CN111273284A - Human body trunk micromotion state change feature extraction method - Google Patents

Human body trunk micromotion state change feature extraction method Download PDF

Info

Publication number
CN111273284A
CN111273284A CN202010152758.2A CN202010152758A CN111273284A CN 111273284 A CN111273284 A CN 111273284A CN 202010152758 A CN202010152758 A CN 202010152758A CN 111273284 A CN111273284 A CN 111273284A
Authority
CN
China
Prior art keywords
human body
doppler
trunk
time
spectrogram
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010152758.2A
Other languages
Chinese (zh)
Other versions
CN111273284B (en
Inventor
周代英
胡晓龙
李粮余
张同梦雪
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN202010152758.2A priority Critical patent/CN111273284B/en
Publication of CN111273284A publication Critical patent/CN111273284A/en
Application granted granted Critical
Publication of CN111273284B publication Critical patent/CN111273284B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/88—Radar or analogous systems specially adapted for specific applications
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/11—Measuring movement of the entire body or parts thereof, e.g. head or hand tremor, mobility of a limb
    • A61B5/1123—Discriminating type of movement, e.g. walking or running
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
    • G01S7/36—Means for anti-jamming, e.g. ECCM, i.e. electronic counter-counter measures
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
    • G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • General Physics & Mathematics (AREA)
  • Pathology (AREA)
  • Surgery (AREA)
  • Dentistry (AREA)
  • Biophysics (AREA)
  • Physiology (AREA)
  • Biomedical Technology (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Medical Informatics (AREA)
  • Molecular Biology (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Animal Behavior & Ethology (AREA)
  • General Health & Medical Sciences (AREA)
  • Public Health (AREA)
  • Veterinary Medicine (AREA)
  • Electromagnetism (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention relates to a method for extracting the change characteristics of the micromotion state of a human body trunk. The method firstly extracts the Doppler frequency change curve of the human body from the micro Doppler spectrogram of the human body action, constructs a Markov model from the curve, and finally extracts the micro-motion state change characteristics by utilizing a Markov state transition matrix and a state distribution probability to realize the classification of the human body action. Compared with other human body motion classification methods, the method mainly utilizes the Doppler characteristic of the trunk part with the strongest energy, reduces the environmental noise interference, and solves the problem that the micro Doppler energy of four limbs of a human body is small and the characteristics are difficult to extract. By constructing a Markov model describing the trunk Doppler change process, the trunk state change in the human body motion process is fully utilized, so that the accuracy of human body motion classification is improved.

Description

Human body trunk micromotion state change feature extraction method
Technical Field
The invention particularly relates to a method for extracting the change characteristics of the micromotion state of a human body trunk.
Background
With the development of society, the protection of personal safety of individuals is concerned by more and more people, and human detection and identification gradually become a main research technology of security and monitoring systems. The human body action classification and identification technology can be used for monitoring people with some abnormal behaviors in public places (such as squares, parks and the like) and maintaining public safety. Meanwhile, the technology can be used for monitoring patients in hospitals so as to cure sudden disease conditions in time. In addition, through human action classification and identification, the safety monitoring of the old can be realized in a nursing home or at home.
In the process of human body movement, different movement modes of each part form different forms and degrees of modulation in millimeter wave radar echo signals, the modulation effect can be observed from a time-frequency spectrogram of the radar echo signals, and the modulation effect is called micro Doppler characteristics of the human body part.
The human body action classification and identification method based on time-frequency analysis is a classic method, and achieves good identification effect. For example, some methods use the envelope variation characteristics of human micro-doppler images for motion classification; and some methods use the characteristics of micro-Doppler bandwidth, four-limb movement periods, trunk average Doppler and the like extracted from a human body micro-Doppler spectrogram to realize classification and identification of human body actions. The classification methods can obtain good human body motion classification effect under the condition of high signal-to-noise ratio. However, these methods use the upper and lower envelopes of the micro doppler spectrum generated by the motion of the limbs of the human body to extract the features, and the energy is small and is easily interfered by background noise, so that the micro doppler features extracted under the condition of low signal to noise ratio are not obvious, thereby reducing the classification performance of the human body actions.
Disclosure of Invention
The main content of the present invention is to provide a method for classifying and recognizing human body motion by using the markov property of human body moving body doppler. Because the micro Doppler energy of the human body trunk part is strongest, the micro Doppler energy is most obvious in spectrogram and easy to extract, and under the condition of large environmental noise interference, the robust classification characteristics can still be extracted, so that the reliable identification of the human body action is realized.
The technical scheme of the invention is as follows: a human body trunk micromotion state change feature extraction method comprises the following steps:
s1, for the continuous wave radar, the received radar echo signal S (t) of the human body trunk is:
Figure BDA0002403018560000021
wherein A is0Proportional to the radar reflection cross section of the trunk, D (t) is the distance between the central position of the trunk and the radar, and lambda is the wavelength of the radar emission wave; s (t) obtaining the time-frequency signal STFT (t, omega) through short-time Fourier transform:
Figure BDA0002403018560000022
Figure BDA0002403018560000023
wherein G (t) is a mean of zero and a variance of σ2ω is angular frequency, and s (τ) is echo signal value at time τ;
taking an absolute value of short-time Fourier transform to obtain a micro Doppler spectrogram S (f, t) of the signal:
S(f,t)＝|STFT(t,ω)|
s2, in order to improve the accuracy of human trunk Doppler curve extraction, image segmentation and morphological processing are required to be carried out on the micro Doppler spectrogram, specifically:
s21, utilizing the background time-frequency spectrogram without the target human body to obtain the energy distribution matrix P belonging to C by statisticsV×TV is the maximum energy value of the spectrogram, and T is the total movement time of the human body;
s22, calculating division threshold th (t) at time t:
Figure BDA0002403018560000024
wherein the content of the first and second substances,S0(f, T) is a background time-frequency spectrum, T is the total time length of human body movement, [ -f [ ]m,fm]Is the range of spectrogram doppler frequencies;
s23, performing threshold segmentation on the micro Doppler spectrogram according to a time sequence:
Figure BDA0002403018560000025
ST(f, t) is a spectrogram obtained after threshold segmentation;
s24, performing morphological processing including corrosion and expansion operation on the segmented micro Doppler spectrogram; through corrosion and expansion treatment, noise can be further eliminated, and simultaneously, micro Doppler signals of human body movement are greatly restored, so that the whole human body trunk Doppler image is more complete;
s3, extracting the Doppler curve of the human body:
in the process of human body movement, the trunk is an energy concentration area of echo signals. As for the micro doppler power distribution result at a certain time, the doppler signal intensity formed by the trunk movement shows the maximum value, so the doppler signal line caused by the trunk movement can be extracted by peak detection, that is, the point of the micro doppler spectrogram processed in step S2 with the maximum power value at time t is extracted, and the doppler frequency m (t) at that time is recorded:
Figure BDA0002403018560000031
