CN106373336A - Fall detection method and device - Google Patents
Fall detection method and device Download PDFInfo
- Publication number
- CN106373336A CN106373336A CN201610756018.3A CN201610756018A CN106373336A CN 106373336 A CN106373336 A CN 106373336A CN 201610756018 A CN201610756018 A CN 201610756018A CN 106373336 A CN106373336 A CN 106373336A
- Authority
- CN
- China
- Prior art keywords
- target image
- detected
- target
- posture
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/04—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons
- G08B21/0407—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons based on behaviour analysis
- G08B21/043—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons based on behaviour analysis detecting an emergency event, e.g. a fall
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/103—Static body considered as a whole, e.g. static pedestrian or occupant recognition
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/04—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons
- G08B21/0438—Sensor means for detecting
- G08B21/0476—Cameras to detect unsafe condition, e.g. video cameras
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Gerontology & Geriatric Medicine (AREA)
- Business, Economics & Management (AREA)
- Emergency Management (AREA)
- Multimedia (AREA)
- Social Psychology (AREA)
- Psychology (AREA)
- Psychiatry (AREA)
- Human Computer Interaction (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Emergency Alarm Devices (AREA)
Abstract
The invention provides a fall detection method and device. The method comprises the following steps: pre-dividing postures of an object to be detected, wherein the postures comprise a standing posture and a lie-down posture; collecting a target image comprising the object to be detected in real time; identifying a human body image of the object to be detected from the current target image; according to the human body image, determining the posture of the object to be detected in the current target image; judging whether the posture is the lie-down posture, and if so, determining a standing target image closest to the current target image in time in a target image before the current target image, wherein the standing target image is a target image of the object to be detected, the posture of which is the standing posture; determining time length between the current target image and the corresponding standing target image; and when the time length is smaller than or equal to a preset time length, judging that the object to be detected in the current target image falls over. The fall detection method and device can detect fall more conveniently.
Description
Technical field
The present invention relates to technical field of image processing, the method and device fallen down particularly to a kind of detection.
Background technology
With the raising of people's safety consciousness, how to ensure the safety of the person, how to find in the hole in time
People, is increasingly taken seriously.For example, with the aging of population, nursing old people is more and more important, and old people falls down right
It is breakneck for old people, how to check that old people falls down in time and be increasingly taken seriously.
In prior art, detect whether target to be detected falls down and mainly realized by photographic head.Clapped by photographic head
Take the photograph target to be detected, to be monitored the situation of target to be detected by the monitor of manual observation photographic head, when discovery target to be detected
When falling down, fall down signal by manually sending, to provide help to the target to be detected fallen down.
Visible by foregoing description, to be detected to be checked in prior art by way of the monitor of manual observation photographic head
Survey whether target falls down, realize finding that target to be detected is fallen down it is necessary to artificial observe monitor, this detection always in time
Mode is very inconvenient.
Content of the invention
Embodiments provide the method and device that a kind of detection is fallen down, can more easily detect and fall down.
On the one hand, embodiments provide a kind of method that detection is fallen down, comprising:
S0: divide the posture of target to be detected in advance, the posture of described target to be detected includes: the posture stood, lie down
Posture;
S1: Real-time Collection comprises the target image of described target to be detected;
S2: identify the human body image of described target to be detected from current target image;
S3: according to the described human body image in described current target image, determine described target to be detected described current
Posture in target image;
S4: whether the posture judging the target described to be detected in described current target image is the posture lied down, if
It is, then execution step s5;
S5: determine in the target image before described current target image in time with described current target image
Close target image of standing, wherein, described target image of standing is the posture of described target to be detected is the posture stood
Target image;
S6: determine described current target image and immediate with described current target image in time target figure of standing
Time span between picture;
S7: judge whether described time span is less than or equal to predetermined time period, if it is, determining described current goal
Target described to be detected in image is fallen down.
Further, described s0, comprising: divide the posture of described target to be detected previously according to division rule;
Wherein, described division rule includes:
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in described target image is more than or equal to the
During one preset value, the posture of the target described to be detected in described target image be described in the posture stood;
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in described target image is less than or equal to the
During two preset values, the posture of the target described to be detected in described target image be described in the posture lied down;
Wherein, described first preset value is more than described second preset value;
Described s3, comprising:
Determine the minimum enclosed rectangle of the described human body image in described current target image;
Determine the ratio of the height and the width of described minimum enclosed rectangle;
According to the ratio of described division rule and described height and the width, determine described target to be detected in described current mesh
Posture in logo image.
Further, the described minimum enclosed rectangle determining the described human body image in described current target image, comprising:
Set up rectangular coordinate system;
Described current target image is loaded in described rectangular coordinate system;
Determine the maximum transverse axis coordinate in described human body image and minimum transverse axis coordinate, and maximum ordinate of orthogonal axes and minimum
Ordinate of orthogonal axes;
According to described maximum transverse axis coordinate, described minimum transverse axis coordinate, described maximum ordinate of orthogonal axes and the described minimum longitudinal axis
Coordinate, determines described minimum enclosed rectangle;
The ratio of the described height and the width determining described minimum enclosed rectangle, comprising:
In described rectangular coordinate system, described minimum enclosed rectangle is made perpendicular to first length on the side of described transverse axis
For described height, using the second length perpendicular to the side of the described longitudinal axis for the described minimum enclosed rectangle as described width, by institute
State the ratio as described height and the width for the ratio of the first length and described second length.
Further, also include:
Determine the relaxing area of described target to be detected;
After judging that described time span is less than or equal to described predetermined time period, in the described current mesh of described determination
Before target described to be detected in logo image is fallen down, also include:
Judge target described to be detected in described current target image whether in described relaxing area, if it is,
Determine that the target described to be detected in described current target image is not fallen down, otherwise, it determines in described current target image
Described target to be detected is fallen down.
Further, after the described target described to be detected determining in described current target image is fallen down, also include:
Send the alarm signal that described target to be detected is fallen down.
On the other hand, embodiments provide the device that a kind of detection is fallen down, comprising:
Division unit, for dividing the posture of target to be detected, the posture of described target to be detected includes: the appearance stood
Gesture, the posture lied down;
Collecting unit, comprises the target image of described target to be detected for Real-time Collection;
Recognition unit, for identifying the human body image of described target to be detected from current target image;
Posture determining unit, for according to the described human body image in described current target image, determining described to be detected
Posture in described current target image for the target;
Posture judging unit, whether the posture for judging the target described to be detected in described current target image is to lie
The posture fallen, if it is, notify image determination unit;
Described image determining unit, for determine in target image before described current target image in time with
The immediate target image of standing of described current target image, wherein, described target image of standing is described target to be detected
Posture is the target image of the posture stood;
Time determining unit, for determining described current target image and connecing most with described current target image in time
The near time span stood between target image;
Fall down judging unit, for judging whether described time span is less than or equal to predetermined time period, if it is, really
Target described to be detected in fixed described current target image is fallen down.
Further, described division unit, for dividing the posture of described target to be detected according to division rule;
Wherein, described division rule includes:
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in described target image is more than or equal to the
During one preset value, the posture of the target described to be detected in described target image be described in the posture stood;
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in described target image is less than or equal to the
During two preset values, the posture of the target described to be detected in described target image be described in the posture lied down;
Wherein, described first preset value is more than described second preset value;
Described posture determining unit, for determining the external square of minimum of the described human body image in described current target image
Shape, determines the ratio of the height and the width of described minimum enclosed rectangle, according to described division rule and described height and the width
Ratio, determines posture in described current target image for the described target to be detected.
Further, described posture determining unit is executing the described described human figure determining in described current target image
During the minimum enclosed rectangle of picture, it is used for setting up rectangular coordinate system, described current target image is loaded into described rectangular coordinate system
In, determine the maximum transverse axis coordinate in described human body image and minimum transverse axis coordinate, and maximum ordinate of orthogonal axes and the minimum longitudinal axis
Coordinate, sits according to described maximum transverse axis coordinate, described minimum transverse axis coordinate, described maximum ordinate of orthogonal axes and the described minimum longitudinal axis
Mark, determines described minimum enclosed rectangle;
Described posture determining unit, when executing the ratio of the described height and the width determining described minimum enclosed rectangle, is used
In in described rectangular coordinate system, using the first length perpendicular to the side of described transverse axis for the described minimum enclosed rectangle as described
Highly, using the second length perpendicular to the side of the described longitudinal axis for the described minimum enclosed rectangle as described width, by described first
The ratio of length and described second length is as the ratio of described height and the width.
Further, also include:
Area determination unit, for determining the relaxing area of described target to be detected;
Described fall down judging unit judge described time span be less than or equal to described predetermined time period after, in institute
State determine described current target image in target described to be detected fall down before, be additionally operable to judge in described current target image
Target described to be detected whether in described relaxing area, if it is, determine described current target image in described in treat
Detection target is not fallen down, otherwise, it determines the target described to be detected in described current target image is fallen down.
Further, also include:
Alarm unit, for determining the target described to be detected in described current target image in described judging unit of falling down
After falling down, send the alarm signal that described target to be detected is fallen down.
In embodiments of the present invention, Real-time Collection comprises the target image of target to be detected, treats in each target image
Detection target carries out falling down inspection, extracts human body image, determine the appearance of human body image in target image from target image
Gesture, when the time span that this posture is the posture lied down and the posture lied down to this from the nearest posture stood is less than or equal to
During predetermined time period, determine that target to be detected is fallen down, whole detection process need not manually participate in, can more easily detect
Fall down.
Brief description
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
Have technology description in required use accompanying drawing be briefly described it should be apparent that, drawings in the following description are the present invention
Some embodiments, for those of ordinary skill in the art, on the premise of not paying creative work, can also basis
These accompanying drawings obtain other accompanying drawings.
Fig. 1 is the flow chart of the method that a kind of detection that one embodiment of the invention provides is fallen down;
Fig. 2 is a kind of schematic diagram of the minimum enclosed rectangle of human body image that one embodiment of the invention provides;
Fig. 3 is the schematic diagram of the minimum enclosed rectangle of another human body image that one embodiment of the invention provides;
Fig. 4 is the flow chart of the method that another kind of detection that one embodiment of the invention provides is fallen down;
Fig. 5 is the schematic diagram of the minimum enclosed rectangle of another human body image that one embodiment of the invention provides;
Fig. 6 is the schematic diagram of the device that a kind of detection that one embodiment of the invention provides is fallen down.
Specific embodiment
Purpose, technical scheme and advantage for making the embodiment of the present invention are clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described it is clear that described embodiment is
The a part of embodiment of the present invention, rather than whole embodiments, based on the embodiment in the present invention, those of ordinary skill in the art
The every other embodiment being obtained on the premise of not making creative work, broadly falls into the scope of protection of the invention.
As shown in figure 1, embodiments providing a kind of method that detection is fallen down, the method can include following walking
Rapid:
Step 101: divide the posture of target to be detected in advance, the posture of described target to be detected includes: the posture stood,
The posture lied down;
Step 102: Real-time Collection comprises the target image of described target to be detected;
Step 103: identify the human body image of described target to be detected from current target image;
Step 104: according to the described human body image in described current target image, determine described target to be detected described
Posture in current target image;
Step 105: whether the posture judging the target described to be detected in described current target image is the posture lied down,
If it is, execution step 106;
Step 106: determine in target image before described current target image in time with described current goal
The immediate target image of standing of image, wherein, described target image of standing is that the posture of described target to be detected is stood
The target image of posture;
Step 107: determine that described current target image is stood with immediate with described current target image in time
Time span between target image;
Step 108: judge whether described time span is less than or equal to predetermined time period, if it is, execution step
109.
Step 109: determine that the target described to be detected in described current target image is fallen down.
In embodiments of the present invention, Real-time Collection comprises the target image of target to be detected, treats in each target image
Detection target carries out falling down inspection, extracts human body image, determine the appearance of human body image in target image from target image
Gesture, when the time span that this posture is the posture lied down and the posture lied down to this from the nearest posture stood is less than or equal to
During predetermined time period, determine that target to be detected is fallen down, whole detection process need not manually participate in, can more easily detect
Fall down.
In order to the posture of to be detected target is determined more accurately out, in an embodiment of the present invention, described divide in advance
The posture of target to be detected, comprising: divide the posture of described target to be detected previously according to division rule;
Wherein, described division rule includes:
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in described target image is more than or equal to the
During one preset value, the posture of the target described to be detected in described target image be described in the posture stood;
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in described target image is less than or equal to the
During two preset values, the posture of the target described to be detected in described target image be described in the posture lied down;
Wherein, described first preset value is more than described second preset value;
Described according to the described human body image in described current target image, determine described target to be detected described current
Posture in target image, comprising:
Determine the minimum enclosed rectangle of the described human body image in described current target image;
Determine the ratio of the height and the width of described minimum enclosed rectangle;
According to the ratio of described division rule and described height and the width, determine described target to be detected in described current mesh
Posture in logo image.
In the present embodiment, because, when standing and lie down, the ratio of the height and the width of human body is that difference is very big
, can distinguish what posture is target to be detected be in exactly by this feature.Here the first preset value and second
Preset value can rule of thumb determine it is also possible to be determined according to experimental data, for example, obtains human body in large quantities in advance
Image, in the ratio of the posture stood and the height and the width of posture lied down, carries out process and obtains first presetting to these ratios
Value and the second preset value.There is difference in the build due to different targets to be detected, treat in order to more accurately determine
The posture of detection target, can be directed to each target to be detected, predefine out the corresponding posture stood and the posture lied down
Height and the width ratio, and then determine corresponding first preset value of each target to be detected and the second preset value.Wherein,
First preset value span can be [1.5,2], and the span of the second preset value can be [0.5,2/3], for example: first
Preset value is 2, and the second preset value is 0.5.Except the posture stood and the posture lied down, middle posture can also be marked off, tool
Body ground, when the ratio of the height and the width of the minimum enclosed rectangle of the human body image in target image is more than the second preset value and little
When the first preset value, the posture of the target to be detected in target image is middle posture.So can be by target to be detected
Posture is divided into: the posture stood, middle posture, the posture lied down.
In addition, minimum enclosed rectangle here refers to the minimum rectangle of the area surrounding human body image.As shown in Fig. 2 figure
In rectangle abcd be human body image 201 minimum enclosed rectangle.In general, minimum enclosed rectangle here is orthogonal
Side be respectively parallel to vertically and horizontally, the height of minimum enclosed rectangle here refers to minimum enclosed rectangle perpendicular
The length on Nogata side upwards, the width of minimum enclosed rectangle refers to the length on minimum enclosed rectangle side in the horizontal direction.
In order to accurately determine minimum enclosed rectangle, described human body image in the described current target image of described determination
Minimum enclosed rectangle, comprising:
Set up rectangular coordinate system;
Described current target image is loaded in described rectangular coordinate system;
Determine the maximum transverse axis coordinate in described human body image and minimum transverse axis coordinate, and maximum ordinate of orthogonal axes and minimum
Ordinate of orthogonal axes;
According to described maximum transverse axis coordinate, described minimum transverse axis coordinate, described maximum ordinate of orthogonal axes and the described minimum longitudinal axis
Coordinate, determines described minimum enclosed rectangle;
The ratio of the described height and the width determining described minimum enclosed rectangle, comprising:
In described rectangular coordinate system, described minimum enclosed rectangle is made perpendicular to first length on the side of described transverse axis
For described height, using the second length perpendicular to the side of the described longitudinal axis for the described minimum enclosed rectangle as described width, by institute
State the ratio as described height and the width for the ratio of the first length and described second length.
In the present embodiment, current target image is loaded in rectangular coordinate system, is each point of current target image
Distribution coordinate, determine in the human body image of current target image institute a little in maximum transverse axis coordinate and minimum transverse axis seat
Mark, and maximum ordinate of orthogonal axes and minimum ordinate of orthogonal axes.Maximum transverse axis coordinate and minimum transverse axis coordinate, and maximum ordinate of orthogonal axes
Determine four points with minimum ordinate of orthogonal axes, the rectangle that this four points are constituted is exactly minimum enclosed rectangle.
For example, as shown in figure 3, setting up rectangular coordinate system, transverse axis is x-axis, and the longitudinal axis is y-axis.Current target image
Maximum transverse axis coordinate in human body image 301 is the x-axis coordinate of point a1, and the x-axis coordinate of point a1 is 4;In human body image 301
Little transverse axis coordinate is the x-axis coordinate of point b1, and the x-axis coordinate of point b1 is 2;Maximum ordinate of orthogonal axes in human body image 301 is point c1
Y-axis coordinate, the y-axis coordinate of point c1 is 6;Minimum ordinate of orthogonal axes in human body image 301 is the y-axis coordinate of point d1, the y of point d1
Axial coordinate is 1.The coordinate of the point a2 that maximum transverse axis coordinate is constituted with maximum ordinate of orthogonal axes is (4,6), minimum transverse axis coordinate and
The coordinate of the point b2 that big ordinate of orthogonal axes is constituted is (2,6), the coordinate of the point c2 that minimum transverse axis coordinate is constituted with minimum ordinate of orthogonal axes
The coordinate of the point d2 constituting with minimum ordinate of orthogonal axes for (2,1), maximum transverse axis coordinate is (4,1), and rectangle a2b2c2d2 is human body
Image 301 minimum enclosed rectangle.First length of the vertical side b2c2 with transverse axis of this minimum enclosed rectangle is 5, and this minimum is external
Rectangular vertical is 2 with second length of the side a2b2 of the longitudinal axis, and the ratio of the height and the width of minimum enclosed rectangle is 2.5.When
When one preset value is 2, the ratio of the height and the width of minimum enclosed rectangle is more than the first preset value, determines current target image
In the posture of target to be detected be the posture stood.
In order to be determined more accurately whether target to be detected falls down, in an embodiment of the present invention, also include:
Determine the relaxing area of described target to be detected;
After judging that described time span is less than or equal to described predetermined time period, in the described current mesh of described determination
Before target described to be detected in logo image is fallen down, also include:
Judge target described to be detected in described current target image whether in described relaxing area, if it is,
Determine that the target described to be detected in described current target image is not fallen down, otherwise, it determines in described current target image
Described target to be detected is fallen down.
Due to when target to be detected needs rest, the posture lied down also being taken to rest, and also may be used when lying down
Can quickly reach the posture lied down, it is, meeting time span to be less than or equal to predetermined time period, to be checked in order to judge
Survey the state that target is in resting and be in the state fallen down, whether can be come in relaxing area by target to be detected true
Fixed.Here relaxing area can be the regions such as sofa, bed.In the present embodiment, determine the relaxing area of target to be detected, tool
Body can be accomplished by: using target to be detected bed in the environment region as relaxing area, or,
Relaxing area is determined in the pickup area of target image.When identifying current target image, in identification current target image it is
Whether no have relaxing area, and judge target to be detected in relaxing area.Judging target to be detected whether in relaxing area
When middle, can be realized by the minimum enclosed rectangle of human body image, when the minimum enclosed rectangle of human body image is in rest area
When the ratio of the area of the part in domain and the gross area is more than or equal to preset ratio, determine target to be detected in relaxing area, no
Then, determine target to be detected not in relaxing area.Here preset ratio can be 0.5.
In order to provide help for the target to be detected fallen down in time, in an embodiment of the present invention, in described determination
After target described to be detected in described current target image is fallen down, also include: send the report that described target to be detected is fallen down
Alert signal.
The related personnel such as caregiver, the guardian of target to be detected, Jin Erji can be notified in time by alarm signal
When provide help for target to be detected.This alarm signal can be note, voice message, chimes of doom, blinking light etc..
As shown in figure 4, embodiments providing a kind of method that detection is fallen down, the method can include following walking
Rapid:
Step 401: previously according to division rule by the posture of target to be detected be divided into the posture stood, middle posture and
The posture lied down.
Here target to be detected can be people to be detected, for example: old man, child etc..
This division rule specifically includes:
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in target image is pre- more than or equal to first
If during value, the posture of the target to be detected in target image is the posture stood;
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in target image is pre- less than or equal to second
If during value, the posture of the target to be detected in target image is the posture lied down;
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in target image is more than the second preset value
And during less than the first preset value, the posture of the target to be detected in target image is middle posture;
Wherein, the first preset value is more than the second preset value, and for example: the first preset value is 2, the second preset value is 0.5.
Step 402: set up rectangular coordinate system.
Specifically, using x-axis as transverse axis, using y-axis as the longitudinal axis.
Step 403: determine the relaxing area of target to be detected.
For example, using the region of the bed in the pickup area of target image as relaxing area.
Step 404: Real-time Collection comprises the target image of target to be detected.
Here target image can be gathered by photographic head.
Step 405: identify the human body image of target to be detected from current target image.
In order to detect to target to be detected in time, need in time to each target image collecting
Reason.When identifying human body image, can be realized by background subtraction.
Step 406: current target image is loaded in rectangular coordinate system.
In this rectangular coordinate system, each pixel in current target image corresponds to a coordinate.
Step 407: determine the maximum transverse axis coordinate in the human body image of current target image and minimum transverse axis coordinate, and
Maximum ordinate of orthogonal axes and minimum ordinate of orthogonal axes.
For example, as shown in figure 5, the maximum transverse axis coordinate in the human body image 501 of current target image is the x of point c3
Axial coordinate, the x-axis coordinate of point c3 is 6;Minimum transverse axis coordinate in human body image 501 is the x-axis coordinate of point d3, the x-axis of point d3
Coordinate is 1;Maximum ordinate of orthogonal axes in human body image 501 is the y-axis coordinate of point b3, and the y-axis coordinate of point b3 is 4;Human body image
Minimum ordinate of orthogonal axes in 501 is the y-axis coordinate of point d3, and the y-axis coordinate of point d3 is 2.
Step 408: according to maximum transverse axis coordinate, minimum transverse axis coordinate, maximum ordinate of orthogonal axes and minimum ordinate of orthogonal axes, determine
Minimum enclosed rectangle.
, the coordinate of the point b4 that maximum transverse axis coordinate is constituted with maximum ordinate of orthogonal axes is taking the human body image 501 in Fig. 5 as a example
(6,4), the coordinate of the point d4 that minimum transverse axis coordinate is constituted with minimum ordinate of orthogonal axes is (1,2), and maximum transverse axis coordinate is indulged with minimum
The coordinate of the point a4 that axial coordinate is constituted is (6,2), the coordinate of the point c4 that minimum transverse axis coordinate is constituted with maximum ordinate of orthogonal axes for (1,
4), rectangle a4b4c4d4 is human body image 501 minimum enclosed rectangle.
Step 409: in rectangular coordinate system, using the first length perpendicular to the side of transverse axis for the minimum enclosed rectangle as height
Degree, using the second length perpendicular to the side of the longitudinal axis for the minimum enclosed rectangle as width, by the ratio of the first length and the second length
Value is as the ratio of height and the width.
, vertical the first length with the side a4b4 of transverse axis of this minimum enclosed rectangle is taking the human body image 501 in Fig. 5 as a example
2, second length of the vertical side a4d4 with the longitudinal axis of this minimum enclosed rectangle is 5, the ratio of the height and the width of minimum enclosed rectangle
It is worth for 0.4.
Step 410: according to the ratio of division rule and height and the width, determine target to be detected in current target image
Posture.
, when the second preset value is 0.5 taking the human body image 501 in Fig. 5 as a example, the height and the width of minimum enclosed rectangle
Ratio be less than the second preset value, according to division rule, determine that the posture of target to be detected in current target image is to lie down
Posture.
Step 411: whether the posture judging the target to be detected in current target image is the posture lied down, if it is,
Then step 412, otherwise, execution step 417.
Specifically, if the posture of target to be detected is the posture lied down in current target image, current target image
In target to be detected it may happen that falling down, but, if the posture of target to be detected is not the appearance lied down in current target image
Gesture, then in current target image, target to be detected does not just send and falls down, and can continue with next target image.
, determine target to be detected in current target image in step 410 taking the human body image 501 in Fig. 5 as a example
Posture is the posture lied down, and this step judges the appearance of target to be detected in current target image according to the result in step 410
Gesture is the posture lied down, then execute execution step 412.
Step 412: determine in the target image before current target image and connect most with current target image in time
Near target image of standing, wherein, target image of standing be target to be detected posture be the posture stood target image.
Due to coherent action when falling down, therefore, when judging whether target to be detected falls down, except judging to treat
The detection current posture of target is it is also contemplated that posture before this posture.Fall down and enter usually within the posture short time stood
To the posture lied down, therefore, in this step, before needing to find the posture lied down at this, the immediate posture stood.
When processing target image, it is to process sequentially in time, first process time of origin target image earlier, after
Process later target image.When processing each target image, target to be detected in each target image all can be determined
Posture.Therefore, this step can find immediate with current target image in time standing according to result before
Target image.
For example, it has been determined that going out the target to be detected in the corresponding target image of Fig. 3 before current target image
Posture be the posture stood.And, determine that the corresponding target image of Fig. 3 is closest with current target image in time
Target image of standing.
Step 413: determine current target image and stand target image immediate with current target image in time
Between time span.
For example, there is 6: 6: 30.52 in the afternoon, Fig. 3 corresponding target figure in the corresponding current target image of Fig. 5
As there are 6: 6: 30.12 in the afternoon.Time span between this two target images is 0.4 second.
Step 414: judge whether time span is less than or equal to predetermined time period, if it is, execution step 415, no
Then, execution step 417.
Usually it is rapidly introduced into, from the state stood, the state lied down due to falling down, the state stood and the state fallen down
Time interval very short, here by limiting this time interval, the value model of this predetermined time period for predetermined time period
Enclosing can be [0.5, the 1.5] second, can be 1.2 seconds, 1 second, 0.9 second, 0.8 second etc..
, between this two target images taking the corresponding current target image of Fig. 5 and the corresponding target image of standing of Fig. 3 as a example
Time span be 0.4 second.When predetermined time period is 0.8 second, the time span between this two target images is less than pre-
If time span, execution step 415.
If this time span is less than or equal to predetermined time period, the target to be detected in current target image may be fallen
Fall, continue through subsequent step to determine, if this time span is more than predetermined time period, target to be detected is probably certainly
Oneself lies down, and does not fall down, and processes next target image.
Step 415: judge target to be detected in current target image whether in relaxing area, if it is, execution
Step 417, otherwise, execution step 416.
Even if time span is less than or equal to predetermined time period, but it is also possible to target to be detected is quickly to lie down
Relaxing area is rested, in order to improve the accuracy of detection, in this step, judges target to be detected whether in rest
In region.
When whether the target to be detected in judging current target image is in relaxing area, can come in the following manner
Realize: whether the human body image judging in current target image is in relaxing area.
Taking the corresponding target to be detected of Fig. 5 as a example, if it is judged that the corresponding target to be detected of Fig. 5 is in relaxing area,
Then determine that target to be detected is not fallen down.
Step 416: determine that the target to be detected in current target image is fallen down, send the alarm signal that target to be detected is fallen down
Number.
In this step, after determining that the target to be detected in current target image is fallen down, send alarm signal, be to treat in time
Detection goal seeking helps.
Step 417: using next target image as current target image, return to step 405.
As shown in fig. 6, the device that a kind of detection provided in an embodiment of the present invention is fallen down, comprising:
Division unit 601, for dividing the posture of target to be detected, the posture of described target to be detected includes: stands
Posture, the posture lied down;
Collecting unit 602, comprises the target image of described target to be detected for Real-time Collection;
Recognition unit 603, for identifying the human body image of described target to be detected from current target image;
Posture determining unit 604, for according to the described human body image in described current target image, determining described to be checked
Survey posture in described current target image for the target;
Posture judging unit 605, for judge the target described to be detected in described current target image posture whether
For the posture lied down, if it is, notifying image determination unit 606;
Described image determining unit 606, for determining in the target image before described current target image in the time
Above with the immediate target image of standing of described current target image, wherein, described target image of standing is described mesh to be detected
Target posture is the target image of the posture stood;
Time determining unit 607, for determine described current target image with time with described current target image
The immediate time span stood between target image;
Fall down judging unit 608, for judging whether described time span is less than or equal to predetermined time period, if it is,
Then determine that the target described to be detected in described current target image is fallen down.
In an embodiment of the present invention, described division unit, for dividing described target to be detected according to division rule
Posture;
Wherein, described division rule includes:
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in described target image is more than or equal to the
During one preset value, the posture of the target described to be detected in described target image be described in the posture stood;
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in described target image is less than or equal to the
During two preset values, the posture of the target described to be detected in described target image be described in the posture lied down;
Wherein, described first preset value is more than described second preset value;
Described posture determining unit, for determining the external square of minimum of the described human body image in described current target image
Shape, determines the ratio of the height and the width of described minimum enclosed rectangle, according to described division rule and described height and the width
Ratio, determines posture in described current target image for the described target to be detected.
In an embodiment of the present invention, described posture determining unit is in executing the described current target image of described determination
During the minimum enclosed rectangle of described human body image, it is used for setting up rectangular coordinate system, described current target image is loaded into described
In rectangular coordinate system, determine the maximum transverse axis coordinate in described human body image and minimum transverse axis coordinate, and maximum ordinate of orthogonal axes
With minimum ordinate of orthogonal axes, according to described maximum transverse axis coordinate, described minimum transverse axis coordinate, described maximum ordinate of orthogonal axes and described
Little ordinate of orthogonal axes, determines described minimum enclosed rectangle;
Described posture determining unit, when executing the ratio of the described height and the width determining described minimum enclosed rectangle, is used
In in described rectangular coordinate system, using the first length perpendicular to the side of described transverse axis for the described minimum enclosed rectangle as described
Highly, using the second length perpendicular to the side of the described longitudinal axis for the described minimum enclosed rectangle as described width, by described first
The ratio of length and described second length is as the ratio of described height and the width.
In an embodiment of the present invention, this device also includes:
Area determination unit, for determining the relaxing area of described target to be detected;
Described fall down judging unit judge described time span be less than or equal to described predetermined time period after, in institute
State determine described current target image in target described to be detected fall down before, be additionally operable to judge in described current target image
Target described to be detected whether in described relaxing area, if it is, determine described current target image in described in treat
Detection target is not fallen down, otherwise, it determines the target described to be detected in described current target image is fallen down.
In an embodiment of the present invention, also include: alarm unit, for described fall down judging unit determine described current
After target described to be detected in target image is fallen down, send the alarm signal that described target to be detected is fallen down.
The device that a kind of detection provided in an embodiment of the present invention is fallen down, the product form of this device can have various ways,
For example: outward appearance is manufactured into the form of similar mirror, can realize as " makeups witch mirror ", this mirror be provided with photographic head and makees
For collecting unit, this mirror can be placed on target to be detected often in the pickup area of activity.
The contents such as the information exchange between each unit in said apparatus, implementation procedure, due to implementing with the inventive method
Example is based on same design, and particular content can be found in the narration in the inventive method embodiment, and here is omitted.
Each embodiment of the present invention at least has the advantages that
1st, in embodiments of the present invention, Real-time Collection comprises the target image of target to be detected, in each target image
Target to be detected carries out falling down inspection, extracts human body image from target image, determines human body image in target image
Posture, when the time span that this posture is the posture lied down and the posture lied down to this from the nearest posture stood is less than etc.
When predetermined time period, determine that target to be detected is fallen down, whole detection process need not manually participate in, can more easily examine
Survey is fallen down.
2nd, because, when standing and lie down, the ratio difference of the height and the width of human body is very big, implement in the present invention
The height under different gestures and the width of human body in example, is represented by the height of the minimum enclosed rectangle of human body image and width
Degree, can distinguish the posture of target to be detected exactly by the height of minimum enclosed rectangle and the ratio of width, and then, energy
Enough detect whether target to be detected falls down exactly.
3rd, in embodiments of the present invention, predefine out the relaxing area of target to be detected, determine target to be detected
After being in the posture lied down, then determine target to be detected whether in relaxing area, when in target to be detected in relaxing area,
Determine that target to be detected is not fallen down, it is to avoid target to be detected is rested in relaxing area and is the feelings fallen down by false judgment
Condition is so that testing result is more accurate.
4th, in embodiments of the present invention, after determining that target to be detected is fallen down, send the warning that target to be detected is fallen down
Signal, notifies the related personnel such as caregiver, the guardian of target to be detected in time by alarm signal, is mesh to be detected in time
Mark provides and helps it is ensured that the safety of personnel to be detected.
It should be noted that herein, such as first and second etc relational terms are used merely to an entity
Or operation is made a distinction with another entity or operation, and not necessarily requires or imply exist between these entities or operation
Any this actual relation or order.And, term " inclusion ", "comprising" or its any other variant are intended to non-
The comprising of exclusiveness, so that including a series of process of key elements, method, article or equipment not only include those key elements,
But also include other key elements being not expressly set out, or also include being consolidated by this process, method, article or equipment
Some key elements.In the absence of more restrictions, the key element being limited by sentence " including ", does not arrange
Remove and also there is other same factor in the process including described key element, method, article or equipment.
One of ordinary skill in the art will appreciate that: all or part of step realizing said method embodiment can be passed through
Completing, aforesaid program can be stored in the storage medium of embodied on computer readable the related hardware of programmed instruction, this program
Upon execution, execute the step including said method embodiment;And aforesaid storage medium includes: rom, ram, magnetic disc or light
Disk etc. is various can be with the medium of store program codes.
Finally it should be understood that the foregoing is only presently preferred embodiments of the present invention, it is merely to illustrate the skill of the present invention
Art scheme, is not intended to limit protection scope of the present invention.All any modifications made within the spirit and principles in the present invention,
Equivalent, improvement etc., are all contained in protection scope of the present invention.
Claims (10)
1. a kind of method that detection is fallen down is it is characterised in that include:
S0: divide the posture of target to be detected in advance, the posture of described target to be detected includes: the posture stood, the appearance lied down
Gesture;
S1: Real-time Collection comprises the target image of described target to be detected;
S2: identify the human body image of described target to be detected from current target image;
S3: according to the described human body image in described current target image, determine described target to be detected in described current goal
Posture in image;
S4: whether the posture judging the target described to be detected in described current target image is the posture lied down, if it is,
Execution step s5;
S5: determine closest with described current target image in time in the target image before described current target image
Target image of standing, wherein, described target image of standing be described target to be detected posture be the posture stood target
Image;
S6: determine described current target image with immediate with described current target image in time stand target image it
Between time span;
S7: judge whether described time span is less than or equal to predetermined time period, if it is, determining described current target image
In target described to be detected fall down.
2. method according to claim 1 it is characterised in that
Described s0, comprising: divide the posture of described target to be detected previously according to division rule;
Wherein, described division rule includes:
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in described target image is pre- more than or equal to first
If during value, the posture of the target described to be detected in described target image be described in the posture stood;
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in described target image is pre- less than or equal to second
If during value, the posture of the target described to be detected in described target image be described in the posture lied down;
Wherein, described first preset value is more than described second preset value;
Described s3, comprising:
Determine the minimum enclosed rectangle of the described human body image in described current target image;
Determine the ratio of the height and the width of described minimum enclosed rectangle;
According to the ratio of described division rule and described height and the width, determine described target to be detected in described current goal figure
Posture in picture.
3. method according to claim 2 it is characterised in that
The described minimum enclosed rectangle determining the described human body image in described current target image, comprising:
Set up rectangular coordinate system;
Described current target image is loaded in described rectangular coordinate system;
Determine the maximum transverse axis coordinate in described human body image and minimum transverse axis coordinate, and maximum ordinate of orthogonal axes and the minimum longitudinal axis
Coordinate;
According to described maximum transverse axis coordinate, described minimum transverse axis coordinate, described maximum ordinate of orthogonal axes and described minimum ordinate of orthogonal axes,
Determine described minimum enclosed rectangle;
The ratio of the described height and the width determining described minimum enclosed rectangle, comprising:
In described rectangular coordinate system, using the first length perpendicular to the side of described transverse axis for the described minimum enclosed rectangle as institute
State height, using the second length perpendicular to the side of the described longitudinal axis for the described minimum enclosed rectangle as described width, by described the
The ratio of one length and described second length is as the ratio of described height and the width.
4. method according to claim 1 is it is characterised in that also include:
Determine the relaxing area of described target to be detected;
After judging that described time span is less than or equal to described predetermined time period, in the described current goal figure of described determination
Before target described to be detected in picture is fallen down, also include:
Judge target described to be detected in described current target image whether in described relaxing area, if it is, determining
Target described to be detected in described current target image is not fallen down, otherwise, it determines described in described current target image
Target to be detected is fallen down.
5. according to described method arbitrary in claim 1-4 it is characterised in that
After the described target described to be detected determining in described current target image is fallen down, also include:
Send the alarm signal that described target to be detected is fallen down.
6. the device that a kind of detection is fallen down is it is characterised in that include:
Division unit, for dividing the posture of target to be detected, the posture of described target to be detected includes: the posture stood, lies
The posture fallen;
Collecting unit, comprises the target image of described target to be detected for Real-time Collection;
Recognition unit, for identifying the human body image of described target to be detected from current target image;
Posture determining unit, for according to the described human body image in described current target image, determining described target to be detected
Posture in described current target image;
Posture judging unit, whether the posture for judging the target described to be detected in described current target image lies down
Posture, if it is, notify image determination unit;
Described image determining unit, for determine in target image before described current target image in time with described
The immediate target image of standing of current target image, wherein, described target image of standing is the posture of described target to be detected
The target image of the posture for standing;
Time determining unit, for determine described current target image with immediate with described current target image in time
The time span stood between target image;
Fall down judging unit, for judging whether described time span is less than or equal to predetermined time period, if it is, determining institute
The target described to be detected stated in current target image is fallen down.
7. device according to claim 6 it is characterised in that
Described division unit, for dividing the posture of described target to be detected according to division rule;
Wherein, described division rule includes:
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in described target image is pre- more than or equal to first
If during value, the posture of the target described to be detected in described target image be described in the posture stood;
When the ratio of the height and the width of the minimum enclosed rectangle of the human body image in described target image is pre- less than or equal to second
If during value, the posture of the target described to be detected in described target image be described in the posture lied down;
Wherein, described first preset value is more than described second preset value;
Described posture determining unit, for determining the minimum enclosed rectangle of the described human body image in described current target image,
Determine the ratio of the height and the width of described minimum enclosed rectangle, according to the ratio of described division rule and described height and the width
Value, determines posture in described current target image for the described target to be detected.
8. device according to claim 7 it is characterised in that
The minimum of described human body image in executing the described current target image of described determination for the described posture determining unit is external
During rectangle, it is used for setting up rectangular coordinate system, described current target image is loaded in described rectangular coordinate system, determine described people
Maximum transverse axis coordinate in body image and minimum transverse axis coordinate, and maximum ordinate of orthogonal axes and minimum ordinate of orthogonal axes, according to described
Maximum transverse axis coordinate, described minimum transverse axis coordinate, described maximum ordinate of orthogonal axes and described minimum ordinate of orthogonal axes, determine described minimum
Boundary rectangle;
Described posture determining unit execute described determine described minimum enclosed rectangle height and the width ratio when, for
In described rectangular coordinate system, using the first length perpendicular to the side of described transverse axis for the described minimum enclosed rectangle as described height
Degree, using the second length perpendicular to the side of the described longitudinal axis for the described minimum enclosed rectangle as described width, long by described first
The ratio of degree and described second length is as the ratio of described height and the width.
9. device according to claim 6 is it is characterised in that also include:
Area determination unit, for determining the relaxing area of described target to be detected;
Described judging unit of falling down after judging that described time span is less than or equal to described predetermined time period, described really
Before target described to be detected in fixed described current target image is fallen down, it is additionally operable to judge the institute in described current target image
State target to be detected whether in described relaxing area, if it is, determining described to be detected in described current target image
Target is not fallen down, otherwise, it determines the target described to be detected in described current target image is fallen down.
10. according to described device arbitrary in claim 6-9 it is characterised in that also including:
In described judging unit of falling down, alarm unit, for determining that the target described to be detected in described current target image is fallen down
Afterwards, send the alarm signal that described target to be detected is fallen down.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610756018.3A CN106373336A (en) | 2016-08-30 | 2016-08-30 | Fall detection method and device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610756018.3A CN106373336A (en) | 2016-08-30 | 2016-08-30 | Fall detection method and device |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106373336A true CN106373336A (en) | 2017-02-01 |
Family
ID=57901064
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610756018.3A Pending CN106373336A (en) | 2016-08-30 | 2016-08-30 | Fall detection method and device |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106373336A (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107016350A (en) * | 2017-04-26 | 2017-08-04 | 中科唯实科技(北京)有限公司 | A kind of Falls Among Old People detection method based on depth camera |
CN107358784A (en) * | 2017-07-25 | 2017-11-17 | 深圳市盛路物联通讯技术有限公司 | A kind of long distance monitoring method and apparatus |
WO2018223812A1 (en) * | 2017-06-07 | 2018-12-13 | 京东方科技集团股份有限公司 | Smart toilet and safety monitoring system based on smart toilet |
CN110261867A (en) * | 2018-03-12 | 2019-09-20 | 松下知识产权经营株式会社 | Tracking device and tracking |
CN110633652A (en) * | 2019-08-27 | 2019-12-31 | 宇龙计算机通信科技(深圳)有限公司 | Unexpected situation determination method and device and electronic equipment |
CN111047827A (en) * | 2019-12-03 | 2020-04-21 | 北京深测科技有限公司 | Intelligent monitoring method and system for environment-assisted life |
CN112819855A (en) * | 2020-12-31 | 2021-05-18 | 广州紫川电子科技有限公司 | Personnel tumbling detection method, equipment and medium based on nursing home |
WO2021217848A1 (en) * | 2020-04-29 | 2021-11-04 | 平安科技(深圳)有限公司 | Human body falling detection method and apparatus, and computer device and storage medium |
CN113990037A (en) * | 2021-11-01 | 2022-01-28 | 珠海华发人居生活研究院有限公司 | Intelligent monitoring system for health and accidents of old people |
CN117116004A (en) * | 2023-10-25 | 2023-11-24 | 深圳市微克科技有限公司 | Tumble alarm method, system and medium of intelligent wearable device |
CN117953306A (en) * | 2024-02-23 | 2024-04-30 | 深圳职业技术大学 | Tumble detection method, tumble detection system, electronic equipment and medium |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN200941648Y (en) * | 2006-08-07 | 2007-08-29 | 扬州大学 | Cell phone having warning function when dumping short messages |
CN102387345A (en) * | 2011-09-09 | 2012-03-21 | 浙江工业大学 | Safety monitoring system based on omnidirectional vision for old people living alone |
US20120086824A1 (en) * | 2010-10-06 | 2012-04-12 | Hoya Corporation | Camera with image stabilization function |
CN103903281A (en) * | 2014-04-04 | 2014-07-02 | 西北工业大学 | Old people tumbling detecting method based on multi-feature analyzing and scene studying |
CN105303563A (en) * | 2015-09-22 | 2016-02-03 | 北京格灵深瞳信息技术有限公司 | Fall-down detection method and device |
CN105718886A (en) * | 2016-01-20 | 2016-06-29 | 武汉日创科技有限公司 | Moving personnel safety abnormity tumbling detection method |
CN105718857A (en) * | 2016-01-13 | 2016-06-29 | 兴唐通信科技有限公司 | Human body abnormal behavior detection method and system |
-
2016
- 2016-08-30 CN CN201610756018.3A patent/CN106373336A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN200941648Y (en) * | 2006-08-07 | 2007-08-29 | 扬州大学 | Cell phone having warning function when dumping short messages |
US20120086824A1 (en) * | 2010-10-06 | 2012-04-12 | Hoya Corporation | Camera with image stabilization function |
CN102387345A (en) * | 2011-09-09 | 2012-03-21 | 浙江工业大学 | Safety monitoring system based on omnidirectional vision for old people living alone |
CN103903281A (en) * | 2014-04-04 | 2014-07-02 | 西北工业大学 | Old people tumbling detecting method based on multi-feature analyzing and scene studying |
CN105303563A (en) * | 2015-09-22 | 2016-02-03 | 北京格灵深瞳信息技术有限公司 | Fall-down detection method and device |
CN105718857A (en) * | 2016-01-13 | 2016-06-29 | 兴唐通信科技有限公司 | Human body abnormal behavior detection method and system |
CN105718886A (en) * | 2016-01-20 | 2016-06-29 | 武汉日创科技有限公司 | Moving personnel safety abnormity tumbling detection method |
Non-Patent Citations (1)
Title |
---|
CHIEN-LIANG LIU等: ""A fall detection system using k-nearest neighbor classifer"", 《EXPERT SYSTEMS WITH APPLICATIONS》 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107016350A (en) * | 2017-04-26 | 2017-08-04 | 中科唯实科技(北京)有限公司 | A kind of Falls Among Old People detection method based on depth camera |
WO2018223812A1 (en) * | 2017-06-07 | 2018-12-13 | 京东方科技集团股份有限公司 | Smart toilet and safety monitoring system based on smart toilet |
US10692347B2 (en) | 2017-06-07 | 2020-06-23 | Boe Technology Group Co., Ltd. | Smart toilet and safety monitoring system based on smart toilet |
CN107358784A (en) * | 2017-07-25 | 2017-11-17 | 深圳市盛路物联通讯技术有限公司 | A kind of long distance monitoring method and apparatus |
CN110261867A (en) * | 2018-03-12 | 2019-09-20 | 松下知识产权经营株式会社 | Tracking device and tracking |
CN110633652A (en) * | 2019-08-27 | 2019-12-31 | 宇龙计算机通信科技(深圳)有限公司 | Unexpected situation determination method and device and electronic equipment |
CN111047827A (en) * | 2019-12-03 | 2020-04-21 | 北京深测科技有限公司 | Intelligent monitoring method and system for environment-assisted life |
WO2021217848A1 (en) * | 2020-04-29 | 2021-11-04 | 平安科技(深圳)有限公司 | Human body falling detection method and apparatus, and computer device and storage medium |
CN112819855A (en) * | 2020-12-31 | 2021-05-18 | 广州紫川电子科技有限公司 | Personnel tumbling detection method, equipment and medium based on nursing home |
CN113990037A (en) * | 2021-11-01 | 2022-01-28 | 珠海华发人居生活研究院有限公司 | Intelligent monitoring system for health and accidents of old people |
CN117116004A (en) * | 2023-10-25 | 2023-11-24 | 深圳市微克科技有限公司 | Tumble alarm method, system and medium of intelligent wearable device |
CN117953306A (en) * | 2024-02-23 | 2024-04-30 | 深圳职业技术大学 | Tumble detection method, tumble detection system, electronic equipment and medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106373336A (en) | Fall detection method and device | |
JP6534499B1 (en) | MONITORING DEVICE, MONITORING SYSTEM, AND MONITORING METHOD | |
WO2021227874A1 (en) | Falling behaviour detection method and device | |
JP6773829B2 (en) | Object recognition device, object recognition method, and object recognition program | |
CN105960663A (en) | Information processing device, information processing method, and program | |
TWI541769B (en) | Falling down detecting systems and method | |
Barabas et al. | Kinect-based platform for movement monitoring and fall-detection of elderly people | |
US20190370996A1 (en) | Image processing apparatus | |
CN110738077B (en) | Foreign matter detection method and device | |
CN110390329A (en) | A kind of terahertz image recognition methods and system based on deep learning | |
CN113384267A (en) | Fall real-time detection method, system, terminal equipment and storage medium | |
JP6822326B2 (en) | Watching support system and its control method | |
CN116758691B (en) | Fire-fighting early warning method and device based on multi-mode data | |
CN116645633A (en) | Method and system for detecting behavior of worker crossing fence based on computer vision | |
US11288817B2 (en) | Image processing device, stationary object tracking system, image processing method, and recording medium | |
CN112149527A (en) | Wearable device detection method and device, electronic device and storage medium | |
CN111191499A (en) | Fall detection method and device based on minimum center line | |
CN113780255B (en) | Danger assessment method, device, equipment and storage medium | |
CN116071696A (en) | Building stair congestion detection method and device based on YOLOv7 | |
CN115050102A (en) | Mobile equipment standard sitting posture analysis processing method and system and storage medium thereof | |
CN105550531B (en) | A kind of intelligent nursing management instrument | |
CN116069596A (en) | Situation awareness method, system, equipment and storage medium | |
Rao et al. | Real time face mask detection and thermal screening with audio response for covid-19 | |
CN113920544A (en) | Safety management system and method for stamping workshop and electronic equipment | |
JP6874838B2 (en) | Information processing equipment, control methods, and programs |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170201 |