CN106557759A - A kind of sign board information getting method and device - Google Patents
A kind of sign board information getting method and device Download PDFInfo
- Publication number
- CN106557759A CN106557759A CN201611063821.5A CN201611063821A CN106557759A CN 106557759 A CN106557759 A CN 106557759A CN 201611063821 A CN201611063821 A CN 201611063821A CN 106557759 A CN106557759 A CN 106557759A
- Authority
- CN
- China
- Prior art keywords
- image
- coordinate points
- target
- mean
- traffic mark
- 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.)
- Granted
Links
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
Abstract
The disclosure is directed to a kind of sign board information getting method and device, are related to technical field of information processing, methods described includes：Hough transformation is carried out to target image, the multiple coordinate points in Hough parameter space are obtained, the target image is the image of sign board information to be obtained；Based on the plurality of coordinate points, the image-region of traffic mark board is determined by mean shift algorithm from the target image；The image-region of the traffic mark board is identified, the sign board information of the traffic mark board is obtained.The embodiment of the present disclosure is detected to target image by Hough transformation, and the image-region of traffic mark board is determined using mean shift algorithm, afterwards, it is identified by the image-region to the traffic mark board, obtain the sign board information of the traffic mark board, the accuracy of detection of the image-region of traffic signss is improve, false drop rate is reduced.
Description
Technical field
It relates to technical field of information processing, more particularly to a kind of sign board information getting method and device.
Background technology
With the development of the information processing technology, intelligent driving technology is increasingly paid close attention to by people.Wherein, realize that intelligence is driven
The important prerequisite sailed is the sign board information that can automatically get accurate traffic mark board.
In correlation technique, terminal carries out image acquisition by image capture device first, afterwards, judges the image for collecting
In the whether image comprising traffic mark board.When the image comprising traffic mark board in the image for collecting, by the traffic mark
The image of will board carries out similarity-rough set one by one with the multiple traffic mark board templates in sample set, when the figure of the traffic mark board
During as being more than predetermined threshold value with the similarity of at least one traffic mark board template, then at least one traffic mark board mould is chosen
In plate, the traffic mark board template maximum with the image similarity of the traffic mark board be used as final recognition result, and obtain should
The sign board information of traffic mark board.
The content of the invention
To overcome problem present in correlation technique, the disclosure to provide a kind of sign board information getting method and device.
According to the first aspect of the embodiment of the present disclosure, there is provided a kind of sign board information getting method, methods described include：
Hough transformation is carried out to target image, the multiple coordinate points in Hough parameter space are obtained, the target image is
The image of sign board information to be obtained；
Based on the plurality of coordinate points, the figure of traffic mark board is determined by mean shift algorithm from the target image
As region；
The image-region of the traffic mark board is identified, the sign board information of the traffic mark board is obtained.
Alternatively, it is described based on the plurality of coordinate points, friendship is determined from the target image by mean shift algorithm
The image-region of logical sign board, including：
From selection target coordinate points in the plurality of coordinate points, and select positioned at target circle from the plurality of coordinate points
Coordinate points in shape region, the target border circular areas are as the center of circle and with preset length as radius with the coordinates of targets point
Border circular areas；
Based on the coordinates of targets point and the coordinate points in the target border circular areas, mean-shift vector is calculated；
Judge the vector value of the mean-shift vector whether less than predetermined threshold value；
When the vector value of the mean-shift vector is not less than predetermined threshold value, by the vector of the mean-shift vector eventually
Point is defined as the coordinates of targets point, and returns the coordinate points selected from the plurality of coordinate points in target border circular areas
The step of, till determination obtains mean-shift vector of the vector value less than the predetermined threshold value；
Based on for being calculated coordinates of targets point of the vector value less than the mean-shift vector of the predetermined threshold value, from institute
The image-region of traffic mark board is determined in stating target image.
Alternatively, it is described based on the coordinates of targets point and the coordinate points in the target border circular areas, calculate equal
Value offset vector, including：
Determine the coordinates of targets point described；
Based on the coordinates of targets point is in the coordinate in the Hough parameter space and is located in the target border circular areas
Coordinate of the coordinate points in the Hough parameter space, calculate the mean-shift vector by following formula；
Alternatively, the image-region to the traffic mark board is identified, and obtains the mark of the traffic mark board
Will board information, including：
The image-region of the traffic mark board is identified by default convolutional network model, obtains the traffic mark
The sign board information of will board, the convolution kernel of the default convolutional network model, the convolution number of plies and the full connection number of plies are respectively less than specified
The convolution kernel of convolutional network model, the convolution number of plies and connect full the number of plies.
Alternatively, the specified convolution network model is AlexNet network modeies.
According to the second aspect of the embodiment of the present disclosure, there is provided a kind of sign board information acquisition device, described device include：
Conversion module, for carrying out Hough transformation to target image, obtains the multiple coordinate points in Hough parameter space, institute
State the image that target image is sign board information to be obtained；
Determining module, for based on the plurality of coordinate points, being determined from the target image by mean shift algorithm
The image-region of traffic mark board；
Identification module, for being identified to the image-region of the traffic mark board, obtains the traffic mark board
Sign board information.
Alternatively, the determining module includes：
Submodule is selected, for from selection target coordinate points in the plurality of coordinate points, and from the plurality of coordinate points
It is middle to select the coordinate points in the target border circular areas, the target border circular areas be with the coordinates of targets point as the center of circle and with
Border circular areas of the preset length for radius；
Calculating sub module, for based on the coordinates of targets point and the coordinate points in the target border circular areas, counting
Calculate mean-shift vector；
Judging submodule, for judging the vector value of the mean-shift vector whether less than predetermined threshold value；
First determination sub-module, for when the mean-shift vector vector value be not less than predetermined threshold value when, will be described
The vectorial terminal of mean-shift vector is defined as the coordinates of targets point, and return is selected positioned at mesh from the plurality of coordinate points
The step of marking the coordinate points in border circular areas, until determining that obtaining vector value less than the mean-shift vector of the predetermined threshold value is
Only；
Second determination sub-module, for based on for be calculated vector value less than the predetermined threshold value mean shift to
The coordinates of targets point of amount, determines the image-region of traffic mark board from the target image.