since the moving speed of the trunk does not change abruptly when the human body moves in actual conditions, the instantaneous change rate of the doppler frequency of the trunk needs to be limited. Then under the constraint condition, defining the effective point of the trunk Doppler at the initial moment as Xtorso(1)＝0，Ytorso(1)＝M(0)，Xtorso(m) is the time of the mth trunk effective point, Ytorso(m) is the doppler frequency of the mth torso significant point, and then the next torso doppler significant point is selected from m (t) using the nearest neighbor peak principle:
Figure BDA0002403018560000032
Ytorso(m)＝M(t),Xtorso(m)＝t
Figure BDA0002403018560000033
β is a limit slope, and all valid points are extracted in an iteration mode;
performing curve fitting on the obtained Doppler effective points of the human body by using a least square method to obtain an error function e (α)1,α2,...,αL) Comprises the following steps:
Figure BDA0002403018560000034
where F (t) is the fitted curve function, α1,α2,...,αLThe polynomial coefficient in F (t) and M is the total effective point number of the trunk Doppler;
to obtain the fitted curve function with minimum error, let e (α)1,α2,...,αL) Are respectively paired with α1,α2,...,αLObtaining the partial derivative, and making the partial derivative be 0, then obtaining L equations, and solving the system of L equations can obtain α1,α2,...,αLObtaining a fitting curve function F (t);
s4, extracting the change characteristics of the jogging state of the human body, wherein the Doppler frequency of the human body has Markov property along with the time change process, so that a Markov model can be constructed by using the extracted Doppler curve function F (t), and the change characteristics of the jogging state of the human body are extracted, specifically as follows:
and F (t) is quantized, and the quantization result is:
Figure BDA0002403018560000041
wherein [. ]]As a rounding function, fdIs a frequency quantization interval; defining the number of frequency quantization intervals as N;
statistics at time intervals of tdIn the case of (1), the number of transitions between the Doppler frequencies of the respective bodies and the number of occurrences of the Doppler frequency of the bodies, if Fq(t) i and Fq(t+td) J, then Wn(i,j)＝Wn(i,j)+1，pn(j)＝pn(j) + 1; wherein i, j is 0,1,2, … N, WnFor Markov state transition times statistical matrix, Wn(i, j) is a matrix WnElement of (5), pnAs a state distribution statistical vector, pn(j) Is pnThe elements of (1);
calculating a state transition matrix W and a state distribution probability vector p:
Figure BDA0002403018560000042
Figure BDA0002403018560000043
where W (i, j) is an element in the state transition matrix W, and p (j) is an element in the state distribution probability vector p;
acquiring the change characteristics of the micromotion state of the human body: calculating by using the state transition matrix W and the state distribution probability vector p to obtain a human trunk micromotion state change characteristic matrix C corresponding to each type of action:
C＝W*diag(p)。
compared with other human body motion classification methods, the method disclosed by the invention has the beneficial effects that the Doppler characteristic of the trunk part with the strongest energy is mainly utilized, the environmental noise interference is reduced, and the problem that the micro Doppler energy of four limbs of a human body is small and the characteristics are difficult to extract is solved; by constructing a Markov model describing the trunk Doppler change process, the trunk state change in the human body motion process is fully utilized, so that the accuracy of human body motion classification is improved.
Detailed Description
The effectiveness of the invention is demonstrated below in connection with a specific simulation example.
The simulation experiment uses a frequency modulation continuous wave radar (AWR1642) of TI company, and radar echo data of 6 motion actions of 3 persons 3.5-5 meters away from the radar are respectively collected. The 6 actions are running, walking, forward tumbling, stooping, squatting, and swing arm in place, respectively. 3072 time-frequency spectrograms are obtained by each action, wherein 1536 spectrograms are used for classifier training, and 1536 spectrograms are used for testing.
Aiming at a training data set and a testing data set, the method and the other two conventional methods are utilized to extract the characteristics corresponding to various human body actions, the characteristics are classified by a minimum distance classifier, and the experimental results are shown in the table 1:
TABLE 1 classification of human body movements by three methods
Figure BDA0002403018560000051
Figure BDA0002403018560000061
The method 1 extracts the envelope variation characteristics of the human body micro Doppler image from radar echo data. The method 2 extracts the characteristics of micro Doppler bandwidth, four limb movement periods, trunk average Doppler and the like from the micro Doppler spectrogram of the human body action.
As can be seen from the classification result data in the table, the method 2 classifies the actions depending on the envelope of the micro-Doppler image of the human body, so that the accuracy of classifying the actions with similar spectrogram envelopes of running, walking and swing arm in place is low. Although the method 1 has the average doppler feature of the trunk, the classification features mainly extracted depend on the envelope, so the classification effect is improved but the classification is not very effective for some actions. The human body action classification method provided by the invention mainly depends on the trunk Doppler with strong radar echo energy and obvious micro Doppler images, and is not easy to be interfered by noise. The classification accuracy in the table shows that the human body motion classification method provided by the invention has good classification effect.

Claims (1)

1. A human body trunk micromotion state change feature extraction method is characterized by comprising the following steps:
s1, for the continuous wave radar, the received radar echo signal S (t) of the human body trunk is:
Figure FDA0002403018550000011
wherein A is0Proportional to the radar reflection cross section of the trunk, D (t) is the distance between the central position of the trunk and the radar, and lambda is the wavelength of the radar emission wave; s (t) obtaining the time-frequency signal STFT (t, omega) through short-time Fourier transform:
Figure FDA0002403018550000012
Figure FDA0002403018550000013
wherein G (t) is a mean of zero and a variance of σ2ω is angular frequency, and s (τ) is echo signal value at time τ;
taking an absolute value of short-time Fourier transform to obtain a micro Doppler spectrogram S (f, t) of the signal:
S(f,t)＝|STFT(t,ω)|
s2, carrying out image segmentation and morphological processing on the micro Doppler spectrogram, specifically:
s21, utilizing the background time-frequency spectrogram without the target human body to obtain the energy distribution matrix P belonging to C by statisticsV×TV is the maximum energy value of the spectrogram, and T is the total movement time of the human body;
s22, calculating division threshold th (t) at time t:
Figure FDA0002403018550000014
wherein S is0(f, T) is a background time-frequency spectrum, T is the total time length of human body movement, [ -f [ ]m,fm]Is the range of spectrogram doppler frequencies;
s23, performing threshold segmentation on the micro Doppler spectrogram according to a time sequence:
Figure FDA0002403018550000015
ST(f, t) is a spectrogram obtained after threshold segmentation;
s24, performing morphological processing including corrosion and expansion operation on the segmented micro Doppler spectrogram;
s3, extracting the Doppler curve of the human body:
extracting the point with the maximum energy value at the time t of the micro-Doppler spectrogram processed in the step S2, and recording the Doppler frequency M (t) at the time:
Figure FDA0002403018550000021
defining the Doppler effective point of the trunk at the initial moment as Xtorso(1)＝0，Ytorso(1)＝M(0)，Xtorso(m) is the time of the mth trunk effective point, Ytorso(m) is the doppler frequency of the mth torso significant point, and then the next torso doppler significant point is selected from m (t) using the nearest neighbor peak principle:
Figure FDA0002403018550000022
Ytorso(m)＝M(t),Xtorso(m)＝t
Figure FDA0002403018550000023
β is a limit slope, and all valid points are extracted in an iteration mode;
performing curve fitting on the obtained Doppler effective points of the human body by using a least square method to obtain an error function e (α)1,α2,...,αL) Comprises the following steps:
Figure FDA0002403018550000024
where F (t) is the fitted curve function, α1,α2,...,αLThe polynomial coefficient in F (t) and M is the total effective point number of the trunk Doppler;
let e (α)1,α2,...,αL) Are respectively paired with α1,α2,...,αLObtaining the partial derivative, and making the partial derivative be 0, then obtaining L equations, and solving the system of L equations can obtain α1,α2,...,αLObtaining a fitting curve function F (t);
s4, extracting the change characteristics of the micro-motion state of the human body trunk:
and F (t) is quantized, and the quantization result is:
Figure FDA0002403018550000031
wherein [. ]]As a rounding function, fdIs a frequency quantization interval; defining the number of frequency quantization intervals as N;
statistics at time intervals of tdIn the case of (1), the number of transitions between the Doppler frequencies of the respective bodies and the number of occurrences of the Doppler frequency of the bodies, if Fq(t) i and Fq(t+td) J, then Wn(i,j)＝Wn(i,j)+1，pn(j)＝pn(j) + 1; wherein i, j is 0,1,2, … N, WnFor Markov state transition times statistical matrix, Wn(i, j) is a matrix WnElement of (5), pnAs a state distribution statistical vector, pn(j) Is pnThe elements of (1);
calculating a state transition matrix W and a state distribution probability vector p:
Figure FDA0002403018550000032
Figure FDA0002403018550000033
where W (i, j) is an element in the state transition matrix W, and p (j) is an element in the state distribution probability vector p;
acquiring the change characteristics of the micromotion state of the human body: calculating by using the state transition matrix W and the state distribution probability vector p to obtain a human trunk micromotion state change characteristic matrix C corresponding to each type of action:
C＝W*diag(p)。
CN202010152758.2A 2020-03-06 2020-03-06 Human body trunk micromotion state change feature extraction method Active CN111273284B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010152758.2A CN111273284B (en) 2020-03-06 2020-03-06 Human body trunk micromotion state change feature extraction method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010152758.2A CN111273284B (en) 2020-03-06 2020-03-06 Human body trunk micromotion state change feature extraction method

Publications (2)

Publication Number Publication Date
CN111273284A true CN111273284A (en) 2020-06-12
CN111273284B CN111273284B (en) 2022-05-17

Family

ID=70997674

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010152758.2A Active CN111273284B (en) 2020-03-06 2020-03-06 Human body trunk micromotion state change feature extraction method

Country Status (1)

Country Link
CN (1) CN111273284B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112163510A (en) * 2020-09-25 2021-01-01 电子科技大学 Human body action classification recognition method based on multi-observation variable HMM model
CN113208566A (en) * 2021-05-17 2021-08-06 深圳大学 Data processing method and device, electronic equipment and storage medium
CN116597521A (en) * 2023-07-19 2023-08-15 山东科技大学 Human track and trunk activity recognition method and system

Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2589979A1 (en) * 2011-11-03 2013-05-08 Thales Nederland B.V. System for characterizing motion of an individual, notably a human individual
CN104267394A (en) * 2014-10-07 2015-01-07 电子科技大学 High-resolution human body target motion feature detecting method
US20160379475A1 (en) * 2015-06-29 2016-12-29 Echocare Technologies Ltd. Human motion feature extraction in personal emergency response systems and methods
CN107238823A (en) * 2017-06-13 2017-10-10 电子科技大学 The state feature extracting method of true and false target
CN107861109A (en) * 2017-10-31 2018-03-30 陕西师范大学 A kind of target micro-doppler curve extracting method based on order particles filtering
CN108388850A (en) * 2018-02-08 2018-08-10 天津大学 A kind of human motion recognition method based on k arest neighbors and micro-Doppler feature
WO2019006473A1 (en) * 2017-06-30 2019-01-03 The Johns Hopkins University Systems and method for action recognition using micro-doppler signatures and recurrent neural networks
CN109738887A (en) * 2018-12-17 2019-05-10 武汉理工大学 A kind of target body moving state identification method suitable for fine motion interference scene
CN110146872A (en) * 2019-05-17 2019-08-20 西安电子科技大学 Human body micro-doppler component extraction method based on inverse Radon transform
CN110414426A (en) * 2019-07-26 2019-11-05 西安电子科技大学 A kind of pedestrian's Approach for Gait Classification based on PC-IRNN
CN110450784A (en) * 2019-07-30 2019-11-15 深圳普捷利科技有限公司 A kind of driver status monitoring method and system based on fmcw radar

Patent Citations (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2589979A1 (en) * 2011-11-03 2013-05-08 Thales Nederland B.V. System for characterizing motion of an individual, notably a human individual
CN104267394A (en) * 2014-10-07 2015-01-07 电子科技大学 High-resolution human body target motion feature detecting method
US20160379475A1 (en) * 2015-06-29 2016-12-29 Echocare Technologies Ltd. Human motion feature extraction in personal emergency response systems and methods
CN107238823A (en) * 2017-06-13 2017-10-10 电子科技大学 The state feature extracting method of true and false target
WO2019006473A1 (en) * 2017-06-30 2019-01-03 The Johns Hopkins University Systems and method for action recognition using micro-doppler signatures and recurrent neural networks
CN107861109A (en) * 2017-10-31 2018-03-30 陕西师范大学 A kind of target micro-doppler curve extracting method based on order particles filtering
CN108388850A (en) * 2018-02-08 2018-08-10 天津大学 A kind of human motion recognition method based on k arest neighbors and micro-Doppler feature
CN109738887A (en) * 2018-12-17 2019-05-10 武汉理工大学 A kind of target body moving state identification method suitable for fine motion interference scene
CN110146872A (en) * 2019-05-17 2019-08-20 西安电子科技大学 Human body micro-doppler component extraction method based on inverse Radon transform
CN110414426A (en) * 2019-07-26 2019-11-05 西安电子科技大学 A kind of pedestrian's Approach for Gait Classification based on PC-IRNN
CN110450784A (en) * 2019-07-30 2019-11-15 深圳普捷利科技有限公司 A kind of driver status monitoring method and system based on fmcw radar

Non-Patent Citations (8)

* Cited by examiner, † Cited by third party
Title
DE WIT, JJM: "Radar Micro-Doppler Feature Extraction Using the Singular Value Decomposition", 《2014 INTERNATIONAL RADAR CONFERENCE (RADAR)》 *
HONG, LING: "Micro-Doppler Curves Extraction Based on High-Order Particle Filter Track-Before Detect", 《REMOTE SENSING LETTERS》 *
PADAR, M.O: "Classification of human motion using radar micro-Doppler signatures with hidden Markov models", 《2016 IEEE RADAR CONFERENCE (RADARCONF)》 *
张群: "微动目标雷达特征提取、成像与识别研究进展", 《雷达学报》 *
张翼: "基于微多普勒特征的人体运动辨识", 《西南大学学报》 *
董亮华: "雷达回波人体特征提取算法研究", 《万方数据库》 *
陈亦望: "于微多普勒信号三维形状特征的人体目标动作分类方法", 《模式识别与人工智能》 *
黄琳琳: "雷达微多普勒特征提取与参数估计算法研究", 《万方数据库》 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112163510A (en) * 2020-09-25 2021-01-01 电子科技大学 Human body action classification recognition method based on multi-observation variable HMM model
CN112163510B (en) * 2020-09-25 2022-04-22 电子科技大学 Human body action classification recognition method based on multi-observation variable HMM model
CN113208566A (en) * 2021-05-17 2021-08-06 深圳大学 Data processing method and device, electronic equipment and storage medium
CN113208566B (en) * 2021-05-17 2023-06-23 深圳大学 Data processing method and device, electronic equipment and storage medium
CN116597521A (en) * 2023-07-19 2023-08-15 山东科技大学 Human track and trunk activity recognition method and system
CN116597521B (en) * 2023-07-19 2023-09-19 山东科技大学 Human track and trunk activity recognition method and system

Also Published As

Publication number Publication date
CN111273284B (en) 2022-05-17

Similar Documents

Publication Publication Date Title
CN111273284B (en) Human body trunk micromotion state change feature extraction method
Kim et al. Human activity classification based on micro-Doppler signatures using an artificial neural network
CN108226892B (en) Deep learning-based radar signal recovery method in complex noise environment
Yang et al. Omnidirectional motion classification with monostatic radar system using micro-Doppler signatures
Fioranelli et al. Classification of unarmed/armed personnel using the NetRAD multistatic radar for micro-Doppler and singular value decomposition features
CN111220958A (en) Radar target Doppler image classification and identification method based on one-dimensional convolutional neural network
CN103310789B (en) A kind of sound event recognition method of the parallel model combination based on improving
Wang et al. Blind source extraction of acoustic emission signals for rail cracks based on ensemble empirical mode decomposition and constrained independent component analysis
CN110045348A (en) A kind of human motion state classification method based on improvement convolutional neural networks
CN112617813B (en) Multi-sensor-based non-invasive fall detection method and system
CN104887263A (en) Identity recognition algorithm based on heart sound multi-dimension feature extraction and system thereof
CN110647788A (en) Human daily behavior classification method based on micro-Doppler characteristics
CN112580486A (en) Human behavior classification method based on radar micro-Doppler signal separation
Zhu et al. Classification of UAV-to-ground targets based on enhanced micro-Doppler features extracted via PCA and compressed sensing
CN113341392B (en) Human behavior classification method based on multi-station radar micro-Doppler motion direction finding
CN116338684A (en) Human body falling detection method and system based on millimeter wave radar and deep learning
Chen et al. Human activity recognition using temporal 3DCNN based on FMCW radar
CN110414426B (en) Pedestrian gait classification method based on PC-IRNN
Xiong et al. An unsupervised dictionary learning algorithm for neural recordings
CN112163510B (en) Human body action classification recognition method based on multi-observation variable HMM model
Wang et al. Rapid recognition of human behavior based on micro-Doppler feature
Tivive et al. Automatic human motion classification from Doppler spectrograms
Avanzato et al. A cnn-based differential image processing approach for rainfall classification
CN108549076B (en) Method for identifying scenes of various unmanned aerial vehicles based on velocity and rhythm diagram
CN109541567A (en) High-speed maneuver object detection method based on deep learning

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant