CN114442079A

CN114442079A - Target object falling detection method and device

Info

Publication number: CN114442079A
Application number: CN202210043513.5A
Authority: CN
Inventors: 贺飞翔; 王泽涛; 丁玉国
Original assignee: Beijing Qinglei Technology Co ltd
Current assignee: Beijing Qinglei Technology Co ltd
Priority date: 2022-01-14
Filing date: 2022-01-14
Publication date: 2022-05-06

Abstract

The invention discloses a method and a device for detecting falling of a target object. Wherein, the method comprises the following steps: receiving a feedback signal generated by a target object transmitting a transmitting signal of a radar; determining corresponding range-doppler information according to the feedback signal; determining human body velocity information of the target object through the range-Doppler information; and under the condition that the human body speed information exceeds a preset speed threshold, determining whether the target object falls down or not according to the distance-Doppler information. The invention solves the technical problems of large calculated amount, high cost, high false alarm rate and false negative rate and poor privacy when the prior art falls into detection.

Description

Target object falling detection method and device

Technical Field

The invention relates to the field of fall detection, in particular to a fall detection method and device for a target object.

Background

With the gradual step-in of China into the aging society, the working pressure of young people is high, the number of empty nesters is continuously increased, and casualties which are finally caused by the fact that the old people accidentally fall down and cannot be timely treated are an important reason for influencing the physical and mental health of the old people. In the bathroom environment, the floor is wet and smooth, the space is narrow, and the falling event is more easily caused by the uncomfortable conditions of the body such as dizziness caused after showering or sitting and the like. The falling detection scheme based on the millimeter wave radar has the advantages of strong penetrability in the environment of a toilet, difficulty in being influenced by environmental factors such as fog and illumination, strong privacy protection and the like, and is a mainstream selection scheme applied at present.

The scheme of the prior art is as follows: the current solutions for detecting a toilet fall event are of four types: the radar equipment carries out a three-dimensional point cloud imaging scheme, a multi-sensor scheme such as infrared and radar, a speed and height information judgment scheme of the radar equipment to a target and a camera imaging scheme.

Problems and drawbacks that exist: the three-dimensional point cloud imaging needs more antennas to have better angular resolution, the equipment cost is higher, the point cloud information cannot be well restored to the human body posture after being clustered, and the problems of erroneous judgment and missing judgment still exist. Infrared devices can fail during showering, and the installation convenience, cost and system robustness of the multi-sensor system all have an influence. In a complex use scene of a toilet, falling detection from speed and height information dimensions easily causes more misjudgments. The camera imaging scheme has the problem of invading the privacy of people in a private space such as a toilet.

In view of the above problems, no effective solution has been proposed.

Disclosure of Invention

The embodiment of the invention provides a method and a device for detecting falling of a target object, which are used for at least solving the technical problems of large calculated amount, high cost, high false alarm rate and false missing report rate and poor privacy in falling detection in the prior art.

According to an aspect of an embodiment of the present invention, there is provided a fall detection method for a target object, including: receiving a feedback signal generated by a target object transmitting a transmitting signal of a radar; determining corresponding range-doppler information according to the feedback signal; determining human body velocity information of the target object through the range-Doppler information; and under the condition that the human body speed information exceeds a preset speed threshold, determining whether the target object falls down or not according to the distance-Doppler information.

Optionally, before receiving a feedback signal generated by the target object transmitting a radar transmission signal, the method further includes: detecting normal attitude parameters of the target object in a detection environment and environment parameters of the detection environment through radar emission signals according to a preset frequency; matching the normal attitude parameters with trusted data in an attitude database, wherein the attitude database stores the trusted normal attitude parameters and the environment parameters; writing the normal attitude parameters into the attitude database under the condition of successful matching; and updating the posture threshold parameter according to the latest data of the posture database, wherein the posture threshold parameter is used for detecting whether the target object falls down.

Optionally, before receiving a feedback signal generated by the target object transmitting a radar transmission signal, the method further includes: detecting an ambient noise power of the target object when the target object is not in a detection environment through a transmission signal of the radar, wherein the attitude threshold parameter comprises the ambient noise power; determining the signal average power of a feedback signal on a floor of the detection environment through the feedback signal; determining whether the target object exists in the detection environment according to the environmental noise power and the signal average power; wherein the target object is determined to be present in the detection environment if the signal average power is greater than the ambient noise power.

Optionally, determining the corresponding range-doppler information according to the feedback signal includes: performing fast Fourier transform after removing direct current from the fast time signal of the feedback signal to obtain distance dimension information; performing fast Fourier transform after removing direct current from the slow time signal of the feedback signal to obtain Doppler information; and accumulating the distance dimension information according to the distance dimension information and the Doppler information to obtain the distance-Doppler information.

Optionally, performing fast fourier transform after removing direct current from the fast time signal of the feedback signal, and obtaining distance dimension information includes: fast Fourier transform is carried out on fast time signals in a plurality of slow time dimensions on each frame of signals of the feedback signals, and a first range profile is obtained; performing fast Fourier transform on the fast time signal of the feedback signal in a frame time dimension to obtain a second range profile; wherein the distance dimension information includes the first range profile and the second range profile.

Optionally, when the human body velocity information exceeds a preset velocity threshold, determining whether the target object falls according to the range-doppler information includes: detecting the height of the target object according to the distance-Doppler information, and determining a first risk probability that the target object falls; detecting a distance gate of a floor of a detection environment where the target object is located according to the distance-Doppler information, and determining a second risk probability of falling of the target object if a human body respiratory frequency characteristic exists; determining envelope graphs of different range gates according to the range-Doppler information, and determining a third risk probability of falling of the target object; determining a fourth risk probability that the target object falls according to whether the distance-Doppler information has weak breathing signal characteristics; determining a comprehensive probability of falling of the target object according to the first risk probability, the second risk probability, the third risk probability, the fourth risk probability and the corresponding weights; and determining that the target object falls down under the condition that the comprehensive probability reaches a preset probability threshold.

Optionally, detecting the height of the target object according to the range-doppler information, and determining the first risk probability that the target object falls includes: determining whether the target object has motion close to the ground according to the first range profile; under the condition of the motion close to the ground, acquiring a corresponding falling height according to the height of the human body in the attitude threshold parameter, wherein the falling height corresponds to the height of the human body; the height of the human body is determined according to the position of a high-power point in the second range profile at a range gate and the installation height of the radar; determining the highest height of the target object according to the first distance image; determining that the target object falls with the first risk probability if the highest height is less than the fall height.

Optionally, detecting, according to the range-doppler information, a range gate on a ground of a detection environment where the target object is located, and whether a human respiratory frequency feature exists, the determining the second risk probability that the target object falls includes: according to the second distance image, carrying out fast Fourier transform on a range gate within a first preset height range of the floor to obtain first target frequency information; determining that the target object falls with the second risk probability if the first target frequency information has a respiratory frequency characteristic.

Optionally, determining envelope graphs of different range gates according to the range-doppler information, and determining a third risk probability that the target object falls includes: determining envelope graphs of different range gates according to the first range profile; determining that the target object has the third risk probability of falling if the envelope graph does not match the envelope graph of the target object for non-falling, wherein the posture threshold parameter comprises the envelope graph of the normal posture of the target object for non-falling.

Optionally, determining a fourth risk probability that the target object falls according to whether the range-doppler information has a signal characteristic of weak breathing includes: according to the second distance image, performing fast Fourier transform on a range gate in a second preset height range to obtain second target frequency information, wherein the second preset height range is higher than the first preset height range; determining that the target object falls with the fourth risk probability if the second target frequency information has a respiratory frequency characteristic.

According to another aspect of the embodiments of the present invention, there is also provided a fall detection apparatus for a target object, including: the receiving module is used for receiving a feedback signal generated by a target object transmitting a transmitting signal of the radar; a first determining module, configured to determine corresponding range-doppler information according to the feedback signal; a second determining module, configured to determine human body velocity information of the target object through the range-doppler information; and the third determining module is used for determining whether the target object falls down or not according to the distance-Doppler information under the condition that the human body speed information exceeds a preset speed threshold.

According to another aspect of the embodiments of the present invention, there is also provided a processor for executing a program, wherein the program is executed to perform the fall detection method for a target object as described in any one of the above.

According to another aspect of the embodiments of the present invention, there is also provided a computer storage medium including a stored program, wherein when the program runs, the apparatus on which the computer storage medium is located is controlled to execute the method for fall detection of a target object.

In the embodiment of the invention, a feedback signal generated by receiving a transmitting signal of a target object transmitting radar is adopted; determining corresponding range-doppler information according to the feedback signal; determining human body velocity information of the target object through the range-Doppler information; under the condition that the human body speed information exceeds a preset speed threshold value, determining whether a target object falls or not according to distance-Doppler information, obtaining distance-Doppler information through a feedback signal of a radar, and under the condition that the human body speed information exceeds the preset speed threshold value, determining whether the target object falls or not, so that the purpose of detecting whether the target object falls or not through a radar monitoring method for obtaining the distance-Doppler information is achieved, the problems that the calculated amount is large in a three-dimensional point cloud mode in radar detection are avoided, the problems that an infrared detection device is high in cost, the false alarm rate and the false alarm rate are high when environmental conditions are harsh, and the privacy of the detection of a camera device is poor are avoided, the technical effects of reducing the calculated amount, reducing the cost, improving the accuracy and improving the privacy are achieved, and the falling detection in the prior art is further solved, the method has the technical problems of large calculation amount, high cost, high false alarm rate and missing report rate and poor privacy.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the invention without limiting the invention. In the drawings:

fig. 1 is a flow chart of a fall detection method for a target object according to an embodiment of the invention;

FIG. 2 is a schematic diagram of a detection system architecture according to an embodiment of the present invention;

FIG. 3 is a flow chart of an overall detection method according to an embodiment of the invention;

fig. 4 is a flow chart of data iteration of a SaaS system according to an embodiment of the present invention;

fig. 5 is a flow chart of a fall detection algorithm according to an embodiment of the invention;

fig. 6 is a schematic view of a fall detection apparatus for a target object according to an embodiment of the present invention.

Detailed Description

In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It should be noted that the terms "first," "second," and the like in the description and claims of the present invention and in the drawings described above are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used is interchangeable under appropriate circumstances such that the embodiments of the invention described herein are capable of operation in sequences other than those illustrated or described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.

According to an embodiment of the invention, there is provided a method embodiment of a fall detection method for a target object, it being noted that the steps illustrated in the flowchart of the drawings may be performed in a computer system such as a set of computer executable instructions, and that although a logical order is illustrated in the flowchart, in some cases the steps illustrated or described may be performed in an order different than here.

Fig. 1 is a flowchart of a fall detection method for a target object according to an embodiment of the present invention, as shown in fig. 1, the method including the steps of:

step S102, receiving a feedback signal generated by a target object transmitting a radar transmitting signal;

step S104, determining corresponding distance-Doppler information according to the feedback signal;

step S106, determining the human body velocity information of the target object through the distance-Doppler information;

and step S108, determining whether the target object falls down or not according to the distance-Doppler information under the condition that the human body speed information exceeds a preset speed threshold.

Through the steps, a feedback signal generated by receiving a transmitting signal of a target object transmitting radar is adopted; determining corresponding range-doppler information according to the feedback signal; determining human body velocity information of the target object through the range-Doppler information; under the condition that the human body speed information exceeds a preset speed threshold value, determining whether a target object falls or not according to distance-Doppler information, obtaining distance-Doppler information through a feedback signal of a radar, and under the condition that the human body speed information exceeds the preset speed threshold value, determining whether the target object falls or not, so that the purpose of detecting whether the target object falls or not through a radar monitoring method for obtaining the distance-Doppler information is achieved, the problems that the calculated amount is large in a three-dimensional point cloud mode in radar detection are avoided, the problems that an infrared detection device is high in cost, the false alarm rate and the false alarm rate are high when environmental conditions are harsh, and the privacy of the detection of a camera device is poor are avoided, the technical effects of reducing the calculated amount, reducing the cost, improving the accuracy and improving the privacy are achieved, and the falling detection in the prior art is further solved, the method has the technical problems of large calculation amount, high cost, high false alarm rate and missing report rate and poor privacy.

The radar may be a millimeter wave radar, and a detection signal, for example, an ultrasonic wave, is transmitted to the target object through the transmitting antenna, and after being reflected by the target object, the reflected feedback signal may be received by the receiving antenna of the radar, and a fast time signal in the feedback signal may be converted into a one-dimensional range profile signal, that is, the above-mentioned range dimension information, where the one-dimensional range profile is a vector sum of a target scattering point sub-echo obtained by the broadband radar according to the signal in a radar ray projection, which is actually a scattering intensity distribution map of each range unit on the target object, and the range dimension information may obtain a height, a speed, and the like of the target object. Specifically, the distance between the target object and the radar can be determined according to the distance dimension information, so that the height of the target object can be determined. The height change of the target image in time can be determined through the distance dimension information of the multi-frame images, and then the speed change of the target object is determined.

In addition, the fast fourier transform is performed on the slow time signal of the feedback signal, and doppler dimensional information can be obtained. The Doppler dimension information can detect weak motion behaviors of the target object, including breathing and heartbeat and the like. On the other hand, the distance dimension information and the Doppler information contain some integral features, for example, a high-power range gate can be fitted with an envelope graph conforming to the posture of the human body.

The human body velocity information of the target object is determined through the distance-Doppler information, and under the condition that the human body velocity information exceeds a preset velocity threshold value, the target object is proved to have obvious height change motion, and a falling behavior possibly exists to a great extent. And then determining other posture parameters including the breathing frequency, the body height, the envelope graph and the like according to the distance-Doppler information to determine whether the target object falls down. Specifically, the detection modes corresponding to each posture parameter are different, and whether the target object falls down or not is comprehensively judged from a plurality of angles of the posture parameters. Not only can directly improve the accuracy of fall detection, but also avoid various false alarms from multiple aspects.

The embodiment can measure the radial relative motion speed of the target to the radar according to the Doppler frequency. A fall is a movement away from the radar, with a negative doppler frequency, beyond which a suitable threshold is selected, a fall movement is considered.

The method and the device have the advantages that the range-Doppler information is obtained through the feedback signal of the radar, whether a target object falls or not is determined according to the range-Doppler information under the condition that the human body speed information exceeds a preset speed threshold value, the purpose of detecting whether the target object falls or not in a mode of obtaining the range-Doppler information through radar monitoring is achieved, the problem that calculated amount is large in a three-dimensional point cloud mode in radar detection is avoided, the problems that an infrared detection device is high in cost, false alarm rate and missing report rate are high when environmental conditions are harsh are avoided, the problem that privacy of detection of a camera device is poor are also avoided, the technical effects of reducing calculated amount, reducing cost, improving accuracy and improving privacy are achieved, and the technical problems that detection falls in the prior art is large in calculated amount, high in cost, high in false alarm rate and missing report rate and poor in privacy are solved.

Optionally, before receiving a feedback signal generated by the target object transmitting a radar transmission signal, the method further includes: detecting normal attitude parameters of a target object in a detection environment and environment parameters of the detection environment through a radar transmitting signal according to a preset frequency; matching the normal attitude parameters with trusted data in an attitude database, wherein the attitude database stores the trusted normal attitude parameters and the environment parameters; writing the normal attitude parameters into an attitude database under the condition of successful matching; and updating the posture threshold parameter according to the latest data of the posture database, wherein the posture threshold parameter is used for detecting whether the target object falls down.

The posture threshold parameters comprise the environmental noise power of a subsequent target object when the target object is not in the detection environment, the height of the target object, and an envelope graph of the target object in a non-falling normal posture, such as a standing posture and a sitting posture. Parameter bases can be provided for subsequent detection through the attitude threshold parameters, and the attitude threshold parameters are required for risk judgment of multiple parameters in subsequent falling detection. The normal attitude parameters of the target object can be determined through the normal attitude parameters and the environmental parameters, and then the attitude parameters under normal conditions can be processed to obtain other threshold parameters, for example, the envelope graph of the normal attitude, the human body height range of the normal attitude and the like can be obtained by processing the normal range-Doppler information.

It should be noted that the attitude threshold parameter of the attitude database may also be updated through a specific drop detection, and after a drop detection is performed once and after user confirmation, the trusted data in the detection process is written into the attitude database, and the attitude threshold parameter is updated. That is, in other embodiments, the posture threshold parameter may further include a posture parameter of the fall, so that the detected posture parameter of the target object may be matched from the viewpoint of the posture parameter of the fall to determine the proximity of the detected posture parameter, and if the detected posture parameter is close enough, the target object may also be determined to fall.

In this embodiment, as shown in fig. 4, the posture database is a SaaS system posture database, and the SaaS system adapts to the information of the human body posture database, and accumulates and iterates posture threshold parameters in a fall detection and recognition algorithm according to data used by a user every day. And step S31, judging whether the data accumulation is full of three days, if not, loading default parameters, and if yes, entering the step S32. And step S32, extracting the credible data in the database through information such as the mean value, the variance and the like, and comparing the credible data with the current storage result. And step S33, if the matching degree of the current data and the historical data in the database is low, the information storage of the calculation is abandoned. If the confidence is high, writing into the database. And step S34, loading parameters such as radar installation height, environmental noise power, user height information, human body low attitude information and the like, and providing the parameters for a subsequent algorithm module to use. The attitude database is updated in time, so that the attitude threshold value parameters in the attitude database are updated in time according to the situation, and the falling accuracy of the target object can be improved.

Optionally, before receiving a feedback signal generated by the target object transmitting a radar transmission signal, the method further includes: detecting the environmental noise power of a target object when the target object is not in a detection environment through a transmitting signal of a radar, wherein the attitude threshold parameter comprises the environmental noise power; determining the signal average power of a feedback signal on a floor by detecting the feedback signal of the floor of the environment; determining whether a target object exists in the detection environment or not according to the environmental noise power and the signal average power; and determining that the target object exists in the detection environment under the condition that the average power of the signal is greater than the power of the environmental noise.

Before receiving a feedback signal generated by a target object transmitting a radar transmitting signal, whether a target object exists in a detection environment can be determined through the radar transmitting signal, specifically, the target object exists in the detection environment is determined according to the comparison between the signal average power of a floor and the ambient noise power, and if the signal average power is larger than the ambient noise power. And thus, in the case where it is determined that the target object enters the detection environment, the feedback signal of the radar is acquired.

In this embodiment, the Micro distance of the feedback signal is used to find the average power from the gate near the floor, and if the difference from the noise power exceeds a predetermined threshold range, it is determined that the vehicle is in a manned state. Therefore, under the condition that a person is determined, whether the person falls down is detected in a mode of obtaining distance-Doppler information through radar monitoring according to the steps, the problem that the accuracy rate is poor when the person falls down is solved, and the problem that operation resources are wasted when the steps are executed under the unmanned condition is also solved.

Optionally, determining the corresponding range-doppler information according to the feedback signal includes: removing direct current from a fast time signal of the feedback signal, and then performing fast Fourier transform to obtain distance dimension information; performing fast Fourier transform after removing direct current from a slow time signal of the feedback signal to obtain Doppler information; and accumulating the distance dimension information according to the distance dimension information and the Doppler information to obtain distance-Doppler information.

Specifically, radar echo signal raw data is read from ADC cache data of the radar. And removing the direct current component of the fast time signal, and then carrying out FFT (fast Fourier transform) to obtain distance dimension information. And performing FFT after the direct current of the slow time signal is removed to obtain Doppler dimension information. And accumulating 20 frames of data of the Micro distance image of each frame, removing direct current, and then calculating the average value to obtain the Micro distance image. The range-Doppler plots were obtained after another fft on the Micro. The Micro one-dimensional range image only contains range and energy information. And range-doppler contains range, velocity, energy information.

For the fast time signal and the slow time signal, the radar periodically sends pulse signals when working, and samples echo signals in the pulse interval time. Although on one time axis, the echo sampling interval and the pulse repetition interval (pulse period) are very different in magnitude, for example, the echo sampling interval is about 10 to the power of-8, and the pulse repetition interval is about 10 to the power of-3, so that the echo sampling interval and the pulse repetition interval are divided into two dimensions, namely, a fast time and a slow time, and their corresponding signals are a fast time signal and a slow time signal.

Optionally, performing fast fourier transform after removing direct current from the fast time signal of the feedback signal, and obtaining distance dimension information includes: fast Fourier transform is carried out on fast time signals in a plurality of slow time dimensions on each frame of signals of the feedback signals, and a first range profile is obtained; performing fast Fourier transform on a fast time signal of the feedback signal in a frame time dimension to obtain a second range profile; the distance dimension information comprises a first distance image and a second distance image.

The distance dimension information may include the first distance image and the second distance image, the feedback signal may include a plurality of frames of signal segments, and the time corresponding to each frame of signal segment may be the same. Each frame of signal of the feedback signal carries out fast Fourier transform on the fast time signal in a plurality of slow time dimensions to obtain a first range profile, namely a Macro range profile, and fast Fourier change is carried out on the fast time signal in a plurality of slow time dimensions of each frame. The fast fourier transform is performed on the fast time signals of the feedback signals in the frame time dimension to obtain the second range profile, that is, the Micro range profile, and the fast fourier transform may be performed on the fast time signals of each frame of the feedback signals, respectively, and then the fast time signals are combined in sequence to obtain the second range profile. Since the second distance is like in the frame time dimension, distance changes with large motion amplitude, for example, height changes of the human body caused by human body motion, can be detected.

Optionally, determining whether the target object falls down according to the range-doppler information when the human body velocity information exceeds the preset velocity threshold includes: detecting the height of the target object according to the distance-Doppler information, and determining a first risk probability of falling of the target object; detecting whether a distance gate of a floor of a detection environment where the target object is located exists or not according to the distance-Doppler information, and determining a second risk probability of falling of the target object; determining envelope graphs of different range gates according to the range-Doppler information, and determining a third risk probability of falling of the target object; determining a fourth risk probability of falling of the target object according to whether the distance-Doppler information has a signal characteristic of weak breathing; determining the comprehensive probability of falling of the target object according to the first risk probability, the second risk probability, the third risk probability, the fourth risk probability and the weights corresponding to the first risk probability, the second risk probability, the third risk probability and the fourth risk probability respectively; and determining that the target object falls down under the condition that the comprehensive probability reaches a preset probability threshold.

The human height information is extracted from the distance dimension information of the distance-Doppler information, and the corresponding falling height is adaptively matched according to different heights, so that the algorithm applicability is improved. The Micro second distance is like extracting human body breathing information near the floor, namely the human body breathing frequency characteristic, and eliminating interference misjudgment caused by complex environments such as a closestool, a water pipe and the like in a toilet. Human body envelope information, namely the envelope graph, is extracted from the Micro second distance image distance dimension, so that misjudgment of scenes such as sitting postures, low postures and the like can be effectively eliminated. The Macro first distance image distance dimension extracts human body activity information, and the false judgment of scenes such as hand washing clothes in a toilet can be eliminated by the aid of the signal characteristics of weak breath.

And integrating the risk judgment, weighting and calculating the falling probability of the human body, and clearing and correcting if an obvious non-falling signal appears in the process of accumulating the falling probability, for example, a signal appears at a higher distance gate, the human body is determined to be in a standing posture, and the accumulated falling probability belongs to a misjudgment. And if the falling probability is accumulated to exceed the threshold value, outputting falling alarm information.

Optionally, detecting the height of the target object according to the range-doppler information, and determining the first risk probability of falling of the target object includes: determining whether the target object has motion close to the ground according to the first range profile; under the condition of movement close to the ground, acquiring a corresponding falling height according to the human body height in the attitude threshold parameter, wherein the falling height corresponds to the human body height, and the human body height is also the human body height of the target object; the height of the human body is determined according to the position of a high-power point in the second range profile at a range gate and the installation height of the radar; determining the highest height of the target object according to the distance dimension information; in the case where the highest height is smaller than the fall height fall distance range, it is determined that the target object falls with a first risk probability.

Whether the human body has a movement far away from a radar or falls to the ground or not is extracted through the range-Doppler image of the Macro range image (namely, the first range image), the Doppler frequency set by the threshold value is not too high, and the situation of missing report of a scene of slowly falling due to dizziness and the like is avoided. After the trend that the human body is far away from the radar is detected, the height information of the human body can be calculated through the position of a high-power point in the one-dimensional range profile at a range gate and the installation height of the radar in the Micro range profile (namely, a second range profile). And matching a falling height descending distance of the human body corresponding to the height after the big data summarization according to the height information of the human body, and considering that the human body has a first risk probability of falling if the non-noise power distance gate closest to the radar is far away from the normal range of the falling height of the human body. The corresponding falling height is adaptively matched according to different heights, and the algorithm applicability is improved.

Optionally, detecting, according to the range-doppler information, whether a human respiratory frequency feature exists in a range gate on the ground of a detection environment where the target object is located, and determining the second risk probability of the target object falling includes: performing fast Fourier transform on a range gate within a preset height range of the floor according to the Micro range profile (second range profile) to obtain target frequency information; and determining that the target object falls has a second risk probability in the case that the target frequency information has clear breathing frequency characteristics.

The human body can lie on the floor after falling down, the FFT is carried out on the sliding window of the distance door near the floor under the condition of the known radar height to extract the frequency information of the signal, when the human body does other normal activities in a toilet, the frequency domain is relatively disordered, when the human body falls down, the frequency characteristic of one breath can be relatively clearly extracted, and the human body is considered to have higher second risk probability of falling down. The interference misjudgment caused by complex environments such as a toilet bowl, a water pipe and the like in the toilet can be effectively eliminated.

Optionally, determining envelope graphs of different range gates according to the range-doppler information, and determining a third risk probability that the target object falls includes: determining envelope graphs of different range gates according to the first range profile; and determining that the target object has a third risk probability of falling under the condition that the envelope graph does not match with the envelope graph of the non-falling target object, wherein the posture threshold parameter comprises the envelope graph of the non-falling normal posture of the target object.

The standing posture and the sitting posture of the human body can present an envelope graph which spans a plurality of range gates and is consistent with the posture of the human body on the one-dimensional range profile of the Macro first range profile. When a person falls, a narrow envelope graph with strong energy only in a few range gates is presented on the one-dimensional range image, and when the characteristics of the envelope graph appear, the person is considered to have a third risk probability of falling. The false judgment of scenes such as low postures of a toilet and the like can be effectively eliminated.

For a scene that people cannot save themselves after falling down, after the people fall down to a toilet or a wash platform, very obvious signals generated by weak movement of breathing can be seen in the Micro one-dimensional distance image (the second distance image), and the people can hardly see the signals due to the absence of quick movement in the Macro one-dimensional distance image (the first distance image), so that the people are considered to have higher fourth risk probability of falling down. The misjudgment in the scenes of washing clothes by hands in a toilet and the like can be eliminated.

It should be noted that this embodiment also provides an alternative implementation, which is described in detail below.

The embodiment provides a toilet fall detection method based on a radar one-dimensional range profile, and solves the problems that the existing fall detection method is large in calculation amount, high in equipment cost, high in false alarm rate and false alarm rate, high in missing report rate, capable of invading privacy, easy to be influenced by the environment and the like.

Fig. 3 is a flowchart of an overall detection method according to an embodiment of the present invention, which includes the following steps, as shown in fig. 3:

and step S1, arranging a radar in the toilet, requiring the radar to be installed in the middle of the suspended ceiling of the toilet, irradiating downwards, mining the environment in space, calculating parameters such as distance information and signal noise information of the radar from the ground, and keeping the radar in the cloud.

Step S2, initializing parameters, and configuring the initial frequency, cut-off frequency, rising edge frequency modulation slope, duration, falling edge duration and rest time of the frequency modulation continuous wave FMCW signal; configuring ADC sampling frequency, fast time sampling point number, chirp number in one frame and frame period;

and step S3, the SaaS system is adapted to the human body posture database information and accumulates and iterates posture threshold parameters in the fall detection and identification algorithm according to data used by the user every day.

And step S4, reading the radar echo signal raw data from the ADC buffer data.

And step S5, removing the direct current component of the fast time signal and then performing FFT to obtain distance dimension information.

And step S6, performing FFT after the slow time signal is subjected to direct current removal, and obtaining Doppler dimension information.

Step S7, perform FFT on the fast time in the frame, and accumulate and smooth the fast time to remove the dc to obtain the Micro one-dimensional range profile, that is, the second range profile. And performing FFT on the fast time between frames, accumulating and smoothing the fast time and removing direct current to obtain a Macro one-dimensional range profile, namely the first range profile.

And step S8, calculating average power of the distance gate of the Micro distance image near the floor, comparing the average power with noise power, and filtering to obtain the state of existence of people in the current toilet. And (4) calculating average power of the Micro distance image near the floor by a distance gate, and judging that the Micro distance image is in a human state if the difference value with the noise power exceeds a set threshold range.

And step S9, judging whether a person exists in the current toilet or not, and returning to the step S4 when no person exists. And if the person is in the fall detection information extraction algorithm, the person enters the fall detection information extraction algorithm.

And step S10, clouding according to the user characteristic information extracted in step S9, such as height, sitting height, washing height and the like, and returning to step S3.

Fig. 4 is a flowchart of data iteration of the SaaS system according to the embodiment of the present invention, and as shown in fig. 4, the SaaS system posture library information of the embodiment includes the following steps:

and step S31, judging whether the data accumulation is full of three days, if not, loading default parameters, and if yes, entering the step S32.

And step S32, extracting the credible data in the database through information such as the mean value, the variance and the like, and comparing the credible data with the current storage result.

And step S33, if the matching degree of the current data and the historical data in the database is low, the information storage of the calculation is abandoned. If the confidence is high, writing into the database.

And step S34, loading parameters such as radar installation height, environmental noise power, user height information, human body low attitude information and the like, and providing the parameters for a subsequent algorithm module to use.

Fig. 5 is a flowchart of a fall detection algorithm according to an embodiment of the invention, and as shown in fig. 5, the fall detection information extraction algorithm of the embodiment includes the following steps:

and step S71, extracting human body speed information from the Macro range-Doppler image, entering step S72 when detecting that the human body speed information exceeds a radar speed threshold, and ending the falling detection information extraction algorithm if the human body speed information exceeds the radar speed threshold.

And step S72, extracting height information of the human body through Micro distance dimension, adaptively matching corresponding falling heights according to different heights, and improving algorithm applicability. Micro distance-Doppler dimension extracts human body respiration information near the floor, and interference misjudgment caused by complex environments such as a closestool, a water pipe and the like in a toilet is eliminated. Micro distance dimension extracts human body envelope information, and misjudgment of scenes such as sitting postures, low postures and the like can be effectively eliminated. The Macro distance dimension extracts the human body activity information, and the misjudgment of scenes such as washing clothes by hands in a toilet can be eliminated.

And step S73, integrating all the information in the step S72, weighting the information results of different dimensions, calculating fall scores, and accumulating the scores of each frame of data.

And step S74, extracting the rising information of the human body in the Micro distance dimension, clearing the falling score counted in the step S73 if the human body stands up, and otherwise, entering the step S75.

And step S75, judging whether the score exceeds a threshold value, outputting fall alarm information if the score exceeds the threshold value, and finishing the fall detection information extraction algorithm if the score does not exceed the threshold value.

In the embodiment, under the complex environment of a toilet, the one-dimensional range profile of the millimeter wave radar is subjected to multi-dimensional human body falling characteristic information extraction, and in the actual use process, similar to scenes that a horse barrel is sitting on, the person bends over to lose things and the like and easily misreport, the situation that the person breathes on the horse barrel, the person bends over to lose things and the like can be effectively misreported and avoided.

The information such as speed, distance and the like obtained by radar calculation is high in accuracy, acceleration is not used as a main falling dependence condition, the method has better adaptability to slow falling caused by dizziness and the like of the old, and the falling identification accuracy rate is effectively improved.

According to the embodiment, multiple receiving antennas are not needed for angle measurement imaging, and the development cost is reduced. The radar is less interfered by environmental factors (such as illumination and fog), does not invade the privacy of a human body, and is very suitable for the private environment of a toilet. And the cross calculation is not needed by a plurality of sensor data, and the system integration level is higher.

In implementation, fig. 2 is a schematic diagram of a detection system architecture according to an embodiment of the present invention, and as shown in fig. 2, the embodiment is a toilet fall detection scheme based on a millimeter wave radar, in which a millimeter wave radar module sends and collects echo signals of a human body, a processor converts the echo signals of a receiving channel into digital signals through an a/D chip, processes the digital signals through an fall detection algorithm module to obtain fall detection results, and outputs the results to a terminal through a data transmission module for display.

The implementation method comprises the following theoretical analysis:

as shown in fig. 3, the millimeter wave radar fall detection system is installed in the middle of a suspended ceiling of a toilet, the millimeter wave radar is an active sensor, and a sensor transmitting antenna transmits pulsed electromagnetic waves at a fixed period; the electromagnetic wave can be uninterruptedly reflected after people enter a toilet, and the reflected electromagnetic wave is transmitted to a receiving antenna of the millimeter wave radar sensor. The radar radio frequency signal input by the receiving antenna is filtered and sampled after the sensor obtains a difference frequency signal through frequency mixing, and the signal is output to the MCU for signal processing.

The MCU firstly carries out fast time FFT (fast Fourier transform) on the input radar signal in a frame time dimension to obtain information of the radar signal in a distance dimension, namely a Macro one-dimensional range profile, and then carries out slow time FFT to obtain the Doppler dimension information of the radar signal. And then FFT is carried out on the fast time for a plurality of slow time dimensions of each frame to obtain a Micro one-dimensional range profile. Wherein the Macro dimension measures rapid movements of the body, such as the fall characteristics of the body; the Micro dimension can measure the slow motion of a human body, FFT is carried out on a specific distance unit of the Micro dimension, and Micro velocity Doppler information, such as the breathing characteristics of the human body, can be obtained.

And judging whether people exist or not by judging the environmental noise and the power difference value at the current moment, and entering a detection and identification algorithm when people exist in the toilet.

As shown in fig. 4, the algorithm extracts whether a motion of falling away from the radar, that is, towards the ground exists in the human body through the range-doppler image of the Micro latitude, and the doppler frequency set by the threshold is not too high, so that the situation of missing report in a slow falling scene due to dizziness and the like is avoided. Specifically, according to the magnitude of the Doppler frequency, the radial relative movement speed of the target to the radar can be measured. A fall is a movement away from the radar, with a negative doppler frequency, beyond which a suitable threshold is selected, a fall movement is considered. After the trend that the human body is far away from the radar is detected, the height information of the human body can be calculated in the Micro dimension through the position of a high-power point in the one-dimensional range profile at a range gate and the installation height of the radar. And matching the falling distance of the falling height of the human body corresponding to the height after the big data summarization according to the height information of the human body, and considering that the person has the falling risk if the non-noise power distance gate closest to the radar is far away from the normal range of the falling height of the human body.

The human body can lie on the floor after falling down, fft is carried out on a distance door sliding window near the floor under the condition that the height of the radar is known to extract frequency information of signals, when the human body does other normal activities in a toilet, the frequency domain is relatively disordered, and when the human body falls down, a respiratory frequency characteristic can be relatively clearly extracted, so that the human body is considered to have higher falling risk.

The standing posture and the sitting posture of the human body can present an envelope figure which spans a plurality of distance gates and is consistent with the posture of the human body on the Micro one-dimensional distance image. When a person falls, a narrow envelope graph with strong energy only in a few distance gates is presented on the one-dimensional distance image, and when the characteristics of the envelope graph appear, the person is considered to have a falling risk.

For the scene that the person can not save oneself after falling down, the very obvious signal generated by the weak movement of breathing can be seen in the Micro one-dimensional distance image after the person falls down, and the signal can not be seen almost because of the absence of the quick movement in the Macro dimension, so that the person is considered to have higher falling risk. The Micro is sensitive to the detection of the weak movement, the weak movement caused by breathing can embody power information on a Micro one-dimensional distance image, and the fallen power information is mainly concentrated near the floor. Macro is sensitive to fast motion detection and can express power or doppler at the moment of a fall, but cannot express power after the fall. The weak movement of breathing, when falling down, the energy is concentrated near the floor (the radar installation height information can be obtained during installation), in other use scenes of a toilet, only the toilet bowl can enable the human body to be relatively static and embody the weak breathing movement, and the energy is concentrated at the position at least 1 m away from the ground. And the weak movement of breathing can not occur in other using scenes.

And (3) integrating the four risk judgments, weighting and calculating the falling probability of the human body, and clearing and correcting if an obvious non-falling signal appears in the process of accumulating the falling probability, for example, a signal appears at a higher distance gate, the human body is determined to be in a standing posture, and the accumulated falling probability belongs to a misjudgment. And if the falling probability is accumulated to exceed the threshold value, outputting falling alarm information.

After the user uses the toilet, the user posture data and the environment data measured by the radar can be clouded to the SaaS system for iterative updating, so that the environment and the human body characteristics of the equipment can be better learned, and the falling detection performance is improved. The database information comprises radar installation height, environmental noise power, user height information calculated by the radar, and low-attitude height information of daily behaviors of the user such as sitting, sitting and the like in use.

In the embodiment, the whole framework based on the millimeter wave radar fall detection algorithm and the fall detection information extraction scheme are key points. In addition, the method is based on a millimeter wave radar fall detection data processing flow, a fall detection algorithm implementation step, a fall detection algorithm engineering implementation mode and a detection scheme for weak movement in Micro dimension, and comprises the steps of determining human body height, energy envelope, extracting respiratory characteristics and the like.

Fig. 6 is a schematic diagram of a fall detection apparatus for a target object according to an embodiment of the present invention, and as shown in fig. 6, according to another aspect of the embodiment of the present invention, there is also provided a fall detection apparatus for a target object, including: the receiving module 62, the first determining module 64, the second determining module 66, and the third determining module 68, which will be described in detail below.

A receiving module 62, configured to receive a feedback signal generated by a target object transmitting a radar transmission signal; a first determining module 64, connected to the receiving module 62, for determining corresponding range-doppler information according to the feedback signal; a second determining module 66, connected to the first determining module 64, for determining the body velocity information of the target object through the range-doppler information; and a third determining module 68, connected to the second determining module 66, for determining whether the target object falls down according to the distance-doppler information if the human body velocity information exceeds the preset velocity threshold.

By the device, a receiving module 62 is adopted to receive a feedback signal generated by the target object transmitting a radar transmitting signal; the first determining module 64 determines corresponding range-doppler information according to the feedback signal; the second determination module 66 determines the body velocity information of the target object through the range-doppler information; the third determining module 68 determines whether the target object falls according to the distance-doppler information when the human body velocity information exceeds the preset velocity threshold, obtains the distance-doppler information by the feedback signal of the radar, determines whether the target object falls when the human body velocity information exceeds the preset velocity threshold, and detects whether the target object falls by monitoring the distance-doppler information by the radar, thereby not only avoiding the problem that the radar detection has large calculation amount by the three-dimensional point cloud, but also avoiding the problems of high cost of the infrared detection device, high false alarm rate and false alarm rate when the environmental conditions are harsh, and poor privacy of the camera device, achieving the technical effects of reducing the calculation amount, reducing the cost, improving the accuracy and improving the privacy, and further solving the falling detection of the prior art, the method has the technical problems of large calculation amount, high cost, high false alarm rate and missing report rate and poor privacy.

According to another aspect of the embodiments of the present invention, there is also provided a processor for executing a program, wherein the program when executed performs the fall detection method for a target object of any one of the above.

According to another aspect of the embodiments of the present invention, there is also provided a computer storage medium including a stored program, wherein when the program runs, an apparatus in which the computer storage medium is located is controlled to execute the fall detection method for a target object of any one of the above.

The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.

In the above embodiments of the present invention, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.

In the embodiments provided in the present application, it should be understood that the disclosed technology can be implemented in other ways. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units may be a logical division, and in actual implementation, there may be another division, for example, multiple units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed coupling or direct coupling or communication connection between each other may be an indirect coupling or communication connection through some interfaces, units or modules, and may be electrical or in other forms.

The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.

In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.

The foregoing is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, various modifications and decorations can be made without departing from the principle of the present invention, and these modifications and decorations should also be regarded as the protection scope of the present invention.

Claims

1. A fall detection method for a target object, comprising:

receiving a feedback signal generated by a target object transmitting a transmitting signal of a radar;

determining corresponding range-doppler information according to the feedback signal;

determining human body velocity information of the target object through the range-Doppler information;

and under the condition that the human body speed information exceeds a preset speed threshold, determining whether the target object falls down or not according to the distance-Doppler information.

2. The method of claim 1, wherein prior to receiving a feedback signal generated by a target object transmitting a signal of the radar, the method further comprises:

detecting normal attitude parameters of the target object in a detection environment and environment parameters of the detection environment through radar emission signals according to a preset frequency;

matching the normal attitude parameters with trusted data in an attitude database, wherein the attitude database stores the trusted normal attitude parameters and the environment parameters;

writing the normal attitude parameters into the attitude database under the condition of successful matching;

and updating the posture threshold parameter according to the latest data of the posture database, wherein the posture threshold parameter is used for detecting whether the target object falls down.

3. The method of claim 2, wherein prior to receiving the feedback signal generated by the target object transmitting the radar's transmission signal, the method further comprises:

detecting an ambient noise power of the target object when the target object is not in a detection environment through a transmission signal of the radar, wherein the attitude threshold parameter comprises the ambient noise power;

determining the signal average power of a feedback signal on a floor of the detection environment through the feedback signal;

determining whether the target object exists in the detection environment according to the environmental noise power and the signal average power;

wherein the target object is determined to be present in the detection environment if the signal average power is greater than the ambient noise power.

4. The method as claimed in claim 1, wherein the corresponding distance is determined based on the feedback signal

The doppler information includes:

performing fast Fourier transform after removing direct current from the fast time signal of the feedback signal to obtain distance dimension information;

performing fast Fourier transform after removing direct current from the slow time signal of the feedback signal to obtain Doppler information;

and accumulating the distance dimension information according to the distance dimension information and the Doppler information to obtain the distance-Doppler information.

5. The method of claim 4, wherein performing fast Fourier transform after de-DC the fast time signal of the feedback signal to obtain distance dimension information comprises:

fast Fourier transform is carried out on fast time signals in a plurality of slow time dimensions on each frame of signals of the feedback signals, and a first range profile is obtained;

performing fast Fourier transform on the fast time signal of the feedback signal in a frame time dimension to obtain a second range profile;

wherein the distance dimension information includes the first range profile and the second range profile.

6. The method of claim 5, wherein determining whether the target object falls according to the range-Doppler information if the human body velocity information exceeds a preset velocity threshold comprises:

detecting the height of the target object according to the distance-Doppler information, and determining a first risk probability that the target object falls;

detecting a distance gate of a floor of a detection environment where the target object is located according to the distance-Doppler information, and determining a second risk probability of falling of the target object if a human body respiratory frequency characteristic exists;

determining envelope graphs of different range gates according to the range-Doppler information, and determining a third risk probability of falling of the target object;

determining a fourth risk probability that the target object falls according to whether the range-Doppler information has a signal characteristic of weak breathing;

determining a comprehensive probability of falling of the target object according to the first risk probability, the second risk probability, the third risk probability, the fourth risk probability and the corresponding weights;

and determining that the target object falls down under the condition that the comprehensive probability reaches a preset probability threshold.

7. The method of claim 6, wherein detecting the height of the target object from the range-Doppler information, wherein determining the first risk probability that the target object falls comprises:

determining whether the target object has motion close to the ground according to the first range profile;

under the condition of the motion close to the ground, acquiring a corresponding falling height according to the height of the human body in the attitude threshold parameter, wherein the falling height corresponds to the height of the human body; the height of the human body is determined according to the position of a high-power point in the second range profile at a range gate and the installation height of the radar;

determining the highest height of the target object according to the first distance image;

determining that the target object falls with the first risk probability if the highest height is less than the fall height.

8. The method of claim 6, wherein the step of detecting whether a human respiratory frequency characteristic exists or not according to the range gate of the ground of the detection environment where the target object is located and the range-Doppler information, and the step of determining the second risk probability that the target object falls comprises the steps of:

according to the second distance image, carrying out fast Fourier transform on a range gate within a first preset height range of the floor to obtain first target frequency information;

determining that the target object falls with the second risk probability if the first target frequency information has a respiratory frequency characteristic.

9. The method of claim 6, wherein determining an envelope graph of different range gates from the range-Doppler information, wherein determining a third risk probability of the target object falling comprises:

determining envelope graphs of different range gates according to the first range profile;

determining that the target object has the third risk probability of falling if the envelope graph does not match the envelope graph of the target object for non-falling, wherein the posture threshold parameter comprises the envelope graph of the normal posture of the target object for non-falling.

10. The method of claim 6, wherein determining a fourth risk probability that the target object falls based on whether the range-Doppler information has a weak breathing signal characteristic comprises:

according to the second distance image, performing fast Fourier transform on a range gate in a second preset height range to obtain second target frequency information, wherein the second preset height range is higher than the first preset height range;

determining that the target object falls with the fourth risk probability if the second target frequency information has a respiratory frequency characteristic.

11. An apparatus for fall detection of a target object, comprising:

the receiving module is used for receiving a feedback signal generated by a target object transmitting a transmitting signal of the radar;

a first determining module, configured to determine corresponding range-doppler information according to the feedback signal;

a second determining module, configured to determine human body velocity information of the target object through the range-doppler information;

and the third determining module is used for determining whether the target object falls down or not according to the distance-Doppler information under the condition that the human body speed information exceeds a preset speed threshold.

12. A processor for running a program, wherein the program is run to perform the method of fall detection of a target object of any one of claims 1 to 10.

13. A computer storage medium comprising a stored program, wherein the program, when executed, controls an apparatus in which the computer storage medium is located to perform a fall detection method for a target object according to any one of claims 1 to 10.