CN116265984A - Method, device and equipment for detecting falling ground - Google Patents
Method, device and equipment for detecting falling ground Download PDFInfo
- Publication number
- CN116265984A CN116265984A CN202111544969.1A CN202111544969A CN116265984A CN 116265984 A CN116265984 A CN 116265984A CN 202111544969 A CN202111544969 A CN 202111544969A CN 116265984 A CN116265984 A CN 116265984A
- Authority
- CN
- China
- Prior art keywords
- target object
- target
- determining
- probability value
- micro
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 66
- 238000001514 detection method Methods 0.000 claims abstract description 54
- 230000009471 action Effects 0.000 claims abstract description 41
- 230000003068 static effect Effects 0.000 claims abstract description 21
- 230000008569 process Effects 0.000 claims description 33
- 239000000758 substrate Substances 0.000 claims 2
- 238000005452 bending Methods 0.000 abstract description 10
- 230000002829 reductive effect Effects 0.000 abstract description 5
- 238000012544 monitoring process Methods 0.000 description 35
- 238000012549 training Methods 0.000 description 17
- 230000006870 function Effects 0.000 description 16
- 238000004422 calculation algorithm Methods 0.000 description 10
- 238000010586 diagram Methods 0.000 description 9
- 238000007781 pre-processing Methods 0.000 description 9
- 230000002441 reversible effect Effects 0.000 description 9
- 238000004458 analytical method Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 7
- 238000004590 computer program Methods 0.000 description 7
- 238000005259 measurement Methods 0.000 description 6
- 230000005283 ground state Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000012545 processing Methods 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000013145 classification model Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000032683 aging Effects 0.000 description 2
- 238000013136 deep learning model Methods 0.000 description 2
- 238000010801 machine learning Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003062 neural network model Methods 0.000 description 2
- 208000026137 Soft tissue injury Diseases 0.000 description 1
- 208000003443 Unconsciousness Diseases 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 230000008030 elimination Effects 0.000 description 1
- 238000003379 elimination reaction Methods 0.000 description 1
- 230000008717 functional decline Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000004630 mental health Effects 0.000 description 1
- 230000029058 respiratory gaseous exchange Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000008733 trauma Effects 0.000 description 1
- 230000003313 weakening effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/02—Systems using reflection of radio waves, e.g. primary radar systems; Analogous systems
- G01S13/50—Systems of measurement based on relative movement of target
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The application provides a method, a device and equipment for detecting falling ground, wherein the method comprises the following steps: acquiring an initial echo signal of a target scene through a radar sensor, and determining a target echo signal based on the initial echo signal; the initial echo signal comprises an echo signal reflected by a static object and an echo signal reflected by a target object, and the target echo signal comprises the echo signal reflected by the target object; determining micro Doppler features corresponding to the target object based on the target echo signals; determining whether the target object is a quasi-static target person based on the micro-Doppler characteristic; if yes, determining the spatial position of the target object based on the target echo signal; determining whether the target object has a ground-down state based on the spatial position. Through the technical scheme of the application, whether the target object has the falling state or not can be accurately detected, normal actions such as bending down, squatting down and the like can be distinguished, the false alarm probability is effectively reduced, and reliable detection of the falling event is realized.
Description
Technical Field
The application relates to the technical field of personnel monitoring, in particular to a method, a device and equipment for detecting falling to the ground.
Background
With the promotion of an aging society, the problem of population aging is more and more serious, and the aged is easy to fall down due to the reasons of body function decline, balance coordination ability weakening, vision deterioration and the like, so that the consequences of soft tissue injury, mind trauma and the like caused by falling down are influenced, the physical and mental health of the aged is influenced, and the burden of families and society is increased.
In order to reduce the loss caused by falling to the ground, the falling to the ground state needs to be found in time, and then the falling to the ground state is timely rescued, so that more serious consequences caused by falling to the ground are avoided. For example, after the old person falls to the ground at home, if the state of falling to the ground can be found in time, medical staff can be informed to the scene in time to rescue the old person.
In order to find the ground-falling state in time, in the related art, a camera may be deployed, an image is acquired by the camera, and whether the ground-falling state exists or not is analyzed based on the image. However, in a scene where the camera is not deployed (the camera cannot be deployed due to privacy requirements), it is also impossible to determine whether or not a reverse state exists based on image analysis.
Disclosure of Invention
The application provides a method for detecting falling ground, which comprises the following steps:
acquiring an initial echo signal of a target scene through a radar sensor, and determining a target echo signal based on the initial echo signal; the initial echo signals comprise echo signals reflected by a static object and echo signals reflected by a target object, and the target echo signals comprise echo signals reflected by the target object;
Determining micro Doppler features corresponding to the target object based on the target echo signals;
determining whether the target object is a quasi-static target person based on the micro-Doppler characteristic;
if so, determining the spatial position of the target object based on the target echo signal, wherein the spatial position comprises the horizontal position of the target object and the vertical height of the target object;
if the horizontal position of the target object is kept unchanged and the vertical height of the target object is smaller than a preset height threshold value in the continuous first time period, determining that the target object has an inverted state;
otherwise, determining that the target object has no ground falling state.
The application provides a fall detection device, the device includes:
the acquisition module is used for acquiring initial echo signals corresponding to a target scene through the radar sensor and determining target echo signals corresponding to the target scene based on the initial echo signals; the initial echo signals comprise echo signals reflected by a static object and echo signals reflected by a target object, and the target echo signals comprise echo signals reflected by the target object;
The determining module is used for determining micro Doppler characteristics corresponding to the target object based on the target echo signal; determining whether the target object is a quasi-static target person based on the micro-Doppler characteristic; if yes, determining the spatial position of the target object based on the target echo signal, wherein the spatial position comprises the horizontal position of the target object and the vertical height of the target object;
the detection module is used for determining that the target object has a ground falling state if the horizontal position of the target object is kept unchanged and the vertical height of the target object is smaller than a preset height threshold value within a continuous first duration; otherwise, determining that the target object has no ground falling state.
The application provides a fall detection device, include: a processor and a machine-readable storage medium storing machine-executable instructions executable by the processor; the processor is configured to execute machine-executable instructions to implement the method for detecting a fall over disclosed in the embodiments of the present application.
As can be seen from the above technical solutions, in the embodiments of the present application, an initial echo signal of a target scene may be acquired by a radar sensor, a target echo signal is determined based on the initial echo signal, a micro-doppler feature corresponding to a target object is determined based on the target echo signal, and whether the target object is a quasi-static target person is determined based on the micro-doppler feature. When the target object is a quasi-static target person, whether the target object has a falling state or not can be determined based on the spatial position of the target object, so that whether the target object has the falling state or not can be accurately detected, normal actions such as bending, squatting and the like can be distinguished, namely misjudgment on the normal actions such as bending, squatting and the like can be avoided, the misinformation probability is effectively reduced, and reliable detection on the falling event is realized. By identifying quasi-static target personnel, dynamic target personnel are excluded from the ground falling monitoring function, so that the problems caused by monitoring based on the ground falling action can be effectively avoided, namely, normal activities such as bending, squatting and the like are prevented from being identified as ground falling. The interference of other moving objects in the target scene on the ground reversing monitoring function can be effectively eliminated, the interference of non-personnel micro-moving targets in the target scene is effectively eliminated, and the monitoring reliability is improved. By identifying the target micro-motion characteristics, whether the micro-motion signals are derived from personnel or non-personnel (such as fans, pets and the like) can be distinguished, so that interference of the non-personnel on monitoring is effectively eliminated, and a more reliable monitoring result is obtained. According to the method, whether the ground falling state exists or not is not analyzed through the image, and whether the ground falling state exists or not can be determined in an application scene without a camera, so that the universality of ground falling detection is improved.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following description will briefly describe the drawings that are required to be used in the embodiments of the present application or the description in the prior art, and it is obvious that the drawings in the following description are only some embodiments described in the present application, and other drawings may also be obtained according to these drawings of the embodiments of the present application for a person having ordinary skill in the art.
FIG. 1 is a flow chart of a method of detecting a fall in the ground in one embodiment of the present application;
FIG. 2 is a schematic diagram of a ground fall detection system in one embodiment of the present application;
FIG. 3 is a flow chart of a method of detecting a fall over in one embodiment of the present application;
FIG. 4 is a schematic structural view of a ground fall detection device in one embodiment of the present application;
fig. 5 is a hardware configuration diagram of the ground fall detection apparatus in one embodiment of the present application.
Detailed Description
The terminology used in the embodiments of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in this application and the claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein refers to any or all possible combinations including one or more of the associated listed items.
It should be understood that although the terms first, second, third, etc. may be used in embodiments of the present application to describe various information, these information should not be limited to these terms. These terms are only used to distinguish one type of information from another. For example, a first message may also be referred to as a second message, and similarly, a second message may also be referred to as a first message, without departing from the scope of the present application. Depending on the context, furthermore, the word "if" used may be interpreted as "at … …" or "at … …" or "in response to a determination".
In an embodiment of the present application, referring to fig. 1, a method for detecting a fall is provided, where the method may include:
And 102, determining micro Doppler characteristics corresponding to the target object based on the target echo signals.
In one possible implementation, determining the micro-doppler characteristics corresponding to the target object based on the target echo signal may include, but is not limited to, the following: performing FFT (Fast Fourier Transform ) on the target echo signal to obtain an FFT result corresponding to the target echo signal; and determining the phase corresponding to the target object based on the FFT result, and determining the micro Doppler characteristic corresponding to the target object based on the phase. Or generating a complex vector based on the numerical value corresponding to the target object in the target echo signal, and performing STFT (Short Time Fourier Transform, short-time Fourier transform) on the complex vector to obtain an STFT result corresponding to the target echo signal; and determining a characteristic spectrogram corresponding to the target object based on the STFT result, and determining micro Doppler features corresponding to the target object based on the characteristic spectrogram.
For example, if the target object is determined to be a person class based on the micro-doppler feature corresponding to the target object, and the target object is determined to be a non-dynamic class based on the micro-doppler feature corresponding to the target object, then the target object is determined to be a quasi-static target person. Or if the target object is determined to be a non-personnel type based on the micro-Doppler characteristic corresponding to the target object, and/or if the target object is determined to be a dynamic type based on the micro-Doppler characteristic corresponding to the target object, determining that the target object is not a quasi-static target personnel.
In one possible implementation, determining that the target object is a non-dynamic class or a dynamic class based on the micro-doppler feature corresponding to the target object may include, but is not limited to: inputting the micro Doppler characteristic into a first classifier to obtain a first probability value corresponding to the micro Doppler characteristic; wherein the first classifier is used for distinguishing dynamic category from non-dynamic category. On the basis, if the first probability value is a probability value that the target object belongs to a dynamic class, the target object can be determined to be the dynamic class when the first probability value is larger than a first judgment threshold value, and the target object can be determined to be the non-dynamic class when the first probability value is not larger than the first judgment threshold value. Or if the first probability value is a probability value that the target object belongs to the non-dynamic class, when the first probability value is greater than the second decision threshold, the target object can be determined to be the non-dynamic class, and when the first probability value is not greater than the second decision threshold, the target object can be determined to be the dynamic class.
In one possible implementation, determining the target object to be a person class or a non-person class based on the micro-doppler feature to which the target object corresponds may include, but is not limited to: inputting the micro Doppler characteristic into a second classifier to obtain a second probability value corresponding to the micro Doppler characteristic; wherein the second classifier is used for distinguishing personnel category from non-personnel category. On the basis, if the second probability value is a probability value that the target object belongs to the personnel category, the target object can be determined to be the personnel category when the second probability value is larger than a third judgment threshold value, and the target object can be determined to be the non-personnel category when the second probability value is not larger than the third judgment threshold value. Or if the second probability value is a probability value that the target object belongs to a non-personnel category, when the second probability value is greater than a fourth decision threshold, the target object can be determined to be the non-personnel category, and when the second probability value is not greater than the fourth decision threshold, the target object can be determined to be the personnel category.
For example, the spatial position of the target object may include a horizontal position of the target object and a vertical height of the target object. On the basis, if the horizontal position of the target object is kept unchanged and the vertical height of the target object is smaller than a preset height threshold value in the continuous first time period, the situation that the target object is in an inverted ground state can be determined. Otherwise, it may be determined that the target object does not have a ground-down state.
In one possible implementation, after determining whether the target object has a ground-falling state based on the spatial position, if the target object has a ground-falling state, ground-falling warning may also be performed on the target object. Before the corresponding early warning time length of the ground fall early warning does not reach the second time length, if the target object is determined to be restored to normal action, the ground fall early warning is eliminated; or when the corresponding early warning time length of the ground falling early warning reaches the second time length, if the target object is determined not to resume normal action, the ground falling alarm is carried out on the target object.
For example, after the pre-warning is performed on the target object, if it is detected that the horizontal position of the target object remains unchanged or the vertical height of the target object is smaller than a preset height threshold, it is determined that the target object does not resume normal action. If the horizontal position of the target object is detected to be changed and the vertical height of the target object is not less than the preset height threshold value, determining that the target object has recovered to normal action.
As can be seen from the above technical solutions, in the embodiments of the present application, an initial echo signal of a target scene may be acquired by a radar sensor, a target echo signal is determined based on the initial echo signal, a micro-doppler feature corresponding to a target object is determined based on the target echo signal, and whether the target object is a quasi-static target person is determined based on the micro-doppler feature. When the target object is a quasi-static target person, whether the target object has a falling state or not can be determined based on the spatial position of the target object, so that whether the target object has the falling state or not can be accurately detected, normal actions such as bending, squatting and the like can be distinguished, namely misjudgment on the normal actions such as bending, squatting and the like can be avoided, the misinformation probability is effectively reduced, and reliable detection on the falling event is realized. By identifying quasi-static target personnel, dynamic target personnel are excluded from the ground falling monitoring function, so that the problems caused by monitoring based on the ground falling action can be effectively avoided, namely, normal activities such as bending, squatting and the like are prevented from being identified as ground falling. The interference of other moving objects in the target scene on the ground reversing monitoring function can be effectively eliminated, the interference of non-personnel micro-moving targets in the target scene is effectively eliminated, and the monitoring reliability is improved. By identifying the target micro-motion characteristics, whether the micro-motion signals are derived from personnel or non-personnel (such as fans, pets and the like) can be distinguished, so that interference of the non-personnel on monitoring is effectively eliminated, and a more reliable monitoring result is obtained. According to the method, whether the ground falling state exists or not is not analyzed through the image, and whether the ground falling state exists or not can be determined in an application scene without a camera, so that the universality of ground falling detection is improved.
The above technical solutions of the embodiments of the present application are described below with reference to specific application scenarios.
According to the ground falling detection equipment, the ground falling detection equipment can be a ground falling detection system, and the ground falling detection system can comprise a radar sensor, a monitoring unit, a communication unit and an alarm unit, and is used for realizing the detection and alarm functions of ground falling personnel, avoiding misjudgment on normal activities such as bending down, squatting down and the like, effectively reducing misinformation probability and realizing reliable detection of ground falling events. The falling detection system can also realize continuous monitoring of vital signs of falling personnel.
In the reverse detection system, the radar sensor may be a millimeter wave radar, or may be other types of radar, without limitation. The radar sensor may include 1 transmitter and 3 receivers, i.e., a 1-transmit-3-receive radar sensor, and of course, the number of transmitters may be greater, and the number of receivers may be greater, without limitation. The radar sensor based on the 1-transmission and 3-reception structure enables the radar sensor to have an angle measurement capability for a horizontal direction and an angle measurement capability for a vertical direction, wherein the angle measurement capability in the horizontal direction is used for measuring the horizontal position of a target object, and the angle measurement capability in the vertical direction is used for measuring the vertical height of the target object. Wherein, for 3 receivers, 1 pair of receivers are horizontally distributed to provide horizontal angular measurement capability, and another 1 pair of receivers are vertically distributed to provide vertical angular measurement capability. These 3 receivers may each correspond to 1 set of receiver links, or may share 1 set of receiver links. The operating band of the radar sensor can be selected arbitrarily, and is not limited thereto. The radar sensor may be mounted in any manner, and is not limited to this, for example, a ceiling mounting manner or a wall mounting manner is adopted to mount the radar sensor in a target scene.
For example, the target scene on which the radar sensor is installed is any scene in which it is necessary to detect whether or not a ground fall state exists, and the target scene is not limited to this, and may be an indoor scene or an outdoor scene.
In the ground falling detection system, the monitoring unit is an entity for realizing detection of ground falling personnel, and can comprise a preprocessing module, a target detection module, a target identification module, a position calculation module, a ground falling identification module, an alarm control module and a sign monitoring module. The preprocessing module is used for realizing a preprocessing process, the target detection module is used for realizing a target detection process, the target identification module is used for realizing a target identification process, the position calculation module is used for realizing a position calculation process, the reverse recognition module is used for realizing a reverse recognition process, the alarm control module is used for realizing an alarm control process and the sign monitoring module is used for realizing a sign monitoring process.
In the inverted detection system, the communication unit can be used to enable communication between the monitoring unit and the alarm unit. For example, the communication unit forwards the message to the alarm unit after receiving the message from the monitoring unit, and the communication unit forwards the message to the monitoring unit after receiving the message from the alarm unit.
In the reverse detection system, the alarm unit may include a remote alarm unit and an indoor alarm unit. The remote alarm unit is used for notifying guardianship of monitored personnel, including but not limited to mobile phones, personal computers and the like, supports the installation and application of the APP, and can be at least one, such as 1 remote alarm unit when 1 guardian exists, 2 remote alarm units when 2 guardianship exist, and the like. The indoor alarm unit is used for prompting the monitored personnel and can have the functions of voice prompt and the like.
In this embodiment, objects in the target scene that may reflect the echo signal are divided into: static objects (i.e., static background objects such as floors, walls, homes, etc.), dynamic objects (i.e., dynamic background objects such as fans, pets, robots, etc.), and personnel. The static object is in an absolute static state, and no large-amplitude motion or small-amplitude inching exists. The dynamic object has obvious motion or inching, and based on the motion amplitude of the dynamic object, the dynamic object can be divided into a dynamic object with large motion and a dynamic object with inching only, the dynamic object with large motion can be called a true dynamic object, and the dynamic object with inching only can be called a quasi-static object or a non-dynamic object. The personnel have obvious movements or micro movements, the personnel can be divided into the personnel with large movements and the personnel with only micro movements based on the movement amplitude of the personnel, the personnel with large movements can be called dynamic personnel, and the personnel with only micro movements can be called quasi-static personnel.
Because the person in the target scene is only a jogging person and is not a person moving greatly when the person is in an inverted state, the quasi-static person is the target to be detected.
Under the above application scenario, the embodiment of the present application provides a method for detecting a reverse land, which may be applied to a reverse land detection system, and is shown in fig. 3, which is a schematic flow chart of the reverse land detection method, where the method includes:
For example, the objects on which the radar wave acts include static objects, dynamic objects and persons, so after the radar sensor transmits the radar wave, the radar sensor may collect an echo signal, which is referred to as an initial echo signal for convenience of distinction, the initial echo signal is a hybrid echo signal, that is, the initial echo signal is a signal returned by a plurality of objects, and the initial echo signal may be expressed as shown in formula (1):
In equation (1), term 1 represents echo signals reflected by all stationary objects, α i Ith reflected signal s representing static object 1i (t) amplitude, item 2 represents echo signals reflected by all dynamic objects, β j Jth reflected signal s representing dynamic object 2j (t) amplitude, item 3 represents echo signals reflected by all persons, γ k Kth reflected signal s representing person 3k The amplitude of (t), w (t) is used to represent noise.
For static objects (such as ground, walls, home furnishings and the like), the static objects are in an absolute static state, no large movement or small micro movement exists, echo signals at all times are highly consistent, and for dynamic objects (such as electric fans, pets and the like) and personnel, obvious movement or micro movement exists, and obvious change in amplitude or phase exists in the echo signals at all times. According to the time invariance of the stationary object reflected signal, the preprocessing module can remove the echo signal reflected by the stationary object from the initial echo signal to obtain a target echo signal (the mixed signal which is remained except the stationary object reflected echo signal is called a target echo signal), the target echo signal can comprise a signal returned by at least one object, and the target echo signal can be expressed as shown in a formula (2):
In equation (2), term 1 represents the echo signals reflected by all dynamic objects, β j Jth reflected signal s representing dynamic object 2j (t) amplitude, item 2 represents echo signals reflected by all persons, γ k Kth reflected signal s representing person 3k The amplitude of (t), w (t) is used to represent noise.
In summary, it can be seen that if the static object exists in the target scene and the static object reflects the echo signal, the initial echo signal includes the echo signal reflected by the static object, but the target echo signal does not include the echo signal reflected by the static object. If a dynamic object exists in the target scene and the echo signal is reflected by the dynamic object, the initial echo signal comprises the echo signal reflected by the dynamic object, and the target echo signal also comprises the echo signal reflected by the dynamic object. If a person exists in the target scene and the person reflects the echo signal, the initial echo signal comprises the echo signal reflected by the person, and the target echo signal also comprises the echo signal reflected by the person.
In this embodiment, the dynamic object and the person are collectively referred to as a target object, the initial echo signal includes an echo signal reflected by the static object and an echo signal reflected by the target object, and the target echo signal includes an echo signal reflected by the target object. In the preprocessing process, the echo signal reflected by the static object is removed from the initial echo signal, and the residual echo signal (namely, the echo signal reflected by the target object) is taken as the target echo signal.
For example, the target object may be a dynamic object or a person, and since the dynamic object may be a true dynamic object or a quasi-static object, and the person may be a dynamic person or a quasi-static person, that is, the target object may be a true dynamic object, a quasi-static object, a dynamic person, or a quasi-static person, it is necessary to determine whether the target object is a quasi-static person based on the target echo signal (i.e., the echo signal reflected by the target object) in the present embodiment, and perform inverse detection on the basis of the quasi-static person.
For example, if the target echo signal includes an echo signal reflected by the target object, based on the target echo signal, a target object may be detected by using a target detection algorithm, and a distance unit where the target object is located is obtained (the distance unit may be a minimum distance length that can be resolved by the radar sensor, which is called range bin). For example, if the target echo signal includes an echo signal reflected by a target object, a target object may be detected by using a target detection algorithm, and a distance unit where the target object is located may be obtained. If the target echo signal includes echo signals reflected by two target objects, then a target detection algorithm may be used to detect the two target objects, and obtain a distance unit where each target object is located, and so on.
The target detection algorithm may include, but is not limited to, a target detection algorithm such as CFAR (Constant False Alarm Rate ), and the present embodiment does not limit the target detection algorithm as long as the target object can be detected based on the target echo signal and a distance unit where the target object is located is obtained.
When the target detection algorithm is adopted to detect the target object, the related information of the target object, such as distance, speed and the like, can be detected based on the target echo signal acquired by the radar sensor, and the detection process of the target detection algorithm is not limited in the embodiment, and is related to the working principle of the target detection algorithm.
For example, when a target object is detected based on a target echo signal, at least one target object may be detected, which may be a real dynamic object, or a quasi-static object, or a dynamic person, or a quasi-static person, for each detected target object. With reference to the above embodiment, the real dynamic object is a dynamic object that moves largely, the quasi-static object is a dynamic object that has only micro-motion, the dynamic person is a person that moves largely, the quasi-static person is a person that has only micro-motion, and the quasi-static person is a target to be detected.
For example, if the total number N of target objects is 0, it means that no target object is detected and no subsequent steps are performed. Alternatively, if the total number N of target objects is greater than 0, indicating that N target objects are detected, for each target object detected, a subsequent step may be performed based on the target echo signal.
By way of example, due to the presence of physiological signs, the characteristics of a quasi-static person present on the radar signal are different from those of a real dynamic object present on the radar signal, the characteristics of a quasi-static person present on the radar signal are different from those of a quasi-static object present on the radar signal, and the quasi-static person can be found out by identifying the differences in the characteristics. For this reason, in this embodiment, for each target object, the micro-doppler characteristic corresponding to the target object may be determined based on the target echo signal, and the quasi-static person is identified based on the micro-doppler characteristic, that is, whether each target object is a quasi-static target person is determined.
In order to obtain the micro Doppler characteristic corresponding to the target object, a certain time t can be taken d And detecting a distance unit where the target object is based on the multi-frame target echo signals in the target object, wherein the distance unit where the target object is can be used for extracting micro Doppler features corresponding to the target object according to the distance unit where the target object is.
In one possible implementation manner, an FFT may be performed on each frame of the target echo signal, so as to obtain an FFT result corresponding to each frame of the target echo signal. Based on the distance unit where the target object is located, determining the complex number at the distance unit (the complex number at the distance unit also represents the complex number of the target object) from the FFT result corresponding to each frame of target echo signal, and calculating the phase (namely the phase corresponding to the target object) based on the complex number at the distance unit, thereby obtaining the micro Doppler characteristic.
In another possible implementation manner, based on a distance unit where the target object is located, a complex vector may be formed by a value at the distance unit (the number of values at the distance unit is also a value corresponding to the target object) in each frame of the target echo signal, and STFT is performed on the complex vector to obtain an STFT result corresponding to the target echo signal; and determining a characteristic spectrogram corresponding to the target object based on the STFT result, and determining a micro Doppler feature corresponding to the target object based on the characteristic spectrogram, namely extracting the characteristic spectrogram as the micro Doppler feature.
Of course, the above is merely an example of acquiring the micro doppler feature, and the acquisition method is not limited.
Illustratively, after the micro-doppler features corresponding to the target object are obtained, it may be determined whether the target object is a quasi-static target person based on the micro-doppler features. In one possible implementation, the first classifier and the second classifier may also be trained, and based on the first classifier and the second classifier, whether the target object is a quasi-static target person may be determined based on the micro-doppler features corresponding to the target object.
Illustratively, the first classifier is configured to distinguish between a dynamic category and a non-dynamic category, where the real dynamic object and the dynamic person belong to the dynamic category, and where the quasi-static object and the quasi-static person belong to the non-dynamic category. The first classifier may be a machine learning model, such as a deep learning model or a neural network model, and the structure of the first classifier is not limited. The first classifier is used for realizing a classification function, the classification result of the first classifier is a dynamic class or a non-dynamic class, and the input of the first classifier is a micro Doppler characteristic.
For training out the first classifier, a large number of training samples can be obtained, and for each training sample, the training sample can comprise a micro-Doppler feature and a label value corresponding to the micro-Doppler feature, and if the micro-Doppler feature is a micro-Doppler feature corresponding to a true dynamic object or a dynamic person, the label value corresponding to the micro-Doppler feature is a first value, and the first value represents a dynamic category corresponding to the micro-Doppler feature; if the micro-Doppler feature is a micro-Doppler feature corresponding to a quasi-static object or a quasi-static person, the label value corresponding to the micro-Doppler feature is a second value, and the second value represents the non-dynamic category corresponding to the micro-Doppler feature.
After a large number of training samples are obtained, a first classifier can be obtained based on the training samples, the training process of the first classifier is not limited, and after training is completed, the first classifier which has completed training can determine the classification result corresponding to the micro Doppler features, namely the dynamic class or the non-dynamic class.
Illustratively, the second classifier is configured to distinguish between a person class and a non-person class, wherein the real dynamic object and the quasi-static object belong to the non-person class, and wherein the dynamic person and the quasi-static person belong to the person class. The second classifier may be a machine learning model, such as a deep learning model or a neural network model, and the structure of the second classifier is not limited. The second classifier is used for realizing the classification function, the classification result of the second classifier is the personnel category or the non-personnel category, and the input of the second classifier is the micro Doppler characteristic.
For training out the second classifier, a large number of training samples can be obtained, and for each training sample, the training sample can comprise a micro-Doppler feature and a label value corresponding to the micro-Doppler feature, and if the micro-Doppler feature is a micro-Doppler feature corresponding to a dynamic person or a quasi-static person, the label value corresponding to the micro-Doppler feature is a first value, and the first value represents the person category corresponding to the micro-Doppler feature; if the micro-Doppler feature is a micro-Doppler feature corresponding to a true dynamic object or a quasi-static object, the tag value corresponding to the micro-Doppler feature is a second value, and the second value represents a non-personnel category corresponding to the micro-Doppler feature.
After a large number of training samples are obtained, a second classifier can be obtained based on the training samples, the training process of the second classifier is not limited, and after the training is finished, the second classifier which is finished to be trained can determine the classification result corresponding to the micro Doppler characteristics, namely the personnel category or the non-personnel category.
The first classifier and the second classifier may be combined into one classification model, or the first classifier may be split into more classification models, or the second classifier may be split into more classification models, which is not limited in this embodiment, as long as the classification of people and non-people, dynamic classification and non-dynamic classification can be distinguished. For convenience of distinction, the first classifier and the second classifier are taken as examples in the following.
Based on the first classifier and the second classifier, for each target object, it is determined whether the target object is a quasi-static target person in the following manner. Of course, the following is merely an example, and is not limited thereto.
In the mode 1, the micro-doppler feature corresponding to the target object is input to the first classifier, and the micro-doppler feature corresponding to the target object is input to the second classifier when the target object is in a non-dynamic class. In mode 1, it is possible to determine whether the target object is a quasi-static target person by the following steps.
And S11, inputting the micro Doppler characteristic corresponding to the target object into a first classifier to obtain a first probability value corresponding to the micro Doppler characteristic. The first probability value may be a probability value that the target object belongs to a dynamic class, or the first probability value may be a probability value that the target object belongs to a non-dynamic class.
For example, since the first classifier is used for distinguishing the dynamic category from the non-dynamic category, after the micro doppler feature corresponding to the target object is input to the first classifier, the first classifier may output a first probability value corresponding to the micro doppler feature, where the first probability value is a probability value that the target object belongs to the dynamic category. On the basis, if the first probability value is larger than a first judgment threshold value (configured according to experience), determining that the target object is a dynamic category, namely marking a dynamic label for the target object. And if the first probability value is not greater than the first judgment threshold value, determining that the target object is of a non-dynamic type, namely marking a non-dynamic label for the target object.
For example, since the first classifier is used for distinguishing the dynamic category from the non-dynamic category, after the micro doppler feature corresponding to the target object is input to the first classifier, the first classifier may output a first probability value corresponding to the micro doppler feature, where the first probability value is a probability value that the target object belongs to the non-dynamic category. On the basis, if the first probability value is larger than the second judgment threshold value, determining that the target object is of a non-dynamic type, namely marking the target object with a non-dynamic label. And if the first probability value is not greater than the second judgment threshold value, determining that the target object is a dynamic class, namely marking a dynamic label for the target object.
In summary, for each target object, it may be determined whether the target object is a dynamic type or a non-dynamic type, if the target object is a dynamic type, the target object is not subjected to subsequent analysis, and if the target object is a non-dynamic type, the target object is continuously subjected to subsequent analysis, and step S12 is performed, that is, the target object marked with the non-dynamic tag is selected from all the target objects, and step S12 is performed.
Step S12, aiming at a target object of a non-dynamic class, inputting the micro Doppler characteristic corresponding to the target object into a second classifier to obtain a second probability value corresponding to the micro Doppler characteristic. The second probability value may be a probability value that the target object belongs to a person class or a probability value that the target object belongs to a non-person class.
For example, since the second classifier is used to distinguish between a person class and a non-person class, after the micro doppler feature corresponding to the target object is input to the second classifier, the second classifier may output a second probability value corresponding to the micro doppler feature, and the second probability value is a probability value that the target object belongs to the person class. On the basis, if the second probability value is larger than the third judgment threshold value, determining that the target object is of a personnel category, namely marking personnel labels for the target object. And if the second probability value is not greater than the third judgment threshold value, determining that the target object is of a non-personnel type, namely marking a non-personnel label for the target object.
For example, since the second classifier is used to distinguish between the person class and the non-person class, after the micro doppler feature corresponding to the target object is input to the second classifier, the second classifier may output a second probability value corresponding to the micro doppler feature, and the second probability value is a probability value that the target object belongs to the non-person class. On the basis, if the second probability value is larger than the fourth judgment threshold value, determining that the target object is of a non-personnel type, namely marking the target object with a non-personnel label. And if the second probability value is not greater than the fourth judgment threshold value, determining that the target object is a personnel category, namely marking personnel labels for the target object.
In summary, for each target object of the non-dynamic class, it may be determined whether the target object is a personnel class or a non-personnel class, if the target object is a non-personnel class, the subsequent analysis is not performed on the target object, and if the target object is a personnel class, the subsequent analysis is continuously performed on the target object, and step S13 is performed, that is, the target object marked as a personnel tag is selected and step S13 is performed.
Step S13, aiming at a target object of a non-dynamic type and a personnel type, determining the target object as a quasi-static target personnel. In summary, it is determined whether the target object is a quasi-static target person, and if the target object is a person class and the target object is a non-dynamic class, the target object is a quasi-static target person. If the target object is a non-personnel category and/or the target object is a dynamic category, the target object is not a quasi-static target person.
And 2, inputting the micro Doppler characteristic corresponding to the target object into the second classifier, and inputting the micro Doppler characteristic corresponding to the target object into the first classifier when the target object is of a person type. In mode 2, it is possible to determine whether the target object is a quasi-static target person by the following steps.
And S21, inputting the micro Doppler characteristic corresponding to the target object into a second classifier to obtain a second probability value corresponding to the micro Doppler characteristic. The second probability value may be a probability value that the target object belongs to a person class, or the second probability value may be a probability value that the target object belongs to a non-person class.
For example, since the second classifier is used to distinguish between a person class and a non-person class, after the micro doppler feature corresponding to the target object is input to the second classifier, the second classifier may output a second probability value corresponding to the micro doppler feature, and the second probability value is a probability value that the target object belongs to the person class. On the basis, if the second probability value is larger than the third judgment threshold value, determining that the target object is of a personnel category, namely marking personnel labels for the target object. And if the second probability value is not greater than the third judgment threshold value, determining that the target object is of a non-personnel type, namely marking a non-personnel label for the target object.
For example, since the second classifier is used to distinguish between the person class and the non-person class, after the micro doppler feature corresponding to the target object is input to the second classifier, the second classifier may output a second probability value corresponding to the micro doppler feature, and the second probability value is a probability value that the target object belongs to the non-person class. On the basis, if the second probability value is larger than the fourth judgment threshold value, determining that the target object is of a non-personnel type, namely marking the target object with a non-personnel label. And if the second probability value is not greater than the fourth judgment threshold value, determining that the target object is a personnel category, namely marking personnel labels for the target object.
In summary, for each target object, it may be determined whether the target object is a person type or a non-person type, if the target object is a non-person type, the subsequent analysis is not performed on the target object, and if the target object is a person type, the subsequent analysis is performed on the target object continuously, and step S22 is performed, that is, the target object marked as a person tag is selected from all the target objects, and step S22 is performed.
Step S22, aiming at a target object of a person class, inputting a micro Doppler characteristic corresponding to the target object into a first classifier to obtain a first probability value corresponding to the micro Doppler characteristic. The first probability value may be a probability value that the target object belongs to a dynamic class or a probability value that the target object belongs to a non-dynamic class.
For example, since the first classifier is used for distinguishing the dynamic category from the non-dynamic category, after the micro doppler feature corresponding to the target object is input to the first classifier, the first classifier may output a first probability value corresponding to the micro doppler feature, where the first probability value is a probability value that the target object belongs to the dynamic category. On the basis, if the first probability value is larger than the first judgment threshold value, determining that the target object is a dynamic class, namely marking the dynamic label for the target object. And if the first probability value is not greater than the first judgment threshold value, determining that the target object is of a non-dynamic type, namely marking a non-dynamic label for the target object.
For example, since the first classifier is used for distinguishing the dynamic category from the non-dynamic category, after the micro doppler feature corresponding to the target object is input to the first classifier, the first classifier may output a first probability value corresponding to the micro doppler feature, where the first probability value is a probability value that the target object belongs to the non-dynamic category. On the basis, if the first probability value is larger than the second judgment threshold value, determining that the target object is of a non-dynamic type, namely marking the target object with a non-dynamic label. And if the first probability value is not greater than the second judgment threshold value, determining that the target object is a dynamic class, namely marking a dynamic label for the target object.
In summary, for each person class of the target object, it may be determined whether the target object is a dynamic class or a non-dynamic class, if the target object is a dynamic class, the subsequent analysis is not performed on the target object, and if the target object is a non-dynamic class, the subsequent analysis is continuously performed on the target object, and step S23 is performed, that is, the target object marked as a non-dynamic label is selected and step S23 is performed.
Step S23, aiming at the target object of the personnel category and the non-dynamic category, determining the target object as a quasi-static target personnel. In summary, it is determined whether the target object is a quasi-static target person, and if the target object is a person class and the target object is a non-dynamic class, the target object is a quasi-static target person. If the target object is a non-personnel category and/or the target object is a dynamic category, the target object is not a quasi-static target person.
In summary, in step 303, it may be determined whether the target object is a quasi-static target person, step 304 is performed on the target object of the quasi-static target person, and other types of target objects do not perform step 304.
For example, for a target object of an quasi-static target person, the DOA (Direction Of Arrival ) estimation may be performed based on the target echo signal, so as to obtain the spatial position of the target object, and in this embodiment, the DOA estimation manner is not limited, for example, the DOA estimation may be performed by using a beam forming or subspace class algorithm, so long as the spatial position of the target object can be obtained.
For example, the falling detection process may be periodically performed, and steps 301 to 304 are performed in each period to obtain the spatial position of the target object, and if M (M is a positive integer) spatial positions of the target object are obtained in the continuous first duration, that is, M horizontal positions and M vertical heights exist, if the M horizontal positions remain unchanged and the M vertical heights are smaller than a preset height threshold (may be configured empirically, and are not limited), it is determined that the target object has a falling state, and a subsequent alarm control process needs to be performed. If the M horizontal positions change (e.g., a certain horizontal position is different from other horizontal positions), and/or the M vertical heights are not smaller than a preset height threshold (e.g., at least one vertical height is greater than the preset height threshold), determining that the target object has no ground falling state, and not executing the subsequent alarm control process.
For example, after the pre-warning is performed on the target object, if it is detected that the horizontal position of the target object remains unchanged or the vertical height of the target object is smaller than a preset height threshold, it is determined that the target object does not resume normal action. If the horizontal position of the target object is detected to be changed and the vertical height of the target object is not less than the preset height threshold value, determining that the target object has recovered to normal action.
For example, when the target object has a ground falling state, the alarm control module can send a ground falling early warning signal to the indoor alarm unit through the communication unit, and the indoor alarm unit starts the ground falling early warning. For example, the indoor warning unit may include a sound pre-warning device (a device capable of generating sound), and the indoor warning unit performs a floor-reversing pre-warning, i.e., a sound pre-warning, on the target object through the sound pre-warning device without limitation to the sound pre-warning device.
When the target object is in a ground-down state, starting from the ground-down early warning of the target object, the warning control module counts the early warning time, for example, starts a timer, the timeout time of the timer is a second time length, the second time length can be configured according to experience, the time of the timer also represents the early warning time length, and the time length from the starting of the ground-down early warning to the current time is represented. And before the corresponding early warning time length of the reverse early warning reaches the second time length (namely before the timer is overtime), judging whether the target object is restored to normal action.
If the pre-warning duration does not reach the second duration, the target object has recovered to normal action, and then the floor-reversing pre-warning can be eliminated, for example, the warning control module can send an elimination signal to the indoor warning unit through the communication unit, and the floor-reversing pre-warning is eliminated by the indoor warning unit, for example, the indoor warning unit stops generating the sound pre-warning through the sound pre-warning device. If the target object has recovered to normal action, the end of the ground falling warning of the target object is indicated, that is, the target object can stand again without the ground falling warning, so that the ground falling warning can be eliminated, the ground falling warning of the target object is not carried out, and the false alarm is reduced.
And if the early warning time length reaches the second time length, the target object does not resume normal action, and the ground-reversing warning is carried out on the target object. For example, the alarm control module can send a remote alarm signal to the remote alarm unit through the communication unit, and the remote alarm unit carries out the ground falling alarm, so that other objects, such as family members, medical staff and the like, except the target object are informed of the ground falling state of the target object, and the ground falling state of the target object is timely rescued. If the target object does not return to normal action, it indicates that the target object is still in the ground-reversing condition (e.g., the target object is not awake or unable to stand, etc.), so that the target object needs to be subjected to ground-reversing alarm, so as to rescue the target object, for example, the remote alarm unit may send an alarm message, dial an alarm call, etc., so that other objects can know the ground-reversing state of the target object, and timely rescue the ground-reversing state of the target object.
In one possible implementation, the determination of whether the target object has recovered to normal action may be as follows: the backward detection process is periodically executed, each period is executed to obtain the spatial position of the target object, on the basis, from the backward early warning of the target object, it is assumed that K (K is a positive integer) spatial positions of the target object are obtained in a continuous second time period, that is, K horizontal positions and K vertical heights exist, and if the K horizontal positions are kept unchanged or the K vertical heights are smaller than a preset height threshold, it is determined that the target object does not return to normal action. If the K horizontal positions change (if a certain horizontal position is different from other horizontal positions) and the K vertical heights are not smaller than the preset height threshold (if at least one vertical height is greater than the preset height threshold), determining that the target object has recovered to normal action.
In a possible implementation manner, after the falling alarm is performed on the target object, the guardian can also release the falling alarm through the user APP on the remote alarm unit, and the process is not repeated.
According to the technical scheme, in the embodiment of the application, when the target object is a quasi-static target person, whether the target object has the ground falling state or not can be determined based on the spatial position of the target object, so that whether the target object has the ground falling state or not can be accurately detected, normal actions such as bending, squatting and the like can be distinguished, and the false alarm probability is effectively reduced. By identifying quasi-static target personnel, dynamic target personnel are excluded from the ground falling monitoring function, so that the problems caused by monitoring based on the ground falling action can be effectively avoided, namely, normal activities such as bending, squatting and the like are prevented from being identified as ground falling. Interference of other moving objects in the target scene on the ground falling monitoring function is effectively eliminated, interference of non-personnel micro-moving targets in the target scene is effectively eliminated, and monitoring reliability is improved. By identifying the target inching characteristics, whether the inching signals are derived from personnel or non-personnel can be distinguished, so that the interference of the non-personnel on monitoring is effectively eliminated, and a more reliable monitoring result is obtained.
When the ground-falling state of the target object is detected, firstly performing ground-falling early warning on the target object (such as performing ground-falling early warning in a sound mode, etc., if the target object does not need to be rescued, false alarm can be eliminated), and if the target object is restored to normal action before the early warning time period reaches the second time period, the target object does not need to be rescued, and the ground-falling early warning is eliminated, so that false alarm is eliminated. When the early warning time length reaches the second time length, if the target object does not return to normal action, the target object is indicated to need rescue (such as a user waiting for rescue in a ground-reversing state), and then the ground-reversing warning is carried out on the target object (such as medical staff is informed to rescue the target object in the field), so that great convenience is brought to the target object waiting for rescue in the ground-reversing state, and particularly for unconscious target objects, the medical staff can be timely informed to rescue the target object in the field.
Based on the same application concept as the above method, an apparatus for detecting a falling over is provided in an embodiment of the present application, as shown in fig. 4, which is a schematic structural diagram of the apparatus for detecting a falling over, and the apparatus may include:
an acquisition module 41, configured to acquire an initial echo signal corresponding to a target scene by using a radar sensor, and determine a target echo signal corresponding to the target scene based on the initial echo signal; the initial echo signals comprise echo signals reflected by a static object and echo signals reflected by a target object, and the target echo signals comprise echo signals reflected by the target object;
A determining module 42, configured to determine a micro-doppler feature corresponding to the target object based on the target echo signal; determining whether the target object is a quasi-static target person based on the micro-Doppler characteristic; if yes, determining the spatial position of the target object based on the target echo signal, wherein the spatial position comprises the horizontal position of the target object and the vertical height of the target object;
the detection module 43 is configured to determine that the target object has an inverted ground state if the horizontal position of the target object remains unchanged and the vertical height of the target object is less than a preset height threshold value within a continuous first duration; otherwise, determining that the target object has no ground falling state.
Illustratively, the determining module 42 is specifically configured to, when determining whether the target object is a quasi-static target person based on the micro-doppler characteristics: if the target object is determined to be a person class based on the micro Doppler feature corresponding to the target object, and the target object is determined to be a non-dynamic class based on the micro Doppler feature corresponding to the target object, determining that the target object is a quasi-static target person; or if the target object is determined to be a non-personnel type based on the micro-Doppler characteristic corresponding to the target object, and/or if the target object is determined to be a dynamic type based on the micro-Doppler characteristic corresponding to the target object, determining that the target object is not a quasi-static target personnel.
Illustratively, the determining module 42 is specifically configured to, when determining that the target object is a non-dynamic class or a dynamic class based on the micro-doppler feature corresponding to the target object: inputting the micro Doppler characteristic into a first classifier to obtain a first probability value corresponding to the micro Doppler characteristic; wherein the first classifier is used for distinguishing dynamic categories from non-dynamic categories; if the first probability value is a probability value that the target object belongs to a dynamic class, determining that the target object is a dynamic class when the first probability value is larger than a first judgment threshold value, and determining that the target object is a non-dynamic class when the first probability value is not larger than the first judgment threshold value; and if the first probability value is a probability value that the target object belongs to a non-dynamic class, determining that the target object is the non-dynamic class when the first probability value is larger than a second judgment threshold value, and determining that the target object is the dynamic class when the first probability value is not larger than the second judgment threshold value.
Illustratively, the determining module 42 is specifically configured to, when determining the target object is of a person class or a non-person class based on the micro-doppler feature corresponding to the target object: inputting the micro Doppler characteristic to a second classifier to obtain a second probability value corresponding to the micro Doppler characteristic; the second classifier is used for distinguishing personnel types from non-personnel types; if the second probability value is a probability value that the target object belongs to a personnel category, determining that the target object is a personnel category when the second probability value is larger than a third judgment threshold value, and determining that the target object is a non-personnel category when the second probability value is not larger than the third judgment threshold value; and if the second probability value is a probability value that the target object belongs to a non-personnel category, determining that the target object is a non-personnel category when the second probability value is larger than a fourth judgment threshold value, and determining that the target object is a personnel category when the second probability value is not larger than the fourth judgment threshold value.
The determining module is specifically configured to, when determining the micro doppler feature corresponding to the target object based on the target echo signal: performing FFT on the target echo signal to obtain an FFT result corresponding to the target echo signal; determining a phase corresponding to the target object based on the FFT result, and determining a micro Doppler characteristic corresponding to the target object based on the phase; or generating a complex vector based on the numerical value corresponding to the target object in the target echo signal, and performing STFT on the complex vector to obtain an STFT result corresponding to the target echo signal; and determining a feature spectrogram corresponding to the target object based on the STFT result, and determining a micro Doppler feature corresponding to the target object based on the feature spectrogram.
Illustratively, the detecting module 43 is further configured to, after determining that the target object has a ground-down state: performing ground fall early warning on the target object; before the early warning time period does not reach the second time period, if the target object is determined to be restored to normal action, eliminating the ground fall early warning; or when the early warning time length reaches the second time length, if the target object is determined not to resume normal action, performing ground-reversing warning on the target object.
Illustratively, the detection module 43 is further configured to: if the horizontal position of the target object is detected to be unchanged, or the vertical height of the target object is smaller than a preset height threshold value, determining that the target object does not return to normal action; if the horizontal position of the target object is detected to be changed and the vertical height of the target object is not smaller than a preset height threshold value, determining that the target object is restored to normal action.
Based on the same application concept as the above method, an apparatus for detecting a falling over is provided in an embodiment of the present application, as shown in fig. 5, and the apparatus for detecting a falling over may include: a processor 51 and a machine-readable storage medium 52, the machine-readable storage medium 52 storing machine-executable instructions executable by the processor 51; the processor 51 is configured to execute machine-executable instructions to implement the fall detection method disclosed in the above examples of the present application.
Based on the same application concept as the above method, the embodiment of the application further provides a machine-readable storage medium, where a plurality of computer instructions are stored, and when the computer instructions are executed by a processor, the method for detecting the falling of the ground disclosed in the above example of the application can be implemented.
Wherein the machine-readable storage medium may be any electronic, magnetic, optical, or other physical storage device that can contain or store information, such as executable instructions, data, or the like. For example, a machine-readable storage medium may be: RAM (Radom Access Memory, random access memory), volatile memory, non-volatile memory, flash memory, a storage drive (e.g., hard drive), a solid state drive, any type of storage disk (e.g., optical disk, dvd, etc.), or a similar storage medium, or a combination thereof.
The system, apparatus, module or unit set forth in the above embodiments may be implemented in particular by a computer chip or entity, or by a product having a certain function. A typical implementation device is a computer, which may be in the form of a personal computer, laptop computer, cellular telephone, camera phone, smart phone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or a combination of any of these devices.
For convenience of description, the above devices are described as being functionally divided into various units, respectively. Of course, the functions of each element may be implemented in one or more software and/or hardware elements when implemented in the present application.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present application may take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Moreover, these computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The foregoing is merely exemplary of the present application and is not intended to limit the present application. Various modifications and changes may be made to the present application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. which are within the spirit and principles of the present application are intended to be included within the scope of the claims of the present application.
Claims (10)
1. A method of detecting a fall, the method comprising:
acquiring an initial echo signal of a target scene through a radar sensor, and determining a target echo signal based on the initial echo signal; the initial echo signals comprise echo signals reflected by a static object and echo signals reflected by a target object, and the target echo signals comprise echo signals reflected by the target object;
determining micro Doppler features corresponding to the target object based on the target echo signals;
determining whether the target object is a quasi-static target person based on the micro-Doppler characteristic;
if so, determining the spatial position of the target object based on the target echo signal, wherein the spatial position comprises the horizontal position of the target object and the vertical height of the target object;
if the horizontal position of the target object is kept unchanged and the vertical height of the target object is smaller than a preset height threshold value in the continuous first time period, determining that the target object has an inverted state;
otherwise, determining that the target object has no ground falling state.
2. The method of claim 1, wherein the determining whether the target object is a quasi-static target person based on the micro-doppler features comprises:
If the target object is determined to be a person class based on the micro Doppler feature corresponding to the target object, and the target object is determined to be a non-dynamic class based on the micro Doppler feature corresponding to the target object, determining that the target object is a quasi-static target person; or alternatively, the process may be performed,
and if the target object is determined to be a non-personnel type based on the micro Doppler characteristic corresponding to the target object, and/or the target object is determined to be a dynamic type based on the micro Doppler characteristic corresponding to the target object, determining that the target object is not a quasi-static target personnel.
3. The method of claim 2, wherein determining that the target object is a non-dynamic class or a dynamic class based on the micro-doppler feature corresponding to the target object comprises:
inputting the micro Doppler characteristic into a first classifier to obtain a first probability value corresponding to the micro Doppler characteristic; wherein the first classifier is used for distinguishing dynamic categories from non-dynamic categories;
if the first probability value is a probability value that the target object belongs to a dynamic class, determining that the target object is a dynamic class when the first probability value is larger than a first judgment threshold value, and determining that the target object is a non-dynamic class when the first probability value is not larger than the first judgment threshold value;
And if the first probability value is a probability value that the target object belongs to a non-dynamic class, determining that the target object is the non-dynamic class when the first probability value is larger than a second judgment threshold value, and determining that the target object is the dynamic class when the first probability value is not larger than the second judgment threshold value.
4. The method of claim 2, wherein determining the target object as a person class or a non-person class based on the micro-doppler feature corresponding to the target object comprises:
inputting the micro Doppler characteristic to a second classifier to obtain a second probability value corresponding to the micro Doppler characteristic; the second classifier is used for distinguishing personnel types from non-personnel types;
if the second probability value is a probability value that the target object belongs to a personnel category, determining that the target object is a personnel category when the second probability value is larger than a third judgment threshold value, and determining that the target object is a non-personnel category when the second probability value is not larger than the third judgment threshold value;
and if the second probability value is a probability value that the target object belongs to a non-personnel category, determining that the target object is a non-personnel category when the second probability value is larger than a fourth judgment threshold value, and determining that the target object is a personnel category when the second probability value is not larger than the fourth judgment threshold value.
5. The method of claim 1, wherein the step of determining the position of the substrate comprises,
the determining the micro doppler feature corresponding to the target object based on the target echo signal includes:
performing Fast Fourier Transform (FFT) on the target echo signal to obtain an FFT result corresponding to the target echo signal; determining a phase corresponding to the target object based on the FFT result, and determining a micro Doppler characteristic corresponding to the target object based on the phase; or alternatively, the process may be performed,
generating a complex vector based on a numerical value corresponding to the target object in the target echo signal, and performing short-time Fourier transform (STFT) on the complex vector to obtain an STFT result corresponding to the target echo signal; and determining a feature spectrogram corresponding to the target object based on the STFT result, and determining a micro Doppler feature corresponding to the target object based on the feature spectrogram.
6. The method of claim 1, wherein the step of determining the position of the substrate comprises,
after the determining that the target object has the ground falling state, the method further comprises:
performing ground fall early warning on the target object;
before the early warning time period does not reach the second time period, if the target object is determined to be restored to normal action, eliminating the ground fall early warning; or when the early warning time length reaches the second time length, if the target object is determined not to resume normal action, performing ground-reversing warning on the target object.
7. The method of claim 6, wherein the step of providing the first layer comprises,
after the pre-warning is performed on the target object, the method further comprises the following steps:
if the horizontal position of the target object is detected to be unchanged or the vertical height of the target object is smaller than a preset height threshold value, determining that the target object does not return to normal action;
if the horizontal position of the target object is detected to be changed and the vertical height of the target object is not smaller than a preset height threshold value, determining that the target object is restored to normal action.
8. A fall detection device, the device comprising:
the acquisition module is used for acquiring initial echo signals corresponding to a target scene through the radar sensor and determining target echo signals corresponding to the target scene based on the initial echo signals; the initial echo signals comprise echo signals reflected by a static object and echo signals reflected by a target object, and the target echo signals comprise echo signals reflected by the target object;
the determining module is used for determining micro Doppler characteristics corresponding to the target object based on the target echo signal; determining whether the target object is a quasi-static target person based on the micro-Doppler characteristic; if yes, determining the spatial position of the target object based on the target echo signal, wherein the spatial position comprises the horizontal position of the target object and the vertical height of the target object;
The detection module is used for determining that the target object has a ground falling state if the horizontal position of the target object is kept unchanged and the vertical height of the target object is smaller than a preset height threshold value within a continuous first duration; otherwise, determining that the target object has no ground falling state.
9. The apparatus of claim 8, wherein the device comprises a plurality of sensors,
the determining module is specifically configured to, when determining whether the target object is a quasi-static target person based on the micro-doppler feature: if the target object is determined to be a person class based on the micro Doppler feature corresponding to the target object, and the target object is determined to be a non-dynamic class based on the micro Doppler feature corresponding to the target object, determining that the target object is a quasi-static target person; or if the target object is determined to be a non-personnel type based on the micro-Doppler characteristic corresponding to the target object, and/or if the target object is determined to be a dynamic type based on the micro-Doppler characteristic corresponding to the target object, determining that the target object is not a quasi-static target personnel;
the determining module is specifically configured to, when determining that the target object is in a non-dynamic class or a dynamic class based on the micro-doppler feature corresponding to the target object: inputting the micro Doppler characteristic into a first classifier to obtain a first probability value corresponding to the micro Doppler characteristic; wherein the first classifier is used for distinguishing dynamic categories from non-dynamic categories; if the first probability value is a probability value that the target object belongs to a dynamic class, determining that the target object is a dynamic class when the first probability value is larger than a first judgment threshold value, and determining that the target object is a non-dynamic class when the first probability value is not larger than the first judgment threshold value; if the first probability value is a probability value that the target object belongs to a non-dynamic class, determining that the target object is a non-dynamic class when the first probability value is larger than a second judgment threshold value, and determining that the target object is a dynamic class when the first probability value is not larger than the second judgment threshold value;
The determining module is specifically configured to, when determining the target object is a person class or a non-person class based on the micro doppler feature corresponding to the target object: inputting the micro Doppler characteristic to a second classifier to obtain a second probability value corresponding to the micro Doppler characteristic; the second classifier is used for distinguishing personnel types from non-personnel types; if the second probability value is a probability value that the target object belongs to a personnel category, determining that the target object is a personnel category when the second probability value is larger than a third judgment threshold value, and determining that the target object is a non-personnel category when the second probability value is not larger than the third judgment threshold value; if the second probability value is a probability value that the target object belongs to a non-personnel category, when the second probability value is larger than a fourth judgment threshold value, determining that the target object is a non-personnel category, and when the second probability value is not larger than the fourth judgment threshold value, determining that the target object is a personnel category;
the determining module is specifically configured to, when determining the micro doppler feature corresponding to the target object based on the target echo signal: performing FFT on the target echo signal to obtain an FFT result corresponding to the target echo signal; determining a phase corresponding to the target object based on the FFT result, and determining a micro Doppler characteristic corresponding to the target object based on the phase; or generating a complex vector based on the numerical value corresponding to the target object in the target echo signal, and performing STFT on the complex vector to obtain an STFT result corresponding to the target echo signal; determining a feature spectrogram corresponding to the target object based on the STFT result, and determining a micro Doppler feature corresponding to the target object based on the feature spectrogram;
Wherein, the detection module is further configured to, after determining that the target object has a ground-down state: performing ground fall early warning on the target object; before the early warning time period does not reach the second time period, if the target object is determined to be restored to normal action, eliminating the ground fall early warning; or when the early warning time length reaches the second time length, if the target object is determined not to resume normal action, performing ground-reversing warning on the target object;
wherein, the detection module is further used for: if the horizontal position of the target object is detected to be unchanged, or the vertical height of the target object is smaller than a preset height threshold value, determining that the target object does not return to normal action; if the horizontal position of the target object is detected to be changed and the vertical height of the target object is not smaller than a preset height threshold value, determining that the target object is restored to normal action.
10. An inversion detection apparatus, characterized by comprising: a processor and a machine-readable storage medium storing machine-executable instructions executable by the processor; the processor is configured to execute machine executable instructions to implement the method steps of any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111544969.1A CN116265984A (en) | 2021-12-16 | 2021-12-16 | Method, device and equipment for detecting falling ground |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111544969.1A CN116265984A (en) | 2021-12-16 | 2021-12-16 | Method, device and equipment for detecting falling ground |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116265984A true CN116265984A (en) | 2023-06-20 |
Family
ID=86743269
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111544969.1A Pending CN116265984A (en) | 2021-12-16 | 2021-12-16 | Method, device and equipment for detecting falling ground |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116265984A (en) |
-
2021
- 2021-12-16 CN CN202111544969.1A patent/CN116265984A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110738095B (en) | Behavior analysis method and device | |
| Gjoreski et al. | Accelerometer placement for posture recognition and fall detection | |
| US9177203B2 (en) | Target detection device and target detection method | |
| WO2016147770A1 (en) | Monitoring system and monitoring method | |
| RU2625370C2 (en) | Method and device for detecting lift movement | |
| CN112184626A (en) | Gesture recognition method, device, equipment and computer readable medium | |
| US20180313950A1 (en) | CNN-Based Remote Locating and Tracking of Individuals Through Walls | |
| Ramezani et al. | Sensing-Fi: Wi-Fi CSI and accelerometer fusion system for fall detection | |
| CN114446026B (en) | Article forgetting reminding method, corresponding electronic equipment and device | |
| Komai et al. | Beacon-based multi-person activity monitoring system for day care center | |
| CN105913614A (en) | Method and device of monitoring tumbling, intelligent wearable device and tumbling alarm system | |
| US20220246015A1 (en) | Fall detection method and apparatus, and wearable device | |
| Mokhtari et al. | Non-wearable UWB sensor to detect falls in smart home environment | |
| CN109561855A (en) | Equipment, system and method for fall detection | |
| CN113945915A (en) | Method and device for detecting falling by millimeter wave radar | |
| CN114895301A (en) | Millimeter wave radar and video-assisted indoor fall detection method and device | |
| AU2018375154A1 (en) | Ultrasound analytics for actionable information | |
| Mohamed et al. | Fall detection systems for elderly care: A survey | |
| CN114469076A (en) | Identity feature fused old solitary people falling identification method and system | |
| CN114509749A (en) | Indoor positioning detection system and method | |
| CN114442079A (en) | Target object falling detection method and device | |
| De Araujo et al. | An algorithm for fall detection using data from smartwatch | |
| US20230410519A1 (en) | Suspicious person alarm notification system and suspicious person alarm notification method | |
| Merrouche et al. | Fall detection using head tracking and centroid movement based on a depth camera | |
| CN116265984A (en) | Method, device and equipment for detecting falling ground |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |