traversal module 151, an identification module 152, an acquisition module 153, a track extraction module 154, and a training module 155. Figure 15

[0191] The traversal module 151 is configured to obtain the first radar echo of the to-be-detected region, and construct a sampling window to traverse the first radar echo, thereby obtaining the second radar echo in the sampling window at each sliding window.

[0192] The identification module 152 is configured to obtain the trained target life body detection model, and obtain the life body detection result of the second radar echo based on the target life body detection model.

[0193] The acquisition module 153 is configured to obtain the target life body detection result of the to-be-detected region corresponding to the first radar echo according to the life body detection result of each of the second radar echoes.

[0194] In the embodiment of the application, the acquisition module 153 is further configured to integrate the life body detection result of each of the second radar echoes, obtain the candidate life body detection result of the first radar echo, determine the misrecognition result in the candidate life body detection result and delete it, and determine the candidate life body detection result after deletion as the target life body detection result of the to-be-detected region corresponding to the first radar echo.

[0195] ​In the embodiment of the present application, the acquisition module 153 is further configured to: splice the life body detection results of all the second radar echoes according to the positions of the second radar echoes in the first radar echoes, and acquire the candidate life body detection result of the first radar echoes according to the spliced life body detection results.

[0196] In the embodiment of the present application, the acquisition module 153 is further configured to: cluster the candidate life body detection results, acquire the maximum Euclidean distance and the average Euclidean distance of each cluster, determine the misrecognition result from the candidate life body detection results according to the maximum Euclidean distance and the average Euclidean distance, delete the misrecognition result in the candidate life body detection results, and acquire the target life body detection result.

[0197] In the embodiment of the present application, the life body detection device 1500 further includes a trajectory extraction module 154, configured to: acquire the detection state value, the prediction state value and the Kalman gain of the life body in the single slow time in the to-be-detected region from the target life body detection result, determine the target state value of the life body in the single slow time according to the detection state value, the prediction state value and the Kalman gain, and acquire the moving trajectory of the life body in the to-be-detected region according to the target state value corresponding to each of all the slow times in the first radar echoes.

[0198] In the embodiment of the present application, the traversal module 151 is further configured to: acquire the initial radar echoes of the to-be-detected region according to the ultra-wideband radar, eliminate the static background wave component and the direct current component in the initial radar echoes to acquire the candidate radar echoes of the to-be-detected region after elimination, and perform gain control on the candidate radar echoes, and take the radar echoes obtained after the gain control as the first radar echoes of the to-be-detected region.

[0199] In the embodiment of the present application, the traversal module 151 is further configured to: eliminate the static background wave component in the initial radar echoes based on the path signal subtraction method, and eliminate the direct current component in the initial radar echoes based on the time average subtraction method.

[0200] In the embodiment of the present application, the life body detection device 1500 further includes a training module 155, configured to: acquire the life body detection model to be trained, acquire the sample life body signal, the sample environment signal and the sample interference signal, generate the training sample of the life body detection model, and train the life body detection model to be trained based on the training sample until a trained target life body detection model is obtained.

[0201] The life body detection device provided in the application obtains first radar echoes of a to-be-detected region, traverses the first radar echoes through a slidingly constructed sampling window, obtains second radar echoes framed in the sampling window, further inputs the second radar echoes into a trained target life body detection model, obtains a life body detection result of the second radar echoes according to an output result of the model, and obtains a life body detection result of the first radar echoes according to respective life body detection results of all the second radar echoes, and determines the life body detection result of the first radar echoes as a target life body detection result of the to-be-detected region. In the application, the second radar echoes are obtained by traversing and sampling the first radar echoes, the life body detection result of the first radar echoes is obtained through the life body detection result of the second radar echoes, the calculation amount of obtaining the life body detection result of the radar echoes is reduced, the efficiency of life body detection is improved, the life body detection result is determined based on the trained target life body detection model, the accuracy of life body detection in the to-be-detected region is improved, and the life body detection algorithm is optimized.

[0202] To achieve the above-mentioned embodiments, the application further provides an electronic device, a computer-readable storage medium and a computer program product.

[0203] Figure 16 The block diagram of the electronic device of an embodiment of the application is shown in FIG. 1. The electronic device shown in FIG. 1 can implement the life body detection method of the embodiments. Figure 16 Figures 1 to 14 The block diagram of the electronic device of an embodiment of the application is shown in FIG. 1. The electronic device shown in FIG. 1 can implement the life body detection method of the embodiments.

[0204] To achieve the above-mentioned embodiments, the application further provides a non-transitory computer-readable storage medium storing computer instructions, the computer instructions being used to make a computer execute the life body detection method of the embodiments. Figures 1 to 14

[0205] To achieve the above-mentioned embodiments, the application further provides a computer program product, when an instruction processor in the computer program product executes, the life body detection method of the embodiments is executed. Figures 1 to 14

[0206] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" and the like means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any suitable manner in any one or more embodiments or examples. In addition, the person skilled in the art can combine and combine the different embodiments or examples described in the present specification and the features of the different embodiments or examples without contradiction. ​​​

[0207] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of such features. Throughout the description of this application, "plurality" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0208] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present application belong.