Alternatively, the calculating sub module is used for：
Determine coordinate of the coordinates of targets point in the Hough parameter space, and be located at the target border circular areas
Coordinate of the interior coordinate points in the Hough parameter space；
Based on the coordinates of targets point is in the coordinate in the Hough parameter space and is located in the target border circular areas
Coordinate of the coordinate points in the Hough parameter space, calculate the mean-shift vector by following formula；
Wherein, in above-mentioned formula, mh,GFor the mean-shift vector, to preset confidence level, g () is gradient probability to q
Density function, x is coordinate of the coordinates of targets point in the Hough parameter space, xjIt is positioned at the target border circular areas
Coordinate of the interior jth coordinate points in the Hough parameter space, h are the preset length, and the n is positioned at the target circle
The total number of the coordinate points in shape region.
Alternatively, the identification module includes：
Identification submodule, for being known to the image-region of the traffic mark board by default convolutional network model
Not, obtain the sign board information of the traffic mark board, the convolution kernel of the default convolutional network model, the convolution number of plies and connect entirely
Connect the number of plies respectively less than to specify the convolution kernel of convolution network model, the convolution number of plies and connect full the number of plies.
Alternatively, the specified convolution network model is AlexNet network modeies.
The technical scheme that embodiment of the disclosure is provided can include following beneficial effect：The embodiment of the present disclosure adopts Hough
Conversion is detected to target image, and reduces phase determining the image-region of traffic mark board by mean shift algorithm
When detecting to target image only with Hough transformation in the technology of pass, due to high light, the impact of external environment condition such as block, it is difficult to
Accurately determine traffic mark board image-region and the caused flase drop to target image, improve the image district of traffic mark board
The accuracy of detection in domain, reduces false drop rate.
It should be appreciated that the general description of the above and detailed description hereinafter are only exemplary and explanatory, not
The disclosure can be limited.
Description of the drawings
During accompanying drawing herein is merged in description and the part of this specification is constituted, show the enforcement for meeting the present invention
Example, and be used for explaining the principle of the present invention together with description.
Fig. 1 is a kind of flow chart of the sign board information getting method according to an exemplary embodiment.
Fig. 2 is a kind of flow chart of the sign board information getting method according to an exemplary embodiment.
Fig. 3 A are a kind of block diagrams of the sign board information acquisition device according to an exemplary embodiment.
Fig. 3 B are a kind of block diagrams of the determining module according to an exemplary embodiment.
Fig. 4 is a kind of block diagram of the sign board information acquisition device according to an exemplary embodiment.
Specific embodiment
Here in detail exemplary embodiment will be illustrated, its example is illustrated in the accompanying drawings.Explained below is related to
During accompanying drawing, unless otherwise indicated, the same numbers in different accompanying drawings represent same or analogous key element.Following exemplary embodiment
Described in embodiment do not represent and the consistent all embodiments of the present invention.Conversely, they be only with as appended by
The example of consistent apparatus and method in terms of some described in detail in claims, the present invention.
Before detailed explanation is carried out to the embodiment of the present disclosure, first the application scenarios of the embodiment of the present disclosure are given
Introduce.With the development of the information processing technology, intelligent driving technology is increasingly paid close attention to by people.Wherein, realize intelligent driving
An important prerequisite be the sign board information that can automatically get accurate traffic mark board.
In the related, terminal can be detected to the image-region of traffic mark board using Hough transformation, but
Due to the impact of the external environment condition such as high light, deformation during photographic subjects image, the inspection for being detected using Hough transformation merely
Survey precision relatively low, false drop rate is high.When the image-region to traffic mark board is identified, terminal can be calculated using template matching
Method or depth convolutional network.Template matching algorithm accuracy of identification is relatively low, when running into high light, blocking, identification
Mistake phenomenon is than more serious, and depth convolutional network model is more due to parameter, and model is complicated, therefore recognition speed is slower, and
Committed memory is serious, is particularly unsuitable for the mobile terminal of similar mobile phone etc.
In order to solve the above problems, the embodiment of the present disclosure provides a kind of sign board information getting method, and the method is right
While target image is detected using Hough transformation, the image district of traffic mark board is determined also using mean shift algorithm
Domain, improves the accuracy of detection of the image-region of traffic mark board, reduces false drop rate.Afterwards, terminal by using convolution kernel,
The convolution number of plies, the full connection number of plies respectively less than specify figure of the default convolutional network model of convolution network model to the traffic mark board
As region is identified, on the premise of accuracy of identification is ensured, the complexity of convolutional network model is reduced, improve identification should
The speed of the image-region of traffic mark board, reduces the memory usage of terminal.
Fig. 1 is a kind of flow chart of the sign board information getting method according to an exemplary embodiment, such as Fig. 1 institutes
Show, the sign board information getting method is used in terminal, comprises the following steps：
In a step 101, Hough transformation is carried out to target image, obtains the multiple coordinate points in Hough parameter space, should
Target image is the image of sign board information to be obtained.
In a step 102, based on the plurality of coordinate points, traffic mark is determined from the target image by mean shift algorithm
The image-region of will board.
In step 103, the image-region of the traffic mark board is identified, obtains the sign board of the traffic mark board
Information.
The embodiment of the present disclosure is determined by mean shift algorithm and is handed over when being detected to target image using Hough transformation
The image-region of logical sign board, when detecting to target image only with Hough transformation in reducing correlation technique, due to
High light, the impact for the external environment condition such as blocking, it is difficult to accurately determine the image-region of traffic mark board and caused to target image
Flase drop, improve the accuracy of detection of the image-region of traffic mark board, reduce false drop rate.
Alternatively, based on the plurality of coordinate points, traffic mark board is determined by mean shift algorithm from the target image
Image-region, including：
From selection target coordinate points in the plurality of coordinate points, and select positioned at target circle from the plurality of coordinate points
Coordinate points in domain, target border circular areas are the border circular areas with coordinates of targets point as the center of circle and with preset length as radius；
Based on the coordinates of targets point and the coordinate points in the target border circular areas, mean-shift vector is calculated；
Judge the vector value of the mean-shift vector whether less than predetermined threshold value；
When the vector value of the mean-shift vector is not less than predetermined threshold value, will be the vectorial terminal of the mean-shift vector true
It is set to coordinates of targets point, and returns the step of coordinate points in the target border circular areas are selected from the plurality of coordinate points,
Till determination obtains mean-shift vector of the vector value less than predetermined threshold value；
Based on for being calculated coordinates of targets point of the vector value less than the mean-shift vector of predetermined threshold value, from target figure
The image-region of traffic mark board is determined as in.
