CN109919132B - Pedestrian falling identification method based on skeleton detection - Google Patents
Pedestrian falling identification method based on skeleton detection Download PDFInfo
- Publication number
- CN109919132B CN109919132B CN201910222523.3A CN201910222523A CN109919132B CN 109919132 B CN109919132 B CN 109919132B CN 201910222523 A CN201910222523 A CN 201910222523A CN 109919132 B CN109919132 B CN 109919132B
- Authority
- CN
- China
- Prior art keywords
- pedestrian
- human body
- fall
- skeleton
- image
- 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.)
- Active
Links
Images
Abstract
A pedestrian fall identification method based on skeleton detection comprises the following steps: s1, acquiring a monitoring area image by using a camera; s2, segmenting the image to obtain a pedestrian human body region image, and detecting to obtain pedestrian skeleton characteristic point distribution information; s3, analyzing the distribution information of the pedestrian skeleton feature points, extracting the joint point coordinates of key parts of the pedestrian human body, and obtaining the spatial position features of the joint points of the pedestrian human body and the pedestrian posture geometric quantity according to the joint point coordinates, wherein the key parts are preset parts on the human body; s4, establishing a falling detection model according to the spatial position characteristics of the human body joint points of the pedestrians and the geometric quantities of the postures of the pedestrians; and S5, judging the postures of the pedestrians by using the falling detection model, wherein the postures of the pedestrians comprise normal walking and falling, and detecting and identifying the postures of the pedestrians according to the judgment result. The invention can actively detect the abnormal conditions of pedestrian falling and the like in the monitoring video, and can improve the monitoring capability of pedestrian safety events by combining with an early warning system.
Description
Technical Field
The invention relates to the field of machine vision, in particular to a pedestrian falling identification method based on skeleton detection through multi-feature comprehensive judgment.
Background
In recent years, the aging problem of the population of China is becoming more and more serious. The number of the old people increases year by year, and the falling is one of the big problems troubling the old people. The fall incidence of old people is high, the problem is often serious, and the problem becomes a serious medical problem and a social problem in the current era. The falling of the old people is detected by a scientific and effective method, and the falling position is positioned by a monitoring video, so that the time for treating the old people is shortened. The characteristics of falling behaviors are utilized to distinguish the falling behaviors from the behaviors common in life. However, at present, an accurate model cannot be provided for the research on the falling behaviors of the old people, and the problem of serious misjudgment is caused on the identification of falling and other similar motion behaviors.
At present, there are two main methods for detecting falling behavior. One is a fall detection technique based on wearable sensors; another is video-based detection techniques. The fall detection technology based on the wearable sensor has high accuracy, but the cost is too high, and the wearable sensor is uncomfortable to wear. The video-based fall detection technology utilizes the human body contour motion features extracted by one or more cameras to perform identification, but is easily influenced by the illumination intensity and the observation visual angle, so that the identification rate is low.
Disclosure of Invention
In order to solve the technical problem, the invention provides a pedestrian falling identification method based on skeleton detection.
In order to solve the technical problems, the invention adopts the following technical scheme:
a pedestrian fall identification method based on skeleton detection comprises the following steps:
s1, acquiring a monitoring area image by using a camera;
s2, segmenting the image to obtain a pedestrian human body region image, and detecting to obtain pedestrian skeleton characteristic point distribution information;
s3, analyzing the distribution information of the skeleton characteristic points of the pedestrian, extracting the joint point coordinates of key parts of the pedestrian human body, and obtaining the spatial position characteristics of the joint points of the pedestrian human body and the posture geometric quantity of the pedestrian according to the joint point coordinates, wherein the key parts comprise a left eye, a right eye, a left ear, a right ear, a mouth, a chest and neck part, a left shoulder, a left elbow, a left hand, a right shoulder, a right elbow, a right hand, a left hip, a left knee, a left foot, a right hip, a right knee and a right foot;
s4, establishing a falling detection model according to the spatial position characteristics of the human body joint points of the pedestrians and the geometric quantities of the postures of the pedestrians;
and S5, judging the postures of the pedestrians by using the falling detection model, wherein the postures of the pedestrians comprise normal walking and falling, and detecting and identifying the postures of the pedestrians according to the judgment result.
The step S2 further includes establishing a pedestrian fall recognition image coordinate system:
taking the coordinates of the upper left corner of the monitored area image acquired by the camera as the origin of coordinates (0,0), the positive direction of an x axis along the horizontal right direction of the image, and the positive direction of a y axis along the vertical downward direction of the image;
and combining the current frame image acquisition time t with the image coordinates (x, y) to form a continuous video space-time characteristic coordinate system (x, y, t), and associating the pedestrian skeleton characteristic point distribution information with time to form a skeleton characteristic point space-time position data set.
The fall detection model has the following fall determination conditions:
wherein i is the number of the pedestrian skeleton characteristic point, t is the current frame image acquisition time, yi(t) is the Y-axis coordinate of the pedestrian skeleton feature point i at the frame time t,is the angle of deviation of the perpendicular bisector of the human body, ktThe slope of the inclination of the perpendicular bisector of the human body is ∈tThe length of the human body captured by the camera within the frame time t is within the range, delta is a preset length threshold, and if the judgment condition 1 or the judgment condition 2 is met, the suspected fall is judged.
The fall detection model comprises the following fall determination methods:
finding the positions of the gravity center point and the geometric center of the lower limbs of the human body, preliminarily judging the suspected falling condition of the pedestrian according to the position information of coordinates of the two points, wherein the coordinates of the gravity center position of the human body are as follows:
the coordinates of the geometric center position of the lower limbs of the human body are as follows:
xi(t) is the X-axis coordinate of the pedestrian skeleton feature point i at the frame time t, if Pf(t) has an X coordinate greater than Pl(t) X coordinate ofThe person is judged to be suspected to fall.
The fall detection model further comprises the following fall determination methods:
connecting lines of the left shoulder and the right shoulder of the human body and the midpoints between the left hip and the right hip are used as the perpendicular bisectors of the human body of the pedestrian, and the included angle between the perpendicular bisectors and the ground is detectedSetting a comparison threshold, judging the relation between the included angle and the comparison threshold, and if the included angle is greater than or equal to the comparison threshold, judging that the person is suspected to fall;
the formula of the included angle between the perpendicular bisector of the human body and the ground is as follows:
wherein k istThe position coordinates of the 4 joint points of the left shoulder, the right shoulder, the left hip and the right hip are obtained, and the concrete formula is as follows:
wherein i is 5, 6, 11, 12 respectively represent the joint point numbers of right shoulder, left hip, right hip, ifOrAnd determining the person is suspected to fall, wherein alpha and beta are set contrast threshold values.
The included angle between the perpendicular bisector of the human body and the groundAnd when the image is shot, calculating the height of the human skeleton of the pedestrian in the image by using the delta value as a set angle value, and calculating the reflection distance of the pedestrian formed by the camera and the head joint point:
wherein HtIs the distance between the head joint point and the foot joint point, d is the height from the camera to the ground, w is the horizontal distance between the human body and the camera, HtThe calculation formula of (2) is as follows:
wherein xhAnd xfIs the x coordinate, y, of the head joint point and the foot joint pointhAnd yfSetting an element for the y coordinates of the head joint point and the foot joint pointt=Ht-λtNamely:
if e ist>δ, it is determined as a suspected fall.
In the process of acquiring the monitoring area image, if the time is t1Detecting that the suspected fall is satisfied at the beginning of the moment, reading a plurality of continuous frames to t2And detecting whether each frame meets the suspected falling judgment condition, and judging that the frame falls if the number of the suspected falling frames is larger than a set fall-like frequency threshold, wherein the formula is as follows:
where K is the time t1To t2Detecting the number of frames fallen, N is the threshold of the suspected falling frequency, if mu<And 0, judging that the person falls.
In the established fall detection model, the priority level of the determination condition 1 is higher than that of the determination condition 2, the determination of the determination condition 1 is performed first, and if the determination condition 1 is satisfied, the calculation of the determination condition 2 is not performed, and if the determination condition 1 is not satisfied, the calculation of the determination condition 2 is performed.
The invention can monitor the pedestrian in real time, recognize the motion posture of the pedestrian and quickly and accurately detect the falling of the pedestrian. The system is particularly suitable for indoor monitoring systems of the old, and the data are collected and analyzed to determine results and are sent to the remote terminal in real time, so that convenience is provided for guardians, and guarantee is provided for life safety of the old.
Drawings
FIG. 1 is a flow chart of a process of the present invention;
FIG. 2 is a diagram of a human skeleton;
fig. 3 is a schematic fall diagram;
FIG. 4 is a view of the vertical mid-line of the body;
fig. 5 is a schematic diagram of the relationship between the camera and the human body position information.
Detailed Description
For a better understanding of the features and technical means of the invention, together with the specific objects and functions attained by the invention, reference is made to the following detailed description taken in conjunction with the accompanying drawings.
As shown in fig. 1-5, the invention discloses a pedestrian fall identification method based on skeleton detection, which comprises the following steps:
and S1, acquiring the monitored area image by using the camera. The type of camera used is not particularly limited, and a camera having a high pixel is preferable as long as the image pickup function is satisfied.
And S2, segmenting the image to obtain a pedestrian human body region image, and detecting to obtain the distribution information of the pedestrian skeleton characteristic points.
And S3, analyzing the distribution information of the skeleton characteristic points of the pedestrian, extracting the joint point coordinates of key parts of the pedestrian human body, and obtaining the spatial position characteristics of the joint points of the pedestrian human body and the posture geometric quantity of the pedestrian according to the joint point coordinates, wherein the key parts comprise a left eye, a right eye, a left ear, a right ear, a mouth, a chest and neck part, a left shoulder, a left elbow, a left hand, a right shoulder, a right elbow, a right hand, a left hip, a left knee, a left foot, a right hip, a right knee and a right foot. As shown in fig. 2, the skeleton information of the human body extracted for obtaining the image specifically includes 18 joint points P of the human bodyi(t)={(xi(t),yi(t)) | i ═ 0,1,2, …,17}, structured feature joint point position information is composed of these 18 key points, and the corresponding labels are left eye 1, right eye 2, left ear 3, right ear 4, mouth 0, thoracic neck 17, left shoulder 6, left elbow 8, left hand 10, right shoulder 5, right elbow 7, right hand 9, left hip 11, left knee 13, left foot 15, right hip 12, right knee 14, and right foot 16, respectively. The human body skeleton information is obtained by detecting a convolutional neural network, and a skeleton model is established by a human body posture estimation AlphaPose method. The human skeleton model detected by the method has the advantages of being free from the influence of illumination conditions and shielding, and provides a reliable foundation for further implementation of the method.
And S4, establishing a falling detection model according to the spatial position characteristics of the human body joint points of the pedestrians and the geometric quantities of the postures of the pedestrians.
And S5, judging the postures of the pedestrians by using the falling detection model, wherein the postures of the pedestrians comprise normal walking and falling, and detecting and identifying the postures of the pedestrians according to the judgment result.
The fall detection model has the following fall determination conditions:
wherein i is the number of the pedestrian skeleton characteristic point, t is the current frame image acquisition time, yi(t) is the Y-axis coordinate of the pedestrian skeleton feature point i at the frame time t,is the angle of deviation of the perpendicular bisector of the human body, ktThe slope of the inclination of the perpendicular bisector of the human body is ∈tThe length of the human body captured by the camera within the frame time t is within the range, delta is a preset length threshold, and if the judgment condition 1 or the judgment condition 2 is met, the suspected fall is judged. The priority level of the determination condition 1 is higher than that of the determination condition 2, and the determination condition 1 is first determined, and if the determination condition 1 is satisfied, the calculation of the determination condition 2 is not performed, and if the determination condition 1 is not satisfied, the calculation of the determination condition 2 is performed.
As a further aspect, the fall detection model includes the following fall determination methods:
(1) establishing a pedestrian falling identification image coordinate system: the upper left corner coordinate of the monitoring area image acquired by the camera is used as the origin of coordinates (0,0), the horizontal right direction of the image is the positive direction of an x axis, and the vertical downward direction of the image is the positive direction of a y axis.
(2) And combining the current frame image acquisition time t with the image coordinates (x, y) to form a continuous video space-time characteristic coordinate system (x, y, t), and associating the pedestrian skeleton characteristic point distribution information with time to form a skeleton characteristic point space-time position data set.
(3) Finding the positions of the gravity center point and the geometric center of the lower limbs of the human body, preliminarily judging the suspected falling condition of the pedestrian according to the position information of coordinates of the two points, wherein the coordinates of the gravity center position of the human body are as follows:
the coordinates of the geometric center position of the lower limbs of the human body are as follows:
xi(t) is the X-axis coordinate of the pedestrian skeleton feature point i at the frame time t, if Pf(t) has an X coordinate greater than Pl(t) X coordinate ofDetermining the person is suspected to fall; if it isIs less than or equal toThe following steps (4) and (5) are continued to judge the falling situation. As shown in fig. 3, which is a schematic diagram of a fall, in the monitored image, the center of gravity of the human body is located above the geometric center of the lower limbs under normal standing or walking conditions, so if the y coordinate of the center of gravity of the human body of the pedestrian in the monitored image is greater than the y coordinate of the geometric center of the lower limbs of the human body of the pedestrian, that is, the center of gravity of the human body is located below the geometric center of the lower limbs, it is determined that the human body is suspected to fall.
(4) As shown in figure 4, connecting lines of the left shoulder and the right shoulder of the human body and the midpoints between the left hip and the right hip are used as a perpendicular bisector AB of the human body of the pedestrian, and an included angle between the perpendicular bisector and the ground is detectedSetting a comparison threshold, judging the relation between the included angle and the comparison threshold, and if the included angle is greater than or equal to the comparison threshold, judging that the person is suspected to fall; and if the included angle is smaller than the comparison threshold value, the walking is normal.
The formula of the included angle between the perpendicular bisector of the human body and the ground is as follows:
wherein k istThe position coordinates of the 4 joint points of the left shoulder, the right shoulder, the left hip and the right hip are obtained, and the concrete formula is as follows:
wherein i is 5, 6, 11, 12 respectively represent the joint point numbers of right shoulder, left hip, right hip, ifOrAnd determining the person is suspected to fall, wherein alpha and beta are set contrast threshold values. The horizontal direction is set to be 0 degrees, when a person is in a standing posture, the horizontal direction is set to be 90 degrees, and the included angle relation satisfies When is at timeOrThe person is judged to be suspected to fall.
The included angle between the perpendicular bisector of the human body and the groundAnd when the image is shot, calculating the height of the human skeleton of the pedestrian in the image by using the delta value as a set angle value, and calculating the reflection distance of the pedestrian formed by the camera and the head joint point:
wherein HtIs the distance between the head joint point and the foot joint point, d is the height from the camera to the ground, w is the horizontal distance between the human body and the camera, HtThe calculation formula of (2) is as follows:
wherein xhAnd xfIs the x coordinate, y, of the head joint point and the foot joint pointhAnd yfFor the joint point of the head and the footThe y coordinate of the node is set as ∈t=Ht-λtNamely:
if e ist>δ, it is determined as a suspected fall.
In the process of acquiring the monitoring area image, if the time is t1Detecting that the suspected fall is satisfied at the beginning of the moment, reading a plurality of continuous frames to t2And detecting whether each frame meets the suspected falling judgment condition, and judging that the frame falls if the number of the suspected falling frames is larger than a set fall-like frequency threshold, wherein the formula is as follows:
where K is the time t1To t2Detecting the number of frames fallen, N is the threshold of the suspected falling frequency, if mu<And 0, judging that the person falls. Due to the complexity of human motion characteristics, such as the occurrence of actions of squatting and tying a lace, sitting down, lying down and the like, the system has the possibility of misjudgment, and the problem is solved through the method.
Although the present invention has been described in detail with reference to the embodiments, it will be apparent to those skilled in the art that modifications, equivalents, improvements, and the like can be made in the technical solutions of the foregoing embodiments or in some of the technical features of the foregoing embodiments, but those modifications, equivalents, improvements, and the like are all within the spirit and principle of the present invention.
Claims (5)
1. A pedestrian fall identification method based on skeleton detection comprises the following steps:
s1, acquiring a monitoring area image by using a camera;
s2, segmenting the image to obtain a pedestrian human body region image, and detecting to obtain pedestrian skeleton characteristic point distribution information;
s3, analyzing the distribution information of the skeleton characteristic points of the pedestrian, extracting the joint point coordinates of key parts of the pedestrian human body, and obtaining the spatial position characteristics of the joint points of the pedestrian human body and the posture geometric quantity of the pedestrian according to the joint point coordinates, wherein the key parts comprise a left eye, a right eye, a left ear, a right ear, a mouth, a chest and neck part, a left shoulder, a left elbow, a left hand, a right shoulder, a right elbow, a right hand, a left hip, a left knee, a left foot, a right hip, a right knee and a right foot;
s4, establishing a falling detection model according to the spatial position characteristics of the human body joint points of the pedestrians and the geometric quantities of the postures of the pedestrians;
s5, judging the postures of the pedestrians by using a falling detection model, wherein the postures of the pedestrians comprise normal walking and falling, and detecting and identifying the postures of the pedestrians according to the judgment result;
the step S2 further includes establishing a pedestrian fall recognition image coordinate system:
taking the coordinates of the upper left corner of the monitored area image acquired by the camera as the origin of coordinates (0,0), the positive direction of an x axis along the horizontal right direction of the image, and the positive direction of a y axis along the vertical downward direction of the image;
combining the current frame image to obtain time t and image coordinates (x, y) to form a continuous video space-time characteristic coordinate system (x, y, t), and associating the pedestrian skeleton characteristic point distribution information with the time to form a skeleton characteristic point space-time position data set;
the fall detection model has the following fall determination conditions:
wherein i is the number of the pedestrian skeleton characteristic point, t is the current frame image acquisition time, yi(t) is the Y-axis coordinate of the pedestrian skeleton feature point i at the frame time t,is the angle of deviation of the perpendicular bisector of the human body, ktThe slope of the inclination of the perpendicular bisector of the human body is ∈tThe length variation range of the human body captured by the camera in the frame time t is delta which is a preset length threshold, if the length variation range meets a judgment condition 1 or a judgment condition 2, the human body is judged to be suspected to fall down, and the human body skeleton information extracted from the image is obtained, wherein the human body skeleton information specifically comprises 18 joint points P of the human bodyi(t)={(xi(t),yi(t)) | i ═ 0,1,2,.. 17}, structured feature joint point position information is composed of the 18 key points, and corresponding marks are left eye 1, right eye 2, left ear 3, right ear 4, mouth 0, chest neck 17, left shoulder 6, left elbow 8, left hand 10, right shoulder 5, right elbow 7, right hand 9, left hip 11, left knee 13, left foot 15, right hip 12, right knee 14 and right foot 16 respectively;
when the included angle between the perpendicular bisector of the human body and the groundAnd when the image is shot, calculating the height of the human skeleton of the pedestrian in the image by using the delta value as a set angle value, and calculating the reflection distance of the pedestrian formed by the camera and the head joint point:
wherein HtIs the distance between the head joint point and the foot joint point, d is the height from the camera to the ground, w is the horizontal distance between the human body and the camera, HtThe calculation formula of (2) is as follows:
wherein xhAnd xfIs the x coordinate, y, of the head joint point and the foot joint pointhAnd yfSetting an element for the y coordinates of the head joint point and the foot joint pointt=Ht-λtNamely:
if e istIf the answer is more than delta, the suspected fall is judged.
2. The pedestrian fall recognition method based on skeleton detection according to claim 1, wherein the fall detection model comprises the following fall determination methods:
finding the positions of the gravity center point and the geometric center of the lower limbs of the human body, preliminarily judging the suspected falling condition of the pedestrian according to the position information of coordinates of the two points, wherein the coordinates of the gravity center position of the human body are as follows:
the coordinates of the geometric center position of the lower limbs of the human body are as follows:
3. The pedestrian fall recognition method based on skeleton detection according to claim 2, wherein the fall detection model further comprises the following fall determination methods:
connecting lines of the left shoulder and the right shoulder of the human body and the midpoints between the left hip and the right hip are used as the perpendicular bisectors of the human body of the pedestrian, and the included angle between the perpendicular bisectors and the ground is detectedSetting contrast thresholdJudging the relation between the included angle and the comparison threshold value, and if the included angle is greater than or equal to the comparison threshold value, judging that the person is suspected to fall;
the formula of the included angle between the perpendicular bisector of the human body and the ground is as follows:
wherein k istThe position coordinates of the 4 joint points of the left shoulder, the right shoulder, the left hip and the right hip are obtained, and the concrete formula is as follows:
4. The pedestrian fall recognition method based on skeleton detection as claimed in claim 3, wherein in the process of acquiring the monitoring area image, if at t1Detecting that the suspected fall is satisfied at the beginning of the moment, reading a plurality of continuous frames to t2And detecting whether each frame meets the suspected falling judgment condition, and judging that the frame falls if the number of the suspected falling frames is larger than a set fall-like frequency threshold, wherein the formula is as follows:
where K is the time t1To t2And detecting the number of frames falling, wherein N is a set suspected falling frequency threshold, and if mu is less than 0, determining that the person falls.
5. The method for recognizing a pedestrian fall based on skeleton detection according to claim 4, wherein in the established fall detection model, the determination condition 1 has a higher priority level than the determination condition 2, the determination of the determination condition 1 is performed first, and if the determination condition 1 is satisfied, the calculation of the determination condition 2 is not performed, and if the determination condition 1 is not satisfied, the calculation of the determination condition 2 is performed.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910222523.3A CN109919132B (en) | 2019-03-22 | 2019-03-22 | Pedestrian falling identification method based on skeleton detection |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910222523.3A CN109919132B (en) | 2019-03-22 | 2019-03-22 | Pedestrian falling identification method based on skeleton detection |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109919132A CN109919132A (en) | 2019-06-21 |
CN109919132B true CN109919132B (en) | 2021-04-23 |
Family
ID=66966467
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910222523.3A Active CN109919132B (en) | 2019-03-22 | 2019-03-22 | Pedestrian falling identification method based on skeleton detection |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109919132B (en) |
Families Citing this family (48)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110378244B (en) * | 2019-05-31 | 2021-12-31 | 曹凯 | Abnormal posture detection method and device |
CN112120703A (en) * | 2019-06-25 | 2020-12-25 | 株式会社日立制作所 | Fall risk assessment method and device |
CN110287923B (en) * | 2019-06-29 | 2023-09-15 | 腾讯科技(深圳)有限公司 | Human body posture acquisition method, device, computer equipment and storage medium |
CN110490070A (en) * | 2019-07-12 | 2019-11-22 | 浙江省北大信息技术高等研究院 | A kind of fall detection method based on human body attitude estimation |
CN110490080B (en) * | 2019-07-22 | 2023-05-09 | 毕昇云(武汉)信息技术有限公司 | Human body falling judgment method based on image |
CN110458061B (en) * | 2019-07-30 | 2022-04-05 | 四川工商学院 | Method for identifying old people falling down and accompanying robot |
CN110598536A (en) * | 2019-07-31 | 2019-12-20 | 杭州连帆科技有限公司 | Falling detection method and system based on human skeleton motion model |
CN110477925B (en) * | 2019-08-23 | 2022-05-24 | 广东省智能制造研究所 | Fall detection and early warning method and system for elderly people in nursing home |
CN110532966A (en) * | 2019-08-30 | 2019-12-03 | 深兰科技(上海)有限公司 | A kind of method and apparatus carrying out tumble identification based on disaggregated model |
CN110688969A (en) * | 2019-09-30 | 2020-01-14 | 上海依图网络科技有限公司 | Video frame human behavior identification method |
CN112699706A (en) * | 2019-10-22 | 2021-04-23 | 广州弘度信息科技有限公司 | Fall detection method, system and storage medium |
CN111046749B (en) * | 2019-11-25 | 2023-05-23 | 西安建筑科技大学 | Human body falling behavior detection method based on depth data |
CN111191511B (en) * | 2019-12-03 | 2023-08-18 | 北京联合大学 | Dynamic real-time behavior recognition method and system for prison |
CN111179281A (en) * | 2019-12-24 | 2020-05-19 | 广东省智能制造研究所 | Human body image extraction method and human body action video extraction method |
CN111243229A (en) * | 2019-12-31 | 2020-06-05 | 浙江大学 | Old people falling risk assessment method and system |
CN111460908B (en) * | 2020-03-05 | 2023-09-01 | 中国地质大学(武汉) | Human body fall recognition method and system based on OpenPose |
CN111461042A (en) * | 2020-04-07 | 2020-07-28 | 中国建设银行股份有限公司 | Fall detection method and system |
CN113470310A (en) * | 2020-04-26 | 2021-10-01 | 海信集团有限公司 | Intelligent household equipment and anti-falling early warning method |
CN111582158A (en) * | 2020-05-07 | 2020-08-25 | 济南浪潮高新科技投资发展有限公司 | Tumbling detection method based on human body posture estimation |
CN113642361B (en) * | 2020-05-11 | 2024-01-23 | 杭州萤石软件有限公司 | Fall behavior detection method and equipment |
CN113705274B (en) * | 2020-05-20 | 2023-09-05 | 杭州海康威视数字技术股份有限公司 | Climbing behavior detection method and device, electronic equipment and storage medium |
CN111967307B (en) * | 2020-07-02 | 2021-11-09 | 广东技术师范大学 | Target exception handling method and device, computer equipment and storage medium |
CN111783702A (en) * | 2020-07-20 | 2020-10-16 | 杭州叙简科技股份有限公司 | Efficient pedestrian tumble detection method based on image enhancement algorithm and human body key point positioning |
CN112164091B (en) * | 2020-08-25 | 2022-08-16 | 南京邮电大学 | Mobile device human body pose estimation method based on three-dimensional skeleton extraction |
CN112036299B (en) * | 2020-08-31 | 2022-12-13 | 山东科技大学 | Examination cheating behavior detection method and system under standard examination room environment |
CN111985443A (en) * | 2020-09-01 | 2020-11-24 | 安徽炬视科技有限公司 | Robust human body posture estimation algorithm |
CN111931733B (en) * | 2020-09-25 | 2021-02-26 | 西南交通大学 | Human body posture detection method based on depth camera |
CN112287759A (en) * | 2020-09-26 | 2021-01-29 | 浙江汉德瑞智能科技有限公司 | Tumble detection method based on key points |
CN112287758B (en) * | 2020-09-26 | 2022-08-26 | 浙江汉德瑞智能科技有限公司 | Climbing identification method based on key point detection |
CN111881898B (en) * | 2020-09-27 | 2021-02-26 | 西南交通大学 | Human body posture detection method based on monocular RGB image |
CN112749642A (en) * | 2020-12-30 | 2021-05-04 | 苏州涟漪信息科技有限公司 | Method and device for identifying falling |
CN112766129A (en) * | 2021-01-12 | 2021-05-07 | 云南电网有限责任公司电力科学研究院 | Power grid worker behavior real-time monitoring method and device |
KR102558054B1 (en) * | 2021-01-22 | 2023-07-19 | 동의대학교 산학협력단 | Method and System for detecting fall situation by using deep learning model |
CN112686211A (en) * | 2021-01-25 | 2021-04-20 | 广东工业大学 | Fall detection method and device based on attitude estimation |
CN112907892A (en) * | 2021-01-28 | 2021-06-04 | 上海电机学院 | Human body falling alarm method based on multiple views |
CN112861686B (en) * | 2021-02-01 | 2022-08-30 | 内蒙古大学 | SVM-based image target detection method |
CN112906548A (en) * | 2021-02-07 | 2021-06-04 | 广东省科学院智能制造研究所 | Fall detection method and system based on edge calculation |
CN112906599A (en) * | 2021-03-04 | 2021-06-04 | 杭州海康威视数字技术股份有限公司 | Gait-based personnel identity identification method and device and electronic equipment |
CN112966597A (en) * | 2021-03-04 | 2021-06-15 | 山东云缦智能科技有限公司 | Human motion action counting method based on skeleton key points |
CN113204989B (en) * | 2021-03-19 | 2022-07-29 | 南京邮电大学 | Human body posture space-time feature extraction method for tumble analysis |
CN112949569B (en) * | 2021-03-25 | 2022-07-29 | 南京邮电大学 | Falling analysis-oriented human body posture point effective extraction method |
CN113450538A (en) * | 2021-06-28 | 2021-09-28 | 杭州电子科技大学 | Warning system based on painful expression discernment and fall action detection |
CN113537005B (en) * | 2021-07-02 | 2023-08-11 | 福州大学 | Online examination student behavior analysis method based on attitude estimation |
CN113792700B (en) * | 2021-09-24 | 2024-02-27 | 成都新潮传媒集团有限公司 | Storage battery car in-box detection method and device, computer equipment and storage medium |
CN113822250A (en) * | 2021-11-23 | 2021-12-21 | 中船(浙江)海洋科技有限公司 | Ship driving abnormal behavior detection method |
CN114694252B (en) * | 2022-03-30 | 2023-04-28 | 电子科技大学 | Old people falling risk prediction method |
CN115273243B (en) * | 2022-09-27 | 2023-03-28 | 深圳比特微电子科技有限公司 | Fall detection method, device, electronic equipment and computer readable storage medium |
CN117278837B (en) * | 2023-11-16 | 2024-01-26 | 新乡天辅电子科技有限公司 | Emergency rescue-oriented imaging equipment control method |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104598896A (en) * | 2015-02-12 | 2015-05-06 | 南通大学 | Automatic human tumble detecting method based on Kinect skeleton tracking |
CN104794463A (en) * | 2015-05-11 | 2015-07-22 | 华东理工大学 | System and method for achieving indoor human body falling detection based on Kinect |
CN107194967A (en) * | 2017-06-09 | 2017-09-22 | 南昌大学 | Human fall detection method and device based on Kinect depth image |
CN107220604A (en) * | 2017-05-18 | 2017-09-29 | 清华大学深圳研究生院 | A kind of fall detection method based on video |
CN107578036A (en) * | 2017-09-28 | 2018-01-12 | 南通大学 | A kind of depth image tumble recognizer based on wavelet moment |
CN108664887A (en) * | 2018-03-20 | 2018-10-16 | 重庆邮电大学 | Prior-warning device and method are fallen down in a kind of virtual reality experience |
CN109389041A (en) * | 2018-09-07 | 2019-02-26 | 南京航空航天大学 | A kind of fall detection method based on joint point feature |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9489570B2 (en) * | 2013-12-31 | 2016-11-08 | Konica Minolta Laboratory U.S.A., Inc. | Method and system for emotion and behavior recognition |
US9892611B1 (en) * | 2015-06-01 | 2018-02-13 | Cerner Innovation, Inc. | Method for determining whether an individual enters a prescribed virtual zone using skeletal tracking and 3D blob detection |
CN108734141B (en) * | 2018-05-28 | 2020-07-03 | 南京邮电大学 | Intelligent carpet falling discrimination method based on machine learning |
CN109460702B (en) * | 2018-09-14 | 2022-02-15 | 华南理工大学 | Passenger abnormal behavior identification method based on human body skeleton sequence |
-
2019
- 2019-03-22 CN CN201910222523.3A patent/CN109919132B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104598896A (en) * | 2015-02-12 | 2015-05-06 | 南通大学 | Automatic human tumble detecting method based on Kinect skeleton tracking |
CN104794463A (en) * | 2015-05-11 | 2015-07-22 | 华东理工大学 | System and method for achieving indoor human body falling detection based on Kinect |
CN107220604A (en) * | 2017-05-18 | 2017-09-29 | 清华大学深圳研究生院 | A kind of fall detection method based on video |
CN107194967A (en) * | 2017-06-09 | 2017-09-22 | 南昌大学 | Human fall detection method and device based on Kinect depth image |
CN107578036A (en) * | 2017-09-28 | 2018-01-12 | 南通大学 | A kind of depth image tumble recognizer based on wavelet moment |
CN108664887A (en) * | 2018-03-20 | 2018-10-16 | 重庆邮电大学 | Prior-warning device and method are fallen down in a kind of virtual reality experience |
CN109389041A (en) * | 2018-09-07 | 2019-02-26 | 南京航空航天大学 | A kind of fall detection method based on joint point feature |
Also Published As
Publication number | Publication date |
---|---|
CN109919132A (en) | 2019-06-21 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109919132B (en) | Pedestrian falling identification method based on skeleton detection | |
CN111931733B (en) | Human body posture detection method based on depth camera | |
CN111881898B (en) | Human body posture detection method based on monocular RGB image | |
CN110706255A (en) | Fall detection method based on self-adaptive following | |
CN110472473A (en) | The method fallen based on people on Attitude estimation detection staircase | |
CN111783702A (en) | Efficient pedestrian tumble detection method based on image enhancement algorithm and human body key point positioning | |
CN109271918B (en) | Method for distinguishing people with balance ability disorder based on gravity center shift model | |
CN110598536A (en) | Falling detection method and system based on human skeleton motion model | |
CN108805021A (en) | The real-time individual tumble behavioral value alarm method of feature based operator | |
CN109101943A (en) | It is a kind of for detecting the machine vision method of Falls Among Old People | |
CN114894337B (en) | Temperature measurement method and device for outdoor face recognition | |
CN115116127A (en) | Fall detection method based on computer vision and artificial intelligence | |
Abd et al. | Human fall down recognition using coordinates key points skeleton | |
JP2006318064A (en) | Image sensor | |
CN113384267A (en) | Fall real-time detection method, system, terminal equipment and storage medium | |
CN112907892A (en) | Human body falling alarm method based on multiple views | |
CN108846996B (en) | Tumble detection system and method | |
US20220366716A1 (en) | Person state detection apparatus, person state detection method, and non-transitory computer readable medium storing program | |
CN108764131A (en) | It is a kind of based on video processing adaptive threshold multiple target fall down detection method | |
KR20230078063A (en) | Server for determining the posture type and operation method thereof | |
CN113221815A (en) | Gait identification method based on automatic detection technology of skeletal key points | |
ShanShan et al. | Fall detection method based on semi-contour distances | |
CN113128330A (en) | Human body falling monitoring system | |
US20210390738A1 (en) | Methods and apparatus to cluster and collect head-toe lines for automatic camera calibration | |
CN108647589B (en) | Tumble detection method based on regularization morphological ratio |
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 | ||
CP01 | Change in the name or title of a patent holder |
Address after: 510000 13 building, 100 martyrs Road, Yuexiu District, Guangzhou, Guangdong. Patentee after: Institute of intelligent manufacturing, Guangdong Academy of Sciences Address before: 510000 13 building, 100 martyrs Road, Yuexiu District, Guangzhou, Guangdong. Patentee before: GUANGDONG INSTITUTE OF INTELLIGENT MANUFACTURING |
|
CP01 | Change in the name or title of a patent holder |