[0209] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and a portable compact disc read-only memory (CDROM). Furthermore, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing it in another suitable manner if necessary, and then storing it in a computer memory.

[0210] It should be understood that parts of the present application can be realized in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be realized as software or firmware stored in a memory and executed by a suitable instruction execution system. As such, if realized in hardware, and in another embodiment, any one or a combination of the following technologies known in the art can be used: discrete logic circuitry having logic gates for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), and the like.

[0211] Those skilled in the art can understand that all or part of the steps carried out by the above-mentioned embodiment methods can be completed by programs instructing relevant hardware, and the programs can be stored in a computer readable storage medium, and when the programs are executed, one or a combination of the steps of the method embodiments is included.

[0212] In addition, each functional unit in each embodiment of the present application can be integrated in one processing module, or each unit can be physically present separately, or two or more units can be integrated in one module. The above-mentioned integrated module can be realized in the form of hardware or in the form of a software functional module. The integrated module, if realized in the form of a software functional module and sold or used as an independent product, can also be stored in a computer readable storage medium.

[0213] The above-mentioned storage medium can be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. A method for detecting a living being, characterized in that: The method comprises: Acquire a first radar echo from the area to be measured, construct a sampling window to traverse the first radar echo, and acquire a second radar echo within the sampling window each time the window is slid; Obtaining a trained target life form detection model, and obtaining a life form detection result of the second radar echo based on the target life form detection model; Obtaining a target living body detection result of the test area corresponding to the first radar echo according to the respective living body detection results of all second radar echoes; The acquiring, based on the life body detection results of all the second radar echoes, the target life body detection result of the test area corresponding to the first radar echo, includes: splicing all the life body detection results of the second radar echoes according to the position of the second radar echo in the first radar echo, and obtaining a candidate life body detection result of the first radar echo based on the spliced ​​life body detection results; Clustering the candidate life form detection results to obtain the maximum Euclidean distance and average Euclidean distance of each cluster; determining a misidentification result from the candidate life body detection results according to the maximum Euclidean distance and the average Euclidean distance; The misidentification result in the candidate life form detection result is deleted, and the target life form detection result is obtained.

2. The method according to claim 1, characterized in that: After obtaining the target life body detection result, the method includes: Obtaining, from the target life body detection result, a detection state value, a predicted state value, and a Kalman gain of the life body in the detection area at a single slow time; determining a target state value of the living being at the single slow time according to the detected state value, the predicted state value, and the Kalman gain; The movement trajectory of the living body in the detection area is obtained according to the target state values ​​corresponding to all the slow times included in the first radar echo.

3. The method according to claim 1, characterized in that The obtaining of the first radar echo of the area to be measured includes: Acquiring an initial radar echo of the area to be measured according to an ultra-wideband radar; Eliminating static background wave components and DC component components in the initial radar echo to obtain candidate radar echoes of the area to be measured after elimination; Gain control is performed on the candidate radar echo, and the radar echo obtained after the gain control is used as the first radar echo of the area to be measured.

4. The method according to claim 3, characterized in that Eliminating the static background wave component and the DC component in the initial radar echo to obtain the candidate radar echo of the area to be measured after the elimination includes: Eliminating the static background wave component in the initial radar echo based on a channel signal subtraction method; Based on time average subtraction, the DC component in the initial radar echo is eliminated.

5. The method according to claim 1, wherein Before acquiring the trained target life form detection model and extracting the life form detection result of the second radar echo based on the target life form detection model, the method includes: Obtaining a life form detection model to be trained; Acquire sample vital sign signals, sample environmental signals, and sample interference signals to generate training samples for the life body detection model; The life form detection model to be trained is trained based on the training samples until the training is completed to obtain the trained target life form detection model.

6. A life detecting device, characterized in that: The device comprises: A traversal module is used to obtain a first radar echo from the area to be measured, and to construct a sampling window to traverse the first radar echo, and to obtain a second radar echo within the sampling window each time the window is slid; an identification module, configured to obtain a trained target life form detection model and obtain a life form detection result of the second radar echo based on the target life form detection model; an acquisition module, configured to acquire a target life body detection result of the test area corresponding to the first radar echo based on the life body detection results of all the second radar echoes; The acquiring, based on the life body detection results of all the second radar echoes, the target life body detection result of the test area corresponding to the first radar echo, includes: splicing all the life body detection results of the second radar echoes according to the position of the second radar echo in the first radar echo, and obtaining a candidate life body detection result of the first radar echo based on the spliced ​​life body detection results; Clustering the candidate life form detection results to obtain the maximum Euclidean distance and average Euclidean distance of each cluster; determining a misidentification result from the candidate life body detection results according to the maximum Euclidean distance and the average Euclidean distance; The misidentification result in the candidate life form detection result is deleted, and the target life form detection result is obtained.

7. A computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images