Alternatively, based on the coordinates of targets point and the coordinate points in the target border circular areas, calculate mean shift to
Amount, including：
Determine the coordinate of the coordinates of targets point in Hough parameter space, and the coordinate in the target border circular areas
Coordinate of the point in Hough parameter space；
Based on the coordinates of targets point in the coordinate in Hough parameter space and the coordinate points in the target border circular areas
Coordinate in Hough parameter space, calculates mean-shift vector by following formula；
Wherein, in above-mentioned formula, mh,GFor mean-shift vector, to preset confidence level, g () is gradient probability density to q
Function, x be coordinate of the coordinates of targets point in Hough parameter space, xjIt is that jth coordinate points in target border circular areas exist
Coordinate in Hough parameter space, h are preset length, and n is the total number of the coordinate points in target border circular areas.
Alternatively, the image-region of the traffic mark board is identified, obtains the sign board information of the traffic mark board,
Including：
The image-region of traffic mark board is identified by default convolutional network model, obtains the mark of traffic mark board
Will board information, convolution kernel, the convolution number of plies and the full connection number of plies for presetting convolutional network model respectively less than specify convolution network model
Convolution kernel, the convolution number of plies and connect full the number of plies.
Alternatively, it is intended that convolutional network model is AlexNet network modeies.
Above-mentioned all optional technical schemes, can be according to the alternative embodiment for arbitrarily combining to form the disclosure, disclosure reality
Apply example no longer to repeat this one by one.
Fig. 2 is a kind of flow chart of the sign board information getting method according to an exemplary embodiment, such as Fig. 2 institutes
Show, the sign board information getting method is used in terminal, comprises the following steps：
In step 201, obtain target image.
Terminal can obtain target image in real time by the photographic head of itself, it is also possible to connect other picture pick-up devices, from which
He obtains target image in picture pick-up device.That is to say, the target image can be that the terminal is entered by self-contained photographic head
Row shoots what is got, the image that such as mobile phone is shot by the photographic head of itself；Can also be other that be connected with the terminal
After picture pick-up device shoots, what the terminal was got from the picture pick-up device.
Further, terminal is after target image is got, can by the target image carry out binary conversion treatment or
It is that the target image is processed into into gray-scale maps.
In step 202., Hough transformation is carried out to target image, obtains the multiple coordinate points in Hough parameter space, should
Target image is the image of sign board information to be obtained.
Alternatively, after terminal target image to be carried out binary conversion treatment or target image is processed into gray-scale maps,
Multiple gray value identical pixels in target image after process can be attached, so as to obtain through the plurality of pixel
The multiple parameters equation of point.Afterwards, the target image after processing is carried out into Hough transformation, so as to obtain in Hough parameter space
Multiple coordinate points.
Wherein, Hough parameter space referred to what the parametric equation of pixel in the target image was obtained Jing after Hough transformation
The space being made up of the parameter in the parametric equation.For example, in target image, the parametric equation of excessively a certain pixel (x, y)
For y=kx+b, then the Hough parameter space after Hough transformation is the space being made up of (k, b).Again for example, in target image
Interior, the parametric equation of excessively a certain pixel (x, y) is (x-a)2+(y-b)2=r2, the Hough parameter space after Hough transformation
It is then the three dimensions being made up of (a, b, r).
It should be noted that Hough transformation is carried out to target image, obtain multiple coordinate points in Hough parameter space
Method may be referred to correlation technique, and the embodiment of the present disclosure is no longer repeated one by one to this.
In step 203, based on the plurality of coordinate points, traffic mark is determined from the target image by mean shift algorithm
The image-region of will board.
Alternatively, this step 203 can be realized by three below step：
(1) from selection target coordinate points in the plurality of coordinate points.
Wherein, terminal can randomly choose a point from the plurality of coordinate points, using the point of the selection as coordinates of targets
Point.
(2) select the coordinate points in the target border circular areas from the plurality of coordinate points, the target border circular areas be with
The coordinates of targets point is the center of circle and the border circular areas with preset length as radius.
After coordinates of targets point is determined, terminal can be with the coordinates of targets point as the center of circle, with preset length h as radius
Border circular areas are drawn, target border circular areas are obtained.Now, select in the target border circular areas from the plurality of coordinate points
Coordinate points.
(3) based on the coordinates of targets point and the coordinate points in the target border circular areas, calculate mean-shift vector.
Alternatively, terminal determines coordinate of the coordinates of targets point in Hough parameter space first, and is located at the target circle
Coordinate of the coordinate points in shape region in the Hough parameter space；Afterwards, based on coordinates of targets point in Hough parameter space
Coordinate, and coordinate of the coordinate points in the target border circular areas in the Hough parameter space, by following formula
Calculate mean-shift vector；
Wherein, in above-mentioned formula, mh,GFor mean-shift vector, to preset confidence level, g () is gradient probability density to q
Function, x be coordinate of the coordinates of targets point in Hough parameter space, xjIt is that jth coordinate points in target border circular areas exist
Coordinate in Hough parameter space, h are preset length, and n is the total number of the coordinate points in target border circular areas.
(4) judge the vector value of the mean-shift vector whether less than predetermined threshold value.
Based on calculated mean-shift vector in step (3), the vector value of the mean-shift vector is calculated, and should
The vector value of mean-shift vector is compared with predetermined threshold value, so that it is determined that whether the vector value of the mean-shift vector is less than
Predetermined threshold value.
(5) when the vector value of the mean-shift vector is not less than predetermined threshold value, by the vector of the mean-shift vector eventually
Point is defined as coordinates of targets point, and return to step (2), until determining the mean-shift vector for obtaining vector value less than predetermined threshold value
Till.
Further, when the vector value of the mean-shift vector is less than the predetermined threshold value, then determine based in step (1)
Coordinates of targets point coordinate, from the target image determine traffic mark board image-region position and size, so as to
To traffic mark board image-region in the target image.
Wherein, the coordinate based on coordinates of targets point, determines the position of the image-region of traffic mark board from the target image
Put and with the process of realizing of size can be：The coordinate of coordinates of targets point is defined as to cross the parameter side of pixel in the target image
Parameter in journey, and the position of the image-region of traffic mark board and big is determined according to the parametric equation from the target image
It is little.
For example, the mean-shift vector is m1, the vector value of the calculated mean-shift vector is | m1|, preset threshold
It is worth for ε, when | m1| during >=ε, by mean-shift vector m1Vectorial terminal as coordinates of targets point, and return to step (2) continues
Mean-shift vector is calculated, till determination obtains mean-shift vector of the vector value less than predetermined threshold value ε.
Further, as | m1| during ＜ ε, then it is used for calculating mean-shift vector m in obtaining step (1)1Coordinates of targets
The coordinate of point, it is assumed that the Hough parameter space is by the parametric equation (x-a) that pixel is crossed in target image2+(y-b)2=r2Enter
Row Hough transformation and come, and the coordinate of the coordinates of targets point be (a0,b0,r0), then the coordinate of the coordinates of targets point is defined as
Parametric equation (x-a)2+(y-b)2=r2In parameter, that is to say a=a0, b=b0, r=r0, now, terminal can be based on the ginseng
Number equations, determine the image-region of the traffic mark board from the target image for circle, and the figure of the circular traffic sign board
As the central coordinate of circle in region is (a0,b0), radius is r0。
(6) based on for be calculated vector value less than the predetermined threshold value mean-shift vector coordinates of targets point, from
The image-region of traffic mark board is determined in the target image.
, will be used for calculating this less than the mean-shift vector of predetermined threshold value based on the final vector value for determining in step (5)
The coordinate of the coordinates of targets point of mean-shift vector is defined as the parameter crossed in the parametric equation of pixel in the target image, and
Position and the size of the image-region of traffic mark board are determined according to the parametric equation from the target image, so as to obtain the friendship
Logical sign board image-region in the target image.
For example, when in step (5) | m1| >=ε, and return to step (2) continues calculating mean-shift vector afterwards, it is final to determine
The vector value for obtaining is m less than the mean-shift vector of predetermined threshold valuejWhen, then based on for calculating mean-shift vector mj's
Coordinates of targets point, determines position and the size of the image-region of traffic mark board from the target image.
The embodiment of the present disclosure is detected to target image by step 202 and step 203, it is determined that traffic mark board
Image-region；Afterwards, the target image can be cut out by terminal, and the figure of the traffic mark board is removed in removing the target image
As other regions outside region, the image-region of the traffic mark board is obtained, and by the figure of the step 204 pair traffic mark board
As region is identified.
In step 204, the image-region of the traffic mark board is identified, obtains the sign board of the traffic mark board
Information.
Alternatively, terminal can be identified to the image-region of traffic mark board by default convolutional network model, obtained
To the sign board information of traffic mark board.Wherein, convolution kernel, the convolution number of plies and the full connection number of plies for presetting convolutional network model is equal
Convolution kernel, the convolution number of plies less than specified convolution network model and connect full the number of plies.
Wherein, the default convolutional network model can reduce specified volume by the convolution kernel to the specified convolution network model
Product check figure, reduces first and specifies numerical value and reduce by the second specified numerical value to full articulamentum and obtain to the convolution number of plies.That is to say,
The default convolutional network model is obtained from by being improved to the specified convolution network model.Wherein, it is intended that convolution kernel
Number, first specifies numerical value and second to specify numerical value to be the numerical value for pre-setting.By this improvement, end can be effectively prevented from
End when using the specified convolution network model as parameter is more, the higher and caused arithmetic speed of model complexity it is relatively slow and
The serious problem of EMS memory occupation.
Further, the specified convolution network model can be AlexNet network modeies.Wherein, AlexNet network modeies
It is a kind of depth convolutional network model proposed by Alex Krizhevsky.The AlexNet network modeies are joined comprising 60,000,000
Number, five layers of convolutional layer and three layers of full articulamentum, the network model be input into foremost be picture original image vegetarian refreshments, it is finally defeated
The result for going out is the recognition result of picture.Generally, the AlexNet network modeies may finally export five recognition results simultaneously,
Wherein, all wrong probability of five recognition results is 15.3%.
For example, the specified convolution network model be AlexNet network modeies, the convolution number of plies of the AlexNet network modeies
For 5, the convolution kernel of the first two convolutional layer is respectively 11 × 11,5 × 5, behind the convolution kernel of three convolutional layers be 3 × 3, Quan Lian
The number of plies is connect for 3；And the default convolutional network model can be the improvement to the AlexNet network modeies, i.e., in the AlexNet nets
On the basis of network model, the convolution number of plies is reduced to into 4, the convolution kernel of each convolutional layer is reduced to 3 × 3, the full connection number of plies is reduced
For 1.As the default convolutional network model employs less convolution kernel, the convolution number of plies and connects full the number of plies, therefore, adopting should
Default convolutional network model is identified to the image-region of the traffic mark board, reduces the complexity of convolutional network model,
The speed of the image-region for recognizing the traffic mark board is improve, the memory usage of terminal is reduced.
In the disclosed embodiments, target image is detected by Hough transformation, and it is true using mean shift algorithm
Determine the image-region of traffic mark board, when detecting to target image only with Hough transformation in reducing correlation technique,
Due to high light, the impact of external environment condition such as block, it is difficult to accurately determine the image-region of traffic mark board and caused to target
The flase drop of image, improves the accuracy of detection of the image-region of traffic mark board, reduces false drop rate.Afterwards, terminal is by adopting
The default convolutional network model of convolution network model is specified respectively less than to the traffic with convolution kernel, the convolution number of plies, the full connection number of plies
The image-region of sign board is identified, and reduces the complexity of convolutional network model, improves and recognizes the traffic mark board
The speed of image-region, reduces the memory usage of terminal.
Fig. 3 A are a kind of 300 block diagrams of sign board information acquisition device according to an exemplary embodiment.Reference picture 3A,
The device includes conversion module 301, determining module 302 and identification module 303.
Conversion module 301, for carrying out Hough transformation to target image, obtains the multiple coordinates in Hough parameter space
Point, the target image are the image of sign board information to be obtained；
Determining module 302, for based on the plurality of coordinate points, friendship being determined from the target image by mean shift algorithm
The image-region of logical sign board；
Identification module 303, for being identified to the image-region of the traffic mark board, obtains the mark of traffic mark board
Board information.
Alternatively, referring to Fig. 3 B, determining module 302 includes：
Submodule 3021 is selected, for from selection target coordinate points in multiple coordinate points, and is selected from multiple coordinate points
The coordinate points in target border circular areas are selected, target border circular areas are with coordinates of targets point as the center of circle and with preset length as half
The border circular areas in footpath；
Calculating sub module 3022, for based on coordinates of targets point and the coordinate points in target border circular areas, calculating equal
Value offset vector；
Judging submodule 3023, for judging the vector value of mean-shift vector whether less than predetermined threshold value；
First determination sub-module 3024, for when the vector value of mean-shift vector is not less than predetermined threshold value, by average
The vectorial terminal of offset vector is defined as coordinates of targets point, and return is selected in target border circular areas from multiple coordinate points
Coordinate points the step of, until determine obtain vector value less than predetermined threshold value mean-shift vector till；
Second determination sub-module 3025, for based on for be calculated vector value less than predetermined threshold value mean shift to
The coordinates of targets point of amount, determines the image-region of traffic mark board from target image.
Alternatively, calculating sub module 3022 is used for：
Determine coordinate of the coordinates of targets point in Hough parameter space, and the coordinate points in target border circular areas exist
Coordinate in Hough parameter space；
Based on coordinates of targets point in the coordinate in Hough parameter space and the coordinate points in target border circular areas suddenly
Coordinate in husband's parameter space, calculates mean-shift vector by following formula；
Wherein, in above-mentioned formula, mh,GFor mean-shift vector, to preset confidence level, g () is gradient probability density to q
Function, x be coordinate of the coordinates of targets point in Hough parameter space, xjIt is that jth coordinate points in target border circular areas exist
Coordinate in Hough parameter space, h are preset length, and n is the total number of the coordinate points in target border circular areas.
Alternatively, identification module 303 includes：
Identification submodule, for being identified to the image-region of traffic mark board by default convolutional network model, is obtained
To the sign board information of traffic mark board, the convolution kernel of default convolutional network model, the convolution number of plies and the full connection number of plies are respectively less than
The convolution kernel of specified convolution network model, the convolution number of plies and connect full the number of plies.
Alternatively, it is intended that convolutional network model is AlexNet network modeies.
In the disclosed embodiments, terminal detected to target image by Hough transformation, and calculated using mean shift
Method determines the image-region of traffic mark board, target image is detected only with Hough transformation in reducing correlation technique
When, due to high light, the impact of external environment condition such as block, it is difficult to accurately determine the image-region of traffic mark board and caused to mesh
The flase drop of logo image, improves the accuracy of detection of the image-region of traffic mark board, reduces false drop rate.Afterwards, terminal passes through
The default convolutional network model of convolution network model is respectively less than specified to the friendship using convolution kernel, the convolution number of plies, the full connection number of plies
The image-region of logical sign board is identified, and reduces the complexity of convolutional network model, improves and recognizes the traffic mark board
Image-region speed, reduce the memory usage of terminal.
With regard to the device in above-described embodiment, wherein modules perform the concrete mode of operation in relevant the method
Embodiment in be described in detail, explanation will be not set forth in detail herein.
Fig. 4 is a kind of block diagram of the device 400 for sign board acquisition of information according to an exemplary embodiment.Example
Such as, device 400 can be mobile phone, and computer, digital broadcast terminal, messaging devices, game console, flat board set
It is standby, armarium, body-building equipment, personal digital assistant etc..
With reference to Fig. 4, device 400 can include following one or more assemblies：Process assembly 402, memorizer 404, power supply
Component 406, multimedia groupware 408, audio-frequency assembly 410, the interface 412 of input/output (I/O), sensor cluster 414, and
Communication component 416.
The integrated operation of 402 usual control device 400 of process assembly, such as with display, call, data communication, phase
Machine operates and records the associated operation of operation.Process assembly 402 can refer to perform including one or more processors 420
Order, to complete all or part of step of above-mentioned method.Additionally, process assembly 402 can include one or more modules, just
Interaction between process assembly 402 and other assemblies.For example, process assembly 402 can include multi-media module, many to facilitate
Interaction between media component 408 and process assembly 402.
Memorizer 404 is configured to store various types of data to support the operation in device 400.These data are shown
Example includes the instruction of any application program or method for operating on device 400, and contact data, telephone book data disappear
Breath, picture, video etc..Memorizer 404 can be by any kind of volatibility or non-volatile memory device or their group
Close and realize, such as static RAM (SRAM), Electrically Erasable Read Only Memory (EEPROM) is erasable to compile
Journey read only memory (EPROM), programmable read only memory (PROM), read only memory (ROM), magnetic memory, flash
Device, disk or CD.
Power supply module 406 provides power supply for the various assemblies of device 400.Power supply module 406 can include power management system
System, one or more power supplys, and other generate, manage and distribute the component that power supply is associated with for device 400.
Multimedia groupware 408 is included in the screen of one output interface of offer between described device 400 and user.One
In a little embodiments, screen can include liquid crystal display (LCD) and touch panel (TP).If screen includes touch panel, screen
Curtain may be implemented as touch screen, to receive the input signal from user.Touch panel includes one or more touch sensings
Device is with the gesture on sensing touch, slip and touch panel.The touch sensor can not only sensing touch or sliding action
Border, but also detect and the touch or slide related persistent period and pressure.In certain embodiments, many matchmakers
Body component 408 includes a front-facing camera and/or post-positioned pick-up head.When device 400 be in operator scheme, such as screening-mode or
During video mode, front-facing camera and/or post-positioned pick-up head can receive outside multi-medium data.Each front-facing camera and
Post-positioned pick-up head can be a fixed optical lens system or have focusing and optical zoom capabilities.
Audio-frequency assembly 410 is configured to output and/or input audio signal.For example, audio-frequency assembly 410 includes a Mike
Wind (MIC), when device 400 is in operator scheme, such as call model, logging mode and speech recognition mode, mike is matched somebody with somebody
It is set to reception external audio signal.The audio signal for being received can be further stored in memorizer 404 or via communication set
Part 416 sends.In certain embodiments, audio-frequency assembly 410 also includes a speaker, for exports audio signal.
, for interface is provided between process assembly 402 and peripheral interface module, above-mentioned peripheral interface module can for I/O interfaces 412
To be keyboard, click wheel, button etc..These buttons may include but be not limited to：Home button, volume button, start button and lock
Determine button.
Sensor cluster 414 includes one or more sensors, and the state for various aspects are provided for device 400 is commented
Estimate.For example, sensor cluster 414 can detect the opening/closed mode of device 400, and the relative localization of component is for example described
Display and keypad of the component for device 400, sensor cluster 414 can be with 400 1 components of detection means 400 or device
Position change, user is presence or absence of with what device 400 was contacted, 400 orientation of device or acceleration/deceleration and device 400
Temperature change.Sensor cluster 414 can include proximity transducer, be configured to detect when not having any physical contact
The presence of object nearby.Sensor cluster 414 can also include optical sensor, such as CMOS or ccd image sensor, for into
As used in application.In certain embodiments, the sensor cluster 414 can also include acceleration transducer, gyro sensors
Device, Magnetic Sensor, pressure transducer or temperature sensor.
Communication component 416 is configured to facilitate the communication of wired or wireless way between device 400 and other equipment.Device
400 can access the wireless network based on communication standard, such as WiFi, 2G or 3G, or combinations thereof.In an exemplary enforcement
In example, communication component 416 receives the broadcast singal or broadcast related information from external broadcasting management system via broadcast channel.
In one exemplary embodiment, the communication component 416 also includes near-field communication (NFC) module, to promote junction service.Example
Such as, NFC module can be based on RF identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra broadband (UWB) technology,
Bluetooth (BT) technology and other technologies are realizing.
In the exemplary embodiment, device 400 can be by one or more application specific integrated circuits (ASIC), numeral letter
Number processor (DSP), digital signal processing appts (DSPD), PLD (PLD), field programmable gate array
(FPGA), controller, microcontroller, microprocessor or other electronic components realizations, for performing said method.
In the exemplary embodiment, a kind of non-transitorycomputer readable storage medium including instruction, example are additionally provided
Such as include the memorizer 404 of instruction, above-mentioned instruction can be performed to complete said method by the processor 420 of device 400.For example,
The non-transitorycomputer readable storage medium can be ROM, random access memory (RAM), CD-ROM, tape, floppy disk
With optical data storage devices etc..
A kind of non-transitorycomputer readable storage medium, when the instruction in the storage medium is by the process of mobile terminal
When device is performed so that mobile terminal is able to carry out a kind of sign board information getting method, and methods described includes：
Hough transformation is carried out to target image, the multiple coordinate points in Hough parameter space are obtained, the target image is
The image of sign board information to be obtained；
Based on the plurality of coordinate points, the figure of traffic mark board is determined by mean shift algorithm from the target image
As region；
The image-region of the traffic mark board is identified, the sign board information of the traffic mark board is obtained.
Alternatively, it is described based on the plurality of coordinate points, friendship is determined from the target image by mean shift algorithm
The image-region of logical sign board, including：
From selection target coordinate points in the plurality of coordinate points, and select positioned at target circle from the plurality of coordinate points
Coordinate points in shape region, the target border circular areas are as the center of circle and with preset length as radius with the coordinates of targets point
Border circular areas；
Based on the coordinates of targets point and the coordinate points in the target border circular areas, mean-shift vector is calculated；
Judge the vector value of the mean-shift vector whether less than predetermined threshold value；
When the vector value of the mean-shift vector is not less than predetermined threshold value, by the vector of the mean-shift vector eventually
Point is defined as the coordinates of targets point, and returns the coordinate points selected from the plurality of coordinate points in target border circular areas
The step of, till determination obtains mean-shift vector of the vector value less than the predetermined threshold value；
Based on for being calculated coordinates of targets point of the vector value less than the mean-shift vector of the predetermined threshold value, from institute
The image-region of traffic mark board is determined in stating target image.
Alternatively, it is described based on the coordinates of targets point and the coordinate points in the target border circular areas, calculate equal
Value offset vector, including：
Determine the coordinates of targets point described；
Based on the coordinates of targets point is in the coordinate in the Hough parameter space and is located in the target border circular areas
Coordinate of the coordinate points in the Hough parameter space, calculate the mean-shift vector by following formula；
Alternatively, the image-region to the traffic mark board is identified, and obtains the mark of the traffic mark board
Will board information, including：
The image-region of the traffic mark board is identified by default convolutional network model, obtains the traffic mark
The sign board information of will board, the convolution kernel of the default convolutional network model, the convolution number of plies and the full connection number of plies are respectively less than specified
The convolution kernel of convolutional network model, the convolution number of plies and connect full the number of plies.
Alternatively, the specified convolution network model is AlexNet network modeies.
In the disclosed embodiments, terminal detected to target image by Hough transformation, and calculated using mean shift
Method determines the image-region of traffic mark board, target image is detected only with Hough transformation in reducing correlation technique
When, due to high light, the impact of external environment condition such as block, it is difficult to accurately determine the image-region of traffic mark board and caused to mesh
The flase drop of logo image, improves the accuracy of detection of the image-region of traffic mark board, reduces false drop rate.Afterwards, terminal passes through
The default convolutional network model of convolution network model is respectively less than specified to the friendship using convolution kernel, the convolution number of plies, the full connection number of plies
The image-region of logical sign board is identified, and reduces the complexity of convolutional network model, improves and recognizes the traffic mark board
Image-region speed, reduce the memory usage of terminal.
Those skilled in the art will readily occur to its of the present invention after considering description and putting into practice invention disclosed herein
Its embodiment.The application is intended to any modification of the present invention, purposes or adaptations, these modifications, purposes or
Person's adaptations follow the general principle of the present invention and including the undocumented common knowledge in the art of the disclosure
Or conventional techniques.Description and embodiments are considered only as exemplary, and true scope and spirit of the invention are by following
Claim is pointed out.
It should be appreciated that the precision architecture for being described above and being shown in the drawings is the invention is not limited in, and
And various modifications and changes can be being carried out without departing from the scope.The scope of the present invention is limited only by appended claim.
Claims (11)
1. a kind of sign board information getting method, it is characterised in that methods described includes：
Hough transformation is carried out to target image, the multiple coordinate points in Hough parameter space are obtained, the target image is to wait to obtain
Take the image of sign board information；
Based on the plurality of coordinate points, the image district of traffic mark board is determined by mean shift algorithm from the target image
Domain；
The image-region of the traffic mark board is identified, the sign board information of the traffic mark board is obtained.
2. method according to claim 1, it is characterised in that described based on the plurality of coordinate points, by mean shift
Algorithm determines the image-region of traffic mark board from the target image, including：
From selection target coordinate points in the plurality of coordinate points, and select positioned at target circle from the plurality of coordinate points
Coordinate points in domain, the target border circular areas are the circle with the coordinates of targets point as the center of circle and with preset length as radius
Region；
Based on the coordinates of targets point and the coordinate points in the target border circular areas, mean-shift vector is calculated；
Judge the vector value of the mean-shift vector whether less than predetermined threshold value；
When the vector value of the mean-shift vector is not less than predetermined threshold value, will be the vectorial terminal of the mean-shift vector true
It is set to the coordinates of targets point, and returns the step that the coordinate points in target border circular areas are selected from the plurality of coordinate points
Suddenly, till determination obtains mean-shift vector of the vector value less than the predetermined threshold value；
Based on for being calculated coordinates of targets point of the vector value less than the mean-shift vector of the predetermined threshold value, from the mesh
The image-region of traffic mark board is determined in logo image.
3. method according to claim 2, it is characterised in that described based on the coordinates of targets point and to be located at the target
Coordinate points in border circular areas, calculate mean-shift vector, including：
Determine coordinate of the coordinates of targets point in the Hough parameter space, and in the target border circular areas
Coordinate of the coordinate points in the Hough parameter space；
Based on the coordinates of targets point in the coordinate in the Hough parameter space and the seat in the target border circular areas
Coordinate of the punctuate in the Hough parameter space, calculates the mean-shift vector by following formula；
Wherein, in above-mentioned formula, mh,GFor the mean-shift vector, to preset confidence level, g () is gradient probability density to q
Function, x is coordinate of the coordinates of targets point in the Hough parameter space, xjIt is in the target border circular areas
Coordinate of the jth coordinate points in the Hough parameter space, h are the preset length, and the n is positioned at the target circle
The total number of the coordinate points in domain.
4. according to the arbitrary described method of claim 1-3, it is characterised in that the image-region to the traffic mark board
It is identified, obtains the sign board information of the traffic mark board, including：
The image-region of the traffic mark board is identified by default convolutional network model, obtains the traffic mark board
Sign board information, the convolution kernel of the default convolutional network model, the convolution number of plies and the full connection number of plies respectively less than specify convolution
The convolution kernel of network model, the convolution number of plies and connect full the number of plies.
5. method according to claim 4, it is characterised in that the specified convolution network model is AlexNet network moulds
Type.
6. a kind of sign board information acquisition device, it is characterised in that described device includes：
Conversion module, for carrying out Hough transformation to target image, obtains the multiple coordinate points in Hough parameter space, the mesh
Logo image is the image of sign board information to be obtained；
Determining module, for based on the plurality of coordinate points, determining traffic by mean shift algorithm from the target image
The image-region of sign board；
Identification module, for being identified to the image-region of the traffic mark board, obtains the mark of the traffic mark board
Board information.
7. device according to claim 6, it is characterised in that the determining module includes：
Submodule is selected, for from selection target coordinate points in the plurality of coordinate points, and is selected from the plurality of coordinate points
The coordinate points in target border circular areas are selected, the target border circular areas are with the coordinates of targets point as the center of circle and with default
Border circular areas of the length for radius；
Calculating sub module, for based on the coordinates of targets point and the coordinate points in the target border circular areas, calculating equal
Value offset vector；
Judging submodule, for judging the vector value of the mean-shift vector whether less than predetermined threshold value；
First determination sub-module, for when the vector value of the mean-shift vector is not less than predetermined threshold value, by the average
The vectorial terminal of offset vector is defined as the coordinates of targets point, and return is selected positioned at target circle from the plurality of coordinate points
The step of coordinate points in shape region, until determining till obtaining mean-shift vector of the vector value less than the predetermined threshold value；
Second determination sub-module, for based on for being calculated mean-shift vector of the vector value less than the predetermined threshold value
Coordinates of targets point, determines the image-region of traffic mark board from the target image.
8. device according to claim 7, it is characterised in that the calculating sub module is used for：
Determine coordinate of the coordinates of targets point in the Hough parameter space, and in the target border circular areas
Coordinate of the coordinate points in the Hough parameter space；
Based on the coordinates of targets point in the coordinate in the Hough parameter space and the seat in the target border circular areas
Coordinate of the punctuate in the Hough parameter space, calculates the mean-shift vector by following formula；
Wherein, in above-mentioned formula, mh,GFor the mean-shift vector, to preset confidence level, g () is gradient probability density to q
Function, x is coordinate of the coordinates of targets point in the Hough parameter space, xjIt is in the target border circular areas
Coordinate of the jth coordinate points in the Hough parameter space, h are the preset length, and the n is positioned at the target circle
The total number of the coordinate points in domain.
9. according to the arbitrary described device of claim 6-8, it is characterised in that the identification module includes：
Identification submodule, for being identified to the image-region of the traffic mark board by default convolutional network model, is obtained
To the sign board information of the traffic mark board, the convolution kernel of the default convolutional network model, the convolution number of plies and full articulamentum
Count and be respectively less than the convolution kernel for specifying convolution network model, the convolution number of plies and connect full the number of plies.
10. device according to claim 9, it is characterised in that the specified convolution network model is AlexNet network moulds
Type.
11. a kind of sign board information acquisition devices, it is characterised in that described device includes：
Processor；
For storing the memorizer of processor executable；
Wherein, the processor is configured to：
Hough transformation is carried out to target image, the multiple coordinate points in Hough parameter space are obtained, the target image is to wait to obtain
Take the image of sign board information；
Based on the plurality of coordinate points, the image district of traffic mark board is determined by mean shift algorithm from the target image
Domain；
The image-region of the traffic mark board is identified, the sign board information of the traffic mark board is obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611063821.5A CN106557759B (en) | 2016-11-25 | 2016-11-25 | Signpost information acquisition method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201611063821.5A CN106557759B (en) | 2016-11-25 | 2016-11-25 | Signpost information acquisition method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106557759A true CN106557759A (en) | 2017-04-05 |
| CN106557759B CN106557759B (en) | 2020-03-17 |
Family
ID=58445199
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201611063821.5A Active CN106557759B (en) | 2016-11-25 | 2016-11-25 | Signpost information acquisition method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106557759B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107122737A (en) * | 2017-04-26 | 2017-09-01 | 聊城大学 | A kind of road signs automatic detection recognition methods |
| CN110163153A (en) * | 2019-05-23 | 2019-08-23 | 百度在线网络技术（北京）有限公司 | The method and device on traffic mark board boundary for identification |
| CN110287767A (en) * | 2019-05-06 | 2019-09-27 | 深圳市华付信息技术有限公司 | Can attack protection biopsy method, device, computer equipment and storage medium |
| CN111275000A (en) * | 2020-02-18 | 2020-06-12 | 广州敏视数码科技有限公司 | Traffic sign board detection method based on historical positioning data |
| CN112036421A (en) * | 2019-05-16 | 2020-12-04 | 搜狗（杭州）智能科技有限公司 | Image processing method and device and electronic equipment |
| CN112348876A (en) * | 2019-08-08 | 2021-02-09 | 北京地平线机器人技术研发有限公司 | Method and device for acquiring space coordinates of signboards |
| CN113361303A (en) * | 2020-03-05 | 2021-09-07 | 百度在线网络技术（北京）有限公司 | Temporary traffic sign board identification method, device and equipment |
| US11636306B2 (en) * | 2018-05-21 | 2023-04-25 | Imagination Technologies Limited | Implementing traditional computer vision algorithms as neural networks |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102008057675A1 (en) * | 2008-11-17 | 2009-07-02 | Daimler Ag | Traffic sign recognizing method for driver of vehicle, involves assigning additional character to traffic sign, searching additional character in given positions, and detecting additional character by edge detector and grey value detector |
| US20110122226A1 (en) * | 2009-03-16 | 2011-05-26 | Siemens Corporation | System and method for robust 2d-3d image registration |
| CN102147857A (en) * | 2011-03-22 | 2011-08-10 | 黄晓华 | Image processing method for detecting similar round by using improved hough transformation |
| CN103854279A (en) * | 2012-12-03 | 2014-06-11 | 美国亚德诺半导体公司 | Hough transform for circles |
| CN105550632A (en) * | 2014-10-23 | 2016-05-04 | 现代摩比斯株式会社 | Traffic sign recognizing apparatus and operating method thereof |
| CN106022300A (en) * | 2016-06-02 | 2016-10-12 | 中国科学院信息工程研究所 | Traffic sign identifying method and traffic sign identifying system based on cascading deep learning |
-
2016
- 2016-11-25 CN CN201611063821.5A patent/CN106557759B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102008057675A1 (en) * | 2008-11-17 | 2009-07-02 | Daimler Ag | Traffic sign recognizing method for driver of vehicle, involves assigning additional character to traffic sign, searching additional character in given positions, and detecting additional character by edge detector and grey value detector |
| US20110122226A1 (en) * | 2009-03-16 | 2011-05-26 | Siemens Corporation | System and method for robust 2d-3d image registration |
| CN102147857A (en) * | 2011-03-22 | 2011-08-10 | 黄晓华 | Image processing method for detecting similar round by using improved hough transformation |
| CN103854279A (en) * | 2012-12-03 | 2014-06-11 | 美国亚德诺半导体公司 | Hough transform for circles |
| CN105550632A (en) * | 2014-10-23 | 2016-05-04 | 现代摩比斯株式会社 | Traffic sign recognizing apparatus and operating method thereof |
| CN106022300A (en) * | 2016-06-02 | 2016-10-12 | 中国科学院信息工程研究所 | Traffic sign identifying method and traffic sign identifying system based on cascading deep learning |
Non-Patent Citations (4)
| Title |
|---|
| ANDRZEJ RUTA等: "In-vehicle camera traffic sign detection and recognition", 《MACHINE VISION AND APPLICATIONS》 * |
| 李斌: "基于DPM的交通标志检测算法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 江春宇: "交通标志的自动检测和识别", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 赵永强 等: "《成像偏振光谱遥感及应用》", 31 October 2011, 国防工业出版社 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107122737A (en) * | 2017-04-26 | 2017-09-01 | 聊城大学 | A kind of road signs automatic detection recognition methods |
| CN107122737B (en) * | 2017-04-26 | 2020-07-31 | 聊城大学 | Automatic detection and identification method for road traffic signs |
| US11636306B2 (en) * | 2018-05-21 | 2023-04-25 | Imagination Technologies Limited | Implementing traditional computer vision algorithms as neural networks |
| CN110287767A (en) * | 2019-05-06 | 2019-09-27 | 深圳市华付信息技术有限公司 | Can attack protection biopsy method, device, computer equipment and storage medium |
| CN112036421A (en) * | 2019-05-16 | 2020-12-04 | 搜狗（杭州）智能科技有限公司 | Image processing method and device and electronic equipment |
| CN110163153A (en) * | 2019-05-23 | 2019-08-23 | 百度在线网络技术（北京）有限公司 | The method and device on traffic mark board boundary for identification |
| CN112348876A (en) * | 2019-08-08 | 2021-02-09 | 北京地平线机器人技术研发有限公司 | Method and device for acquiring space coordinates of signboards |
| CN111275000A (en) * | 2020-02-18 | 2020-06-12 | 广州敏视数码科技有限公司 | Traffic sign board detection method based on historical positioning data |
| CN111275000B (en) * | 2020-02-18 | 2023-05-02 | 广州敏视数码科技有限公司 | Traffic sign board detection method based on historical positioning data |
| CN113361303A (en) * | 2020-03-05 | 2021-09-07 | 百度在线网络技术（北京）有限公司 | Temporary traffic sign board identification method, device and equipment |
| CN113361303B (en) * | 2020-03-05 | 2023-06-23 | 百度在线网络技术（北京）有限公司 | Temporary traffic sign board identification method, device and equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106557759B (en) | 2020-03-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106557759A (en) | A kind of sign board information getting method and device | |
| CN106503617A (en) | Model training method and device | |
| CN106651955A (en) | Method and device for positioning object in picture | |
| CN106295511B (en) | Face tracking method and device | |
| CN107798669A (en) | Image defogging method, device and computer-readable recording medium | |
| CN108010060A (en) | Object detection method and device | |
| CN108629354A (en) | Object detection method and device | |
| CN107832741A (en) | The method, apparatus and computer-readable recording medium of facial modeling | |
| CN106355573A (en) | Target object positioning method and device in pictures | |
| CN107909113A (en) | Traffic-accident image processing method, device and storage medium | |
| CN105528607A (en) | Region extraction method and model training method and device | |
| CN107480665A (en) | Character detecting method, device and computer-readable recording medium | |
| CN104484871B (en) | edge extracting method and device | |
| CN104077597B (en) | Image classification method and device | |
| CN106548201A (en) | The training method of convolutional neural networks, image-recognizing method and device | |
| CN106228158A (en) | The method and apparatus of picture detection | |
| CN106225764A (en) | Based on the distance-finding method of binocular camera in terminal and terminal | |
| CN104268864B (en) | Card edge extracting method and device | |
| CN106228556A (en) | Image quality analysis method and device | |
| CN107563994A (en) | The conspicuousness detection method and device of image | |
| CN109978891A (en) | Image processing method and device, electronic equipment and storage medium | |
| CN106372603A (en) | Shielding face identification method and shielding face identification device | |
| CN109360197A (en) | Processing method, device, electronic equipment and the storage medium of image | |
| CN107766820A (en) | Image classification method and device | |
| CN106446946A (en) | Image recognition method and device |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |