CN103268468B - Automatic detection method for fastening of safety belts by front sitting persons on motor vehicle - Google Patents
Automatic detection method for fastening of safety belts by front sitting persons on motor vehicle Download PDFInfo
- Publication number
- CN103268468B CN103268468B CN201210233380.4A CN201210233380A CN103268468B CN 103268468 B CN103268468 B CN 103268468B CN 201210233380 A CN201210233380 A CN 201210233380A CN 103268468 B CN103268468 B CN 103268468B
- Authority
- CN
- China
- Prior art keywords
- face
- window
- area
- image
- license plate
- 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
Abstract
The invention discloses an automatic detection method for the fastening of safety belts by front sitting persons on a motor vehicle. The automatic detection method comprises the following steps of detecting faces by selecting a road monitoring picture based on haar-like, and identifying whether the front sitting person is a driver or a passenger; and respectively performing safety belt identification on the driver and the passenger. According to the automatic detection method, the front sitting persons on the motor vehicle can be quickly and accurately identified, and the fastening of the safety belts by the front sitting persons can be detected respectively; and the automatic detection method can be conveniently applied to the real life.
Description
Technical field
The present invention relates to traffic safety technology field, particularly to a kind of motor vehicles front row occupant whether cingulum seat belt
Automatic testing method.
Background technology
At present, in the middle of the image photographing, by human resourcess, communication differentiates that whether passenger, driver are a sheet by a sheet
Upper seat belt.However, because amount of images is huge, identification in this respect needs substantial amounts of human and material resources.Therefore, using machine
Device replaces the mankind, improves efficiency, increases productivity, exactly the meaning of this software.
Existing main flow seat belt recognition methodss are by positioning vehicle license plate;Then estimate the position of main driver;Again
Go the presence of detection seat belt using Hough straight line.First, this method can not find out the position of driver exactly, accurately
Degree is not high;And, this method is only capable of finding out the seat belt that master drives seat, and the seat belt of None- identified front passenger's seat.
Content of the invention
The goal of the invention of the present invention is the technical deficiency for existing seat belt automatic testing method, provides one kind accurate
The really automatic testing method of detection motor vehicles front row occupant whether cingulum seat belt.
For achieving the above object, the technical solution used in the present invention is:
There is provided a kind of automatic testing method of motor vehicles front row occupant whether cingulum seat belt, based on road monitoring figure
Picture, comprises the steps:
A () reading format is the monitoring image of JPG, BMP or PNG;
B (), according to hsv color model, carries out blue region and yellow area to the image file reading in step (a)
Colouring information is analyzed and is combined the structure characteristic analysis such as the polygonal shape of car plate, the ratio of width to height, then the position of positioning licence plate;
(c) according to the image that processed in the car plate position of positioning in step (b) and step (b), by window locations with
The dimensional parameters of the relative geometrical relation of car plate position and input picture are choosing local window, and the vehicle window position as positioning
Put;
D (), according to the result of vehicle window positioning in step (c), carries out Face datection based on haar-like in the range of vehicle window,
And judge that the face detecting is driver or passenger;
E (), according to the judged result in step (d), carries out window division to the position of driver and passenger, and to window first
Afterwards using canny rim detection and Hough straight-line detection, then screening specified angle in the range of straight line and labelling driver with take advantage of
The region of visitor, is finally made whether the differentiation of wear safety belt respectively.
Preferably, comprise the steps in described step (b):
(b-1) monitoring image is changed into hsv color model by equation below from RGB:
V=max (r, g, b) (3);
Wherein h represents tone, and s represents saturation, and v represents lightness, and max represents the maximum taking wherein parameter, and min represents
Take the minima of wherein parameter.
(b-2) following one of which steps will be executed through after (b-1) conversion to image:
1) according to blue color value standard:
Traversing graph is as each pixel, blue region in labelling image;
2) the color value standard according to yellow:
Traversing graph is as each pixel, yellow area in labelling image;
(b-3) region that (b-2) is drawn, first carries out 1 corrosion treatmentCorrosion Science in image procossing, at least carries out 4 expansions
Process;
(b-4) region being drawn using step (b-3), is approached with the precision that area circumference is multiplied by 0.04, produces polygon
Curve;
(b-5) set license plate area as R;Candidate license plate peak width is length, candidate license plate region height is height,
Picture traverse is s.length, and picture altitude is s.height;Candidate license plate region area is area, and image area is
s.area;Candidate license plate regional edge number is sides;The side number according to Polygonal Curves and area from step (b-4), and root
According to equation below:
s.area*0.00176<area<s.area*0.00502 (11)
4<sides<10 (12)
Find out license plate area;
(b-6) if from the result that (b-5) obtains, can not find license plate area, then return (b-2), execute wherein another step
Suddenly;If according to above-mentioned 1) and 2) step all can not find license plate area, and this image is as cannot find license plate image INo car plateAndIf finding license plate area, this license plate area is designated as R=R*.
Preferably, described step (c) comprises the following steps:
(c-1) ifObtain the square m of license plate area10, m00, m01;And then draw the center of license plate area ROtherwise execution step (b-3);
(c-2a) set the width of input picture as src.length, the height of input picture is src.height;Set a trap portion
The width of window is winL, and the height of local window is winH;The x being obtained using (c-1)cRAnd ycRIf, in license plate area R
Heart point horizontal coordinate is centre.x, makes centre.x=xcR, vertical coordinate is centre.y, makes centre.y=ycR;Its
In, x is this local window upper left corner horizontal coordinate in the input image, y be this local window upper left corner in the input image
Vertical coordinate;By equation below:
WinL=src.length*0.43 (14)
WinH=src.height*0.33 (15)
X=centre.x-0.215*src.length (16)
Y=centre.y-0.45*src.height (17);
Find out local window;
(c-2b) portion's window size of setting a trap is the size of input picture, that is,
WinL=src.length
WinH=src.height
X=0
Y=0.
Step (c) terminates, and enters step (d).
Preferably, described step (d) comprises the following steps:
(d-1), in the window obtained in step c, set using the increase income Haar-like face classification device in storehouse of OpenCV
Determine the parameter of minimum window;Wherein, if minimum retrieval window is 10 × 10;
(d-2) set the implementing result as (d-1) for the faces, ifDirectly execution step (d-3);Otherwise,
According to following standard screening face, if head portrait face ∈ faces, head portrait radius is face.radius, the level of head portrait central point
Coordinate is face.x, and vertical coordinate is face.y, and the horizontal coordinate of car plate central point is centre.x, and vertical coordinate is
Centre.y, wherein:
face.radius<0.05*s.length (18)
height<|face.x-xcR|<height*3 (19);
If the face being unsatisfactory for above-mentioned condition is faces-, reject from faces;And now faces is updated to faces=
faces-faces-;Then execution step (c-3);
(d-3) ifThen this image is the image of None- identified;Algorithm flow terminates;Otherwise, carry out as follows
Judge:
Condition 1:If the Horizontal component values of this face are more than the Horizontal component values of car plate, this face is belonging to driver,
Otherwise belong to front passenger;
Or condition 2:If the right region D of this face has vehicle window frame then it is assumed that face is belonging to driver's, otherwise belong to
Front passenger;
The window width of setting regions D is vWinL, and the window height of region D is vWinH;The window upper left corner top of region D
The horizontal component of point coordinates is vWinX, and vertical component is head portrait coordinate vertical component is vWinY;Then pass through equation below:
VWinL=face.radius*6 (22)
VWinH=face.radius*4 (23)
VWinX=face.x+face.radius (24)
VWinY=face.y-face.radius (25);
Indicate region D;
(d-4) result according to (d-3), passenger's window is designated as P(face), driver's window is designated as D(face).
Preferably, described step (e) comprises the following steps:
(e-1) R is designated as the local window region below face1, the width of this window area is sWinL, this window
The height in region is sWinH, and horizontal component size in original image for this window area upper left corner is sWinX, this window upper left
Vertical component size in original image for the angle is sWinY;
SWinL=face.radius*6 (26)
SWinH=face.radius*4 (27)
SWinX=face.x-face.radius*2 (28)
SWinY=face.y (29);
Carry out following operation in this window area:Carry out histogram equalization processing using the increase income function in storehouse of OpenCV,
Gaussian smoothing is carried out for 17 × 17 with core;Then using canny rim detection, wherein in the parameter of canny rim detection,
High threshold value is 60, and low threshold value is 0;Carry out Hough straight-line detection again, in the parameter of Hough straight-line detection, line of shortest length section threshold value
For 5, accumulative total threshold value of voting is 35.Obtain window area R1In straight line, be denoted as lines;
(e-2) result according to (e-1), according to following rule, screens to straight line lines:
1) for driver window D(face):By driver window D(face)In with x-axis positive direction angle not 40 degree with 70 degree it
Between straight line delete, be designated as driver region;
2) for passenger window P(face), passenger window P(face)In with x-axis positive direction angle not 110 degree with 140 degree
Between straight line delete, be designated as passenger area;
According to above-mentioned 1) and 2) result, if find straight line in driver and passenger area simultaneously, or look in driver region
To straight line and passenger do not exist when, then this monitoring image is the legal image passing through;Otherwise, it is violation image.
The present invention, with respect to prior art, has the advantages that:
1st, the present invention can be carried out fast and accurately to motor vehicles front row passenger (including driver and copilot passenger)
Identification, detects their whether cingulum seat belt respectively, if the examination meeting real traffic method defined requires;This automatic detection
Method can conveniently be applied among actual life;
2nd, the present invention can process large nuber of images, can be to the road monitoring image under different kinds of roads, various weather conditions
Carry out correlation analysiss;
3rd, the present invention utilizes the position of face to detect the position of seat belt, more accurate than additive method, and then improves safety
The accuracy of identification of band;
4th, the present invention can accurately distinguish front-seat passenger is passenger or driver, and accordingly makes identifying processing.
Brief description
Fig. 1 is the automatic testing method of motor vehicles of the present invention front row occupant whether cingulum seat belt.
Specific embodiment
With specific embodiment, the goal of the invention of the present invention is described in further detail below in conjunction with the accompanying drawings, embodiment is not
Here can repeat one by one, but embodiments of the present invention are not therefore defined in following examples.Unless stated otherwise, the present invention
Using material and processing method be the art conventional material and processing method.
As shown in figure 1, a kind of automatic testing method of motor vehicles front row occupant whether cingulum seat belt, based on road prison
Control image, comprises the steps:
A () reading format is the monitoring image of JPG, BMP or PNG;
B (), according to hsv color model, carries out blue region and yellow area to the image file reading in step (a)
Colouring information is analyzed and is combined the structure characteristic analysis such as the polygonal shape of car plate, the ratio of width to height, then the position of positioning licence plate;
(c) according to the image that processed in the car plate position of positioning in step (b) and step (b), by window locations with
The dimensional parameters of the relative geometrical relation of car plate position and input picture are choosing local window, and the vehicle window position as positioning
Put;
D (), according to the result of vehicle window positioning in step (c), carries out Face datection based on haar-like in the range of vehicle window,
And judge that the face detecting is driver or passenger;
E (), according to the judged result in step (d), carries out window division to the position of driver and passenger, and to window first
Afterwards using canny rim detection and Hough straight-line detection, then screening specified angle in the range of straight line and labelling driver with take advantage of
The region of visitor, is finally made whether the differentiation of wear safety belt respectively.
Comprise the steps in above-mentioned steps (b):
(b-1) monitoring image is changed into hsv color model by equation below from RGB:
V=max (r, g, b) (3);
Wherein h represents tone, and s represents saturation, and v represents lightness, and max represents the maximum taking wherein parameter, and min represents
Take the minima of wherein parameter.
(b-2) following one of which steps will be executed through after (b-1) conversion to image:
1) according to blue color value standard:
Traversing graph is as each pixel, blue region in labelling image;
2) the color value standard according to yellow:
Traversing graph is as each pixel, yellow area in labelling image;
(b-3) region that (b-2) is drawn, first carries out 1 corrosion treatmentCorrosion Science in image procossing, at least carries out 4 expansions
Process;
(b-4) region being drawn using step (b-3), is approached with the precision that area circumference is multiplied by 0.04, produces polygon
Curve;
(b-5) set license plate area as R;Candidate license plate peak width is length, candidate license plate region height is height,
Picture traverse is s.length, and picture altitude is s.height;Candidate license plate region area is area, and image area is
s.area;Candidate license plate regional edge number is sides;The side number according to Polygonal Curves and area from step (b-4), and root
According to equation below:
s.area*0.00176<area<s.area*0.00502 (11)
4<sides<10 (12)
Find out license plate area;
(b-6) if from the result that (b-5) obtains, can not find license plate area, then return (b-2), execute wherein another step
Suddenly;If according to above-mentioned 1) and 2) step all can not find license plate area, and this image is as cannot find license plate image INo car plateAndIf finding license plate area, this license plate area is designated as R=R*.
Above-mentioned steps (c) comprise the following steps:
(c-1) ifObtain the square m of license plate area10, m00, m01;And then draw the center of license plate area ROtherwise execution step (b-3);
(c-2a) set the width of input picture as src.length, the height of input picture is src.height;Set a trap portion
The width of window is winL, and the height of local window is winH;The x being obtained using (c-1)cRAnd ycRIf, in license plate area R
Heart point horizontal coordinate is centre.x, makes centre.x=xcR, vertical coordinate is centre.y, makes centre.y=ycR;Its
In, x is this local window upper left corner horizontal coordinate in the input image, y be this local window upper left corner in the input image
Vertical coordinate;By equation below:
WinL=src.length*0.43 (14)
WinH=src.height*0.33 (15)
X=centre.x-0.215*src.length (16)
Y=centre.y-0.45*src.height (17);
Find out local window;
(c-2b) portion's window size of setting a trap is the size of input picture, that is,
WinL=src.length
WinH=src.height
X=0
Y=0.
Step (c) terminates, and enters step (d).
Above-mentioned steps (d) comprise the following steps:
(d-1), in the window obtained in step c, set using the increase income Haar-like face classification device in storehouse of OpenCV
Determine the parameter of minimum window;Wherein, if minimum retrieval window is 10 × 10;
(d-2) set the implementing result as (d-1) for the faces, ifDirectly execution step (d-3);Otherwise,
According to following standard screening face, if head portrait face ∈ faces, head portrait radius is face.radius, the level of head portrait central point
Coordinate is face.x, and vertical coordinate is face.y, and the horizontal coordinate of car plate central point is centre.x, and vertical coordinate is
Centre.y, wherein:
face.radius<0.05*s.length (18)
height<|face.x-xcR|<height*3 (19);
If the face being unsatisfactory for above-mentioned condition is faces-, reject from faces;And now faces is updated to faces=
faces-faces-;Then execution step (c-3);
(d-3) ifThen this image is the image of None- identified;Algorithm flow terminates;Otherwise, carry out as follows
Judge, wherein:
Condition 1:
If the Horizontal component values of this face are more than the Horizontal component values of car plate, this face is belonging to driver's;Otherwise
Belong to front passenger;
Or condition 2:
The window width of setting regions D is vWinL, and the window height of region D is vWinH;The window upper left corner top of region D
The horizontal component of point coordinates be vWinX, vertical component be head portrait coordinate vertical component be vWinY;Then pass through equation below:
VWinL=face.radius*6 (22)
VWinH=face.radius*4 (23)
VWinX=face.x+face.radius (24)
VWinY=face.y-face.radius (25);Indicate region D;
(d-4) result according to (d-3), passenger's window is designated as P(face), driver's window is designated as D(face).
Above-mentioned steps (e) comprise the following steps:
(e-1) R is designated as the local window region below face1, the width of this window area is sWinL, this window
The height in region is sWinH, and horizontal component size in original image for this window area upper left corner is sWinX, this window upper left
Vertical component size in original image for the angle is sWinY;
SWinL=face.radius*6 (26)
SWinH=face.radius*4 (27)
SWinX=face.x-face.radius*2 (28)
SWinY=face.y (29);
Carry out following operation in this window area:Carry out histogram equalization processing using the increase income function in storehouse of OpenCV,
Gaussian smoothing is carried out for 17 × 17 with core;Then using canny rim detection, wherein in the parameter of canny rim detection,
High threshold value is 60, and low threshold value is 0;Carry out Hough straight-line detection again, in the parameter of Hough straight-line detection, line of shortest length section threshold value
For 5, accumulative total threshold value of voting is 35;Through above 4 steps, obtain window area R1In straight line, be denoted as lines;
(e-2) result according to (e-1), according to following rule, screens to straight line lines:
1) for driver window D(face):By driver window D(face)In with x-axis positive direction angle not 40 degree with 70 degree it
Between straight line delete, be designated as driver region;
2) for passenger window P(face), passenger window P(face)In with x-axis positive direction angle not 110 degree with 140 degree
Between straight line delete, be designated as passenger area;
According to above-mentioned 1) and 2) result, if find straight line in driver and passenger area simultaneously, or look in driver region
To straight line and passenger do not exist when, then this monitoring image is the legal image passing through;Otherwise, it is violation image.
I.e.:
Wherein, { l1|l1∈lines,l1And x-axis positive direction angle is not between 110 degree and 140 degree };
{l2|l2∈lines,l2And x-axis positive direction angle is not between 40 degree and 70 degree };
If meeting following condition then it is assumed that this monitoring image file is the legal image passing through:
Condition 1:IfDriver, passenger have cingulum seat belt;
Or condition 2:IfDriver has
Cingulum seat belt, and front passenger do not exist;
Above-described embodiment is only presently preferred embodiments of the present invention, is not used for limiting the practical range of the present invention.I.e. all according to
Impartial change and modification that present invention is made, are all covered by the claims in the present invention scope required for protection.
Claims (5)
1. the automatic testing method of a kind of motor vehicles front row occupant whether cingulum seat belt, based on road monitoring image, it is special
Levy and be to comprise the steps:
A () reading format is the monitoring image of JPG, BMP or PNG;
B (), according to hsv color model, carries out the color of blue region and yellow area to the image file reading in step (a)
Information analysiss simultaneously combine the polygonal shape of car plate, aspect ratio structures feature analysiss, the then position of positioning licence plate;
(c) according to the image that processed in the car plate position of positioning in step (b) and step (b), by window locations and car plate
The dimensional parameters of the relative geometrical relation of position and input picture are choosing local window, and the window locations as positioning;
D (), according to the result of vehicle window positioning in step (c), carries out Face datection based on haar-like in the range of vehicle window, and sentences
The face of disconnected detection is driver or passenger;
E (), according to the judged result in step (d), carries out window division to the position of driver and passenger, and window priority is made
With canny rim detection and Hough straight-line detection, the then straight line in the range of screening specified angle labelling driver is with passenger's
Region, is finally made whether the differentiation of wear safety belt respectively.
2. the automatic testing method of motor vehicles according to claim 1 front row occupant whether cingulum seat belt, its feature
It is:Comprise the steps in described step (b):
(b-1) monitoring image is changed into hsv color model by equation below from RGB:
V=max (r, g, b) (3);
Wherein h represents tone, and s represents saturation, and v represents lightness, and max represents and takes wherein r, the maximum of g, b parameter, min table
Show and take wherein r, the minima of g, b parameter;
(b-2) following one of which steps will be executed through the image after (b-1) conversion:
1) according to blue color value standard:
Traversing graph is as each pixel, blue region in labelling image;
2) the color value standard according to yellow:
Traversing graph is as each pixel, yellow area in labelling image;
(b-3) region that (b-2) is drawn, first carries out 1 corrosion treatmentCorrosion Science in image procossing, at least carries out 4 expansion process;
(b-4) region being drawn using step (b-3), is approached with the precision that area circumference is multiplied by 0.04, produces Polygonal Curves;
(b-5) set license plate area as R;Candidate license plate peak width is length, candidate license plate region height is height, image
Width is s.length, and picture altitude is s.height;Candidate license plate region area is area, and image area is s.area;Wait
License plate area side number is selected to be sides;The side number according to Polygonal Curves and area from step (b-4), and according to following public affairs
Formula:
s.area*0.00176<area<s.area*0.00502 (11)
4<sides<10 (12)
Find out license plate area;
(b-6) if from the result that (b-5) obtains, can not find license plate area, then return (b-2), execute wherein another step;As
Fruit is according to above-mentioned 1) and 2) step all can not find license plate area, and this image is as cannot find license plate image INo car plateAnd
If finding license plate area, this license plate area is designated as R=R*.
3. the automatic testing method of motor vehicles according to claim 1 front row occupant whether cingulum seat belt, its feature
It is:Described step (c) comprises the following steps:
(c-1) ifObtain the square m of license plate area10, m00, m01;And then draw the center of license plate area ROtherwise execution step (b-3);
(c-2a) set the width of input picture as src.length, the height of input picture is src.height;If local window
Width be winL, the height of local window is winH;The x being obtained using (c-1)cRAnd ycRIf, the central point of license plate area R
Horizontal coordinate is centre.x, makes centre.x=xcR, vertical coordinate is centre.y, makes centre.y=ycR;Wherein, x is
This local window upper left corner horizontal coordinate in the input image, y are this local window upper left corner in the input image vertical
Coordinate;By equation below:
WinL=src.length*0.43 (14)
WinH=src.height*0.33 (15)
X=centre.x-0.215*src.length (16)
Y=centre.y-0.45*src.height (17);
Find out local window;
(c-2b) portion's window size of setting a trap is the size of input picture, that is,
WinL=src.length
WinH=src.height
X=0
Y=0;
Step (c) terminates, and enters step (d).
4. the automatic testing method of motor vehicles front row occupant whether cingulum seat belt according to claim 3, its feature exists
In:Described step (d) comprises the following steps:
(d-1) in the window obtained in step c, using OpenCV increase income storehouse Haar-like face classification device set
The parameter of wicket;Wherein, if minimum retrieval window is 10 × 10;
(d-2) set the implementing result as (d-1) for the faces, ifDirectly execution step (d-3);Otherwise, according to
Lower standard screening face, if head portrait face ∈ faces, head portrait radius is face.radius, and the horizontal coordinate of head portrait central point is
Face.x, vertical coordinate is face.y, and the horizontal coordinate of car plate central point is centre.x, and vertical coordinate is centre.y, its
In:
face.radius<0.05*s.length (18)
height<|face.x-xcR| < height*3 (19);
If being unsatisfactory for formula (18) and the face of (19) condition is faces-, reject from faces;And now faces is updated to
Faces=faces-faces-;Then execution step (d-3);
(d-3) ifThen this image is the image of None- identified;Algorithm flow terminates;Otherwise, sentenced as follows
Disconnected:
Condition 1:If the Horizontal component values of this face are more than the Horizontal component values of car plate, this face is belonging to driver, otherwise
Belong to front passenger;
Or condition 2:If the right region D of this face has vehicle window frame then it is assumed that face is belonging to driver's, otherwise belong to front-seat
Passenger;
The window width of setting regions D is vWinL, and the window height of region D is vWinH;The window top left corner apex of region D is sat
Target horizontal component is vWinX, and vertical component is head portrait coordinate vertical component is vWinY;Then pass through equation below:
VWinL=face.radius*6 (22)
VWinH=face.radius*4 (23)
VWinX=face.x+face.radius (24)
VWinY=face.y-face.radius (25);
Indicate region D;
(d-4) result according to (d-3), passenger's window is designated as P(face), driver's window is designated as D(face).
5. the automatic testing method of motor vehicles according to claim 4 front row occupant whether cingulum seat belt, its feature
It is:Described step (e) comprises the following steps:
(e-1) R is designated as the local window region below face1, the width of this window area is sWinL, this window area
It is highly sWinH, horizontal component size in original image for this window area upper left corner is sWinX, and this window upper left corner is former
Vertical component size in image is sWinY;
SWinL=face.radius*6 (26)
SWinH=face.radius*4 (27)
SWinX=face.x-face.radius*2 (28)
SWinY=face.y (29);
Carry out following operation in this window area:Carry out histogram equalization processing using the increase income function in storehouse of OpenCV, with core
Carry out Gaussian smoothing for 17 × 17;Then using canny rim detection, wherein in the parameter of canny rim detection, high
Threshold value is 60, and low threshold value is 0;Carry out Hough straight-line detection again, in the parameter of Hough straight-line detection, line of shortest length section threshold value is 5,
Ballot accumulative total threshold value is 35;Obtain window area R1In straight line, be denoted as lines;
(e-2) result according to (e-1), according to following rule, screens to straight line lines:
1) for driver window D(face), by driver window D(face)In and x-axis positive direction angle not between 40 degree and 70 degree
Straight line is deleted, and is designated as driver region;
2) for passenger window P(face), passenger window P(face)In and x-axis positive direction angle not between 110 degree and 140 degree
Straight line delete, be designated as passenger area;
According to above-mentioned 1) and 2) result, if find straight line in driver and passenger area simultaneously, or find directly in driver region
When line and passenger do not exist, then this monitoring image is the legal image passing through;Otherwise, it is violation image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210233380.4A CN103268468B (en) | 2012-07-06 | 2012-07-06 | Automatic detection method for fastening of safety belts by front sitting persons on motor vehicle |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201210233380.4A CN103268468B (en) | 2012-07-06 | 2012-07-06 | Automatic detection method for fastening of safety belts by front sitting persons on motor vehicle |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103268468A CN103268468A (en) | 2013-08-28 |
CN103268468B true CN103268468B (en) | 2017-02-22 |
Family
ID=49012096
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201210233380.4A Active CN103268468B (en) | 2012-07-06 | 2012-07-06 | Automatic detection method for fastening of safety belts by front sitting persons on motor vehicle |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103268468B (en) |
Families Citing this family (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103522982B (en) * | 2013-10-25 | 2016-09-07 | 公安部第三研究所 | Bikini vehicle safety belt detection method based on graphical analysis and device |
CN103955704B (en) * | 2014-04-26 | 2018-08-24 | 合肥工业大学 | A kind of Safe belt detection method based on Adaboost |
CN104376561B (en) * | 2014-11-19 | 2017-04-12 | 青岛海信网络科技股份有限公司 | Safety belt detection method |
CN105809099B (en) * | 2014-12-31 | 2019-12-13 | 成都理想境界科技有限公司 | Safety belt detection method based on monitoring image |
CN105809171B (en) * | 2014-12-31 | 2019-12-13 | 成都理想境界科技有限公司 | Safety belt detection method based on monitoring image |
CN104657752B (en) * | 2015-03-17 | 2018-09-07 | 银江股份有限公司 | A kind of seatbelt wearing recognition methods based on deep learning |
CN104966066A (en) * | 2015-06-26 | 2015-10-07 | 武汉大学 | Traffic block port monitoring oriented in-car human face detection method and system |
CN106570444B (en) * | 2015-10-10 | 2019-04-12 | 腾讯科技(深圳)有限公司 | The vehicle intelligent based reminding method and system of Behavior-based control identification |
CN105373783A (en) * | 2015-11-17 | 2016-03-02 | 高新兴科技集团股份有限公司 | Seat belt not-wearing detection method based on mixed multi-scale deformable component model |
CN106503673A (en) * | 2016-11-03 | 2017-03-15 | 北京文安智能技术股份有限公司 | A kind of recognition methodss of traffic driving behavior, device and a kind of video acquisition device |
CN107766802B (en) * | 2017-09-29 | 2020-04-28 | 广州大学 | Self-adaptive detection method for unbuckled safety belt of front-row driver and passenger of motor vehicle |
CN107944341A (en) * | 2017-10-27 | 2018-04-20 | 荆门程远电子科技有限公司 | Driver based on traffic monitoring image does not fasten the safety belt automatic checkout system |
US20220083786A1 (en) * | 2019-01-17 | 2022-03-17 | Jungo Connectivity Ltd. | Method and system for monitoring a person using infrared and visible light |
CN110443901A (en) * | 2019-06-26 | 2019-11-12 | 成都信息工程大学 | In conjunction with the ticket checking method, device and storage medium of multiple view |
CN112132040B (en) * | 2020-09-24 | 2024-03-15 | 明见(厦门)软件开发有限公司 | Vision-based safety belt real-time monitoring method, terminal equipment and storage medium |
CN113870555B (en) * | 2021-09-08 | 2023-11-03 | 南京静态交通产业技术研究院 | Man-car collaborative identification method based on mobile phone IMSI code and electronic license plate |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101196980A (en) * | 2006-12-25 | 2008-06-11 | 四川川大智胜软件股份有限公司 | Method for accurately recognizing high speed mobile vehicle mark based on video |
CN101398894A (en) * | 2008-06-17 | 2009-04-01 | 浙江师范大学 | Automobile license plate automatic recognition method and implementing device thereof |
CN102521986A (en) * | 2011-12-05 | 2012-06-27 | 沈阳聚德视频技术有限公司 | Automatic detection system for fake plate vehicle and control method for automatic detection system |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20070195990A1 (en) * | 2006-02-16 | 2007-08-23 | Uri Levy | Vision-Based Seat Belt Detection System |
-
2012
- 2012-07-06 CN CN201210233380.4A patent/CN103268468B/en active Active
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101196980A (en) * | 2006-12-25 | 2008-06-11 | 四川川大智胜软件股份有限公司 | Method for accurately recognizing high speed mobile vehicle mark based on video |
CN101398894A (en) * | 2008-06-17 | 2009-04-01 | 浙江师范大学 | Automobile license plate automatic recognition method and implementing device thereof |
CN102521986A (en) * | 2011-12-05 | 2012-06-27 | 沈阳聚德视频技术有限公司 | Automatic detection system for fake plate vehicle and control method for automatic detection system |
Non-Patent Citations (2)
Title |
---|
image-based seat belt detection;huiwen guo et al;《2011 IEEE international conference on vehicular electronics and safety》;20110710;第161-163页 * |
occupant detection through near-infrared imaging;xiaoli hao et al;《tamkang journal of science and engineering》;20111231;第14卷(第3期);第277、279-281页 * |
Also Published As
Publication number | Publication date |
---|---|
CN103268468A (en) | 2013-08-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103268468B (en) | Automatic detection method for fastening of safety belts by front sitting persons on motor vehicle | |
CN101334836B (en) | License plate positioning method incorporating color, size and texture characteristic | |
CN100463000C (en) | Human eye state detection method based on cascade classification and hough circle transform | |
CN103824081B (en) | Method for detecting rapid robustness traffic signs on outdoor bad illumination condition | |
US9633267B2 (en) | Robust windshield detection via landmark localization | |
CN105809184B (en) | Method for real-time vehicle identification and tracking and parking space occupation judgment suitable for gas station | |
CN105373779A (en) | Intelligent vehicle safety belt detection method and intelligent detection system | |
CN105488453A (en) | Detection identification method of no-seat-belt-fastening behavior of driver based on image processing | |
CN104778444A (en) | Method for analyzing apparent characteristic of vehicle image in road scene | |
CN102880863B (en) | Method for positioning license number and face of driver on basis of deformable part model | |
CN103679168A (en) | Detection method and detection device for character region | |
CN103903018A (en) | Method and system for positioning license plate in complex scene | |
CN112712057B (en) | Traffic signal identification method and device, electronic equipment and storage medium | |
CN103455820A (en) | Method and system for detecting and tracking vehicle based on machine vision technology | |
CN105447503A (en) | Sparse-representation-LBP-and-HOG-integration-based pedestrian detection method | |
CN102999749A (en) | Intelligent safety belt regulation violation event detecting method based on face detection | |
CN106446792A (en) | Pedestrian detection feature extraction method in road traffic auxiliary driving environment | |
CN109726717A (en) | A kind of vehicle comprehensive information detection system | |
CN106650752B (en) | A kind of body color recognition methods | |
CN107704853A (en) | A kind of recognition methods of the traffic lights based on multi-categorizer | |
CN103440503A (en) | Vehicle body color detection and identification method | |
CN106407951A (en) | Monocular vision-based nighttime front vehicle detection method | |
CN107292933A (en) | A kind of vehicle color identification method based on BP neural network | |
CN111523415A (en) | Image-based two-passenger one-dangerous vehicle detection method and device | |
Sangnoree et al. | Thermal-image processing and statistical analysis for vehicle category in nighttime traffic |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant |