CN103745598B - Based on the model recognizing method of front face feature - Google Patents
Based on the model recognizing method of front face feature Download PDFInfo
- Publication number
- CN103745598B CN103745598B CN201410009098.7A CN201410009098A CN103745598B CN 103745598 B CN103745598 B CN 103745598B CN 201410009098 A CN201410009098 A CN 201410009098A CN 103745598 B CN103745598 B CN 103745598B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- image
- front face
- spot
- model
- 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.)
- Expired - Fee Related
Links
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention provides a kind of model recognizing method based on front face feature, comprise with lower part: S01, a kind of road vehicles extraction method based on image histogram information: analysis road submits the pavement image that logical bayonet socket is passed back, adopt monocular image analytical approach, extract the vehicle region that may exist in pavement image; S02, a kind of front face intercept method having merged color and gradient information: by the color of evaluating objects and gradient information in the vehicle region image adopting S01 to obtain, complete the intercepting of front face; S03, the vehicle model on-line training carrying out based on foreign peoples's sample analysis, set up the car modal of various vehicle; S04, a kind of vehicle model method of discrimination based on front face proper subspace: the vehicle template matches obtained based on face and S03 before the vehicle that S02 intercepts, draws the differentiation decision-making of vehicle model.The present invention can carry out the automatic identification of vehicle vehicle accurately, greatly facilitates the routine work of the relevant departments needing vehicle information.
Description
Technical field
The present invention relates to image identification technical field, especially a kind of model recognizing method based on face feature before vehicle.
Background technology
Along with the development of artificial intelligence, the automatically field such as control and pattern-recognition, intelligent transportation system is arisen at the historic moment, and obtains great development.The Intelligent Recognition classification of type of vehicle is the important component part of intelligent transportation system always.The accurate identification of type of vehicle has very important meaning for the determination of expressway tol lcollection volume, the management of large parking lot and highway communication Monitor and Control etc.
Traditional vehicle classification detecting device mainly contains piezoelectric type detecting device, infrared detector, magnetic inductive detecting device and ultrasonic detector, is subject to the impact of other outside labile factors, and is mostly substantially estimate Vehicle length and accurately can not estimate type of vehicle.Vision monitor compared with the detecting device of other types, have install simple, easy to maintenance, cost is low and the advantage such as informative, has become the focus of intelligent transportation system research at present.
Video monitoring is the important research contents of of the fields such as computer vision, pattern-recognition and artificial intelligence, has a wide range of applications in security monitoring, intelligent transportation, military navigation etc.In current type of vehicle differentiates, mostly by noise reduction process, rim detection, extract from video image automobile overall length, width, highly, the not characteristic quantity such as bending moment, then pattern-recognition is carried out, judge the vehicle of vehicle, but can only judge the bulk species of vehicle, as lorry, car etc., and the differentiation of the concrete model corresponding to vehicle cannot be realized.
Model corresponding to congener vehicle has nearly ten thousand kinds, and the work differentiated the concrete model of vehicle is very challenging, and current correlative study is very few.
Summary of the invention
The object of the present invention is to provide a kind of model recognizing method based on front face feature, by obtaining traffic block port image, image analysis technology is utilized quick and precisely to obtain front face image, automatically the front face feature recognition template collection of different automobile types is set up, searched for by multilayer and neighbour's comparison technology, and then complete the differentiation work of vehicle of the vehicle by each traffic block port in real time, for traffic administration and criminal investigation system provide technical support.The technical solution used in the present invention is:
Based on a model recognizing method for front face feature, comprise with lower part:
S01, a kind of road vehicles extraction method based on image histogram information: analysis road submits the pavement image that logical bayonet socket is passed back, adopts monocular image analytical approach, extracts the vehicle region that may exist in pavement image;
S02, a kind of front face intercept method having merged color and gradient information: by the color of evaluating objects and gradient information in the vehicle region image adopting S01 to obtain, complete the intercepting of front face;
S03, the vehicle model on-line training carrying out based on foreign peoples's sample analysis, set up the car modal of various vehicle;
S04, a kind of vehicle model method of discrimination based on front face proper subspace: the vehicle template matches obtained based on face and S03 before the vehicle that S02 intercepts, draws the differentiation decision-making of vehicle model.
The invention has the advantages that, the method that the present invention proposes, for the traffic block port image of catching, can quick obtaining front face image, and the automatic identification of vehicle vehicle can be carried out accurately, greatly facilitate the routine work of the relevant departments needing vehicle information.
Accompanying drawing explanation
Fig. 1 is main flow figure of the present invention.
Fig. 2 is detail flowchart of the present invention.
Embodiment
Below in conjunction with concrete drawings and Examples, the invention will be further described.
As shown in Figure 1 and Figure 2:
Model recognizing method based on front face feature proposed by the invention, comprises with lower part:
S01, a kind of road vehicles extraction method based on image histogram information, analysis road submits the pavement image that logical bayonet socket is passed back, adopts monocular image analytical approach, extracts the vehicle region that may exist in pavement image;
Described S01 is specially:
(1) for input picture I, its grey level histogram I is set up
hist, extract the gray scale that frequency of occurrence is maximum, this gray-scale value is designated as B, the corresponding frequency is P;
(2) set gray scale interval threshold T=P/3, set up road surface candidate target mask bianry image M=(I>B+T) | (I<B-T);
(3) medium filtering process is carried out to target mask bianry image M;
(4) according to the number of white pixel in target mask bianry image M, expansion process is carried out to M;
(5) target mask bianry image M is marked, obtains spot set R={R (t) in M, t=1,2, Λ, N
m, N
mfor spot number in M;
(6) analyze the size and dimension of each spot in spot set R, reject the spot that area is too small and length breadth ratio is excessive or too small; Area is too small refers to that the area of spot is less than 50 pixels, and the excessive finger length breadth ratio of length breadth ratio is greater than 5, and length breadth ratio is crossed little finger of toe length breadth ratio and is less than 0.98;
(7) to often couple of spot R (t in spot set R
1), R (t
2), according to this to the space distribution of spot at the plane of delineation, judge whether to belong to same car, if so, then merge two speckle regions, and then obtain speckle regions set P={P (c), c=1,2, Λ, N
p, N
pfor the number in region in regional ensemble P;
(8) size and dimension in the every block region in analyzed area set P and position in the picture, reject the region of too small excessive or the too small and too close image border with length breadth ratio of area, obtain revised vehicle target regional ensemble P'={P'(c) }, c=1,2, Λ, N'
p, N'
pfor the number in region in regional ensemble P'; The undue region near image border refers to that the center in region is less than 1/5 of entire image width far from the distance of image border.Because may have many vehicles in image, what obtain so last is a set that possible comprise multiple vehicle target region.
Concrete, the specific implementation step of the step (4) of described S01 is as follows:
A in () statistics target mask image M, the number of white pixel, is designated as N;
If (b) N<800, be the rectangular element of 15 × 7 sized by choice structure element se; If N>6000, be the rectangular element of 7 × 3 sized by choice structure element se; Otherwise, be the rectangular element of 9 × 5 sized by choice structure element se;
If (c) N<23000, se is adopted to carry out mathematical morphology expansion process to M, otherwise, do not do expansion process;
S02, a kind of front face intercept method having merged color and gradient information, by the color of evaluating objects and gradient information in the vehicle region image adopting S01 to obtain, complete the intercepting of front face.
In the present embodiment, described step S02 is specially:
(1) according to the target area set P'={P'(c obtained) }, in input picture I, intercept corresponding target area image I
p'(c);
(2) at target area image I
p'(c)in, according to car plate colouring information, set up car plate position candidate regions set { R
l(j)=and [x (s), y (s)] | b (s) min (r (s), g (s)) } >T
b, j=1,2, Λ, N
l, N
lfor car plate position candidate regions number, the red, green, blue color component that (r (s), g (s), b (s)) is pixel s, T
bfor color threshold;
(3) by each car plate position candidate regions R
lk the pixel value of the coordinate figure corresponding to () composes 1, rest of pixels composes 0, forms bianry image bw;
(4) doing patch indicia after carrying out medium filtering to bianry image bw, calculate the area of each spot, length breadth ratio, rectangular degree, can not be the spot of car plate according to the geometric properties deletion of spot;
(5) the spot number N of statistics after deletion action
bif, N
b>=1, then this spot is carried out license plate area confirmation;
(6) if confirmed by region, this spot centers [x is exported
l, y
l], and determine Qian Lian area size and position according to position in the picture, center, obtain face image F before in current goal region
c, otherwise, delete this spot, and N
b=N
b-1;
(7) if N
b=0, calculate target area image I
p'(c)gradient image I
g (c), and carry out expansive working;
(8) to I
g (c)mark, find the spot that wherein area is maximum, and to its area, position, symmetry judges, determines front face candidate;
(9) face height before calculating according to front face candidate blob, and determine Qian Lian area size according to speckle displacement, obtain face image F before in current goal region
c.
Concrete, the specific implementation step of the step (5) of described S02 is as follows:
The spot number N of (a) statistics after deletion action
b;
B () is by N
bindividual spot sorts according to the degree of closeness of the size and shape with car plate and size respectively, obtains the spot set D after sorting
band D
s;
If c the maximal value of the degree of approach of the size and shape of () spot and car plate and the difference of second largest value are less than 0.02, then according to D
bsort order carry out license plate area confirmation; Otherwise, according to D
bsort order carry out license plate area confirmation;
D () confirms link, by target area image I at car plate
p'(c)be converted into gray level image, and carry out thresholding process, obtain car plate bianry image I
bw;
E () from top to bottom, from left to right scans car plate bianry image I
bw, record from 0 to 1 and from 1 to 0 number of pixels, and divided by I
bwthe shared height of middle white portion, obtains N
bw;
If (f) N
bw>T
bw, then this region is confirmed by car plate, wherein T
bwfor threshold value;
S03, the vehicle model on-line training carrying out based on foreign peoples's sample analysis, set up the car modal of various vehicle; Specifically comprise the steps:
(1) for given vehicle model G, the front face image collection F corresponding to and intercept in the image of this model is formed
g={ f (i
1), i
1=1,2, Λ, N
train}
g, wherein N
trainfor training image number, and ask for every piece image f (i
1) oriented histogram of gradients feature, formed and correspond to the characteristic set F of this vehicle model
g';
(2) the front face image collection F that the image that formation corresponds to hybrid intercepts
m={ f (j
1), j
1=1,2, Λ, N
m}
m, wherein N
mfor the number that hybrid image is total, and ask for every piece image f (j
1) oriented histogram of gradients feature, formed and correspond to the characteristic set F of hybrid vehicle model
m'; Hybrid vehicle model refers to multiple different automobile types.
(3) by the characteristic set F of hybrid vehicle model
m' in data carry out principal component analysis (PCA), form face proper subspace S before vehicle;
(4) by the characteristic set F of hybrid vehicle model
m' in data projection to face proper subspace S before vehicle, form hybrid template M
m={ d
m(k
1), k
1=1,2, Λ, N
train;
(5) the characteristic set F of given vehicle model G will be corresponded to
g' in data projection to face proper subspace S before vehicle, be formed in the point set D in S
g={ d
g(k
2), k
2=1,2, Λ, N'
f;
(6) to point set D
gcarry out histogram analysis, from F
gthe front face image corresponding to data that the middle deletion frequency of occurrences is less than 0.1, is defined as foreign peoples's sample, records data and data amount check N that the frequency of occurrences is greater than 0.1 simultaneously
g, data are saved as the vehicle template M of given vehicle model G
g={ d
g(k'), k'=1,2, Λ, N
g.
S04, a kind of vehicle model Quick method based on front face proper subspace, based on the vehicle template matches that face before the vehicle that S02 intercepts and S03 obtain, draw the differentiation decision-making of vehicle model.
In the present embodiment, the process of the template matching in described step S04 is specially:
(1) car modal set M={M is formed
g, G=1,2, Λ, n, wherein n is vehicle number to be identified, and calculates the characteristic mean set E={E of each template
g, G=1,2, Λ, n;
(2) by the front face intercept method having merged color and gradient information, front face intercepting is carried out to input picture J, obtain the front face image collection F corresponding to image J
j={ f
j(t) }, t=1,2, Λ, N'
f;
(3) by f
jthe proper subspace S that t () is projected to hybrid before, face image is formed, obtains projection coefficient d (t);
(4) calculate projection coefficient d (t) and distance vector ds (g) of vehicle characteristic mean template E=| d (t)-E
g|
2;
(5) to the sequence that the data in ds are carried out from small to large, get the type corresponding to first 500 and types index thereof, set up coarse search index set ID={i}, and then obtain gross index vehicle template set M
sub={ D
i;
(6) projection coefficient d (t) and gross index vehicle template set M is calculated
subin each data distance matrix dis (i, k)=| d (t)-D
i(k) |
2;
(7) ask for the minimum value of dis, if its minimum value is less than a setting threshold value, then thinks that the vehicle of this input vehicle fails to identify, and return; Otherwise enter (8);
(8) data in dis are sorted from small to large, and searching is worth corresponding vehicle with first 50, and the number of record often kind of vehicle, vehicles maximum for corresponding number is set as final vehicle cab recognition result.
Claims (6)
1. based on a model recognizing method for front face feature, it is characterized in that, comprise with lower part:
S01, a kind of road vehicles extraction method based on image histogram information: analysis road submits the pavement image that logical bayonet socket is passed back, adopts monocular image analytical approach, extracts the vehicle region that may exist in pavement image;
S02, a kind of front face intercept method having merged color and gradient information: by the color of evaluating objects and gradient information in the vehicle region image adopting S01 to obtain, complete the intercepting of front face;
S03, the vehicle model on-line training carrying out based on foreign peoples's sample analysis, set up the car modal of various vehicle;
S04, a kind of vehicle model method of discrimination based on front face proper subspace: the vehicle template matches obtained based on face and S03 before the vehicle that S02 intercepts, draws the differentiation decision-making of vehicle model;
Described S01 specifically comprises the steps:
(1) for input picture I, its grey level histogram I is set up
hist, extract the gray scale that frequency of occurrence is maximum, this gray-scale value is designated as B, the corresponding frequency is P;
(2) set gray scale interval threshold T=P/3, set up road surface candidate target mask bianry image M=(I>B+T) | (I<B-T);
(3) medium filtering process is carried out to target mask bianry image M;
(4) according to the number of white pixel in target mask bianry image M, expansion process is carried out to M;
(5) target mask bianry image M is marked, obtains spot set R={R (t) in M, t=1,2 ..., N
m, N
mfor spot number in M;
(6) analyze the size and dimension of each spot in spot set R, reject the spot that area is too small and length breadth ratio is excessive or too small;
(7) to often couple of spot R (t in spot set R
1), R (t
2), according to this to the space distribution of spot at the plane of delineation, judge whether to belong to same car, if so, then merge two speckle regions, and then obtain speckle regions set P={P (c), c=1,2 ..., N
p, N
pfor the number in region in regional ensemble P;
(8) size and dimension in the every block region in analyzed area set P and position in the picture, reject the region of too small excessive or the too small and too close image border with length breadth ratio of area, obtain revised vehicle target regional ensemble P'={P'(c) }, c=1,2,, N'
p, N'
pfor the number in region in regional ensemble P'.
2., as claimed in claim 1 based on the model recognizing method of front face feature, it is characterized in that: the specific implementation step of the step (4) of described S01 is as follows:
A in () statistics target mask image M, the number of white pixel, is designated as N;
If (b) N<800, be the rectangular element of 15 × 7 sized by choice structure element se; If N>6000, be the rectangular element of 7 × 3 sized by choice structure element se; Otherwise, be the rectangular element of 9 × 5 sized by choice structure element se;
If (c) N<23000, se is adopted to carry out mathematical morphology expansion process to M, otherwise, do not do expansion process.
3., as claimed in claim 1 or 2 based on the model recognizing method of front face feature, it is characterized in that: described S02 specifically comprises the steps:
(1) according to the target area set P'={P'(c obtained) }, in input picture I, intercept corresponding target area image I
p'(c);
(2) at target area image I
p'(c)in, according to car plate colouring information, set up car plate position candidate regions set { R
l(j)=and [x (s), y (s)] | b (s)/min (r (s), g (s)) } >T
b, j=1,2 ..., N
l, N
lfor car plate position candidate regions number, the red, green, blue color component that (r (s), g (s), b (s)) is pixel s, T
bfor color threshold;
(3) by each car plate position candidate regions R
lk the pixel value of the coordinate figure corresponding to () composes 1, rest of pixels composes 0, forms bianry image bw;
(4) doing patch indicia after carrying out medium filtering to bianry image bw, calculate the area of each spot, length breadth ratio, rectangular degree, can not be the spot of car plate according to the geometric properties deletion of spot;
(5) the spot number N of statistics after deletion action
bif, N
b>=1, then this spot is carried out license plate area confirmation;
(6) if confirmed by region, this spot centers [x is exported
l, y
l], and determine Qian Lian area size and position according to position in the picture, center, obtain face image F before in current goal region
c, otherwise, delete this spot, and N
b=N
b-1;
(7) if N
b=0, calculate target area image I
p'(c)gradient image I
g (c), and carry out expansive working;
(8) to I
g (c)mark, find the spot that wherein area is maximum, and to its area, position, symmetry judges, determines front face candidate;
(9) face height before calculating according to front face candidate blob, and determine Qian Lian area size according to speckle displacement, obtain face image F before in current goal region
c.
4., as claimed in claim 3 based on the model recognizing method of front face feature, it is characterized in that: the specific implementation step of the step (5) of described S02 is as follows:
The spot number N of (a) statistics after deletion action
b;
B () is by N
bindividual spot sorts according to the degree of closeness of the size and shape with car plate and size respectively, obtains the spot set D after sorting
band D
s;
If c the maximal value of the degree of closeness of the size and shape of () spot and car plate and the difference of second largest value are less than 0.02, then according to D
bsort order carry out license plate area confirmation; Otherwise, according to D
ssort order carry out license plate area confirmation;
D () confirms link, by target area image I at car plate
p'(c)be converted into gray level image, and carry out thresholding process, obtain car plate bianry image I
bw;
E () from top to bottom, from left to right scans car plate bianry image I
bw, record from 0 to 1 and from 1 to 0 number of pixels, and divided by I
bwthe shared height of middle white portion, obtains N
bw;
If (f) N
bw>T
bw, then this region is confirmed by car plate, wherein T
bwfor threshold value.
5., as claimed in claim 4 based on the model recognizing method of front face feature, it is characterized in that: described S03 specifically comprises the steps:
(1) for given vehicle model G, the front face image collection F corresponding to and intercept in the image of this model is formed
g={ f (i
1), i
1=1,2 ..., N
train}
g, wherein N
trainfor training image number, and ask for every piece image f (i
1) oriented histogram of gradients feature, formed and correspond to the characteristic set F ' of this vehicle model
g;
(2) the front face image collection F that the image that formation corresponds to hybrid intercepts
m={ f (j
1), j
1=1,2 ..., N
m}
m, wherein N
mfor the number that hybrid image is total, and ask for every piece image f (j
1) oriented histogram of gradients feature, formed and correspond to the characteristic set F ' of hybrid vehicle model
m;
(3) by the characteristic set F ' of hybrid vehicle model
min data carry out principal component analysis (PCA), form face proper subspace S before vehicle;
(4) by the characteristic set F ' of hybrid vehicle model
min data projection to face proper subspace S before vehicle, form hybrid template M
m={ d
m(k
1), k
1=1,2 ..., N
train;
(5) the characteristic set F ' of given vehicle model G will be corresponded to
gin data projection to face proper subspace S before vehicle, be formed in the point set D in S
g={ d
g(k
2), k
2=1,2 ..., N'
f;
(6) to point set D
gcarry out histogram analysis, from F
gthe front face image corresponding to data that the middle deletion frequency of occurrences is less than 0.1, is defined as foreign peoples's sample, records data and data amount check N that the frequency of occurrences is greater than 0.1 simultaneously
g, data are saved as the vehicle template M of given vehicle model G
g={ d
g(k'), k'=1,2 ..., N
g.
6., as claimed in claim 5 based on the model recognizing method of front face feature, it is characterized in that, described S04 specifically comprises the steps:
(1) car modal set M={M is formed
g, G=1,2 ..., n, wherein n is vehicle number to be identified, and calculates the characteristic mean set E={E of each template
g, G=1,2 ..., n;
(2) by the front face intercept method having merged color and gradient information, front face intercepting is carried out to input picture J, obtain the front face image collection F corresponding to image J
j={ f
j(t) }, t=1,2 ..., N'
f;
(3) by f
jthe proper subspace S that t () is projected to hybrid before, face image is formed, obtains projection coefficient d (t);
(4) calculate projection coefficient d (t) and distance vector ds (g) of vehicle characteristic mean template E=| d (t)-E
g|
2;
(5) to the sequence that the data in ds are carried out from small to large, get the type corresponding to first 500 and types index thereof, set up coarse search index set ID={i}, and then obtain gross index vehicle template set M
sub={ D
i;
(6) projection coefficient d (t) and gross index vehicle template set M is calculated
subin each data distance matrix dis (i, k)=| d (t)-D
i(k) |
2;
(7) ask for the minimum value of dis, if its minimum value is less than a setting threshold value, then thinks that the vehicle of input vehicle fails to identify, and return; Otherwise enter (8);
(8) data in dis are sorted from small to large, and searching is worth corresponding vehicle with first 50, and the number of record often kind of vehicle, vehicles maximum for corresponding number is set as final vehicle cab recognition result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410009098.7A CN103745598B (en) | 2014-01-09 | 2014-01-09 | Based on the model recognizing method of front face feature |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201410009098.7A CN103745598B (en) | 2014-01-09 | 2014-01-09 | Based on the model recognizing method of front face feature |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103745598A CN103745598A (en) | 2014-04-23 |
CN103745598B true CN103745598B (en) | 2015-12-30 |
Family
ID=50502612
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201410009098.7A Expired - Fee Related CN103745598B (en) | 2014-01-09 | 2014-01-09 | Based on the model recognizing method of front face feature |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103745598B (en) |
Families Citing this family (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE102015209391A1 (en) * | 2015-05-22 | 2016-11-24 | Robert Bosch Gmbh | Method and device for generating a masking rule as well as for masking an image information of a camera |
CN105354533B (en) * | 2015-09-28 | 2018-11-09 | 江南大学 | A kind of unlicensed vehicle model recognizing method of bayonet based on bag of words |
CN105206060B (en) * | 2015-10-14 | 2017-12-29 | 武汉烽火众智数字技术有限责任公司 | A kind of vehicle type recognition device and its method based on SIFT feature |
CN105335757A (en) * | 2015-11-03 | 2016-02-17 | 电子科技大学 | Model identification method based on local characteristic aggregation descriptor |
CN105678911A (en) * | 2016-02-04 | 2016-06-15 | 彭晓梅 | Electric automobile charging pile with WIFI hot spot charging function |
CN105515129A (en) * | 2016-02-04 | 2016-04-20 | 董岩 | Charging pile with electrocar information collecting function |
CN105807249A (en) * | 2016-02-04 | 2016-07-27 | 李良 | Charging pile AC electric energy meter calibration apparatus |
CN108171203B (en) * | 2018-01-17 | 2020-04-17 | 百度在线网络技术(北京)有限公司 | Method and device for identifying vehicle |
CN109299729B (en) * | 2018-08-24 | 2021-02-23 | 四川大学 | Vehicle detection method and device |
CN110211387A (en) * | 2019-04-24 | 2019-09-06 | 湖北工业大学 | A kind of collision based reminding method and device based on vehicle cab recognition |
CN110400464B (en) * | 2019-07-15 | 2021-06-01 | 杭州亿圣信息技术有限公司 | Method for monitoring illegal parking vehicles in community |
CN110533034B (en) * | 2019-08-24 | 2023-03-17 | 大连理工大学 | Automobile front face brand classification method |
CN112183256A (en) * | 2020-09-16 | 2021-01-05 | 广东中鹏热能科技有限公司 | Method for identifying opening degree of single-photo multi-gate valve based on graphic image identification technology |
CN112529558B (en) * | 2020-11-06 | 2021-11-19 | 山东众志电子有限公司 | Follow-up prevention control platform utilizing cloud storage |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN2343635Y (en) * | 1998-01-22 | 1999-10-13 | 北京师范大学 | Motor-vehicle type identifier |
CN1655208B (en) * | 2004-12-31 | 2010-08-11 | 华南理工大学 | Automatic recognition method for motor vehicle model |
KR100939962B1 (en) * | 2008-05-02 | 2010-02-04 | 한국건설기술연구원 | Apparatus and method for classifying vehicle type |
CN201796395U (en) * | 2010-08-19 | 2011-04-13 | 佳木斯大学 | Ground induction coil type vehicle type identification and detection device |
CN103390166B (en) * | 2013-07-17 | 2016-08-17 | 中科联合自动化科技无锡有限公司 | Vehicle model consistency discrimination method based on front face feature |
-
2014
- 2014-01-09 CN CN201410009098.7A patent/CN103745598B/en not_active Expired - Fee Related
Also Published As
Publication number | Publication date |
---|---|
CN103745598A (en) | 2014-04-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103745598B (en) | Based on the model recognizing method of front face feature | |
Huang et al. | On-board vision system for lane recognition and front-vehicle detection to enhance driver's awareness | |
Hadi et al. | Vehicle detection and tracking techniques: a concise review | |
CN104021574B (en) | Pavement disease automatic identifying method | |
CN109670376B (en) | Lane line identification method and system | |
Belaroussi et al. | Road sign detection in images: A case study | |
CN104778444B (en) | The appearance features analysis method of vehicle image under road scene | |
CN102509098B (en) | Fisheye image vehicle identification method | |
CN109101924A (en) | A kind of pavement marking recognition methods based on machine learning | |
CN102880863B (en) | Method for positioning license number and face of driver on basis of deformable part model | |
US20090309966A1 (en) | Method of detecting moving objects | |
Chen et al. | Multi-shape descriptor vehicle classification for urban traffic | |
CN105160309A (en) | Three-lane detection method based on image morphological segmentation and region growing | |
KR20070027768A (en) | Method for traffic sign detection | |
CN105488453A (en) | Detection identification method of no-seat-belt-fastening behavior of driver based on image processing | |
Chen | Nighttime vehicle light detection on a moving vehicle using image segmentation and analysis techniques | |
CN107705301A (en) | A kind of highway graticule damage testing method based on unmanned plane highway map picture | |
CN103390166B (en) | Vehicle model consistency discrimination method based on front face feature | |
CN107180230B (en) | Universal license plate recognition method | |
CN104978567A (en) | Vehicle detection method based on scenario classification | |
CN103871077A (en) | Extraction method for key frame in road vehicle monitoring video | |
CN105139011A (en) | Method and apparatus for identifying vehicle based on identification marker image | |
Balci et al. | Front-View Vehicle Damage Detection using Roadway Surveillance Camera Images. | |
KR101347886B1 (en) | Method and Apparatus for Road Lane Recognition by Surface Region and Geometry Information | |
Munajat et al. | Road detection system based on RGB histogram filterization and boundary classifier |
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 | ||
TR01 | Transfer of patent right |
Effective date of registration: 20190516 Address after: 214000 Yinxiu Road, Binhu District, Wuxi City, Jiangsu Province 813-1401 Patentee after: JIANGSU BOSHI CONSTRUCTION Co.,Ltd. Address before: 214135 Internet of Things Research and Development Center E2-509, 200 Linghu Avenue, Wuxi New District, Jiangsu Province Patentee before: ZHONGKE UNITED AUTOMATION SCIENCE & TECHNOLOGY WUXI Co.,Ltd. |
|
TR01 | Transfer of patent right | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20151230 Termination date: 20220109 |
|
CF01 | Termination of patent right due to non-payment of annual fee |