CN104881856B - The method and its device of regular shape in detection image - Google Patents
The method and its device of regular shape in detection image Download PDFInfo
- Publication number
- CN104881856B CN104881856B CN201410069286.9A CN201410069286A CN104881856B CN 104881856 B CN104881856 B CN 104881856B CN 201410069286 A CN201410069286 A CN 201410069286A CN 104881856 B CN104881856 B CN 104881856B
- Authority
- CN
- China
- Prior art keywords
- shape
- regular shape
- image
- regular
- pixel
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 149
- 238000000034 method Methods 0.000 title description 39
- 238000009827 uniform distribution Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 27
- 230000006870 function Effects 0.000 description 15
- 238000012545 processing Methods 0.000 description 15
- 238000012360 testing method Methods 0.000 description 11
- 238000004590 computer program Methods 0.000 description 8
- 238000003860 storage Methods 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 5
- 238000001914 filtration Methods 0.000 description 4
- 238000007689 inspection Methods 0.000 description 4
- 230000002093 peripheral effect Effects 0.000 description 4
- 230000005291 magnetic effect Effects 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- PCTMTFRHKVHKIS-BMFZQQSSSA-N (1s,3r,4e,6e,8e,10e,12e,14e,16e,18s,19r,20r,21s,25r,27r,30r,31r,33s,35r,37s,38r)-3-[(2r,3s,4s,5s,6r)-4-amino-3,5-dihydroxy-6-methyloxan-2-yl]oxy-19,25,27,30,31,33,35,37-octahydroxy-18,20,21-trimethyl-23-oxo-22,39-dioxabicyclo[33.3.1]nonatriaconta-4,6,8,10 Chemical compound C1C=C2C[C@@H](OS(O)(=O)=O)CC[C@]2(C)[C@@H]2[C@@H]1[C@@H]1CC[C@H]([C@H](C)CCCC(C)C)[C@@]1(C)CC2.O[C@H]1[C@@H](N)[C@H](O)[C@@H](C)O[C@H]1O[C@H]1/C=C/C=C/C=C/C=C/C=C/C=C/C=C/[C@H](C)[C@@H](O)[C@@H](C)[C@H](C)OC(=O)C[C@H](O)C[C@H](O)CC[C@@H](O)[C@H](O)C[C@H](O)C[C@](O)(C[C@H](O)[C@H]2C(O)=O)O[C@H]2C1 PCTMTFRHKVHKIS-BMFZQQSSSA-N 0.000 description 2
- 230000003321 amplification Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000010354 integration Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000003199 nucleic acid amplification method Methods 0.000 description 2
- 241000406668 Loxodonta cyclotis Species 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 210000003205 muscle Anatomy 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000013316 zoning Methods 0.000 description 1
Landscapes
- Image Analysis (AREA)
Abstract
The regular shape detection method and device for regular shape in detection image are provided, the regular shape detection method can include：Obtain view data；Candidate rule shape area in detection image；For the pixel in candidate rule shape area, gradient orientation histogram is generated；And the gradient orientation histogram based on generation, detect the regular shape in the candidate rule shape area., can be based on the gradient orientation histogram for the follow-up regular shape Area generation in image, the regular shape for any kind that one-time detection goes out in each candidate rule shape area using regular shape detection method according to embodiments of the present invention and device.Moreover, regular shape detection method according to embodiments of the present invention, can suitable for from black white image, gray level image, coloured image to anaglyph any kind of image detected rule shape.
Description
Technical field
The present invention relates generally to image procossing, the method and apparatus of regular shape more particularly in detection image.
Background technology
The various objects with regular shape, such as the various traffic signs shown in Fig. 1 in real world be present.
Regular shape refers to the shape that each side is equal and each interior angle number is equal.
Some detected rule shapes have been proposed or the method for the shape of certain symmetry be present.
For example, in U.S. Patent Application Publication in 2006 discloses US2006098877A1, a kind of SHAPE DETECTION is introduced
Technology, the technology obtain gradient image from input image data, and regular shape is obtained using ballot method using gradient intensity vector
Possibility center, and by the angle of gradient vector be multiplied by regular polygon while number so as to will be all while gradient direction rotate to
In same direction, then carry out ballot method and vector corresponding to each gradient direction is added up to determine corresponding regular polygon
Center.
For another example in U.S. Patent Application Publication in 2009 discloses US2009110286A1, a kind of shape is described
Detection technique, the technology detects symmetrical image-region along specific symmetrical line, only in the symmetrical image-region detected
Middle object of the detection with known form.
In the presence of the needs of the technology to can more rapidly detect desired one or more or even whole regular shapes.
The content of the invention
A kind of according to an aspect of the invention, there is provided regular shape detection side for being used for regular shape in detection image
Method, regular shape are the shape that each side is equal and each interior angle number is equal, and the regular shape detection method can include：Obtain
Obtain view data；Candidate rule shape area in detection image；For the pixel in candidate rule shape area, gradient is generated
Direction histogram；And the gradient orientation histogram based on generation, detect the regular shape in the candidate rule shape area.
According to another aspect of the present invention, there is provided a kind of regular shape detection dress for being used for regular shape in detection image
Put, regular shape is the shape that each side is equal and each interior angle number is equal, and the regular shape detection means can include：Figure
As obtaining part, view data is obtained；Candidate rule shape area detection part, the candidate rule shape area in detection image
Domain；Gradient orientation histogram obtains part, for the pixel in candidate rule shape area, generates gradient orientation histogram；With
And regular shape detection part, the gradient orientation histogram based on generation, detect the regular shape in the candidate rule shape area
Shape.
, can be based on for follow-up in image using regular shape detection method according to embodiments of the present invention and device
The gradient orientation histogram of regular shape Area generation, any kind that one-time detection goes out in each candidate rule shape area
Regular shape.Moreover, regular shape detection method according to embodiments of the present invention, can be applied to from black white image, gray-scale map
As, coloured image to anaglyph any kind of image in detected rule shape.
According to another aspect of the present invention, there is provided a kind of regular shape detection side for being used for regular shape in detection image
Method, wherein regular shape are the shape that each side is equal and each interior angle number is equal, and the regular shape detection method can wrap
Include：Obtain the edge image obtained by the edge pixel in image；Based on edge image, integrogram is calculated；Utilize integrogram, meter
Calculate the radial symmetric degree in the toroidal region centered on each pixel；Based on the toroidal region centered on each pixel
Radial symmetric degree, detect candidate rule shape area；And in the candidate rule shape area, detected rule shape.
Above-mentioned regular shape detection method is introduced into integrogram and calculates regional area in image based on toroidal
Radial symmetric degree, to obtain the regular shape region of candidate.The introducing of integrogram can greatly improve the calculating of radial symmetric degree
Speed, and toroidal selects any regular shape profile region that may be present particularly suitable for frame, thus, it is possible to improve rule
Then detection ratio of the shape in candidate rule shape, computational efficiency is improved, save computing resource.
Brief description of the drawings
From the detailed description to the embodiment of the present invention below in conjunction with the accompanying drawings, of the invention these and/or other side and
Advantage will become clearer and be easier to understand, wherein：
Fig. 1 shows the traffic sign example with regular shape.
Fig. 2 shows some examples of regular shape.
Fig. 3 shows the regular shape detection dress according to an embodiment of the invention for being used for regular shape in detection image
Put 1000 functional configuration block diagram.
Fig. 4 shows the regular shape detection side according to an embodiment of the invention for being used for regular shape in detection image
The overview flow chart of method 2000.
Fig. 5 shows an example of input picture.
Fig. 6 shows the image-region higher by radial direction symmetry in detection image according to an embodiment of the invention
The flow chart of the illustrative methods 2200 for the candidate rule shape area come in detection image.
Fig. 7 shows the example of the edge image for being obtained after the image progress rim detection shown in Fig. 5.
Fig. 8 shows the toroidal window that can apply to detect candidate rule shape area in a present embodiment
Schematic diagram.
Fig. 9 shows the schematic diagram for the integral image being calculated from the edge image shown in Fig. 7.
Figure 10 show by the use of integral image calculate by pixel A, B, C, D as four corner pixels square rectangular frame
The schematic diagram of the number method of pixel in 2.
Figure 11 (a), (b), (c) are shown based on integral image and toroidal to calculate the annulus centered on each pixel
Radial symmetric degree in shape area (shown in such as Figure 11 (a)), and obtained radial symmetric degree image（Figure 11 (c)）Signal
Figure, the brightness instruction of each pixel 3 is in the toroidal region centered on the pixel wherein in radial symmetric degree image
Radial symmetric degree, radial symmetric degree is higher, and the brightness of the pixel is higher.
Figure 12 shows the schematic diagram of radial symmetric degree figure, and it is Figure 11 (c) amplification.
Figure 13 (a) shows the radial symmetric degree figure after threshold filtering, Figure 13 (b) show with after threshold filtering
Gray-scale map corresponding to radial symmetric degree figure.
The gradient direction that Figure 14 (a), (b), (c) show a candidate rule shape area in image and generated to it
The schematic diagram of histogram.
Figure 15 shows a kind of signal of the illustrative methods for the validity for verifying the pixel in candidate rule shape area
Figure.
Figure 16 shows the gradient orientation histogram according to an embodiment of the invention based on generation, detects candidate rule
The flow chart of the illustrative methods 2400 of regular shape in shape area.
Figure 17 shows the visual display example of regular shape testing result.
Figure 18 shows the regular shape detection according to an embodiment of the invention for being used for regular shape in detection image
The functional configuration block diagram of device 3000.
Figure 19 shows the regular shape detection according to an embodiment of the invention for being used for regular shape in detection image
The overview flow chart of method 4000.
Figure 20 shows the block diagram suitable for being used for the exemplary computer system 600 for realizing embodiment of the present invention.
Embodiment
In order that those skilled in the art more fully understand the present invention, with reference to the accompanying drawings and detailed description to this hair
It is bright to be described in further detail.
The present invention is directed to the detection of regular shape.Regular shape is the shape that each side is equal and each interior angle number is equal
Shape.Regular shape herein includes circle.Circle can be considered as with the increase of side number, the limitation of regular polygon sequence.Fig. 2
Some examples of regular shape are shown, wherein top is divided into the profile of regular shape, and bottom is divided into the entirety of regular shape.This
A little regular shapes can come across many occasions in real world, such as the road traffic scene shown in Fig. 1.
It will be described in the following order：
1st, the first embodiment of regular shape detection means
2nd, the first embodiment of regular shape detection method
3rd, the second embodiment of regular shape detection means
4th, the second embodiment of regular shape detection method
1st, the first embodiment of regular shape detection means
Fig. 3 shows the regular shape detection dress according to an embodiment of the invention for being used for regular shape in detection image
Put 1000 functional configuration block diagram.
As shown in figure 3, regular shape detection means 1000 can include：Image obtains part 1100, candidate rule shape
Region detection part 1200, gradient orientation histogram obtain part 1300, regular shape detection part 1400.
Image obtains part 1100 and is configured to obtain view data.
Image obtains part 1100 can be with itself equipped with image capturing component（Such as camera）, can also be obtained from outside
Image is taken, image can include gray level image and/or anaglyph, and gray level image here is the concept of broad sense, and scope covers
From black white image to coloured image, the present invention is not limited the type of image.For example, image acquisition part 1100 can be with list
Mesh camera, binocular camera, more mesh cameras etc. wired or wireless connection, to receive the image from its transmission.
As a kind of example of application scenarios, binocular camera can be installed on to the position for example near vehicle mirrors
On, it can shoot to obtain one of left-eye image and eye image, such left-eye image and eye image simultaneously may be used as
Gray level image, and anaglyph can be calculated by left-eye image and eye image.It is hereby based on gray-scale map and/or disparity map
The regular shape detection means of picture, vehicle interior or exterior arrangement can carry out road traffic sign detection.In addition, vehicle interior configures
Message processing device can also configure pedestrian detection, vehicle detection etc..Vehicle control module can receive traffic sign inspection
Survey and/or pedestrian detection, the result of vehicle detection, and corresponding control signal is sent to vehicle.
It should be noted that binocular camera or other cameras are installed on vehicle position and number can be come as needed
Set.For example, binocular camera can be equipped on vehicle front, to shoot vehicle front scene.But, can also be additionally in car
Rear portion placement camera, to shoot rear view of vehicle scene.Alternatively, camera can also be disposed in vehicle left side or right side, with
Shoot left and right vehicle wheel both sides scene.In another example, wide-angle and/or image mosaic function can be incorporated in the camera, with
Just so that the angle of a phase function shooting is wider, it might even be possible to realize 360 degree of full-shape shoot functions.
Candidate rule shape area detection part 1200 is configured to the candidate rule shape area in detection image.
In one example, it can detect candidate's by embodying the regional area of stronger symmetry in detection image
Regular shape region.
As the method example for the regional area for detecting stronger symmetry, such as such as US2009110286A1 can be used
Middle introduction detects the technology of symmetrical image-region by detecting on the symmetry of particular line.
According to one embodiment of present invention, symmetrical part can be detected by detecting the radial symmetry of regional area
Region.Hereinafter by provide in this respect realize being discussed in detail for example.
Gradient orientation histogram obtains part 1300 and is configured to, for the pixel in candidate rule shape area, generate gradient
Direction histogram.
Regular shape detection part 1400 is configured to the gradient orientation histogram of generation, detects the candidate rule shape
Regular shape in region.
The testing result of regular shape detection part 1400 can include the center of regular shape, regular shape type,
It is at least one in the direction of regular shape.
Such testing result can export in a variety of manners.For example, in the example of detection traffic sign, can be with can
It is shown in depending on form on the display screen of the convenient viewing of driver, driver is reminded with speech form.In addition can also be by the testing result
Information be output to vehicle control module, the type by vehicle control module according to the traffic sign recognized, driven automatically
Sail control, such as "（Stop）In the case that stop " indicates, automatically control deceleration or even stop.In addition, such detection knot
Fruit can also be output to follow-up processing module, be provided for example come such as color combining identification information, Letter identification information etc.
Further testing result.
It should be noted that the part of above-mentioned regular shape detection means 1000 can be realized with software program, such as
Realized by CPU combinations RAM and ROM in all-purpose computer etc. and the software code wherein run.Software program can be with
It is stored in the storage mediums such as flash memory, floppy disk, hard disk, CD, is operationally loaded into such as random access storage device RAM
On performed by CPU.In addition, except on all-purpose computer, can also by the cooperation between application specific integrated circuit and software come
Realize.The integrated circuit is included for example, by MPU（Microprocessing unit）、DSP（Digital signal processor）、FPGA（Scene can
Program gate array）、ASIC（Application specific integrated circuit）At least one in realizes.Such all-purpose computer or special collection
It can be for example loaded in into circuit etc. on vehicle, and with leading to installed in such as camera of the imaging device for example on vehicle
Letter, so that the gray level image and/or stereo-picture that are obtained to camera shooting are handled to obtain Traffic Sign Recognition or inspection
Result is surveyed, and can also be alternatively controlled according to the driving of Traffic Sign Recognition or testing result to vehicle, such as to
Go out warning message, self-actuating brake or start emergency protecting equipment etc..In addition, all parts of regular shape detection means 1000 can
To be realized with special hardware, such as specific field programmable gate array, application specific integrated circuit etc..In addition, regular shape
The all parts of detection means 1000 can also be realized using the combination of software and hardware.
The structure and quantity of unit in above-mentioned regular shape detection means 1000 are not formed to the scope of the present invention
Limitation.According to one embodiment of present invention, image obtains part 1100, candidate rule shape area detection part 1200, ladder
Degree direction histogram obtain part 1300, regular shape detection part 1400 can be merged into an independent part performing and
Realize corresponding function and operation, or image can be obtained part 1100, candidate rule shape area detection part 1200,
Gradient orientation histogram obtains part 1300, regular shape detection part 1400 is further broken into smaller unit to realize him
Respective function and operation.
In addition, it is necessary to explanation, the structure shown in Fig. 3 is not exclusive formula, opposite regular shape detection means
1000 can include miscellaneous part, such as display unit, for showing the testing result of regular shape detection part 1400, and
Such as communication component, for will for information about and/or intermediate processing results are delivered to outside etc..
2nd, the first embodiment of regular shape detection method
Fig. 4 shows the regular shape detection side according to an embodiment of the invention for being used for regular shape in detection image
The overview flow chart of method 2000.Each step of the regular shape detection method 2000 can such as shown in Figure 3 corresponding
Part performs, naturally it is also possible to is performed by universal or special computing device.
As shown in figure 4, in step S2100, view data is obtained.
It should be noted that the present invention does not limit for the form or acquisition channel of view data.
For example, the view data can be original shooting to obtain by digital camera or be clapped by analogue camera
The photo taken the photograph handles to obtain by analog-digital commutator.Alternatively, it is also possible to being by image-scanning device such as scanner scanning paper
What part etc. obtained.For another example can also be the image generated on the computing device by word or image procossing application software.
In addition, view data can be raw image data or the view data by pretreatment, such as can be with
It is the view data after noise reduction or the binary image data after binary conversion treatment, or by edge extracting
Edge image afterwards etc..
Fig. 5 shows an example of input picture.
In step S2200, the candidate rule shape area in detection image.
According to one embodiment of present invention, the candidate rule shape area in detection image can be by detection image
Radial symmetric degree higher image-region is realized.
Fig. 6 shows the image-region higher by radial direction symmetry in detection image according to an embodiment of the invention
The flow chart of the illustrative methods 2200 for the candidate rule shape area come in detection image.The illustrative methods 2200 can be used
Step S2200 shown in Fig. 6.
In step S2210, the edge pixel in detection image, edge image is obtained.
In one example, image gradient for example can be calculated using Sobel operators, so as to obtain the ladder of input picture
Spend image.It may then based on gradient image and carry out detection of edge pixels, the introduction about Sobel operators may be referred to ME Sobel
Sociological methodology entitled " Asymptotic confidence intervals are published in nineteen eighty-two
For indirect effects in structural equation models " article.
Can be using any method for being adapted to detect for edge in image come detection of edge pixels.In one example, may be used also
To realize the rim detection in image using canny detectors, the image of edge pixel composition is obtained, hereinafter can abbreviation
For edge image.
Fig. 7 shows the example of the edge image for being obtained after the image progress rim detection shown in Fig. 5.
Edge image shown in Fig. 7 is bianry image, i.e., edge pixel value is 1, and non-edge pixels value is 0.No
Cross, be only for example, actually edge image can also be gray level image or coloured image or anaglyph, wherein non-edge picture
The pixel value of element is 0, and the value of edge pixel can be in certain span.
For ease of description, unless specifically stated otherwise, otherwise will be hereinafter described by taking binary edge map as an example.
Fig. 6 is returned to, in step S2220, in edge image, calculates the toroidal region centered on each pixel
Interior radial symmetric degree.
The thought of this step is that regular shape has preferable radial symmetric degree generally relative to its center.Therefore can
To detect candidate rule shape area by the radial symmetric degree of detection zone.It is further contemplated that arrive, regular shape
Side can be placed within annulus, such as within inscribed circle and circumscribed circle.Candidate rule shape is detected using this toroidal
Region, influence of the noise for detection edge shape in image inside regular shape can be reduced.
Fig. 8 shows the toroidal window that can apply to detect candidate rule shape area in a present embodiment
Schematic diagram.
In practical application, and the annulus being made up of the inscribed circle and circumscribed circle of shape to be detected need not be strictly utilized, and
As long as it is the side that there is shape to be detected in annulus geosphere residence as far as possible.
In one example, if the priori of the size with regular shape to be detected, the priori can be utilized
Knowledge limits the cylindrical and/or inner circle radius of annulus.
If for example, knowing the side size range of the triangle traffic sign in Fig. 5 in advance, it can calculate or estimate outside it
The radius r of round radius R and inscribed circle is met, so as to be subject to certain nargin respectively（δOutsideAnd δIt is interior）In the case of obtain annulus
Exradius（Such as R+ δOutside）And inner circle radius（Such as r- δIt is interior）.
In one example, by the annulus of predefined size in the enterprising line slip of edge image, add up using each pixel in
The toroidal of the heart（The situation that a part for toroidal is fallen into outside image can be excluded）The number of interior pixel, accumulative picture
The number of element is more, represents that the radial symmetric degree in toroidal delineation region is higher.
In order to quickly calculate the number of pixels in certain region, according to a preferred embodiment of the present invention, integrogram is introduced.
Integral image（integral image）, also known as summation area table（Summed area table, abbreviation SAT）, it is
One rapidly and effectively to the Data Structure and Algorithm of calculating sum in the sub-rectangular areas of a grid.Relevant integral image
The following webpage zh.wikipedia.org/zh-cn/ integrograms that may be referred in wikipedia are discussed in detail.
The value at any point (x, y) in integral image refers to the rectangle formed from the upper left corner of image to this point
The gray value sum of all points in region, in the case of bianry image, i.e., the pixel in the rectangular area（Have
Imitate the pixel of pixel value）Number.
Fig. 9 shows the schematic diagram for the integral image being calculated from the edge image shown in Fig. 7.
Using integral image, the number of the pixel in image regional area can be quickly calculated.
Figure 10 show by the use of integral image calculate by pixel A, B, C, D as four corner pixels square rectangular frame
The schematic diagram of the number method of pixel in 2.
The integrated value V of pixel AAIt is the number of the pixel in region 1.
By the number V of pixel A, B, C, D as the pixel in the square rectangular frame 2 of four corner pixels{ABCD}Can profit
Calculated with formula the following (1).
V{ABCD}=VC+VA–VB–VD (1)
Wherein, VB、VC、VDPixel B, C, D integrated value are represented respectively.
Figure 11 (a), (b), (c) are shown based on integral image and toroidal to calculate the annulus centered on each pixel
Radial symmetric degree in shape area (shown in such as Figure 11 (a)), and obtained radial symmetric degree image（Figure 11 (c)）Signal
Figure, the brightness instruction of each pixel 3 is in the toroidal region centered on the pixel wherein in radial symmetric degree image
Radial symmetric degree, radial symmetric degree is higher, and the brightness of the pixel is higher.
As shown in Figure 11 (b), in one example, the radial symmetric degree in each toroidal region can be lived with outer circle
Region 1 pixel number and the difference of the number of the pixel in region 2 lived of interior circle indicate.
Figure 12 shows the schematic diagram of radial symmetric degree figure, and it is Figure 11 (c) amplification.
Integral image and corresponding radial symmetric degree figure, Bu Guoben corresponding to above being illustrated by taking binary edge map as an example
Inventive embodiments can also be by simply changing the situation applied to gray level image, coloured image and anaglyph, such as in ash
In the case of spending image, the value at any point (x, y) in integral image refers to institute's structure from the upper left corner of image to this point
Into rectangular area in all point gray value sum, the radial symmetric in toroidal region centered on each pixel
Degree can using in toroidal region gray value sum a little represent.
In addition, in another example, for example the radial symmetric degree calculated in above-mentioned example can be normalized
The scaling processing of processing etc..
Fig. 6 is returned to, in step S2230, based on the radial symmetric degree in the toroidal region centered on each pixel, inspection
Regular shape region is selected in astronomical observation.
In one example, can be by the radial symmetric degree in the toroidal region centered on each pixel and symmetry threshold
Value is compared, and radial symmetric degree is exceeded into the toroidal region of symmetry threshold value as candidate rule shape area.
On symmetry threshold value, it can rule of thumb set or learn to obtain by learning algorithm.
In one example, can the radial symmetric angle value based on all pixels point in radial symmetric degree image come adaptive
Should ground setting symmetry threshold value.
For example, symmetry threshold value can be set according to following formula (2).
Tthresh=f (the radial symmetric angle value of all pixels point in radial symmetric degree image) (2)
For example, the average value of the radial symmetric angle value of all pixels point in radial symmetric degree image can be calculated, with this
The ratio of average value（The ratio can greater than, equal to or be less than）As symmetry threshold value.
For another example can calculate the radial symmetric angle value of all pixels point in radial symmetric degree image average value and
Variance, and symmetry threshold value is calculated based on the average value and mean square deviation.
Figure 13 (a) shows the radial symmetric degree figure after threshold filtering, each pixel in the radial symmetric degree figure
Point（Pixel with valid pixel value, corresponding bright spot therein）Corresponding to a toroidal（It is used as subsequent analysis pair
The follow-up regular shape region of elephant）Center.Figure 13 (b) shows gray scale corresponding with the radial symmetric degree figure after threshold filtering
Figure.
Fig. 6 is returned to, after the candidate rule shape area in detecting image, processing terminates.
The illustrative methods of the candidate rule shape area in detection image are described with reference to figure 6- Figure 13 above.
It should be noted that the detection method of the candidate rule shape area in Fig. 6-Figure 13 is used as example, without that should manage
Solve as limitation of the present invention.Other any methods that can detect candidate rule shape area may be used to the present invention.
Fig. 4 is returned to, in step S2200, after detection obtains the candidate rule shape area in image, proceeds to step
S2300。
In step S2300, for the pixel in candidate rule shape area, gradient orientation histogram is generated.
The gradient direction that Figure 14 (a), (b), (c) show a candidate rule shape area in image and generated to it
The schematic diagram of histogram.
For a candidate rule shape area, such as by the dotted circle area in the gray-scale map shown in Figure 14 (a)
Take out, obtain Figure 14（b）Shown single candidate rule shape area, for the candidate rule shape area, zoning
The gradient of interior each pixel, obtains the histogram in the direction of the gradient of each pixel（Gradient side is simply referred to as herein
To histogram）.
As it was previously stated, the method for the gradient about calculating pixel can for example use sobel operators.
In one example, during gradient orientation histogram is calculated or before, candidate rule shape can be verified
The validity of pixel in region, and remove the pixel for being unsatisfactory for validity requirement.
Figure 15 shows a kind of signal of the illustrative methods for the validity for verifying the pixel in candidate rule shape area
Figure.
In the illustrative methods, for each pixel such as P1 in candidate rule shape area, the pixel is calculated
Gradient direction（The direction shown in arrow 1 indicated such as label 1）.
Gradient direction 1 based on the pixel, it is determined that by the pixel and with the straight line of the gradient direction tangent, such as label 2
Shown in the dotted line of instruction.
The center C of the candidate rule shape area is calculated to the distance D1 of the straight line 2.
Whether within a predetermined range distance D1 is determined, for example, if the inscribed circle radius for knowing shape to be detected in advance is
R, then may determine that whether distance D1 is less than R+ δ, and wherein δ is a small positive numerical value.
If the distance in preset range, does not consider the pixel when generating gradient orientation histogram.
In Figure 15 example showns, distance D1 associated with pixel P1 within a predetermined range, therefore verifies that pixel P1 is
Valid pixel, and distance D2 associated with pixel P2 is accordingly regarded as inactive pixels outside preset range, in production gradient
Pixel P2 will not be considered during direction histogram.
In Figure 15, reference direction is horizontal direction to the right, will be rotated clockwise to from reference direction where gradient direction
The angle passed through during straight line, the angle as the gradient direction for indicating the pixel.The associated angle of pixel P1 gradient direction
Spend for θ, and the associated angle of pixel P2 gradient direction is 0.In this example, ballot direction is conllinear with gradient direction, root
According to gradient direction and the relative space relation of centroid, shape is pointed to from pixel along gradient direction or opposite direction, selection
The direction at shape center is ballot direction.
Figure 14 (c) shows the gradient direction Nogata for the candidate rule shape area shown in Figure 14 (b), being calculated
The schematic diagram of figure, wherein transverse axis represent the angle value of gradient direction, and the axle of another dimension represents that value is each angle value on transverse axis
Pixel number.
Fig. 4 is returned to, for the pixel in candidate rule shape area in step S2300, generates gradient orientation histogram
Afterwards, step S2400 is proceeded to.
In step S2400, the gradient orientation histogram based on generation, the rule in the candidate rule shape area is detected
Shape.
It should be noted that " detecting the regular shape in the candidate rule shape area " here should do the reason of broad sense
Solution, can refer to detect the candidate rule shape area in whether there is some ad hoc rules shape（Such as triangle）It is or some specific
Regular shape（Triangle as shown in Figure 1, circle, square, hexagon）, can also refer to and detect the candidate rule shape area
It is interior to whether there is any regular shape, which kind of regular shape be present, the type of regular shape existing for detection（Side number）, centre bit
Put, direction（Such as the angle for reference direction）, certain part in detection image can be referred to（Such as right side）In
Regular shape in all candidate rule shape areas, all candidate rule interest regions in detection whole image can also be referred to
Interior regular shape, the regular shape detected in some particular candidate regular shape region can also be referred to.
It will be described below exemplified by detecting any regular shape in the candidate rule shape area.
Figure 16 shows the gradient orientation histogram according to an embodiment of the invention based on generation, detects candidate rule
The flow chart of the illustrative methods 2400 of regular shape in shape area.
As shown in figure 16, in step S2410, whether uniform distribution is had based on gradient orientation histogram, is judged
With the presence or absence of circle in the candidate rule shape area.It is that the gradient direction distribution based on circle should be uniform original in step
Then carry out.
In step S2420, the peak value in gradient orientation histogram is determined.For example, it can be based on and intended pixel number
The comparison of threshold value, to determine peak value.
In step S2430, the minimum angles interval between peak value is calculated.
In step S2440, based on the minimum angles interval between peak value, the most probable number MPN on the side of regular shape is determined
Mesh.
For a regular shape, the peak value of its gradient direction should correspond to represent the gradient direction on each bar side
Angle, the interval between peak value, it should which it is side number to meet 360 degree/k, wherein k.
For example, for equilateral triangle, the interval between the peak value of corresponding gradient orientation histogram should be 360/3=
120.For square, the interval between the peak value of corresponding gradient orientation histogram should be 360/4=90.
Thus, under conditions of the minimum angles interval β between obtaining peak value, the largest possible number thereof on the side of regular shape
It may be calculated the result that rounds up to real-value obtained by 360/ β.
In step S2450, for any one in regular shape of the side number from 3 to maximum number k, by each of peak value
Individual angle position is matched with the angle of the regular shape of corresponding sides number, to determine whether there is the regular shape of the side number.
In addition, in one example, can be by checking gradient orientation histogram after it is determined that ad hoc rules shape be present
In minimum angles value in the angle associated with the given shape, determine rotation of the ad hoc rules shape relative to reference direction
Gyration.
Specifically, for example, for the regular shape that side number is 3, t+120 position can for each peak t, be checked
Put place, t+240 opening position whether there is peak value.If peak value all be present in t, t+120, t+240 opening position, in the presence of three
It is angular.And it can determine that the triangle has turned t degree clockwise relative to being oriented for reference direction.
Similarly, for example, for the regular shape that side number is 4, t+90, t+ can be checked for each peak t
180th, t+270 opening position whether there is peak value.If peak value all be present in t, t+90, t+180, t+270 opening position, deposit
In square.And it can determine that the square has turned t degree clockwise relative to being oriented for reference direction.
By that analogy, for example, for the regular shape that side number is k, t+360/ can be checked for each peak t
K, t+2*360/k ..., t+ (k-1) * 360/k opening position whether there is peak value.If in t, t+360/k, t+2*360/
All there is peak value in k ..., t+ (k-1) * 360/k opening position, then in the presence of positive k sides shape.And it can determine that the positive k sides shape is relative
T degree is turned clockwise in being oriented for reference direction.
After regular shape is detected from a candidate rule shape area, the candidate rule shape area can be utilized
Center instruction regular shape position in the picture.
In figure 16, after step S2450, processing can terminate.
It should be noted that as needed, various modifications can be carried out to the example shown in Figure 16, for example, clearly examining
In the case that the target of survey is circle, the processing shown in Figure 16 can directly terminate after step S2410.
For another example in the case where the target clearly detected is triangle, square and hexagon, can be without performing step
Rapid S2430, S2440, and directly for any one in the regular shape of side number 3,4,6, by all angles position of peak value with
The angle of the regular shape of corresponding sides number is matched, to determine whether there is the regular shape of the side number.
Fig. 4 is returned to, after step S2400, processing can terminate.
In one example, regular shape testing result can be exported in a variety of manners.For example, in road traffic sign detection
In example, it can be shown in visual form on the display screen of the convenient viewing of driver, the image of display as shown in figure 17, wherein
Label 1,2,3 indicates the object with regular shape detected being highlighted.Or can be with speech form prompting department
Machine.In addition the information of the testing result can also be output to vehicle control module, by vehicle control module according to recognizing
The type of traffic sign, Driving control is carried out automatically, such as in " stop（Stop）" in the case of mark, automatically controlling deceleration is
To stopping.In addition, such testing result can also be output to follow-up processing module, come such as color combining identification information,
Letter identification information etc. provides for example further testing result.
Regular shape detection method according to embodiments of the present invention, can be based on for the follow-up regular shape area in image
The gradient orientation histogram of domain generation, the regular shape for any kind that one-time detection goes out in each candidate rule shape area.
Moreover, regular shape detection method according to embodiments of the present invention, can be applied to from black white image, gray-scale map
As, coloured image to anaglyph any kind of image in detected rule shape.
3rd, the second embodiment of regular shape detection means
Figure 18 shows the regular shape detection according to an embodiment of the invention for being used for regular shape in detection image
The functional configuration block diagram of device 3000.
As shown in figure 18, regular shape detection means 3000 can include：Edge image obtains part 3100, integrogram meter
Calculate part 3200, radial symmetric degree obtains part 3300, candidate rule shape area detection part 3400 and regular shape and detected
Part 3500.
Edge image obtains part 3100 and is configured to obtain the edge image obtained by the edge pixel in image.Here
Edge image should be interpreted broadly, it is understood that the set of edge pixel, and and do not need separate visual turn to edge graph
Picture.Similarly, radial symmetric degree image mentioned in this article should also be widely understood, it is understood that the radial symmetric of pixel
The set of degrees of data, again without being individually visualized as radial symmetric degree image.Similarly, integrogram should also do broad sense reason
Solution, it is understood that the set of the integration Value Data of each pixel, again without being individually visualized as radial symmetric degree image.
Integrogram calculating unit 3200 is configured to edge image, calculates integrogram.
Radial symmetric degree obtains part 3300 and is configured to utilize integrogram, calculates the toroidal centered on each pixel
The radial symmetric degree in region.
Candidate rule shape area detection part 3400 is configured to the toroidal region centered on each pixel
Radial symmetric degree, detect candidate rule shape area.
Regular shape detection part 3500 is configured in the candidate rule shape area, detected rule shape.
Function and operation about all parts, the description carried out above in conjunction with Fig. 4-Figure 17 is may be referred to, here no longer
Repeat.
It should be noted that the detected rule shape in the candidate rule shape area of regular shape detection part 3500
Function step S2300 and step S2400 can both be realized like that using gradient orientation histogram as shown in Figure 4.Can also
Realized using the method for other any detected rule shapes, such as in foregoing US2006098877A1 and US2009110286A1
Regular shape detection method.
4th, the second embodiment of regular shape detection method
Figure 19 shows the regular shape detection according to an embodiment of the invention for being used for regular shape in detection image
The overview flow chart of method 4000.
In step S4100, the edge image obtained by the edge pixel in image is obtained.
In step S4200, based on edge image, integrogram is calculated.
In step S4300, using integrogram, the radial symmetric in the toroidal region centered on each pixel is calculated
Degree.
In step S4400, based on the radial symmetric degree in the toroidal region centered on each pixel, detection candidate's rule
Then shape area.
In step S4500, in the candidate rule shape area, detected rule shape.
Detailed realization about each step, the description carried out above in conjunction with Fig. 4-Figure 17 is may be referred to, it is no longer superfluous here
State.
It should be noted that the detected rule shape in the candidate rule shape area in step S3500 both can be as
Step S2300 shown in Fig. 4 and step S2400 are realized using gradient orientation histogram like that.Other any inspections can also be utilized
The method for surveying regular shape is realized, such as the regular shape detection in foregoing US2006098877A1 and US2009110286A1
Method.
The regular shape detection method of second embodiment shown in Figure 18 and Figure 19 and the key of device are to introduce integration
Scheme and the radial symmetric degree of regional area in image is calculated based on toroidal, obtain the regular shape region of candidate, product
The introducing of component can greatly improve the calculating speed of radial symmetric degree, and toroidal selects any regular shape particularly suitable for frame
Shape profile region that may be present, thus, it is possible to improve detection ratio of the regular shape in candidate rule shape, improve and calculate
Efficiency, save computing resource.
5th, for carrying out the computing system of regular shape detection
The present invention can also be implemented by a kind of computing system for being used to carry out regular shape detection.Figure 20 shows suitable
In for realizing the block diagram of the exemplary computer system 600 of embodiment of the present invention.As shown in figure 20, computing system 600 can be with
Including：CPU（CPU）601、RAM（Random access memory）602、ROM（Read-only storage）603rd, system bus
604th, hard disk controller 605, KBC 606, serial interface controller 607, parallel interface controller 608, display control
Device 69, hard disk 610, keyboard 611, serial peripheral equipment 612, concurrent peripheral equipment 613 and display 614.In such devices,
What is coupled with system bus 604 has CPU601, RAM602, ROM603, hard disk controller 605, KBC 606, serial control
Device 607, parallel controller 608 and display controller 609 processed.Hard disk 610 couples with hard disk controller 605, keyboard 611 and keyboard
Controller 606 couples, and serial peripheral equipment 612 couples with serial interface controller 607, and concurrent peripheral equipment 613 with connecing parallel
Mouth controller 608 couples, and display 614 couples with display controller 609.It should be appreciated that the structured flowchart described in Figure 20
Just for the sake of the purpose of example, without limiting the scope of the present invention.In some cases, can increase as the case may be
Add deduct less some equipment.
Person of ordinary skill in the field knows that the present invention can be implemented as system, device, method or computer program
Product.Therefore, the present invention can be implemented as following form, i.e.,：It can be complete hardware, can also be complete software
（Including firmware, resident software, microcode etc.）, can also be hardware and software combine form, referred to generally herein as " circuit ",
" module ", " device " or " system ".In addition, in certain embodiments, the present invention is also implemented as calculating in one or more
The form of computer program product in machine computer-readable recording medium, computer-readable program generation is included in the computer-readable medium
Code.
Any combination of one or more computer-readable mediums can be used.Computer-readable medium can be computer
Readable signal medium or computer-readable recording medium.Computer-readable recording medium can for example be but not limited to electricity, magnetic,
Optical, electromagnetic, the system of infrared ray or semiconductor, device or device, or any combination above.Computer-readable storage medium
The more specifically example of matter（Non exhaustive list）Including：Electrical connection with one or more wires, portable computer magnetic
Disk, hard disk, random access memory（RAM）, read-only storage (ROM), erasable programmable read only memory (EPROM or sudden strain of a muscle
Deposit), optical fiber, portable compact disc read-only storage (CD-ROM), light storage device, magnetic memory device or above-mentioned appoint
The suitable combination of meaning.In this document, computer-readable recording medium can be it is any include or the tangible medium of storage program,
The program can be commanded the either device use or in connection of execution system, device.
Computer-readable signal media can include in a base band or as carrier wave a part propagation data-signal,
Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, including but unlimited
In electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer can
Any computer-readable medium beyond storage medium is read, the computer-readable medium, which can send, propagates or transmit, to be used for
By instruction execution system, device either device use or program in connection.
The program code included on computer-readable medium can use any appropriate medium to transmit, including but not limited to without
Line, electric wire, optical cable, RF etc., or above-mentioned any appropriate combination.
It can be write with one or more programming languages or its combination for performing the computer that operates of the present invention
Program code, described program design language include object oriented program language-such as Java, Smalltalk, C++,
Also include conventional procedural programming language-such as " C " language or similar programming language.Program code can be with
Fully perform, partly perform on the user computer on the user computer, the software kit independent as one performs, portion
Divide and partly perform or performed completely on remote computer or server on the remote computer on the user computer.
It is related in the situation of remote computer, remote computer can be by the network of any kind-include LAN (LAN) or wide
Domain net (WAN)-be connected to subscriber computer, or, it may be connected to outer computer（Such as provided using Internet service
Business passes through Internet connection）.
Above with reference to the method, apparatus of the embodiment of the present invention（System）With the flow chart and/or frame of computer program product
Figure describes the present invention.It should be appreciated that each square frame in each square frame and flow chart and/or block diagram of flow chart and/or block diagram
Combination, can be realized by computer program instructions.These computer program instructions can be supplied to all-purpose computer, special
The processor of computer or other programmable data processing units, so as to produce a kind of machine, these computer program instructions
Performed by computer or other programmable data processing units, generate and advised in the square frame in implementation process figure and/or block diagram
The device of fixed function/operation.
These computer program instructions can also be stored in can cause computer or other programmable data processing units
In the computer-readable medium to work in a specific way, so, the instruction being stored in computer-readable medium just produces one
Command device (the instruction of function/operation specified in the individual square frame including in implementation process figure and/or block diagram
Means manufacture)（manufacture）.
Computer program instructions can also be loaded into computer, other programmable data processing units or miscellaneous equipment
On so that series of operation steps is performed on computer, other programmable data processing units or miscellaneous equipment, in terms of producing
The process that calculation machine is realized, so that the instruction performed on computer or other programmable devices can provide implementation process figure
And/or the process of function/operation specified in the square frame in block diagram.
Flow chart and block diagram in accompanying drawing show system, method and the computer journey of multiple embodiments according to the present invention
Architectural framework in the cards, function and the operation of sequence product.At this point, each square frame in flow chart or block diagram can generation
The part of one module of table, program segment or code, a part for the module, program segment or code include one or more use
In the executable instruction of logic function as defined in realization.It should also be noted that marked at some as in the realization replaced in square frame
The function of note can also be with different from the order marked in accompanying drawing generation.For example, two continuous square frames can essentially base
Originally it is performed in parallel, they can also be performed in the opposite order sometimes, and this is depending on involved function.It is also noted that
It is the combination of each square frame and block diagram in block diagram and/or flow chart and/or the square frame in flow chart, can uses and perform rule
Fixed function or the special hardware based system of operation are realized, or can use the group of specialized hardware and computer instruction
Close to realize.
It is described above only illustrative, much it can be changed and/or be replaced.
For example, in description above, regular shape detection method and device are applied to vehicle assistant drive or vehicle
Automatically control the Traffic Sign Recognition in scene.But this is only example, the invention is not limited in this, but be can apply to
The object of any identification has a case that regular shape, such as identifying having in the situations such as airport, railway station, market
The object of regular shape.
In addition, in description above, for the calculating of image local area symmetry, toroidal window is employed.
Because toroidal be suitable for any direction, any regular shape candidate contours detection, and particularly suitable for one
The situation of secondary detection a variety of objects of different shapes.But, in some cases, such as in object shapes and direction to be detected
In the case of determining, the ring-type window of the shape can also be directly used, for example with two sizes with convenience center
The ring-type window that different squares surrounds detects the symmetry of square candidate region, similarly, is for object to be detected
The situation of positive k sides shape, positive k can be detected using the ring-type window of two positive k sides shapes of different sizes with convenience center
Deform the symmetry of candidate region.
In addition, in description above, during the candidate rule shape area in detection image, image is carried out first
Rim detection.But in some cases, this step and may need not be carried out, such as in view data inherently edge graph
In the case of as data, or in the case where view data is sparse disparities diagram data, the pixel now in view data is big
Part has been edge pixel, therefore rim detection may be no longer necessary.
It is described above various embodiments of the present invention, described above is exemplary, and non-exclusive, and
It is not limited to disclosed each embodiment.In the case of without departing from the scope and spirit of illustrated each embodiment, for this skill
Many modifications and changes will be apparent from for the those of ordinary skill in art field.The selection of term used herein, purport
The principle of each embodiment, practical application or improvement to the technology in market are best being explained, or is making the art
Other those of ordinary skill are understood that each embodiment disclosed herein.
Claims (9)
1. a kind of regular shape detection method for being used for regular shape in detection image, regular shape is that each side is equal and each
The equal shape of interior angle number, the regular shape detection method include：
Obtain view data；
Candidate rule shape area in detection image；
For the pixel in candidate rule shape area, gradient orientation histogram is generated；And
Gradient orientation histogram based on generation, the regular shape in the candidate rule shape area is detected,
The wherein described gradient orientation histogram based on generation, the regular shape detected in the candidate rule shape area include：
Determine the peak value in gradient orientation histogram；
Calculate the minimum angles interval between peak value；
Based on the minimum angles interval between peak value, the largest possible number thereof on the side of regular shape is determined；
For side number from 3 to any one in the maximum number of regular shape, by all angles position of peak value and corresponding sides
The angle of several regular shapes is matched, to determine whether there is the regular shape of the side number.
2. regular shape detection method according to claim 1, the gradient orientation histogram based on generation, detect the candidate
Regular shape in regular shape region includes：
The gradient orientation histogram of each candidate rule shape area based on generation, one-time detection go out each candidate rule shape
The regular shape of any kind in region.
3. regular shape detection method according to claim 1, the candidate rule shape area in the detection image includes：
Edge pixel in detection image, obtains edge image；
In edge image, the radial symmetric degree in the toroidal region centered on each pixel is calculated；And
Based on the radial symmetric degree in the toroidal region centered on each pixel, candidate rule shape area is detected.
4. regular shape detection method according to claim 3, described in edge image, calculate centered on each pixel
The radial symmetric degree in toroidal region includes：
Based on edge image, integrogram is calculated；And
Using integrogram, the radial symmetric degree in the toroidal region centered on each pixel is calculated.
5. regular shape detection method according to claim 1, the pixel in candidate rule shape area, generation ladder
Degree direction histogram includes：
For each pixel in candidate rule shape area, the gradient direction of the pixel is calculated；
Gradient direction based on the pixel, it is determined that by the pixel and with the straight line of the gradient direction tangent；
The center of the candidate rule shape area is calculated to the distance of the straight line；
Determine the distance whether within a predetermined range；And
If the distance in preset range, does not consider the pixel when generating gradient orientation histogram.
6. regular shape detection method according to claim 1, the gradient orientation histogram based on generation, detect the candidate
Regular shape in regular shape region includes：
Whether uniform distribution is had based on gradient orientation histogram, is judged in the candidate rule shape area with the presence or absence of circle
Shape.
7. regular shape detection method according to claim 1, the gradient orientation histogram based on generation, detect the candidate
Regular shape in regular shape region includes：
After it is determined that ad hoc rules shape be present, by checking angle associated with the given shape in gradient orientation histogram
In minimum angles value, determine the anglec of rotation of the ad hoc rules shape relative to reference direction.
8. a kind of regular shape detection means for being used for regular shape in detection image, regular shape is that each side is equal and each
The equal shape of interior angle number, the regular shape detection means include：
Image obtains part, obtains view data；
Candidate rule shape area detection part, the candidate rule shape area in detection image；
Gradient orientation histogram obtains part, for the pixel in candidate rule shape area, generates gradient orientation histogram；With
And
Regular shape detection part, the gradient orientation histogram based on generation, detect the rule in the candidate rule shape area
Shape, the regular shape detection part：
Determine the peak value in gradient orientation histogram；
Calculate the minimum angles interval between peak value；
Based on the minimum angles interval between peak value, the largest possible number thereof on the side of regular shape is determined；
For side number from 3 to any one in the maximum number of regular shape, by all angles position of peak value and corresponding sides
The angle of several regular shapes is matched, to determine whether there is the regular shape of the side number.
9. a kind of regular shape detection method for being used for regular shape in detection image, regular shape is that each side is equal and each
The equal shape of interior angle number, the regular shape detection method include：
Obtain the edge image obtained by the edge pixel in image；
Based on edge image, integrogram is calculated；
Using integrogram, the radial symmetric degree in the toroidal region centered on each pixel is calculated；
Based on the radial symmetric degree in the toroidal region centered on each pixel, candidate rule shape area is detected；And
In the candidate rule shape area, detected rule shape,
Wherein in the candidate rule shape area, detected rule shape includes：
Determine the peak value in gradient orientation histogram；
Calculate the minimum angles interval between peak value；
Based on the minimum angles interval between peak value, the largest possible number thereof on the side of regular shape is determined；
For side number from 3 to any one in the maximum number of regular shape, by all angles position of peak value and corresponding sides
The angle of several regular shapes is matched, to determine whether there is the regular shape of the side number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410069286.9A CN104881856B (en) | 2014-02-27 | 2014-02-27 | The method and its device of regular shape in detection image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410069286.9A CN104881856B (en) | 2014-02-27 | 2014-02-27 | The method and its device of regular shape in detection image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104881856A CN104881856A (en) | 2015-09-02 |
| CN104881856B true CN104881856B (en) | 2018-01-16 |
Family
ID=53949341
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410069286.9A Expired - Fee Related CN104881856B (en) | 2014-02-27 | 2014-02-27 | The method and its device of regular shape in detection image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104881856B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105260713B (en) * | 2015-10-09 | 2019-06-28 | 东方网力科技股份有限公司 | A kind of method for detecting lane lines and device |
| CN105631881B (en) * | 2015-12-30 | 2019-02-12 | 四川华雁信息产业股份有限公司 | Object detection method and device |
| CN107113421B (en) * | 2016-08-09 | 2018-07-06 | 深圳市瑞立视多媒体科技有限公司 | The detection method and device of a kind of optical system imaging quality |
| CN108734131B (en) * | 2018-05-22 | 2021-08-17 | 杭州电子科技大学 | Method for detecting symmetry of traffic sign in image |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102799859A (en) * | 2012-06-20 | 2012-11-28 | 北京交通大学 | Method for identifying traffic sign |
| CN102870138A (en) * | 2010-03-15 | 2013-01-09 | 美国亚德诺半导体公司 | Edge orientation for second derivative edge detection methods |
| CN103390167A (en) * | 2013-07-18 | 2013-11-13 | 奇瑞汽车股份有限公司 | Multi-characteristic layered traffic sign identification method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8447139B2 (en) * | 2010-04-13 | 2013-05-21 | International Business Machines Corporation | Object recognition using Haar features and histograms of oriented gradients |
-
2014
- 2014-02-27 CN CN201410069286.9A patent/CN104881856B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102870138A (en) * | 2010-03-15 | 2013-01-09 | 美国亚德诺半导体公司 | Edge orientation for second derivative edge detection methods |
| CN102799859A (en) * | 2012-06-20 | 2012-11-28 | 北京交通大学 | Method for identifying traffic sign |
| CN103390167A (en) * | 2013-07-18 | 2013-11-13 | 奇瑞汽车股份有限公司 | Multi-characteristic layered traffic sign identification method |
Non-Patent Citations (1)
| Title |
|---|
| 基于稳定特征的交通标志检测与识别技术研究;胡锦城;《中国优秀硕士学位论文全文数据库信息科技辑》;20130615(第06期);第2页、第7-8页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104881856A (en) | 2015-09-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11354827B2 (en) | Methods and systems for fusion display of thermal infrared and visible image | |
| JP6384182B2 (en) | Method and apparatus for detecting linear indicating sign on road | |
| US9851208B2 (en) | Apparatus and method for providing guidance information using crosswalk recognition result | |
| US10032082B2 (en) | Method and apparatus for detecting abnormal situation | |
| JP6088792B2 (en) | Image detection apparatus, control program, and image detection method | |
| CN104881856B (en) | The method and its device of regular shape in detection image | |
| US9141852B1 (en) | Person detection and pose estimation system | |
| CN109409186B (en) | Driver assistance system and method for object detection and notification | |
| JP2005159392A (en) | Vehicle periphery monitoring device and method | |
| EP3702957A1 (en) | Target detection method and apparatus, and computer device | |
| KR101738129B1 (en) | Apparatus and method for recognizing automatic number plate on long distance image using one camera | |
| US11340700B2 (en) | Method and apparatus with image augmentation | |
| JPWO2012014627A1 (en) | Vehicle periphery monitoring device | |
| US11416707B2 (en) | Information processing method, information processing system, and information processing apparatus | |
| CN109242796A (en) | Character image processing method, device, electronic equipment and computer storage medium | |
| JP2009025910A (en) | Obstacle detection device, obstacle detection system, and obstacle detection method | |
| CN111861956A (en) | Picture processing method and device, electronic equipment and medium | |
| CN107958202B (en) | Human body detection apparatus, control method of human body detection apparatus, and storage medium | |
| US20150125082A1 (en) | Method and apparatus for generating hierarchical saliency images detection with selective refinement | |
| KR101558487B1 (en) | Method for providing tire information | |
| JP2007004721A (en) | Object detecting device and object detecting method | |
| Nguyen et al. | An effective implementation of gaussian of gaussian descriptor for person re-identification | |
| JP2020052476A (en) | Object detection device and object detection program | |
| KR101556353B1 (en) | Method for providing tire information using a portable terminal | |
| JP5484118B2 (en) | Vehicle periphery monitoring device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180116 |