CN107704858A - A kind of detection algorithm of license plate sloped angle - Google Patents
A kind of detection algorithm of license plate sloped angle Download PDFInfo
- Publication number
- CN107704858A CN107704858A CN201711051233.4A CN201711051233A CN107704858A CN 107704858 A CN107704858 A CN 107704858A CN 201711051233 A CN201711051233 A CN 201711051233A CN 107704858 A CN107704858 A CN 107704858A
- Authority
- CN
- China
- Prior art keywords
- max
- matrix
- angle
- license plate
- edge
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/63—Scene text, e.g. street names
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/48—Extraction of image or video features by mapping characteristic values of the pattern into a parameter space, e.g. Hough transformation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/625—License plates
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of detection algorithm of license plate sloped angle, comprise the following steps：Input original image matrix binaryzation Tb[W, H]；Handle to obtain horizontal edge matrix T by edge algorithmse[W, H], if edge, Te[x, y]=1；Otherwise Te[x, y]=0；The vertical coordinate ρ in polar coordinate system θ ρ is sought according to the high H of the wide W of imageMAX, ρMAX 2=W2+H2；Generation conversion ballot matrix T [ρMAX, θMAX], θMAXTake 90 °；By T [ρMAX, θMAX] in all elements set to 0；For horizontal edge matrix TeEach element in [W, H], if Te[x, y]==1, from θMAXTo θMAXCalculate ρ values ρ=y*sin (θ)+x*cos (θ), ballot matrix T [ρ, θ]=T [ρ, θ]+1；It is inclined angle to take the θ in ballot matrix T [ρ, θ] corresponding to ρ element maximums.The present invention improves the accuracy of angle of inclination detection, and detection error lifts calculating speed, be not only applicable to car plate within ± 1 °, can more be used in OCR and daily image editing software.
Description
Technical field
The present invention relates to field of license plate recognition, specifically a kind of detection algorithm of license plate sloped angle.
Background technology
License Plate refers to rapidly find out the position where car plate in piece image, typically passes through coarse positioning and fine positioning
The method that is combined is realized.One good algorithm of locating license plate of vehicle, it should be respectively provided with necessarily to tilting car plate and non-inclined car plate
Robustness.For non-inclined car plate, existing many ripe methods realize fine positioning, and for inclined car plate, due to
Existing correlation technique can not be directly used, the fine positioning of car plate can not be realized.The unique method for solving the problem is exactly thick
After positioning, inclination car plate is corrected to non-inclined car plate, then carries out fine positioning again.
For common license plate sloped mode because the setting angle difference of video camera is divided into two classes, one kind is horizontal vertical direction
Tilt simultaneously as shown in Figure 1.The shear that another kind is vertically oriented tilts, as shown in Figure 2.The Recognition Algorithm of License Plate of main flow at present
Usual procedure has：Car plate detection coarse positioning, Slant Rectify, Character segmentation, character recognition etc., its Slant Rectify is to two after it
Individual step serves vital effect, and appropriate sloped correcting method can effectively reduce different by shooting angle and make word
Deformation caused by symbol, therefore the discrimination of car plate can be significantly improved and reduce false drop rate.But to enter line tilt correction and first have to
The detection at angle of inclination is carried out, could be rotated picture to offset the inclination in shooting process according to this angle.In the past
Decades in, people have invented many methods to solve the tilt problem of character, wherein most popular surely belong to Hough transform,
But its detection accuracy is low, error is larger.
The content of the invention
It is an object of the invention to provide a kind of detection algorithm of license plate sloped angle, to solve to carry in above-mentioned background technology
The problem of going out.
To achieve the above object, the present invention provides following technical scheme：
A kind of detection algorithm of license plate sloped angle, comprises the following steps：
1) original image matrix binaryzation T is inputtedb[W, H]；
2) handle to obtain horizontal edge matrix T by edge algorithmse[W, H], if edge, Te[x, y]=1；Otherwise Te
[x, y]=0；
3) the vertical coordinate ρ in polar coordinate system θ-ρ is sought according to the high H of the wide W of imageMAX, ρMAX 2=W2+H2；
4) generation conversion ballot matrix T [ρMAX, θMAX], θMAXTake 90 °；
5) by T [ρMAX, θMAX] in all elements set to 0；
6) for horizontal edge matrix TeEach element in [W, H], if Te[x, y]==1, from-θMAXTo θMAXMeter
Calculate ρ values ρ=y*sin (θ)+x*cos (θ), ballot matrix T [ρ, θ]=T [ρ, θ]+1；
7) it is inclined angle to take the θ in ballot matrix T [ρ, θ] corresponding to ρ element maximums.
As the further scheme of the present invention：In step 6), y*sin (θ)+x*cos (θ) makes integer table, and θ's is incremented by
1°。
As the further scheme of the present invention：Corrected when detecting that horizontal tilt angle is θ with shear, specific steps
As described below：
A. known image width is W, then the maximum displacement of vertical direction is：offvmax=W*tan (θ)；
B. because triangle etc. compares, it is pixel level coordinate to calculate each column vertical shift j：offv=(W-j) * offvmax/
W；
C. shear amount is subtracted by column, obtains the row coordinate after each column is corrected：inew=i-offvj.
Compared with prior art, the beneficial effects of the invention are as follows：
The present invention is based on the theory of Hough transform, and inclination is improved with reference to the method for horizontal rim detection in image
The accuracy of angle detection, randomly generates the text images of some rotations, and detection error is within ± 1 °, and in calculating process
Floating number can be optimized in advance with look-up table, to lift calculating speed, be more suitable for embedded system.The present invention is that one kind passes through
Input picture estimates the angle of inclination in the image level direction in itself, has compared with strongly-adaptive and robustness.It can intert
In the calculating context at any required estimation image inclination angle, car plate is not only applicable to, more OCR can be used in and daily image is compiled
Collect in software.
Brief description of the drawings
Fig. 1 is common license plate sloped mode：Horizontal vertical direction tilts schematic diagram simultaneously；
Fig. 2 is common license plate sloped mode：The shear of vertical direction tilts schematic diagram；
Fig. 3 is schematic diagram of the Hough transform in straight-line detection.
Fig. 4 is to represent straight line accompanying drawings in a manner of y=ax+b in two-dimensional coordinate system.
Fig. 5 is a, and b regards variable mappings as to another coordinate system a-b schematic diagram.
Fig. 6 is the schematic diagram that Fig. 5 is mapped to by Fig. 4.
Fig. 7 is to be substituted for a-b coordinate systems and polar coordinate system θ-ρ schematic diagrames.
Fig. 8 is the original image containing two orthogonal straight lines.
Fig. 9 is the original image containing two orthogonal straight lines by obtaining horizontal coordinate after Hough transform as θ, vertical
Coordinate is ρ schematic diagram.
Figure 10 is inventive algorithm flow chart.
Figure 11 is the schematic diagram after binaryzation in the present invention.
Figure 12 is the schematic diagram after horizontal edge detects in the present invention.
Figure 13 is the schematic diagram after vertical edge detects in the present invention.
Figure 14 is the shear amount schematic diagram that vertical direction each column is calculated in the present invention.
Figure 15 is the shear amount method schematic diagram that vertical direction each column is calculated in the present invention.
Figure 16 is the result schematic diagram after the present invention corrects.
Figure 17 is to do horizontal cutting by inventive algorithm to become to obtain remedial frames schematic diagram.
Embodiment
Below in conjunction with the embodiment of the present invention, the technical scheme in the embodiment of the present invention is clearly and completely described,
Obviously, described embodiment is only part of the embodiment of the present invention, rather than whole embodiments.Based in the present invention
Embodiment, the every other embodiment that those of ordinary skill in the art are obtained under the premise of creative work is not made, all
Belong to the scope of protection of the invention.
Hough transform has several purposes, is commonly used in straight-line detection, and its principle is as follows：
It is well known that representing straight line in a manner of y=ax+b in two-dimensional coordinate system, wherein a and b determine as parameter
The slope (angle) and intercept of this straight line are determined.
And straight line is made up of infinite multiple points, in this straight line to taking two points (xi, yi), (xj, yj) obvious xi,
Yi, xj, yj are constant and meet yi=axi+b, yj=axj+b, (xi, yi) are now regarded as one group of parameter, (xj, yj) regards as
Another group, and a, b regards variable as, will be mapped to another coordinate system：
So far, two straight lines on a-b coordinate systems are mapped to by two points on the straight line of x-y coordinate system
B=-xi+yi
B=-xj+yj
It is clear that the coordinate of this two straight-line intersections is the parameter that (a, b) is just the straight line in x-y coordinate system,
Straight line y=ax+b is mapped as a point in coordinate system a-b.
Similarly, a-b coordinate systems before are substituted for and polar coordinate system θ-ρ is then represented by：
Xcos θ+ysin θ=ρ
The original image that two orthogonal straight lines are contained on the left side is by obtaining right figure wherein horizontal coordinate after Hough transform
θ, vertical coordinate ρ, the point of two extreme values of circle mark are two straight lines of the corresponding left figure of the most point of curve intersection, angle
Directly can directly it be read from horizontal coordinate.
During Recognition Algorithm of License Plate, due to having already passed through coarse positioning, selected region often includes car plate
The maximum region of area, and the horizontal edge of characters on license plate has obvious directive property so foundation can be used as to detect car plate
It is equally effective that angle of inclination and Hough are interrupted straight line to detection.
Embodiment 1
In the embodiment of the present invention, a kind of license plate sloped angle detection algorithm, comprise the following steps：
1) original image matrix binaryzation T is inputtedb[W, H]；
2) x ∈ [0, W-1], y ∈ [0, H-1], a kind of edge algorithms are selected to handle to obtain horizontal edge matrix Te[W, H],
If edge, Te[x, y]=1；Otherwise Te[x, y]=0.
Wherein x ∈ [0, W-1], y ∈ [0, H-1], put Te[x, y]=0
If Tb[x, y]==1 and Tb[x, y+1]==0 Te[x, y+1]=1
If Tb[x, y]==0 and Tb[x, y+1]==1 Te[x, y+1]=1
3) ρ is asked according to the high H of the wide W of imageMAX, ρMAX 2=W2+H2；
4) generation conversion ballot matrix T [ρMAX, θMAX], θMAXIt is desirable 90 °；
5) by T [ρMAX, θMAX] in all elements set to 0；
6) for horizontal edge matrix TeEach element in [W, H], if Te[x, y]==1, from-θMAXTo θMAXMeter
Calculate ρ values ρ=y*sin (θ)+x*cos (θ), ballot matrix T [ρ, θ]=T [ρ, θ]+1；(wherein " T [ρ, θ]=T [ρ, θ]+1 "
Mean that T [ρ, θ] value adds 1 certainly, "==" represent that logic " being equal to " "=" represents the value on right side to be assigned to the left side)
7) it is inclined angle to take the θ in ballot matrix T [ρ, θ] corresponding to ρ element maximums.
After the present invention realizes, specific implementation process CIMS is simple, as shown in Figure 10.
Y*sin (θ)+x*cos (θ) can make integer table in advance to reduce operand wherein in above-mentioned 6th step, and θ's passs
Increase to 1 ° and be enough to ensure that precision.
It is the effect implemented using above-mentioned algorithm below：
Original image is as shown in Figure 1.
Binaryzation result such as Figure 11.
Horizontal edge testing result such as Figure 12.
Vertical edge testing result such as Figure 13.
Horizontal edge and vertical edge are implemented hough become scaling method detect horizontal tilt angle into -14 °, vertically incline
Rake angle is that the simplification of 167 ° of consideration algorithms replaces rotation to correct, it is necessary to calculate the shear of vertical direction each column with shear herein
Amount such as Figure 14.
Method such as Figure 15.Comprise the following steps that described：
A. known image width is W, then the maximum displacement of vertical direction is：offvmax=W*tan (- 14 °)；
B. because triangle etc. compares, it is pixel level coordinate that can calculate each column vertical shift j：offv=(W-j) *
offvmax/W；
C. shear amount is subtracted by column, obtains the row coordinate after each column is corrected：inew=i-offvj.
Result such as Figure 16 after correction.
Similarly, it is also necessary to the detection of vertical edge angle is done, due to preceding step detection vertical tilt angle for 167 ° by upper
State algorithm and do horizontal shear, finally give remedial frames, such as Figure 17.
It is obvious to a person skilled in the art that the invention is not restricted to the details of above-mentioned one exemplary embodiment, Er Qie
In the case of without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, no matter
From the point of view of which point, embodiment all should be regarded as exemplary, and be nonrestrictive, the scope of the present invention is by appended power
Profit requires rather than described above limits, it is intended that all in the implication and scope of the equivalency of claim by falling
Change is included in the present invention.
Moreover, it will be appreciated that although the present specification is described in terms of embodiments, not each embodiment is only wrapped
Containing an independent technical scheme, this narrating mode of specification is only that those skilled in the art should for clarity
Using specification as an entirety, the technical solutions in the various embodiments may also be suitably combined, forms those skilled in the art
It is appreciated that other embodiment.
Claims (3)
1. a kind of detection algorithm of license plate sloped angle, it is characterised in that comprise the following steps：
1) original image matrix binaryzation T is inputtedb[W, H]；
2) handle to obtain horizontal edge matrix T by edge algorithmse[W, H], if edge, Te[x, y]=1；Otherwise Te[x, y]
=0；
3) the vertical coordinate ρ in polar coordinate system θ-ρ is sought according to the high H of the wide W of imageMAX, ρMAX 2=W2+H2；
4) generation conversion ballot matrix T [ρMAX, θMAX], θMAXTake 90 °；
5) by T [ρMAX, θMAX] in all elements set to 0；
6) for horizontal edge matrix TeEach element in [W, H], if Te[x, y]==1, from-θMAXTo θMAXCalculate ρ
Value ρ=y*sin (θ)+x*cos (θ), ballot matrix T [ρ, θ]=T [ρ, θ]+1；
7) it is inclined angle to take the θ in ballot matrix T [ρ, θ] corresponding to ρ element maximums.
2. the detection algorithm of license plate sloped angle according to claim 1, it is characterised in that in step 6), y*sin (θ)+
X*cos (θ) makes integer table, and θ's is incremented by 1 °.
3. the detection algorithm of license plate sloped angle according to claim 1, it is characterised in that when detecting horizontal tilt angle
Spend to be corrected during θ with shear, comprise the following steps that described：
A. known image width is W, then the maximum displacement of vertical direction is：offvmax=W*tan (θ)；
B. because triangle etc. compares, it is pixel level coordinate to calculate each column vertical shift j：offv=(W-j) * offvmax/W；
C. shear amount is subtracted by column, obtains the row coordinate after each column is corrected：inew=i-offvj.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711051233.4A CN107704858A (en) | 2017-10-31 | 2017-10-31 | A kind of detection algorithm of license plate sloped angle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711051233.4A CN107704858A (en) | 2017-10-31 | 2017-10-31 | A kind of detection algorithm of license plate sloped angle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107704858A true CN107704858A (en) | 2018-02-16 |
Family
ID=61176541
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711051233.4A Pending CN107704858A (en) | 2017-10-31 | 2017-10-31 | A kind of detection algorithm of license plate sloped angle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107704858A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110309828A (en) * | 2019-06-27 | 2019-10-08 | 浙江工业大学 | A kind of inclination license plate antidote |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101118596A (en) * | 2007-09-04 | 2008-02-06 | 西安理工大学 | License plate sloped correcting method based on supporting vector machines |
| CN101315668A (en) * | 2008-07-01 | 2008-12-03 | 上海大学 | Automatic detection method for test paper form |
| CN101334836A (en) * | 2008-07-30 | 2008-12-31 | 电子科技大学 | License plate positioning method incorporating color, size and texture characteristic |
| CN101625760A (en) * | 2009-07-28 | 2010-01-13 | 谭洪舟 | Method for correcting certificate image inclination |
| CN102496018A (en) * | 2011-12-08 | 2012-06-13 | 方正国际软件有限公司 | Document skew detection method and system |
| CN103279736A (en) * | 2013-04-27 | 2013-09-04 | 电子科技大学 | License plate detection method based on multi-information neighborhood voting |
| CN105447852A (en) * | 2015-11-12 | 2016-03-30 | 四川浩特通信有限公司 | Vehicle license plate detection and identification method based on Hough transform |
| CN106485248A (en) * | 2016-10-20 | 2017-03-08 | 江苏商贸职业学院 | A kind of slag-soil truck license plate sloped correcting method |
-
2017
- 2017-10-31 CN CN201711051233.4A patent/CN107704858A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101118596A (en) * | 2007-09-04 | 2008-02-06 | 西安理工大学 | License plate sloped correcting method based on supporting vector machines |
| CN101315668A (en) * | 2008-07-01 | 2008-12-03 | 上海大学 | Automatic detection method for test paper form |
| CN101334836A (en) * | 2008-07-30 | 2008-12-31 | 电子科技大学 | License plate positioning method incorporating color, size and texture characteristic |
| CN101625760A (en) * | 2009-07-28 | 2010-01-13 | 谭洪舟 | Method for correcting certificate image inclination |
| CN102496018A (en) * | 2011-12-08 | 2012-06-13 | 方正国际软件有限公司 | Document skew detection method and system |
| CN103279736A (en) * | 2013-04-27 | 2013-09-04 | 电子科技大学 | License plate detection method based on multi-information neighborhood voting |
| CN105447852A (en) * | 2015-11-12 | 2016-03-30 | 四川浩特通信有限公司 | Vehicle license plate detection and identification method based on Hough transform |
| CN106485248A (en) * | 2016-10-20 | 2017-03-08 | 江苏商贸职业学院 | A kind of slag-soil truck license plate sloped correcting method |
Non-Patent Citations (1)
| Title |
|---|
| 周俊: "《计算机控制技术》", 31 December 2016 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110309828A (en) * | 2019-06-27 | 2019-10-08 | 浙江工业大学 | A kind of inclination license plate antidote |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103116896B (en) | Automatic detection tracking method based on visual saliency model | |
| US11838606B2 (en) | Methods and systems for large-scale determination of RGBD camera poses | |
| EP3576017A1 (en) | Method, apparatus, and device for determining pose of object in image, and storage medium | |
| CN105488501B (en) | The method of license plate sloped correction based on rotation projection | |
| CN108537782B (en) | Building image matching and fusing method based on contour extraction | |
| CN109886896A (en) | A kind of blue License Plate Segmentation and antidote | |
| CN102331795B (en) | Method for controlling sunlight reflecting device to automatically track sun based on facula identification | |
| JPH10208056A (en) | Line detection method | |
| CN107622271A (en) | Handwriting text lines extracting method and system | |
| CN111860498B (en) | Method, device and storage medium for generating antagonism sample of license plate | |
| CN105590112B (en) | Text judgment method is tilted in a kind of image recognition | |
| KR20200071853A (en) | System and method for detecting and recognizing license plate | |
| US20230095142A1 (en) | Method and apparatus for improving object image | |
| CN105184294B (en) | It is a kind of based on pixel tracking inclination text judge recognition methods | |
| CN107992869A (en) | For tilting the method, apparatus and electronic equipment of word correction | |
| CN106228531B (en) | Automatic vanishing point calibration method and system based on horizon line search | |
| CN104537627B (en) | A kind of post-processing approach of depth image | |
| CN102073862B (en) | Method for quickly calculating layout structure of document image | |
| CN103679700A (en) | Bill image inversion detection system | |
| CN109858484B (en) | Multi-class transformation license plate correction method based on deflection evaluation | |
| CN107704858A (en) | A kind of detection algorithm of license plate sloped angle | |
| CN110991440A (en) | Pixel-driven mobile phone operation interface text detection method | |
| CN114511865A (en) | Method and device for generating structured information and computer readable storage medium | |
| CN108961182A (en) | Vertical direction vanishing point detection method and video positive twist method for video image | |
| CN108256475B (en) | Bill image inversion detection method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180216 |