Human body posture recognition method and system based on millimeter wave radar detection

Abstract

The invention relates to the technical field of millimeter-wave radar detection, and particularly discloses a human body posture recognition method and system based on millimeter-wave radar detection. The method comprises the following steps: S1, constructing a human body posture information database; S2, identifying target human body information; S3, analyzing dynamic characteristics of a target human body; S4, judging a comprehensive coefficient of the dynamic characteristics of the target human body; the advanced millimeter wave radar technology and a human body posture information database are integrated, efficient and accurate recognition of human body postures can be achieved, a high-quality posture recognition result is provided for a user, in addition, a dynamic feature comprehensive coefficient DCC and a motion situation evaluation coefficient SEC are introduced, and a preset threshold value is combined, so that the human body posture recognition accuracy is improved. The intelligent self-adaptive judgment of the human body state is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of millimeter wave radar detection, and in particular to a human body posture recognition method based on millimeter wave radar detection and a system thereof. Background Art

[0002] Millimeter-wave radar is a sensor that uses electromagnetic waves with a frequency range of 30 GHz to 300 GHz for detection. The wavelength of this frequency band is between 1 mm and 10 mm, so it is called millimeter wave. Compared with traditional optical and infrared sensors, millimeter-wave radar has significant advantages, such as not being restricted by lighting conditions, being able to work in adverse weather conditions, and being able to penetrate non-metallic materials.

[0003] In a Chinese invention application with application publication number CN115345908B, a human posture recognition method based on millimeter wave radar is disclosed, and the method comprises the following steps: step 1, transmitting electromagnetic wave signals to the measured space range through the millimeter wave radar, and obtaining the target point trace based on the echo signal; step 2, for the target point trace, condensing the point trace into different areas, matching the known track according to conditions, and obtaining the motion trajectory of the target; step 3, according to the coordinate position (x, y, z) of the target point cloud in 3D space, and the signal-to-noise ratio intensity (l) of each point; step 4, respectively calculating the speed, distance and height change of the dynamic Doppler spectrum, and extracting the motion characteristics of the dynamic Doppler spectrum; step 5, proposing a sliding window target area detection algorithm, and performing target detection in the target area; step 6, inputting the acquired motion characteristics, RGB of the point trace intensity and its posture label into the neural network for training and learning for recognition.

[0004] In the above invention application, it is necessary to first obtain the motion trajectory through complex operations such as point condensation and conditional track matching, and then calculate the dynamic Doppler spectrum related parameters after multiple steps of processing such as 3D spatial coordinate position and signal-to-noise ratio intensity. The process is cumbersome, not only the calculation is time-consuming, but the cumulative error in the intermediate links may also increase, reducing the accuracy and real-time performance of the final posture recognition; in addition, the disclosed patent is not comprehensive enough in feature mining, lacks the use of key features such as human body center of gravity offset, and does not mention the construction and use of a similar complete posture map database, which increases the risk of misjudgment of posture recognition, especially in complex scenes, the accuracy of static posture recognition is easily affected. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a human posture recognition method based on millimeter wave radar detection and a system thereof to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a method for human posture recognition based on millimeter wave radar detection, comprising the following steps: S1: constructing a human posture information database, S2: identifying target human body information, S3: dynamic feature analysis of the target human body, S4: judging the comprehensive coefficient of the dynamic feature of the target human body, S5: outputting the human posture recognition result in a static state, and S6: outputting the human posture recognition result in a moving state;

[0007] S1: Building a human posture information database: Obtain various posture graphs of human bodies in a static or moving state to form a human posture information database;

[0008] S2: Identify target human body information: The millimeter wave radar transmits electromagnetic wave signals within the target detection range and receives echo signals, and pre-processes the echo signals, and identifies and screens the target human body information based on the pre-processed echo signals;

[0009] S3: Dynamic characteristic analysis of the target human body: The displacement change, azimuth angle deviation and velocity change rate of the target human body are detected, and the dynamic characteristic comprehensive coefficient DCC of the target human body is obtained based on the millimeter wave radar echo signal parameters;

[0010] S4: judging the comprehensive coefficient of the dynamic characteristics of the target human body: judging based on the comprehensive coefficient of the dynamic characteristics of the target human body and the preset threshold value, if it is judged that the target human body is in a static state, then further executing S5, otherwise further executing S6;

[0011] S5: Output of static human posture recognition result: Scan the target human body in a static state, construct a posture map of key nodes of the human body, compare it with the posture map in the human posture information database, select the map that meets the static human posture, and output the static human posture recognition result;

[0012] S6: Output of human posture recognition results in motion state: Detect the acceleration value, center of gravity deviation rate and intensity change rate of the echo signal when the target human body is in motion, analyze and obtain the motion status evaluation coefficient SEC of the target human body, and compare it with the preset motion posture threshold range to analyze the human body posture, and output the human posture recognition results in motion state.

[0013] Preferably, the human posture information database includes but is not limited to static human posture maps of standing, sitting, lying on the side, and squatting, including but not limited to moving human posture maps of walking, running, and running up stairs, including but not limited to human posture maps of key parts and key nodes of the human body, including but not limited to action labels, timestamps and posture-related metadata.

[0014] Preferably, the target detection range is specifically the living space of the target person, the living space of the target person is determined as the target detection range, and the posture of the human body is recognized within the target detection range by millimeter wave radar.

[0015] Preferably, the specific content of the dynamic characteristic analysis of the target human body is:

[0016] S31: Obtain the first distance D between the target human body and the millimeter wave radar 1 , the second distance D after time Δt 2 , substituting it into the formula Obtain the displacement change ΔD of the target human body;

[0017] Get the orientation angle θ of the target human body and substitute it into the formula Get the azimuth angle deviation Ao of the target human body, where θ 0 It is expressed as the initial orientation angle of the target human body;

[0018] Get the first speed V of the target body 1 , the second speed V after time Δt 2 , substituting it into the formula Get the speed change rate ΔV of the target human body;

[0019] S32: Obtain millimeter wave radar echo signal parameters, which include the intensity St, basic frequency f and phase of the millimeter wave radar echo signal

[0020] S33: Extract the displacement change ΔD of the target body, the azimuth deviation Ao of the target body, the velocity change rate ΔV of the target body, the intensity St of the echo signal, the basic frequency f and the phase Substitute them into the formula The dynamic characteristic comprehensive coefficient DCC of the target human body is obtained, where e is a natural constant and α is an error factor.

[0021] Preferably, the specific content of judging the comprehensive coefficient of the target human body dynamic characteristics is:

[0022] The dynamic feature comprehensive coefficient of the target human body is extracted and compared with the preset threshold. If the dynamic feature comprehensive coefficient of the target human body is less than the preset threshold, it is judged that the target human body is in a static state, and further S5 is executed. If the dynamic feature comprehensive coefficient of the target human body is greater than the preset threshold, it is judged that the target human body is in a moving state, and further S6 is executed.

[0023] Preferably, the specific content of the static human body posture recognition result output is:

[0024] Based on the judgment result of S4, an infrared sensor is used to perform a comprehensive scan of the target human body in a static state, and the key nodes of the human body are extracted from the scan data;

[0025] Establish a three-dimensional space coordinate system, that is, the three axes XYZ in the rectangular coordinate system, where the X axis is the vertical direction, that is, the standing direction of the target body, the Y axis is the front-back direction of the target body, and the Z axis is the left-right direction of the target body. The key nodes of the human body extracted from the scanning data are projected into the three-dimensional space coordinate system, and the coordinate values ​​of the key nodes are obtained. Based on the coordinate values, the connection relationship and relative position between the key nodes are described to construct a posture map of the target body;

[0026] The constructed target human body posture map is compared with the map in the pre-constructed human body posture information database. In the comparison results, the target map that matches the static human body posture map in the database is screened out, and the similarity between the target human body posture map and the static state map in the database is compared with the preset similarity threshold. When the similarity between the target human body posture map and the static state map in the database is greater than the preset similarity threshold, it is judged that the target map meets the static state, and the static state human body posture recognition result is output.

[0027] Preferably, the calculation steps of the target human body's motion posture evaluation coefficient are specifically as follows:

[0028] S61: Calculate the acceleration value of the target human body when it is moving, take the ground as the reference, establish the spatial coordinate system 0XYZ, and the coordinate system of the target human body when it is moving is OA X A Y A Z , the coordinate system is transformed by rotating β degrees around the X axis, γ degrees around the Y axis, and α around the Z axis. The coordinate calculation formula is as follows:

[0029]

[0030] , where A X , A Y , A Z They are respectively represented as the acceleration values ​​on the X, Y, and Z axes when the target human body is moving;

[0031] S62: Obtain the initial coordinates of the center of gravity of the human body (x 0 ,y 0 ,z 0 ), after time t′, the coordinates of the center of gravity become (x 1 ,y 1 ,z 1 ), substitute it into the formula Obtain the target human body's center of gravity offset rate Gsr, where L is the target human body's height;

[0032] S63: Obtain the echo signal strength S(t) received at time t, in the time interval [t 0 ,t 1 ], the echo signal strength changes from S(t0) to S(t1), and substituting it into the formula Obtain the echo signal intensity change rate Sr of the target human body;

[0033] S63: Calculate the target human body's motion status evaluation coefficient SEC, the calculation model is specifically:

[0034] Among them, e represents a natural constant, and β represents the angle between the radar beam and the direction of human body movement.

[0035] Preferably, the specific content of the output of the motion state human posture recognition result is:

[0036] The motion posture evaluation coefficient of the target human body is extracted, and the motion posture evaluation coefficient of the target human body is compared with the preset motion posture threshold interval one by one using the cluster analysis method, and the human body posture recognition results corresponding to the preset motion posture threshold interval are output respectively.

[0037] To achieve the above object, the present invention provides the following technical solution: a human posture recognition system based on millimeter wave radar detection, implementing the above human posture recognition method based on millimeter wave radar detection, comprising:

[0038] Constructing a human posture information database module: obtaining various posture graphs of human bodies in a static or moving state to form a human posture information database;

[0039] Target human body information identification module: The millimeter-wave radar transmits electromagnetic wave signals and receives echo signals within the target detection range, and pre-processes the echo signals, and identifies and screens the target human body information based on the pre-processed echo signals;

[0040] Dynamic characteristic analysis module of the target human body: detects the displacement change, azimuth deviation and velocity change rate of the target human body, and obtains the dynamic characteristic comprehensive coefficient DCC of the target human body based on the millimeter wave radar echo signal parameters;

[0041] Target human body dynamic feature comprehensive coefficient judgment module: based on the target human body dynamic feature comprehensive coefficient and the preset threshold value, if the target human body is judged to be in a static state, then further execute S5, otherwise, then further execute S6;

[0042] Static human posture recognition result output module: scans the target human body in a static state, constructs a posture map of the key nodes of the human body, compares it with the posture map in the human posture information database, selects the map that meets the static human posture, and outputs the static human posture recognition result;

[0043] The module outputs the result of human posture recognition in motion: detects the acceleration value, center of gravity deviation rate and intensity change rate of the echo signal when the target human body is in motion, analyzes the motion status evaluation coefficient SEC of the target human body, and compares it with the preset motion posture threshold interval to analyze the human posture, and outputs the recognition result of human posture in motion.

[0044] Technical effects and advantages of the present invention:

[0045] 1. The present invention can accurately capture the tiny dynamic changes of the human body by detecting and analyzing the displacement change, azimuth deviation, velocity change rate and other multi-dimensional information of the target human body, and calculate the dynamic feature comprehensive coefficient DCC based on it, providing more sophisticated data support for posture recognition;

[0046] 2. The present invention can automatically determine whether the target human body is in a static state or a moving state by comparing the target human body dynamic feature comprehensive coefficient DCC with a preset threshold, and select a corresponding posture recognition strategy accordingly, thereby realizing intelligent and adaptive recognition of human body posture. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0048] Figure 1 The present invention is a flow chart of a method for human posture recognition based on millimeter wave radar detection.

[0049] Figure 2 The present invention is a module schematic diagram of a human posture recognition system based on millimeter wave radar detection. DETAILED DESCRIPTION

[0050] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] Example 1

[0052] See also Figure 1 As shown, the present invention provides a human posture recognition method based on millimeter wave radar detection, comprising the following steps: S1: constructing a human posture information database, S2: identifying target human body information, S3: dynamic feature analysis of the target human body, S4: judging the comprehensive coefficient of the dynamic feature of the target human body, S5: outputting the human posture recognition result in a static state, and S6: outputting the human posture recognition result in a moving state;

[0053] S1: Building a human posture information database: Obtain various posture graphs of human bodies in a static or moving state to form a human posture information database;

[0054] In this embodiment, it should be specifically explained that the human posture information database includes but is not limited to static human posture maps of standing, sitting, lying on the side, and squatting, includes but is not limited to moving human posture maps of walking, running, and running up stairs, includes but is not limited to human posture maps of key parts and key nodes of the human body, includes but is not limited to action labels, timestamps and posture-related metadata.

[0055] S2: Identify target human body information: The millimeter wave radar transmits electromagnetic wave signals within the target detection range and receives echo signals, and pre-processes the echo signals, and identifies and screens the target human body information based on the pre-processed echo signals;

[0056] In this embodiment, it should be specifically explained that the target detection range is specifically the living space of the target person, the living space of the target person is determined as the target detection range, and the human body posture is recognized within the target detection range by the millimeter wave radar.

[0057] In this embodiment, it should be specifically explained that the specific content of the identified target human body information is:

[0058] Using the convolutional neural network (CNN) in deep learning technology, the echo signal received by the millimeter-wave radar is converted into two-dimensional image data, in which the distance and angle information are mapped to the coordinates of the image, and the echo intensity is used as the pixel value; by constructing a CNN model containing convolutional layers, pooling layers and fully connected layers, and using a large amount of millimeter-wave radar data containing human and non-human targets for training, the model can automatically learn and extract complex features related to the human body in the radar signal, and ultimately achieve accurate recognition and classification of the target human body.

[0059] S3: Dynamic characteristic analysis of the target human body: The displacement change, azimuth angle deviation and velocity change rate of the target human body are detected, and the dynamic characteristic comprehensive coefficient DCC of the target human body is obtained based on the millimeter wave radar echo signal parameters;

[0060] In this embodiment, it should be specifically explained that the specific content of the dynamic characteristic analysis of the target human body is:

[0061] S31: Obtain the first distance D between the target human body and the millimeter wave radar 1 , the second distance D after time Δt 2 , substituting it into the formula Obtain the displacement change ΔD of the target human body;

[0062] Get the orientation angle θ of the target human body and substitute it into the formula Get the azimuth angle deviation Ao of the target human body, where θ 0 It is expressed as the initial orientation angle of the target human body;

[0063] Get the first speed V of the target body 1 , the second speed V after time Δt 2 , substituting it into the formula Get the speed change rate ΔV of the target human body;

[0064] S32: Obtain millimeter wave radar echo signal parameters, which include the intensity St, basic frequency f and phase of the millimeter wave radar echo signal

[0065] In this embodiment, it should be specifically explained that the intensity, frequency and phase of the radar echo signal can be directly obtained through the radar system; the intensity information is obtained by measuring the amplitude of the received signal; the basic frequency can be directly obtained from the radar system; and the phase information needs to be obtained through a phase detector and a coherent processing method.

[0066] S33: Extract the displacement change ΔD of the target body, the azimuth deviation Ao of the target body, the velocity change rate ΔV of the target body, the intensity St of the echo signal, the basic frequency f and the phase Substitute them into the formula The dynamic characteristic comprehensive coefficient DCC of the target human body is obtained, where e is a natural constant and α is an error factor.

[0067] In this embodiment, it should be specifically explained that the specific method for obtaining the distance between the target human body and the millimeter wave radar, the azimuth angle of the target human body, and the speed of the target human body is:

[0068] Get the millimeter wave radar receiving signal frequency f 2 and the transmitted signal frequency f 1 , substituting it into the formula Δf=f 2 -f 1 , the signal frequency difference Δf between the millimeter wave radar receiving signal frequency and the transmitting signal frequency is obtained;

[0069] Calculate the flight time tf of the millimeter wave radar signal in space. The calculation model is as follows: Where, k represents the frequency modulation slope;

[0070] Based on the signal frequency difference Δf between the millimeter-wave radar receiving signal frequency and the transmitting signal frequency, and the flight time tf of the millimeter-wave radar signal in space, substitute it into the formula Get the distance D between the target human body and the millimeter wave radar, where Sl represents the speed of light;

[0071] Obtain the distance d between millimeter-wave radar antennas, the phase difference ω between millimeter-wave radar antennas, and the wavelength λ, and substitute them into the formula Get the orientation angle θ of the target human body;

[0072] Get the sweep frequency period Tc of the millimeter wave radar and substitute it into the formula Get the velocity V of the target human body.

[0073] S4: judging the comprehensive coefficient of the dynamic characteristics of the target human body: judging based on the comprehensive coefficient of the dynamic characteristics of the target human body and the preset threshold value, if it is judged that the target human body is in a static state, then further executing S5, otherwise further executing S6;

[0074] In this embodiment, it should be specifically explained that the specific content of the determination of the comprehensive coefficient of the target human body dynamic characteristics is:

[0075] The dynamic feature comprehensive coefficient of the target human body is extracted and compared with the preset threshold. If the dynamic feature comprehensive coefficient of the target human body is less than the preset threshold, it is judged that the target human body is in a static state, and further S5 is executed. If the dynamic feature comprehensive coefficient of the target human body is greater than the preset threshold, it is judged that the target human body is in a moving state, and further S6 is executed.

[0076] S5: Output of static human posture recognition result: Scan the target human body in a static state, construct a posture map of key nodes of the human body, compare it with the posture map in the human posture information database, select the map that meets the static human posture, and output the static human posture recognition result;

[0077] In this embodiment, it should be specifically explained that the specific content of the static human body posture recognition result output is:

[0078] Based on the judgment result of S4, an infrared sensor is used to perform a comprehensive scan of the target human body in a static state, and key nodes of the human body are extracted from the scan data; these nodes usually include joint points (such as shoulder joints, elbow joints, hip joints, knee joints, etc.) and feature points such as the head, neck, shoulders, elbows, wrists, waist, knees, ankles, etc.;

[0079] Establish a three-dimensional space coordinate system, that is, the three axes XYZ in the rectangular coordinate system, where the X axis is the vertical direction, that is, the standing direction of the target body, the Y axis is the front-back direction of the target body, and the Z axis is the left-right direction of the target body. The key nodes of the human body extracted from the scanning data are projected into the three-dimensional space coordinate system, and the coordinate values ​​of the key nodes are obtained. Based on the coordinate values, the connection relationship and relative position between the key nodes are described to construct a posture map of the target body;

[0080] The constructed target human body posture map is compared with the map in the pre-constructed human body posture information database. In the comparison results, the target map that matches the static human body posture map in the database is screened out, and the similarity between the target human body posture map and the static state map in the database is compared with the preset similarity threshold. When the similarity between the target human body posture map and the static state map in the database is greater than the preset similarity threshold, it is judged that the target map meets the static state, and the static state human body posture recognition result is output.

[0081] S6: Output of human posture recognition results in motion state: Detect the acceleration value, center of gravity deviation rate and intensity change rate of the echo signal when the target human body is in motion, analyze and obtain the motion status evaluation coefficient SEC of the target human body, and compare it with the preset motion posture threshold range to analyze the human body posture, and output the human posture recognition results in motion state.

[0082] In this embodiment, it should be specifically explained that the calculation steps of the target human body's motion posture evaluation coefficient are specifically as follows:

[0083] S61: Calculate the acceleration value of the target human body when it is moving, take the ground as the reference, establish the spatial coordinate system 0XYZ, and the coordinate system of the target human body when it is moving is OA X A Y A Z , the coordinate system is transformed by rotating β degrees around the X axis, γ degrees around the Y axis, and α around the Z axis. The coordinate calculation formula is as follows:

[0084]

[0085] , where A X , A Y , A Z They are respectively represented as the acceleration values ​​on the X, Y, and Z axes when the target human body is moving;

[0086] S62: Obtain the initial coordinates of the center of gravity of the human body (x 0 ,y 0 ,z 0 ), after time t′, the coordinates of the center of gravity become (x 1 ,y1 ,z 1 ), substitute it into the formula Obtain the target human body's center of gravity offset rate Gsr, where L is the target human body's height;

[0087] S63: Obtain the echo signal strength S(t) received at time t, in the time interval [t 0 ,t 1 ], the echo signal strength changes from S(t0) to S(t1), and substituting it into the formula Obtain the echo signal intensity change rate Sr of the target human body;

[0088] S63: Calculate the target human body's motion status evaluation coefficient SEC, the calculation model is specifically:

[0089] Among them, e represents a natural constant, β represents the angle between the radar beam and the direction of human movement;

[0090] In this embodiment, it should be specifically explained that the specific content of the output of the motion state human posture recognition result is:

[0091] The motion posture evaluation coefficient of the target human body is extracted, and the motion posture evaluation coefficient of the target human body is compared with the preset motion posture threshold interval one by one using the cluster analysis method, and the human body posture recognition results corresponding to the preset motion posture threshold interval are output respectively.

[0092] Example 2

[0093] See also Figure 2 As shown, the present invention provides a human posture recognition system based on millimeter wave radar detection, comprising:

[0094] Constructing a human posture information database module: obtaining various posture graphs of human bodies in a static or moving state to form a human posture information database;

[0095] Target human body information identification module: The millimeter-wave radar transmits electromagnetic wave signals and receives echo signals within the target detection range, and pre-processes the echo signals, and identifies and screens the target human body information based on the pre-processed echo signals;

[0096] Dynamic characteristic analysis module of the target human body: detects the displacement change, azimuth deviation and velocity change rate of the target human body, and obtains the dynamic characteristic comprehensive coefficient DCC of the target human body based on the millimeter wave radar echo signal parameters;

[0097] Target human body dynamic feature comprehensive coefficient judgment module: based on the target human body dynamic feature comprehensive coefficient and the preset threshold value, if the target human body is judged to be in a static state, then further execute S5, otherwise, then further execute S6;

[0098] Static human posture recognition result output module: scans the target human body in a static state, constructs a posture map of the key nodes of the human body, compares it with the posture map in the human posture information database, selects the map that meets the static human posture, and outputs the static human posture recognition result;

[0099] The module outputs the result of human posture recognition in motion: detects the acceleration value, center of gravity deviation rate and intensity change rate of the echo signal when the target human body is in motion, analyzes the motion status evaluation coefficient SEC of the target human body, and compares it with the preset motion posture threshold interval to analyze the human posture, and outputs the recognition result of human posture in motion.

[0100] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0101] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A human body posture recognition method based on millimeter wave radar detection, characterized in that: The following steps are involved: S1: Building a human posture information database: Obtain various posture graphs of human bodies in a static or moving state to form a human posture information database; S2: Identify target human body information: The millimeter wave radar transmits electromagnetic wave signals within the target detection range and receives echo signals, and pre-processes the echo signals, and identifies and screens the target human body information based on the pre-processed echo signals; S3: Dynamic characteristic analysis of the target human body: The displacement change, azimuth angle deviation and velocity change rate of the target human body are detected, and the dynamic characteristic comprehensive coefficient DCC of the target human body is obtained based on the millimeter wave radar echo signal parameters; S4: judging the comprehensive coefficient of the dynamic characteristics of the target human body: judging based on the comprehensive coefficient of the dynamic characteristics of the target human body and the preset threshold value, if it is judged that the target human body is in a static state, then further executing S5, otherwise further executing S6; S5: Output of static human posture recognition result: Scan the target human body in a static state, construct a posture map of key nodes of the human body, compare it with the posture map in the human posture information database, select the map that meets the static human posture, and output the static human posture recognition result; S6: Output of human posture recognition results in motion state: Detect the acceleration value, center of gravity deviation rate and intensity change rate of the echo signal when the target human body is in motion, analyze and obtain the motion status evaluation coefficient SEC of the target human body, and compare it with the preset motion posture threshold range to analyze the human body posture, and output the human posture recognition results in motion state.

2. The method for human body posture recognition based on millimeter wave radar detection according to claim 1, characterized in that: The human posture information database includes but is not limited to static human posture maps of standing, sitting, lying on the side, and squatting, including but not limited to dynamic human posture maps of walking, running, and running up stairs, including but not limited to human posture maps of key parts and key nodes of the human body, including but not limited to action tags, timestamps and posture-related metadata.

3. The method for human body posture recognition based on millimeter wave radar detection according to claim 1, characterized in that: The target detection range is specifically the living space of the target person. The living space of the target person is determined as the target detection range, and the posture of the human body is recognized within the target detection range by the millimeter wave radar.

4. The method for human body posture recognition based on millimeter wave radar detection according to claim 1, characterized in that: The specific content of the dynamic characteristic analysis of the target human body is: S31: Obtain the first distance D1 between the target human body and the millimeter wave radar, and the second distance D2 after a time Δt, and substitute them into the formula Obtain the displacement change ΔD of the target human body; Get the orientation angle θ of the target human body and substitute it into the formula The azimuth angle deviation Ao of the target human body is obtained, where θ0 represents the initial azimuth angle of the target human body; Obtain the first speed V1 of the target human body and the second speed V2 after time Δt, and substitute them into the formula Obtain the speed change rate ΔV of the target human body; S32: Obtain millimeter wave radar echo signal parameters, which include the intensity St, basic frequency f and phase of the millimeter wave radar echo signal S33: Extract the displacement change ΔD of the target body, the azimuth deviation Ao of the target body, the velocity change rate ΔV of the target body, the intensity St of the echo signal, the basic frequency f and the phase Substitute them into the formula The dynamic characteristic comprehensive coefficient DCC of the target human body is obtained, where e is a natural constant and α is an error factor.

5. The method for human body posture recognition based on millimeter wave radar detection according to claim 1, characterized in that: The specific content of the judgment of the comprehensive coefficient of the target human body dynamic characteristics is: The dynamic feature comprehensive coefficient of the target human body is extracted and compared with the preset threshold. If the dynamic feature comprehensive coefficient of the target human body is less than the preset threshold, it is judged that the target human body is in a static state, and further S5 is executed. If the dynamic feature comprehensive coefficient of the target human body is greater than the preset threshold, it is judged that the target human body is in a moving state, and further S6 is executed.

6. The method for human posture recognition based on millimeter wave radar detection according to claim 1, characterized in that: The specific content of the static human body posture recognition result output is: Based on the judgment result of S4, an infrared sensor is used to perform a comprehensive scan of the target human body in a static state, and the key nodes of the human body are extracted from the scan data; Establish a three-dimensional space coordinate system, that is, the three axes XYZ in the rectangular coordinate system, where the X axis is the vertical direction, that is, the standing direction of the target body, the Y axis is the front-back direction of the target body, and the Z axis is the left-right direction of the target body. The key nodes of the human body extracted from the scanning data are projected into the three-dimensional space coordinate system, and the coordinate values ​​of the key nodes are obtained. Based on the coordinate values, the connection relationship and relative position between the key nodes are described to construct a posture map of the target body; The constructed target human body posture map is compared with the map in the pre-constructed human body posture information database. In the comparison results, the target map that matches the static human body posture map in the database is screened out, and the similarity between the target human body posture map and the static state map in the database is compared with the preset similarity threshold. When the similarity between the target human body posture map and the static state map in the database is greater than the preset similarity threshold, it is judged that the target map meets the static state, and the static state human body posture recognition result is output.

7. The method for human posture recognition based on millimeter wave radar detection according to claim 1, characterized in that: The calculation steps of the target human body's motion posture evaluation coefficient are specifically as follows: S61: Calculate the acceleration value of the target human body when it is moving, take the ground as the reference, establish the spatial coordinate system 0XYZ, and the coordinate system of the target human body when it is moving is OA X A Y A Z , the coordinate system is transformed by rotating β degrees around the X axis, γ degrees around the Y axis, and α around the Z axis. The coordinate calculation formula is as follows: Among them, A X , A Y , A Z They are respectively represented as the acceleration values ​​on the X, Y, and Z axes when the target human body is moving; S62: The initial coordinates of the center of gravity of the human body are obtained as (x0, y0, z0). After time t′, the coordinates of the center of gravity become (x1, y1, z1). Substitute them into the formula Obtain the target human body's center of gravity offset rate Gsr, where L is the target human body's height; S63: Obtain the echo signal strength S(t) received at time t. After the time interval [t0, t1], the echo signal strength changes from S(t0) to S(t1). Substitute it into the formula Obtain the echo signal intensity change rate Sr of the target human body; S63: Calculate the target human body's motion status evaluation coefficient SEC, the calculation model is specifically: Among them, e represents a natural constant, and β represents the angle between the radar beam and the direction of human body movement.

8. The method for human posture recognition based on millimeter wave radar detection according to claim 1, characterized in that: The specific content of the output of the motion state human posture recognition result is: The motion posture evaluation coefficient of the target human body is extracted, and the motion posture evaluation coefficient of the target human body is compared with the preset motion posture threshold interval one by one using the cluster analysis method, and the human body posture recognition results corresponding to the preset motion posture threshold interval are output respectively.

9. A human posture recognition system based on millimeter wave radar detection, used to implement a human posture recognition method based on millimeter wave radar detection as described in any one of claims 1 to 8, characterized in that: include: Constructing a human posture information database module: obtaining various posture graphs of human bodies in a static or moving state to form a human posture information database; Target human body information identification module: The millimeter-wave radar transmits electromagnetic wave signals and receives echo signals within the target detection range, and pre-processes the echo signals, and identifies and screens the target human body information based on the pre-processed echo signals; Dynamic characteristic analysis module of the target human body: detects the displacement change, azimuth deviation and velocity change rate of the target human body, and obtains the dynamic characteristic comprehensive coefficient DCC of the target human body based on the millimeter wave radar echo signal parameters; Target human body dynamic feature comprehensive coefficient judgment module: based on the target human body dynamic feature comprehensive coefficient and the preset threshold value, if the target human body is judged to be in a static state, then further execute S5, otherwise, then further execute S6; Static human posture recognition result output module: scans the target human body in a static state, constructs a posture map of the key nodes of the human body, compares it with the posture map in the human posture information database, selects the map that meets the static human posture, and outputs the static human posture recognition result; The module outputs the result of human posture recognition in motion: detects the acceleration value, center of gravity deviation rate and intensity change rate of the echo signal when the target human body is in motion, analyzes the motion status evaluation coefficient SEC of the target human body, and compares it with the preset motion posture threshold interval to analyze the human posture, and outputs the recognition result of human posture in motion.

Images