CN109214256A - A kind of communication chart object detection method, device and vehicle - Google Patents
A kind of communication chart object detection method, device and vehicle Download PDFInfo
- Publication number
- CN109214256A CN109214256A CN201710554363.3A CN201710554363A CN109214256A CN 109214256 A CN109214256 A CN 109214256A CN 201710554363 A CN201710554363 A CN 201710554363A CN 109214256 A CN109214256 A CN 109214256A
- Authority
- CN
- China
- Prior art keywords
- projection matrix
- candidate
- gradient
- radius
- object detection
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 42
- 238000004891 communication Methods 0.000 title claims abstract description 33
- 239000011159 matrix material Substances 0.000 claims abstract description 88
- 238000000034 method Methods 0.000 claims abstract description 34
- 230000009977 dual effect Effects 0.000 claims abstract description 17
- 239000000284 extract Substances 0.000 claims abstract description 6
- 238000000605 extraction Methods 0.000 claims description 3
- 230000007613 environmental effect Effects 0.000 abstract description 2
- 238000010168 coupling process Methods 0.000 description 5
- 238000005859 coupling reaction Methods 0.000 description 5
- 230000008878 coupling Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 3
- 238000005086 pumping Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 1
- 210000004027 cell Anatomy 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000003550 marker Substances 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000012797 qualification Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000011895 specific detection Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/56—Context or environment of the image exterior to a vehicle by using sensors mounted on the vehicle
- G06V20/58—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
- G06V20/582—Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads of traffic signs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/245—Aligning, centring, orientation detection or correction of the image by locating a pattern; Special marks for positioning
Abstract
The invention discloses a kind of communication chart object detection method, device and vehicles, which comprises obtains original image to be detected, and extracts multiple marginal points at the edge of the original image；The gradient information of the multiple marginal point is calculated, and N number of radius r is preset according to gradient information calculatingnProjection matrix；To each radius rnProjection matrix screened to determine candidate projection matrix, and dual threshold method use to remove the noise spot in candidate's projection matrix to determine candidate target pointss；Round candidate region is determined according to the candidate target pointss, and the round candidate region is identified to obtain the corresponding traffic icon type in the round candidate region.The present invention removes noise spot by using dual threshold method, reduces the influence of environmental factor, improves the accuracy rate of icon-based programming.
Description
Technical field
The present invention relates to technical field of vehicle, in particular to a kind of communication chart object detection method, device and vehicle.
Background technique
With increasingly showing for traffic safety problem, the research of intelligent DAS (Driver Assistant System) and pilotless automobile is received
More and more concerns.In traffic environment, traffic sign provides important Traffic Information, is to implement traffic administration, be
Pedestrian and driver etc. provide traffic behavior specification to guarantee the unimpeded critical facility with traffic safety of road traffic.Traffic sign
It is automatic detection with identification when intelligence DAS (Driver Assistant System) and pilotless automobile carry out traffic environment recognize indispensable group
At part.And circular traffic sign includes prohibitory sign, Warning Mark as the big classification in traffic sign, therefore is accurately examined
Circle marker in mapping piece plays the role of most important in improving Traffic Sign Recognition performance.
The position for mainly determining circle using radial object method is identified for circular diagram target, that is, utilizes the rotational symmetry of circle
Property detection circular central position so that it is determined that circle position.But due to the changeable progress edge extracting of environment in actual scene after
It will appear the traffic sign etc. in many interference informations, such as complicated City scenarios, grove, the complexity of these texture-rich
The direction projection figure that scene carries out being obtained with radial symmetric again after marginalisation again can make center of circle candidate point increase, to increase mistake
Inspection rate.Simultaneously as need to seek N number of radius value respectively direction projection figure and carry out ballot sequence to select candidate centre point,
Calculating cost can be greatly improved in this way, reduce speed.
Thus the prior art could be improved and improve.
Summary of the invention
The technical problem to be solved in the present invention is that in view of the deficiencies of the prior art, providing a kind of communication chart target detection
Method, apparatus and vehicle, easy environment factor interference when solving existing road traffic sign detection, and the problem for causing false detection rate high.
In order to solve the above-mentioned technical problem, the technical solution adopted in the present invention is as follows:
A kind of communication chart object detection method comprising:
Original image to be detected is obtained, and extracts multiple marginal points at the edge of the original image；
The gradient information of the multiple marginal point is calculated, and N number of radius r is preset according to gradient information calculatingnThrowing
Shadow matrix, wherein n=1,2 ..., N；
To each radius rnProjection matrix screened to determine candidate projection matrix, and removed using dual threshold method
Noise spot in candidate's projection matrix is to determine candidate target pointss；
Round candidate region is determined according to the candidate target pointss, and the round candidate region is identified to obtain
The corresponding traffic icon type in the circle candidate region.
The communication chart object detection method, wherein it is described to obtain original image to be detected, and extract the original graph
Multiple marginal points at the edge of picture specifically include:
Original image to be detected is obtained, and using the edge of original image described in Canny operator or Sobel operator extraction
Multiple marginal points.
The communication chart object detection method, wherein the gradient information for calculating the multiple marginal point, and according to institute
It states gradient information calculating and presets N number of radius rnProjection matrix specifically include:
The gradient information of each marginal point is calculated, and determines the gradient direction of each marginal point according to the gradient information；
By each marginal point along its gradient direction positive direction and opposite direction respectively to distance be radius rnBallot click through
Row ballot, and each radius r is determined according to the ballotnProjection matrix, wherein the n=1,2 ..., N.
The communication chart object detection method, wherein be calculated by the following formula the gradient information of each marginal point:
Gx=(Z7+2z8+Z9)-(Z1+2z2+Z3)
Gy=(Z3+2z6+Z9)-(Z1+2z4+Z7)
Wherein, pixel Z1-Z9For pixel Z53*3 face domain pixel, the pixel Z5For marginal point, Gx is the pixel Z5's
The gradient value of horizontal direction, Gy are the pixel Z5Vertical direction gradient value.
The communication chart object detection method, wherein be calculated by the following formula the direction of each polling place:
Wherein, P is marginal point, P+ve(P) polling place of the P point along gradient positive direction, P are indicated-ve(P) indicate P point along gradient
The polling place of negative direction, g (P) are the gradient direction of P point, and round () indicates to be rounded.
The communication chart object detection method, wherein described to each radius rnProjection matrix screened with determination
Candidate projection matrix, and dual threshold method is used to remove the noise spot in the candidate projection matrix to determine that candidate target pointss have
Body includes:
Obtain each radius rnProjection matrix in vote the maximum values of points, and maximum value is less than to the throwing of first threshold
Shadow matrix is gone divided by determining candidate projection matrix；
Dual threshold method is used to remove the noise spot in the candidate projection matrix to determine candidate target pointss.
The communication chart object detection method, wherein described using in the dual threshold method removal candidate projection matrix
Noise spot to determine that candidate target pointss specifically include:
The interference that votes in candidate projection matrix are compared with second threshold respectively, and second threshold will be less than
Point removal；
The sum of votes in pre-determined distance is calculated for the polling place that candidate projection matrix retains, and uses the votes
And votes that update the polling place；
Calculate the corresponding radius r of the second maximum value of votes in each updated candidate projection matrixnWeek
Long ratio；
The ratio is compared with third threshold value respectively, and the interior satisfaction of the ballot matrix that will be greater than third threshold value is default
The polling place of condition is determined as candidate target pointss.
The communication chart object detection method, wherein the preset condition are as follows:
Wherein, the th2 indicates third threshold value, and the pt (i, j) indicates updated votes, and i indicates projection matrix
Row coordinate, j indicate projection matrix column coordinate.
A kind of communication chart target detection device comprising:
Processor is adapted for carrying out each instruction；And
Equipment is stored, is suitable for storing a plurality of instruction, described instruction is suitable for being loaded by processor and being executed as above any described
Communication chart object detection method.
A kind of vehicle comprising the detection device of traffic sign as described above.
The utility model has the advantages that compared with prior art, the present invention provides a kind of communication chart object detection method, device and vehicles
, which comprises original image to be detected is obtained, and extracts multiple marginal points at the edge of the original image；Meter
The gradient information of the multiple marginal point is calculated, and N number of radius r is preset according to gradient information calculatingnProjection matrix；To every
A radius rnProjection matrix screened to determine candidate projection matrix, and using the dual threshold method removal candidate projection
Noise spot in matrix is to determine candidate target pointss；Round candidate region is determined according to the candidate target pointss, and to the circle
Shape candidate region is identified to obtain the corresponding traffic icon type in the round candidate region.The present invention is by using dual threashold
Value method removes noise spot, reduces the influence of environmental factor, improves the accuracy rate of icon-based programming.
Detailed description of the invention
Fig. 1 is the flow chart of communication chart object detection method preferred embodiment provided by the invention.
Fig. 2 is the schematic diagram of the polling place of the both direction according to an embodiment of the invention according to P point gradient.
Fig. 3 is the structure principle chart of communication chart target detection device embodiment provided by the invention.
Specific embodiment
The present invention provides a kind of control method and system using self-starting, for make the purpose of the present invention, technical solution and
Effect is clearer, clear, and the present invention is described in more detail as follows in conjunction with drawings and embodiments.It should be appreciated that herein
Described specific examples are only used to explain the present invention, is not intended to limit the present invention.
Those skilled in the art of the present technique are appreciated that unless expressly stated, singular " one " used herein, " one
It is a ", " described " and "the" may also comprise plural form.It is to be further understood that being arranged used in specification of the invention
Diction " comprising " refer to that there are the feature, integer, step, operation, element and/or component, but it is not excluded that in the presence of or addition
Other one or more features, integer, step, operation, element, component and/or their group.It should be understood that when we claim member
Part is " connected " or when " coupled " to another element, it can be directly connected or coupled to other elements, or there may also be
Intermediary element.In addition, " connection " used herein or " coupling " may include being wirelessly connected or wirelessly coupling.It is used herein to arrange
Diction "and/or" includes one or more associated wholes for listing item or any cell and all combinations.
Those skilled in the art of the present technique are appreciated that unless otherwise defined, all terms used herein (including technology art
Language and scientific term), there is meaning identical with the general understanding of those of ordinary skill in fields of the present invention.Should also
Understand, those terms such as defined in the general dictionary, it should be understood that have in the context of the prior art
The consistent meaning of meaning, and unless idealization or meaning too formal otherwise will not be used by specific definitions as here
To explain.
With reference to the accompanying drawing, by the description of the embodiment, further explanation of the contents of the invention are made.
Fig. 1 is please referred to, Fig. 1 is the flow chart of the preferred embodiment of the control method provided by the invention using self-starting.
The described method includes:
S100, original image to be detected is obtained, and extracts multiple marginal points at the edge of the original image.
Specifically, the vehicle front road scene to be detected that can be shot by being mounted on the camera at windshield
Information is as original image to be detected, for example, the size of every frame image is 720*480, frame per second is 30 frames/second.In reality
In the application of border, in order to prevent vehicle jolt in the process of moving and the situation that makes image apprehensive, and effectively reduce calculation amount,
The pumping figure for the original image procession that shooting obtains can be handled, to obtain original image to be detected.At the pumping figure
Reason refers to the even number ranks or odd number ranks of abstract image, can effectively prevent the apprehensive effect of the image for generation of jolting in this way
Fruit.Meanwhile after taking out figure processing, the picture size of original image to be detected can be reduced into one times of acquisition image.
For example, the size of the image of acquisition is 720*480, the size of the original image to be detected after scheming by pumping is
360*240。
The edge is the place of brightness change in image, and the discontinuity based on image brightness values can be partitioned into image
Marginal information.Correspondingly, the marginal information of the available original image, i.e. multiple sides at the edge of acquisition original image
Edge point.The multiple marginal point can use Canny operator
Or Sobel operator detects to obtain.In the present embodiment, the Sobel operator used is
S200, the gradient information for calculating the multiple marginal point, and N number of radius r is preset according to gradient information calculatingn
Projection matrix, wherein n=1,2 ..., N.
Specifically, N number of radius rnIt is the gradient direction of such jointing edge point and described to be pre-set
N number of radius rnIt can determine specific circular diagram target center location.The N is pre-set, can be 2,3,4 etc..?
In the present embodiment, in order to improve round icon center location constant speed degree really, can according to the fixed-size situation of speed(-)limit sign,
It determines the range of a round radius, i.e. maximum radius and least radius, can quickly determine specific round icon in this way
The center of circle position.
Illustratively, the gradient information for calculating the multiple marginal point, and default N is calculated according to the gradient information
A radius rnProjection matrix can specifically include:
S201, the gradient information for calculating each marginal point, and determine according to the gradient information gradient of each marginal point
Direction；
S202, by each marginal point along its gradient direction positive direction and opposite direction respectively to distance be radius rnBallot
Point is voted, and determines each radius r according to the ballotnProjection matrix, wherein the n=1,2 ..., N.
Specifically, in the step S201, the gradient information refers to the horizontal direction of marginal point and vertical
The gradient value in direction can calculate gradient magnitude and gradient angle according to the gradient information.The horizontal direction of the marginal point
It can be calculated by the following formula with the gradient value of vertical direction:
Gx=(Z7+2z8+Z9)-(Z1+2z2+Z3)
Gy=(Z3+2z6+Z9)-(Z1+2z4+Z7) (1)
Wherein, pixel Z1-Z9For pixel Z53*3 face domain pixel, the pixel Z5(x, y) is marginal point, and Gx is the picture
Plain Z5Horizontal direction gradient value, Gy be the pixel Z5Vertical direction gradient value.
It is described to determine that the gradient direction of each marginal point specifically be calculated using following formula according to the gradient information
Gradient direction:
Wherein, α indicates marginal point Z5(x, y) gradient angle.
Further, since traffic sign is different from the bright-dark degree of background, it is possible to which traffic sign is brighter, it is also possible to
Background is brighter, this results in there are two possible gradient directions, is denoted as gradient positive direction and gradient negative direction respectively.Accordingly
, the gradient angle includes positive and negative two values, that is to say, that after obtaining gradient angle α according to formula (2), take respectively α positive value and
Negative value is as marginal point Z5The gradient positive direction and gradient negative direction of (x, y).
In order to illustrate the calculating process of the gradient information, it is assumed that 3*3 Image neighborhood are as follows:
| Z1 | Z2 | Z3 |
| Z4 | Z5 | Z6 |
| Z7 | Z8 | Z9 |
So 3*3 Image neighborhood central point Z can be calculated according to formula (1)5The gradient value Gx of horizontal direction and vertical
The gradient value Gy in direction.The gradient magnitude and gradient angle of 3*3 Image neighborhood central point, institute can be calculated further according to the Gx and Gy
Stating gradient direction can be calculated by following formula:
Wherein, the center gradient magnitude of the ▽ f Image neighborhood, α indicate the center gradient angle of Image neighborhood.
In the step S202, the projection matrix is 2*N, and the line number and columns of each projection matrix are equal.
The size of the projection matrix is equal in magnitude with original image to be detected.For example, the size of the band detection original image is
720*480, then the size of the projection matrix is 480*720.That is, pixel and projection in the original image
Element corresponds in matrix.In practical applications, the 2*N projection matrix can be pre-set, and the 2*N
The element of a projection matrix is 0, and the subsequent number being voted-for according to each point is corresponding in the 2*N projection matrix to update
The value of position to obtain final projection matrix, and can determine circular diagram according to the ballot points of projection matrix record
Target center location.Correspondingly, each projection matrix represents each radius rnPerspective view, each of described projection matrix
Element represents marginal point along the positive negative direction of its gradient with rnBallot when being voted for radius, with the element opposite position
Number.
It is described by each marginal point along its gradient direction positive direction and opposite direction respectively to distance be radius rnBallot
Point is voted, and determines each radius r according to the ballotnProjection matrix be specifically as follows firstly for each of extraction
Marginal point is radius r along gradient positive direction and opposite direction distancenPoint vote respectively, square corresponding to the marginal point
Corresponding position point value adds one in battle array N, and saves each radius rnThe maximum value of ballot points in direction projection figure.
As shown in Fig. 2, the position of the corresponding polling place of marginal point P can be calculated using following formula:
Wherein, P is marginal point, P+ve(P) polling place of the P point along gradient positive direction, P are indicated-ve(P) indicate P point along gradient
The polling place of negative direction, g (P) are the gradient direction of P point, and round () indicates that vector product is rounded.
In one embodiment of the invention, 2*N pair and the congenial image (N of detection image size can be constructed in advance
It is radius number to be detected), every sub-picture represents the corresponding counter of each radius, and there is representative in each position of the image
The counter in the corresponding center of circle in the position, this, simplifies the designs of counter, greatly reduce calculation amount.This method
Circular detection can be not only carried out, the polygon of any rule can also be detected.
S300, to each radius rnProjection matrix screened to determine candidate projection matrix, and use dual threshold side
Method removes the noise spot in the candidate projection matrix to determine candidate target pointss.
Specifically, described to each radius rnProjection matrix carry out screening and refer to votes maximum in projection matrix
Projection matrix less than first threshold is deleted, and can be reduced the number in the candidate center of circle in this way, and then reduce calculation amount, be improved identification
Efficiency.In the present embodiment, the candidate projection matrix can be one, or multiple.Certainly, when candidate projection matrix
When being multiple, successively to each candidate projection matrix use dual threshold method remove the noise spot in candidate's projection matrix with
It determines candidate target pointss, and determines round traffic icon according to candidate target pointss.
Illustratively, described to each radius rnProjection matrix screened to determine candidate projection matrix, and use
Dual threshold method removes the noise spot in the candidate projection matrix to determine that candidate target pointss can specifically include:
S301, each radius r is obtainednProjection matrix in vote the maximum values of points, and by maximum value less than the first threshold
The projection matrix of value is gone divided by determining candidate projection matrix；
S302, dual threshold method is used to remove the noise spot in the candidate projection matrix to determine candidate target pointss.
Specifically, the maximum value is the maximum value of each element in the projection matrix, that is, obtains polling place most
More points.The first threshold be it is pre-set, can be user's self-setting, be also possible to default setting.
Further, described that dual threshold method is used to remove the noise spot in the candidate projection matrix to determine target candidate
Point can specifically include:
S3021, the votes in candidate projection matrix are compared with second threshold respectively, and second threshold will be less than
Noise spot removal.
Specifically, the second threshold is pre-set, for removing the noise spot in candidate projection matrix, that is, is voted
Number is less than the point of second threshold.The votes refer to the size of the corresponding element value of point described in candidate projection matrix.
S3022, the sum that votes in pre-determined distance are calculated for the polling place that candidate projection matrix retains, and described in use
Votes and the update polling place votes.
Specifically, it is described for candidate projection matrix retain polling place calculate pre-determined distance votes and can use
The method of cluster obtains, it is, being clustered to the point in adjacent pre-determined distance to obtain the sum of votes, and described in use
Votes and the update polling place votes.Since circular target can all have higher polling place in its center of circle adjacent
Number, and the high noise spot of votes generally all can be more dispersed, to can reduce the influence of noise spot in this way, while can dash forward
Target is detected out.
Further, the cluster summation can use following formula:
Wherein, the pt (i, j) is the ballot points of corresponding ballot matrix position, and i, j indicate line number and columns, x1,y1Table
Show the point in neighbor distance.
The corresponding radius of S3023, the second maximum value for calculating votes in each updated candidate projection matrix
rnPerimeter ratio.
S3024, the ratio is compared with third threshold value respectively, and will be greater than full in the ballot matrix of third threshold value
The polling place of sufficient preset condition is determined as candidate target pointss.
Specifically, the preset condition are as follows:
Wherein, the th2 indicates third threshold value, and the pt (i, j) indicates updated votes, i, j indicate line number and
Columns.
In one embodiment of the invention, right in order to avoid the case where multiple candidate points occurs in a detection target circle
In same radius rnLower candidate point alternate position spike is less than rnAll candidate points only retain the point more than votes, if there is votes
Equal, then the candidate matrices are deleted.
S400, round candidate region is determined according to the candidate target pointss, and the round candidate region is identified
To obtain the corresponding traffic icon type in the round candidate region.
Specifically, determining the time of the center point according to specific qualifications after the candidate target pointss that the center of circle is chosen in ballot
Traffic sign is selected, no through traffic, restricted speed, No farm vehicles, limitation for example, circular candidate's traffic sign has
Quality, Max. Clearance _M., no left turn, no horn.It is worth explanation, the determining traffic icon type can use
Clustering method, or using the method etc. of neuron convolution, just do not illustrate one by one here.
The present invention also provides a kind of communication chart target detection devices, as shown in figure 3, comprising:
Processor 100 is adapted for carrying out each instruction；And
Equipment 200 is stored, is suitable for storing a plurality of instruction, described instruction is suitable for being loaded by processor and being executed any institute as above
State communication chart object detection method.
The present invention also provides a kind of vehicles comprising the detection device of traffic sign as described above.
Above-mentioned communication chart target detection device and the specific detection process of vehicle have been described in detail in the above-mentioned methods, at this
In just no longer state one by one.
In embodiment provided by the present invention, it should be understood that disclosed system and method can pass through others
Mode is realized.For example, the apparatus embodiments described above are merely exemplary, for example, the division of the module, only
A kind of logical function partition, there may be another division manner in actual implementation, for example, multiple units or components can combine or
Person is desirably integrated into another system, or some features can be ignored or not executed.Another point, shown or discussed is mutual
Between coupling, direct-coupling or communication connection can be through some interfaces, the INDIRECT COUPLING or communication link of device or unit
It connects, can be electrical property, mechanical or other forms.
The unit as illustrated by the separation member may or may not be physically separated, aobvious as unit
The component shown may or may not be physical unit, it can and it is in one place, or may be distributed over multiple
In network unit.It can select some or all of unit therein according to the actual needs to realize the mesh of this embodiment scheme
's.
It, can also be in addition, the functional units in various embodiments of the present invention may be integrated into one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.Above-mentioned integrated list
Member both can take the form of hardware realization, can also realize in the form of hardware adds SFU software functional unit.
The above-mentioned integrated unit being realized in the form of SFU software functional unit can store and computer-readable deposit at one
In storage media.Above-mentioned SFU software functional unit is stored in a storage medium, including some instructions are used so that a computer
It is each that equipment (can be personal computer, server or the network equipment etc.) or processor (processor) execute the present invention
The part steps of embodiment the method.And storage medium above-mentioned includes: USB flash disk, mobile hard disk, read-only memory (Read-
Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic or disk etc. it is various
It can store the medium of program code.
Finally, it should be noted that the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although
Present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that: it still may be used
To modify the technical solutions described in the foregoing embodiments or equivalent replacement of some of the technical features；
And these are modified or replaceed, technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution spirit and
Range.
Claims (10)
1. a kind of communication chart object detection method, characterized in that it comprises:
Original image to be detected is obtained, and extracts multiple marginal points at the edge of the original image；
The gradient information of the multiple marginal point is calculated, and N number of radius r is preset according to gradient information calculatingnProjection square
Battle array, wherein n=1,2 ..., N；
To each radius rnProjection matrix screened to determine candidate projection matrix, and using described in dual threshold method removal
Noise spot in candidate projection matrix is to determine candidate target pointss；
Round candidate region is determined according to the candidate target pointss, and the round candidate region is identified described in acquisition
The corresponding traffic icon type in round candidate region.
2. communication chart object detection method according to claim 1, which is characterized in that described to obtain original graph to be detected
Picture, and the multiple marginal points for extracting the edge of the original image specifically include:
Original image to be detected is obtained, and using the more of the edge of original image described in Canny operator or Sobel operator extraction
A marginal point.
3. communication chart object detection method according to claim 1, which is characterized in that the multiple marginal point of calculating
Gradient information, and N number of radius r is preset according to gradient information calculatingnProjection matrix specifically include:
The gradient information of each marginal point is calculated, and determines the gradient direction of each marginal point according to the gradient information；
By each marginal point along its gradient direction positive direction and opposite direction respectively to distance be radius rnPolling place thrown
Ticket, and each radius r is determined according to the ballotnProjection matrix, wherein the n=1,2 ..., N.
4. communication chart object detection method according to claim 3, which is characterized in that be calculated by the following formula each edge
The gradient information of point:
Gx=(Z7+2z8+Z9)-(Z1+2z2+Z3)
Gy=(Z3+2z6+Z9)-(Z1+2z4+Z7)
Wherein, pixel Z1-Z9For pixel Z53*3 face domain pixel, the pixel Z5For marginal point, Gx is the pixel Z5Level
The gradient value in direction, Gy are the pixel Z5Vertical direction gradient value.
5. communication chart object detection method according to claim 3, which is characterized in that be calculated by the following formula each ballot
The direction of point:
Wherein, P is marginal point, P+ve(P) polling place of the P point along gradient positive direction, P are indicated-ve(P) indicate P point along gradient losing side
To polling place, g (P) be P point gradient direction, round () indicate be rounded.
6. communication chart object detection method according to claim 1, which is characterized in that described to each radius rnProjection square
Battle array is screened to determine candidate projection matrix, and use dual threshold method remove the noise spot in candidate's projection matrix with
Determine that candidate target pointss specifically include:
Obtain each radius rnProjection matrix in vote the maximum values of points, and maximum value is less than to the projection square of first threshold
Battle array is gone divided by determining candidate projection matrix；
Dual threshold method is used to remove the noise spot in the candidate projection matrix to determine candidate target pointss.
7. communication chart object detection method according to claim 6, which is characterized in that described to remove institute using dual threshold method
The noise spot in candidate projection matrix is stated to determine that candidate target pointss specifically include:
Votes in candidate projection matrix are compared with second threshold respectively, and the noise spot for being less than second threshold is gone
It removes；
The sum of votes in pre-determined distance is calculated for the polling place that candidate projection matrix retains, and using the sum of the votes
Update the votes of the polling place；
Calculate the corresponding radius r of the second maximum value of votes in each updated candidate projection matrixnPerimeter ratio
Value；
The ratio is compared with third threshold value respectively, and will be greater than meeting preset condition in the ballot matrix of third threshold value
Polling place be determined as candidate target pointss.
8. communication chart object detection method according to claim 7, which is characterized in that the preset condition are as follows:
Wherein, the th2 indicates third threshold value, and the pt (i, j) indicates updated votes, and i indicates the row of projection matrix
Coordinate, j indicate the column coordinate of projection matrix.
9. a kind of communication chart target detection device, characterized in that it comprises:
Processor is adapted for carrying out each instruction；And
Equipment is stored, is suitable for storing a plurality of instruction, described instruction is suitable for being loaded by processor and being executed as claim 1-8 is any
The communication chart object detection method.
10. a kind of vehicle, which is characterized in that it includes the detection device of traffic sign as claimed in claim 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710554363.3A CN109214256A (en) | 2017-07-07 | 2017-07-07 | A kind of communication chart object detection method, device and vehicle |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710554363.3A CN109214256A (en) | 2017-07-07 | 2017-07-07 | A kind of communication chart object detection method, device and vehicle |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109214256A true CN109214256A (en) | 2019-01-15 |
Family
ID=64991339
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710554363.3A Pending CN109214256A (en) | 2017-07-07 | 2017-07-07 | A kind of communication chart object detection method, device and vehicle |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109214256A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111079723A (en) * | 2020-03-23 | 2020-04-28 | 杭州汇萃智能科技有限公司 | Target positioning method and device, computer equipment and storage medium |
| CN114087989A (en) * | 2021-11-19 | 2022-02-25 | 江苏理工学院 | Method and system for measuring three-dimensional coordinates of circle center of workpiece positioning hole of automobile cylinder |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH08210811A (en) * | 1995-02-03 | 1996-08-20 | Mazda Motor Corp | Method and device for detecting circle or ellipse |
| US20050105827A1 (en) * | 2003-09-09 | 2005-05-19 | Fuji Photo Film Co., Ltd. | Method and apparatus for detecting positions of center points of circular patterns |
| CN104657734A (en) * | 2015-03-05 | 2015-05-27 | 东方网力科技股份有限公司 | Traffic sign detecting method and device |
| CN105740877A (en) * | 2014-12-09 | 2016-07-06 | 比亚迪股份有限公司 | Traffic sign recognition method and device, and vehicle |
| US20180047150A1 (en) * | 2015-02-18 | 2018-02-15 | Siemens Healthcare Diagnostics Inc. | Image-based tube slot circle detection for a vision system |
-
2017
- 2017-07-07 CN CN201710554363.3A patent/CN109214256A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH08210811A (en) * | 1995-02-03 | 1996-08-20 | Mazda Motor Corp | Method and device for detecting circle or ellipse |
| US20050105827A1 (en) * | 2003-09-09 | 2005-05-19 | Fuji Photo Film Co., Ltd. | Method and apparatus for detecting positions of center points of circular patterns |
| CN105740877A (en) * | 2014-12-09 | 2016-07-06 | 比亚迪股份有限公司 | Traffic sign recognition method and device, and vehicle |
| US20180047150A1 (en) * | 2015-02-18 | 2018-02-15 | Siemens Healthcare Diagnostics Inc. | Image-based tube slot circle detection for a vision system |
| CN104657734A (en) * | 2015-03-05 | 2015-05-27 | 东方网力科技股份有限公司 | Traffic sign detecting method and device |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111079723A (en) * | 2020-03-23 | 2020-04-28 | 杭州汇萃智能科技有限公司 | Target positioning method and device, computer equipment and storage medium |
| CN114087989A (en) * | 2021-11-19 | 2022-02-25 | 江苏理工学院 | Method and system for measuring three-dimensional coordinates of circle center of workpiece positioning hole of automobile cylinder |
| CN114087989B (en) * | 2021-11-19 | 2023-09-22 | 江苏理工学院 | Method and system for measuring three-dimensional coordinates of circle center of positioning hole of automobile cylinder workpiece |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108681994B (en) | Image processing method and device, electronic equipment and readable storage medium | |
| Amato et al. | Deep learning for decentralized parking lot occupancy detection | |
| US20180158210A1 (en) | Synthesizing training data for broad area geospatial object detection | |
| CN108229523B (en) | Image detection method, neural network training method, device and electronic equipment | |
| CN109359520B (en) | Crowd counting method, system, computer readable storage medium and server | |
| CN112997190B (en) | License plate recognition method and device and electronic equipment | |
| CN105631418A (en) | People counting method and device | |
| US11227374B2 (en) | Method and apparatus for realtime object detection in unmanned aerial vehicle image | |
| CN108647587B (en) | People counting method, device, terminal and storage medium | |
| CN110598621A (en) | Vehicle brand identification method, device, equipment and computer readable storage medium | |
| KR20110058262A (en) | Apparatus and method for extracting vehicle | |
| CN109214256A (en) | A kind of communication chart object detection method, device and vehicle | |
| CN111127358B (en) | Image processing method, device and storage medium | |
| CN113160272B (en) | Target tracking method and device, electronic equipment and storage medium | |
| EP3798897A1 (en) | Methods for artificial neural networks | |
| CN112862856A (en) | Method, device and equipment for identifying illegal vehicle and computer readable storage medium | |
| CN114648709A (en) | Method and equipment for determining image difference information | |
| CN110705134A (en) | Driving test method, device, equipment and computer readable storage medium | |
| CN114724119B (en) | Lane line extraction method, lane line detection device, and storage medium | |
| CN111104965A (en) | Vehicle target identification method and device | |
| CN115527187A (en) | Method and device for classifying obstacles | |
| CN113428177B (en) | Vehicle control method, device, equipment and storage medium | |
| CN113177917B (en) | Method, system, equipment and medium for optimizing snap shot image | |
| KR102547405B1 (en) | Method and apparatus for object re-identification | |
| CN113963178A (en) | Method, device, equipment and medium for detecting infrared dim and small target under ground-air background |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20190115 |
|
| WD01 | Invention patent application deemed withdrawn after publication |