CN106557759B - Signpost information acquisition method and device - Google Patents
Signpost information acquisition method and device Download PDFInfo
- Publication number
- CN106557759B CN106557759B CN201611063821.5A CN201611063821A CN106557759B CN 106557759 B CN106557759 B CN 106557759B CN 201611063821 A CN201611063821 A CN 201611063821A CN 106557759 B CN106557759 B CN 106557759B
- Authority
- CN
- China
- Prior art keywords
- target
- coordinate point
- mean shift
- image
- traffic sign
- 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.)
- Active
Links
Images
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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Traffic Control Systems (AREA)
- Image Analysis (AREA)
Abstract
The disclosure relates to a signboard information acquisition method and a device, relating to the technical field of information processing, wherein the method comprises the following steps: carrying out Hough transformation on a target image to obtain a plurality of coordinate points in Hough parameter space, wherein the target image is an image of signboard information to be obtained; determining an image area of a traffic sign from the target image through a mean shift algorithm based on the plurality of coordinate points; and identifying the image area of the traffic sign board to obtain the sign board information of the traffic sign board. According to the traffic sign information detection method and device, the target image is detected through Hough transform, the image area of the traffic sign is determined through the mean shift algorithm, then the sign information of the traffic sign is obtained through recognition of the image area of the traffic sign, the detection precision of the image area of the traffic sign is improved, and the false detection rate is reduced.
Description
Technical Field
The present disclosure relates to the field of information processing technologies, and in particular, to a signboard information acquiring method and apparatus.
Background
With the development of information processing technology, intelligent driving technology is more and more concerned by people. One important premise for realizing intelligent driving is that accurate signboard information of the traffic signboard can be automatically acquired.
In the related art, a terminal firstly performs image acquisition through an image acquisition device, and then judges whether the acquired image contains an image of a traffic sign. When the collected images contain the images of the traffic signs, the images of the traffic signs are compared with the plurality of traffic sign templates in the sample set one by one in similarity, and when the similarity between the images of the traffic signs and at least one traffic sign template is larger than a preset threshold value, the traffic sign template with the maximum similarity to the images of the traffic signs in the at least one traffic sign template is selected as a final recognition result, and the sign information of the traffic signs is obtained.
Disclosure of Invention
In order to overcome the problems in the related art, the present disclosure provides a signboard information acquisition method and apparatus.
According to a first aspect of embodiments of the present disclosure, there is provided a signboard information acquisition method, including:
carrying out Hough transformation on a target image to obtain a plurality of coordinate points in Hough parameter space, wherein the target image is an image of signboard information to be obtained;
determining an image area of a traffic sign from the target image through a mean shift algorithm based on the plurality of coordinate points;
and identifying the image area of the traffic sign board to obtain the sign board information of the traffic sign board.
Optionally, the determining, from the target image, an image area of the traffic sign through a mean shift algorithm based on the plurality of coordinate points includes:
selecting a target coordinate point from the plurality of coordinate points, and selecting a coordinate point located within a target circular area from the plurality of coordinate points, wherein the target circular area is a circular area with the target coordinate point as a circle center and a preset length as a radius;
calculating a mean shift vector based on the target coordinate point and coordinate points located within the target circular region;
judging whether the vector value of the mean shift vector is smaller than a preset threshold value or not;
when the vector value of the mean shift vector is not smaller than a preset threshold value, determining the vector end point of the mean shift vector as the target coordinate point, and returning to the step of selecting the coordinate point located in the target circular area from the plurality of coordinate points until the mean shift vector with the vector value smaller than the preset threshold value is determined;
and determining the image area of the traffic sign from the target image based on the target coordinate point of the mean shift vector with the vector value smaller than the preset threshold value.
Optionally, the calculating a mean shift vector based on the target coordinate point and the coordinate points located within the target circular region includes:
determining that the target coordinate point is at the target coordinate point;
calculating the mean shift vector by the following formula based on the coordinates of the target coordinate point in the Hough parameter space and the coordinates of the coordinate points located within the target circular region in the Hough parameter space;
optionally, the recognizing the image area of the traffic sign to obtain the sign information of the traffic sign includes:
and identifying the image area of the traffic sign board through a preset convolution network model to obtain sign board information of the traffic sign board, wherein the convolution kernel, the number of convolution layers and the number of full-connection layers of the preset convolution network model are all smaller than those of the specified convolution network model.
Optionally, the specified convolutional network model is an AlexNet network model.
According to a second aspect of the embodiments of the present disclosure, there is provided a signboard information acquisition apparatus including:
the conversion module is used for carrying out Hough conversion on a target image to obtain a plurality of coordinate points in Hough parameter space, wherein the target image is an image of signboard information to be acquired;
the determining module is used for determining an image area of the traffic sign from the target image through a mean shift algorithm based on the plurality of coordinate points;
and the identification module is used for identifying the image area of the traffic sign board to obtain the sign board information of the traffic sign board.
Optionally, the determining module includes:
the selection submodule is used for selecting a target coordinate point from the coordinate points and selecting a coordinate point located in a target circular area from the coordinate points, and the target circular area is a circular area which takes the target coordinate point as a circle center and takes a preset length as a radius;
a calculation submodule for calculating a mean shift vector based on the target coordinate point and coordinate points located within the target circular region;
the judgment submodule is used for judging whether the vector value of the mean shift vector is smaller than a preset threshold value or not;
a first determining sub-module, configured to determine, when a vector value of the mean shift vector is not less than a preset threshold, a vector end point of the mean shift vector as the target coordinate point, and return to the step of selecting a coordinate point located in a target circular area from the plurality of coordinate points until the mean shift vector having a vector value less than the preset threshold is determined;
and the second determining submodule is used for determining the image area of the traffic sign from the target image based on the target coordinate point of the mean shift vector of which the vector value is smaller than the preset threshold value.
Optionally, the computing sub-module is configured to:
determining coordinates of the target coordinate point in the Hough parameter space and coordinates of coordinate points located in the target circular area in the Hough parameter space;
calculating the mean shift vector by the following formula based on the coordinates of the target coordinate point in the Hough parameter space and the coordinates of the coordinate points located within the target circular region in the Hough parameter space;
wherein, in the above formula, mh,GFor the mean shift vector, q is a preset confidence coefficient, g (-) is a gradient probability density function, x is the coordinate of the target coordinate point in the Hough parameter space, xjThe coordinates of the j-th coordinate point located in the target circular area in the Hough parameter space are obtained, h is the preset length, and n is the total number of coordinate points located in the target circular area.
Optionally, the identification module comprises:
the identification submodule is used for identifying the image area of the traffic sign board through a preset convolution network model to obtain the sign board information of the traffic sign board, and the convolution kernel, the number of the convolution layers and the number of the full-connection layers of the preset convolution network model are all smaller than those of the specified convolution network model.
Optionally, the specified convolutional network model is an AlexNet network model.
The technical scheme provided by the embodiment of the disclosure can have the following beneficial effects: according to the traffic sign image detection method and device, the target image is detected through Hough transform, the image area of the traffic sign is determined through the mean shift algorithm, false detection of the target image caused by the fact that the image area of the traffic sign is difficult to accurately determine due to the influence of external environments such as strong light and shielding when the target image is detected through Hough transform in the related technology is reduced, the detection precision of the image area of the traffic sign is improved, and the false detection rate is reduced.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.
Fig. 1 is a flowchart illustrating a signboard information acquisition method according to an exemplary embodiment.
Fig. 2 is a flowchart illustrating a signboard information acquisition method according to an exemplary embodiment.
Fig. 3A is a block diagram illustrating a signboard information acquisition apparatus according to an exemplary embodiment.
FIG. 3B is a block diagram illustrating a determination module in accordance with an exemplary embodiment.
Fig. 4 is a block diagram illustrating a signboard information acquisition apparatus according to an exemplary embodiment.
Detailed Description
Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the invention, as detailed in the appended claims.
Before explaining the embodiments of the present disclosure in detail, an application scenario of the embodiments of the present disclosure will be described. With the development of information processing technology, intelligent driving technology is more and more concerned by people. One important premise for realizing intelligent driving is that accurate signboard information of the traffic signboard can be automatically acquired.
In the related art, the terminal can detect the image area of the traffic sign by using the hough transform, but due to the influence of external environments such as strong light, deformation and the like when the target image is shot, the detection accuracy by simply adopting the hough transform is low, and the false detection rate is high. When the image area of the traffic sign is identified, the terminal can adopt a template matching algorithm or a deep convolution network. The template matching algorithm has low identification precision, and has serious identification error phenomenon when encountering strong light, shielding and other conditions, while the deep convolution network model has slow identification speed and serious memory occupation because of more parameters and complex model, and is particularly not suitable for mobile terminals such as mobile phones.
In order to solve the above problem, the embodiments of the present disclosure provide a method for acquiring information of a traffic sign, where the method determines an image area of the traffic sign by using a mean shift algorithm while detecting a target image by using hough transform, so as to improve the detection accuracy of the image area of the traffic sign and reduce the false detection rate. Then, the terminal identifies the image area of the traffic sign board by adopting a preset convolution network model with a convolution kernel, the number of convolution layers and the number of full-connection layers all being smaller than the number of appointed convolution network models, so that the complexity of the convolution network model is reduced on the premise of ensuring the identification precision, the speed of identifying the image area of the traffic sign board is improved, and the memory occupancy rate of the terminal is reduced.
Fig. 1 is a flowchart illustrating a signboard information acquisition method according to an exemplary embodiment, which is used in a terminal, as shown in fig. 1, and includes the steps of:
in step 101, hough transformation is performed on a target image to obtain a plurality of coordinate points in a hough parameter space, wherein the target image is an image of signboard information to be acquired.
In step 102, an image area of the traffic sign is determined from the target image by a mean shift algorithm based on the plurality of coordinate points.
In step 103, the image area of the traffic signboard is recognized, and the signboard information of the traffic signboard is obtained.
According to the embodiment of the invention, when the target image is detected by adopting Hough transform, the image area of the traffic sign board is determined by the mean shift algorithm, so that the false detection of the target image caused by the fact that the image area of the traffic sign board is difficult to be accurately determined due to the influence of external environments such as strong light, shielding and the like when the target image is detected by only adopting Hough transform in the related art is reduced, the detection precision of the image area of the traffic sign board is improved, and the false detection rate is reduced.
Optionally, determining an image area of the traffic sign from the target image by a mean shift algorithm based on the plurality of coordinate points includes:
selecting a target coordinate point from the plurality of coordinate points, and selecting a coordinate point located within a target circular area from the plurality of coordinate points, wherein the target circular area is a circular area with the target coordinate point as a circle center and a preset length as a radius;
calculating a mean shift vector based on the target coordinate point and coordinate points located within the target circular region;
judging whether the vector value of the mean shift vector is smaller than a preset threshold value or not;
when the vector value of the mean shift vector is not less than a preset threshold value, determining the vector end point of the mean shift vector as a target coordinate point, and returning to the step of selecting the coordinate point located in the target circular area from the plurality of coordinate points until the mean shift vector with the vector value less than the preset threshold value is determined;
and determining the image area of the traffic sign from the target image based on the target coordinate point for calculating the mean shift vector with the vector value smaller than the preset threshold value.
Optionally, calculating a mean shift vector based on the target coordinate point and coordinate points located within the target circular region, including:
determining the coordinates of the target coordinate point in the Hough parameter space and the coordinates of the coordinate point located in the target circular area in the Hough parameter space;
calculating a mean shift vector by the following formula based on the coordinates of the target coordinate point in the Hough parameter space and the coordinates of the coordinate point located in the target circular area in the Hough parameter space;
wherein, in the above formula, mh,GIs a mean shift vector, q is a preset confidence coefficient, g (-) is a gradient probability density function, x is the coordinate of a target coordinate point in the Hough parameter space, xjThe coordinates of the j-th coordinate point in the target circular area in the Hough parameter space are shown, h is a preset length, and n is the total number of the coordinate points in the target circular area.
Optionally, recognizing the image area of the traffic signboard to obtain the signboard information of the traffic signboard, including:
and identifying the image area of the traffic sign board through a preset convolution network model to obtain sign board information of the traffic sign board, wherein the convolution kernel, the number of convolution layers and the number of full-connection layers of the preset convolution network model are all smaller than those of the specified convolution network model.
Optionally, the convolutional network model is designated as an AlexNet network model.
All the above optional technical solutions can be combined arbitrarily to form optional embodiments of the present disclosure, and the embodiments of the present disclosure are not described in detail again.
Fig. 2 is a flowchart illustrating a signboard information acquisition method according to an exemplary embodiment, which is used in a terminal, as shown in fig. 2, and includes the steps of:
in step 201, a target image is acquired.
The terminal can acquire the target image in real time through a camera of the terminal, and can also be connected with other camera equipment to acquire the target image from the other camera equipment. That is, the target image may be obtained by shooting through a camera carried by the terminal, for example, an image shot by a mobile phone through the camera of the mobile phone; the terminal may acquire the image from another image capturing apparatus connected to the terminal after capturing the image.
Further, after acquiring the target image, the terminal may perform binarization processing on the target image or process the target image into a gray scale image.
In step 202, hough transform is performed on a target image to obtain a plurality of coordinate points in a hough parameter space, where the target image is an image of signboard information to be acquired.
Optionally, after the terminal performs binarization processing on the target image or processes the target image into a gray-scale image, a plurality of pixel points with the same gray value in the processed target image may be connected, so as to obtain a plurality of parameter equations passing through the plurality of pixel points. And then, carrying out Hough transformation on the processed target image so as to obtain a plurality of coordinate points in Hough parameter space.
The Hough parameter space is a space formed by parameters in a parameter equation obtained by Hough transformation of the parameter equation of a pixel point in the target image. For example, in the target image, if the parameter equation of a certain pixel (x, y) is y ═ kx + b, then the hough parameter space after hough transform is a space composed of (k, b). As another example, in the target image, the parameter equation passing a certain pixel point (x, y) is (x-a)2+(y-b)2＝r2The hough parameter space after hough transform is a three-dimensional space composed of (a, b, r).
It should be noted that, reference may be made to related technologies for a method of performing hough transformation on a target image to obtain a plurality of coordinate points in a hough parameter space, and details of this method are not repeated in this disclosure.
In step 203, an image area of the traffic sign is determined from the target image by a mean shift algorithm based on the plurality of coordinate points.
Alternatively, this step 203 may be implemented by the following three steps:
(1) a target coordinate point is selected from the plurality of coordinate points.
Wherein the terminal may randomly select one point from the plurality of coordinate points, and set the selected point as the target coordinate point.
(2) Selecting a coordinate point located in a target circular area from the plurality of coordinate points, wherein the target circular area is a circular area which takes the target coordinate point as a circle center and takes a preset length as a radius.
After the target coordinate point is determined, the terminal may draw a circular area with the target coordinate point as a center of a circle and a preset length h as a radius to obtain a target circular area. At this time, a coordinate point located within the target circular area is selected from the plurality of coordinate points.
(3) Based on the target coordinate point and coordinate points located within the target circular region, a mean shift vector is calculated.
Optionally, the terminal first determines coordinates of a target coordinate point in a hough parameter space and coordinates of a coordinate point located in the target circular area in the hough parameter space; then, calculating a mean shift vector by the following formula based on the coordinates of the target coordinate point in the Hough parameter space and the coordinates of the coordinate point located in the target circular area in the Hough parameter space;
wherein, in the above formula, mh,GIs a mean shift vector, q is a preset confidence coefficient, g (-) is a gradient probability density function, x is the coordinate of a target coordinate point in the Hough parameter space, xjTo be located at a target circleAnd the coordinates of the j-th coordinate point in the shape area in the Hough parameter space, h is the preset length, and n is the total number of coordinate points in the target circular area.
(4) And judging whether the vector value of the mean shift vector is smaller than a preset threshold value.
Calculating a vector value of the mean shift vector based on the mean shift vector calculated in the step (3), and comparing the vector value of the mean shift vector with a preset threshold value, thereby determining whether the vector value of the mean shift vector is less than the preset threshold value.
(5) And (3) when the vector value of the mean shift vector is not less than a preset threshold value, determining the vector end point of the mean shift vector as a target coordinate point, and returning to the step (2) until the mean shift vector with the vector value less than the preset threshold value is determined.
Further, when the vector value of the mean shift vector is smaller than the preset threshold, determining the position and size of the image area of the traffic sign from the target image based on the coordinates of the target coordinate point determined in the step (1), so as to obtain the image area of the traffic sign in the target image.
Based on the coordinates of the target coordinate point, the implementation process of determining the position and the size of the image area of the traffic sign from the target image may be as follows: and determining the coordinates of the target coordinate points as parameters in a parameter equation of pixel points in the target image, and determining the position and the size of the image area of the traffic sign from the target image according to the parameter equation.
For example, the mean shift vector is m1The vector value of the mean shift vector is calculated as | m1If the preset threshold is epsilon, then m is1When | ≧ epsilon, shift the mean value by vector m1And (3) taking the vector end point as a target coordinate point, returning to the step (2) to continue calculating the mean shift vector until the mean shift vector with the vector value smaller than a preset threshold epsilon is determined.
Further, when | m1If | < epsilon, obtaining the mean shift vector m used for calculating in step (1)1Of target coordinate pointCoordinates, assuming that the Hough parameter space is defined by the parameter equation (x-a) of the pixel points in the target image2+(y-b)2＝r2Is obtained by Hough transformation, and the coordinate of the target coordinate point is (a)0,b0,r0) Determining the coordinates of the target coordinate point as a parametric equation (x-a)2+(y-b)2＝r2Is a ═ a0，b＝b0，r＝r0At this time, the terminal may determine that the image area of the traffic signboard is a circle from the target image based on the parametric equation, and the coordinates of the center of the circle of the image area of the circular traffic signboard are (a)0,b0) Radius r0。
(6) And determining the image area of the traffic sign from the target image based on the target coordinate point for calculating the mean shift vector with the vector value smaller than the preset threshold value.
And (4) determining the coordinates of a target coordinate point for calculating the mean shift vector as parameters in a parameter equation of pixel points in the target image based on the mean shift vector of which the finally determined vector value is smaller than a preset threshold in the step (5), and determining the position and the size of the image area of the traffic signboard from the target image according to the parameter equation, so as to obtain the image area of the traffic signboard in the target image.
For example, when | m in step (5)1| ≧ epsilon, and continue to calculate the mean shift vector after returning to step (2), finally determining the mean shift vector with the obtained vector value smaller than the preset threshold value as mjThen based on the mean shift vector m used to calculate the mean shift vectorjThe position and size of the image area of the traffic sign are determined from the target image.
The embodiment of the disclosure detects the target image through the steps 202 and 203, and determines the image area of the traffic sign board; then, the terminal may crop the target image, remove other areas except the image area of the traffic sign in the target image, obtain the image area of the traffic sign, and identify the image area of the traffic sign through step 204.
In step 204, the image area of the traffic sign is identified to obtain sign information of the traffic sign.
Optionally, the terminal may identify the image area of the traffic sign through a preset convolutional network model to obtain sign information of the traffic sign. The convolution kernel, the number of convolution layers and the number of full-connection layers of the preset convolution network model are all smaller than those of the specified convolution network model.
The preset convolution network model can be obtained by reducing the number of specified convolution kernels of the specified convolution network model, reducing the number of convolution layers by a first specified value and reducing the number of full connection layers by a second specified value. That is, the predetermined convolutional network model is obtained by modifying the specified convolutional network model. Wherein, the number of the convolution kernels is appointed, and the first appointed value and the second appointed value are preset values. Through the improvement, the problems of low operation speed and serious memory occupation caused by more parameters and higher model complexity when the terminal uses the appointed convolutional network model can be effectively solved.
Further, the specified convolutional network model may be an AlexNet network model. The AlexNet network model is a deep convolution network model proposed by Alex Krizhevsky. The AlexNet network model comprises sixty million parameters, five convolution layers and three full-connection layers, wherein the initial pixel points of the picture are input at the forefront end of the network model, and the final output result is the recognition result of the picture. In general, the AlexNet network model can finally output five recognition results simultaneously, wherein the probability of all errors of the five recognition results is 15.3%.
For example, the specified convolutional network model is an AlexNet network model, the number of convolutional layers of the AlexNet network model is 5, the convolutional kernels of the first two convolutional layers are respectively 11 × 11 and 5 × 5, the convolutional kernels of the last three convolutional layers are all 3 × 3, and the number of fully-connected layers is 3; the preset convolution network model can be an improvement of the AlexNet network model, that is, on the basis of the AlexNet network model, the number of convolution layers is reduced to 4, the convolution kernel of each convolution layer is reduced to 3 × 3, and the number of full connection layers is reduced to 1. Because the preset convolution network model adopts smaller convolution kernels, convolution layer number and full-connection layer number, the image area of the traffic sign is identified by adopting the preset convolution network model, the complexity of the convolution network model is reduced, the speed of identifying the image area of the traffic sign is improved, and the memory occupancy rate of the terminal is reduced.
In the embodiment of the disclosure, the target image is detected through Hough transform, and the image area of the traffic sign is determined by adopting the mean shift algorithm, so that false detection of the target image caused by difficulty in accurately determining the image area of the traffic sign due to the influence of external environments such as strong light, shielding and the like when the target image is detected by only Hough transform in the related art is reduced, the detection precision of the image area of the traffic sign is improved, and the false detection rate is reduced. Then, the terminal identifies the image area of the traffic sign board by adopting a preset convolution network model with a convolution kernel, the number of convolution layers and the number of full-connection layers all being smaller than the number of appointed convolution network models, so that the complexity of the convolution network model is reduced, the speed of identifying the image area of the traffic sign board is improved, and the memory occupancy rate of the terminal is reduced.
Fig. 3A is a block diagram illustrating a sign information acquisition device 300 according to an example embodiment. Referring to fig. 3A, the apparatus includes a transformation module 301, a determination module 302, and an identification module 303.
The transformation module 301 is configured to perform hough transformation on a target image to obtain a plurality of coordinate points in a hough parameter space, where the target image is an image of signboard information to be acquired;
a determining module 302, configured to determine an image area of the traffic sign from the target image through a mean shift algorithm based on the plurality of coordinate points;
the identifying module 303 is configured to identify the image area of the traffic sign to obtain sign information of the traffic sign.
Optionally, referring to fig. 3B, the determining module 302 includes:
a selection submodule 3021 configured to select a target coordinate point from the plurality of coordinate points, and select a coordinate point located within a target circular area from the plurality of coordinate points, the target circular area being a circular area having the target coordinate point as a center and a preset length as a radius;
a calculation submodule 3022 configured to calculate a mean shift vector based on the target coordinate point and the coordinate points located within the target circular area;
a judging submodule 3023, configured to judge whether a vector value of the mean shift vector is smaller than a preset threshold;
a first determining submodule 3024, configured to determine, when the vector value of the mean shift vector is not less than a preset threshold, a vector end point of the mean shift vector as a target coordinate point, and return to the step of selecting a coordinate point located in the target circular area from the plurality of coordinate points until the mean shift vector having a vector value less than the preset threshold is determined;
a second determining submodule 3025 configured to determine an image area of the traffic sign from the target image based on the target coordinate point of the mean shift vector calculated to have a vector value smaller than the preset threshold value.
Optionally, the calculation submodule 3022 is configured to:
determining the coordinates of a target coordinate point in a Hough parameter space and the coordinates of a coordinate point located in a target circular area in the Hough parameter space;
calculating a mean shift vector by the following formula based on the coordinates of the target coordinate point in the Hough parameter space and the coordinates of the coordinate point located in the target circular area in the Hough parameter space;
wherein, in the above formula, mh,GIs a mean shift vector, q is a preset confidence coefficient, g (-) is a gradient probability density function, x is the coordinate of a target coordinate point in the Hough parameter space, xjAs the j-th coordinate point located in the target circular areaAnd h is a preset length, and n is the total number of coordinate points located in the target circular area.
Optionally, the identifying module 303 comprises:
and the identification submodule is used for identifying the image area of the traffic sign board through a preset convolution network model to obtain the sign board information of the traffic sign board, and the convolution kernel, the number of the convolution layers and the number of the full-connection layers of the preset convolution network model are all smaller than those of the specified convolution network model.
Optionally, the convolutional network model is designated as an AlexNet network model.
In the embodiment of the disclosure, the terminal detects the target image through Hough transform, and determines the image area of the traffic sign board by adopting a mean shift algorithm, so that false detection of the target image caused by difficulty in accurately determining the image area of the traffic sign board due to the influence of external environments such as strong light, shielding and the like when the target image is detected by only Hough transform in the related art is reduced, the detection precision of the image area of the traffic sign board is improved, and the false detection rate is reduced. Then, the terminal identifies the image area of the traffic sign board by adopting a preset convolution network model with a convolution kernel, the number of convolution layers and the number of full-connection layers all being smaller than the number of appointed convolution network models, so that the complexity of the convolution network model is reduced, the speed of identifying the image area of the traffic sign board is improved, and the memory occupancy rate of the terminal is reduced.
With regard to the apparatus in the above-described embodiment, the specific manner in which each module performs the operation has been described in detail in the embodiment related to the method, and will not be elaborated here.
Fig. 4 is a block diagram illustrating an apparatus 400 for sign information acquisition according to an example embodiment. For example, the apparatus 400 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, an exercise device, a personal digital assistant, and the like.
Referring to fig. 4, the apparatus 400 may include one or more of the following components: processing components 402, memory 404, power components 406, multimedia components 408, audio components 410, input/output (I/O) interfaces 412, sensor components 414, and communication components 416.
The processing component 402 generally controls overall operation of the apparatus 400, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. The processing component 402 may include one or more processors 420 to execute instructions to perform all or a portion of the steps of the methods described above. Further, the processing component 402 can include one or more modules that facilitate interaction between the processing component 402 and other components. For example, the processing component 402 can include a multimedia module to facilitate interaction between the multimedia component 408 and the processing component 402.
The memory 404 is configured to store various types of data to support operations at the apparatus 400. Examples of such data include instructions for any application or method operating on the device 400, contact data, phonebook data, messages, pictures, videos, and so forth. The memory 404 may be implemented by any type or combination of volatile or non-volatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disks.
The multimedia component 408 includes a screen that provides an output interface between the device 400 and the user. In some embodiments, the screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation. In some embodiments, the multimedia component 408 includes a front facing camera and/or a rear facing camera. The front camera and/or the rear camera may receive external multimedia data when the apparatus 400 is in an operation mode, such as a photographing mode or a video mode. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.
The audio component 410 is configured to output and/or input audio signals. For example, audio component 410 includes a Microphone (MIC) configured to receive external audio signals when apparatus 400 is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals may further be stored in the memory 404 or transmitted via the communication component 416. In some embodiments, audio component 410 also includes a speaker for outputting audio signals.
The I/O interface 412 provides an interface between the processing component 402 and peripheral interface modules, which may be keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to: a home button, a volume button, a start button, and a lock button.
The sensor component 414 includes one or more sensors for providing various aspects of status assessment for the apparatus 400. For example, the sensor assembly 414 may detect an open/closed state of the apparatus 400, the relative positioning of the components, such as a display and keypad of the apparatus 400, the sensor assembly 414 may also detect a change in the position of the apparatus 400 or a component of the apparatus 400, the presence or absence of user contact with the apparatus 400, orientation or acceleration/deceleration of the apparatus 400, and a change in the temperature of the apparatus 400. The sensor assembly 414 may include a proximity sensor configured to detect the presence of a nearby object without any physical contact. The sensor assembly 414 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 414 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.
The communication component 416 is configured to facilitate wired or wireless communication between the apparatus 400 and other devices. The apparatus 400 may access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In an exemplary embodiment, the communication component 416 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 416 further includes a Near Field Communication (NFC) module to facilitate short-range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, Ultra Wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.
In an exemplary embodiment, the apparatus 400 may be implemented by one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, micro-controllers, microprocessors or other electronic components for performing the above-described methods.
In an exemplary embodiment, a non-transitory computer-readable storage medium comprising instructions, such as the memory 404 comprising instructions, executable by the processor 420 of the apparatus 400 to perform the above-described method is also provided. For example, the non-transitory computer readable storage medium may be a ROM, a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.
A non-transitory computer-readable storage medium in which instructions, when executed by a processor of a mobile terminal, enable the mobile terminal to perform a signboard information acquisition method, the method comprising:
carrying out Hough transformation on a target image to obtain a plurality of coordinate points in Hough parameter space, wherein the target image is an image of signboard information to be obtained;
determining an image area of a traffic sign from the target image through a mean shift algorithm based on the plurality of coordinate points;
and identifying the image area of the traffic sign board to obtain the sign board information of the traffic sign board.
Optionally, the determining, from the target image, an image area of the traffic sign through a mean shift algorithm based on the plurality of coordinate points includes:
selecting a target coordinate point from the plurality of coordinate points, and selecting a coordinate point located within a target circular area from the plurality of coordinate points, wherein the target circular area is a circular area with the target coordinate point as a circle center and a preset length as a radius;
calculating a mean shift vector based on the target coordinate point and coordinate points located within the target circular region;
judging whether the vector value of the mean shift vector is smaller than a preset threshold value or not;
when the vector value of the mean shift vector is not smaller than a preset threshold value, determining the vector end point of the mean shift vector as the target coordinate point, and returning to the step of selecting the coordinate point located in the target circular area from the plurality of coordinate points until the mean shift vector with the vector value smaller than the preset threshold value is determined;
and determining the image area of the traffic sign from the target image based on the target coordinate point of the mean shift vector with the vector value smaller than the preset threshold value.
Optionally, the calculating a mean shift vector based on the target coordinate point and the coordinate points located within the target circular region includes:
determining that the target coordinate point is at the target coordinate point;
calculating the mean shift vector by the following formula based on the coordinates of the target coordinate point in the Hough parameter space and the coordinates of the coordinate points located within the target circular region in the Hough parameter space;
optionally, the recognizing the image area of the traffic sign to obtain the sign information of the traffic sign includes:
and identifying the image area of the traffic sign board through a preset convolution network model to obtain sign board information of the traffic sign board, wherein the convolution kernel, the number of convolution layers and the number of full-connection layers of the preset convolution network model are all smaller than those of the specified convolution network model.
Optionally, the specified convolutional network model is an AlexNet network model.
In the embodiment of the disclosure, the terminal detects the target image through Hough transform, and determines the image area of the traffic sign board by adopting a mean shift algorithm, so that false detection of the target image caused by difficulty in accurately determining the image area of the traffic sign board due to the influence of external environments such as strong light, shielding and the like when the target image is detected by only Hough transform in the related art is reduced, the detection precision of the image area of the traffic sign board is improved, and the false detection rate is reduced. Then, the terminal identifies the image area of the traffic sign board by adopting a preset convolution network model with a convolution kernel, the number of convolution layers and the number of full-connection layers all being smaller than the number of appointed convolution network models, so that the complexity of the convolution network model is reduced, the speed of identifying the image area of the traffic sign board is improved, and the memory occupancy rate of the terminal is reduced.
Other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the invention and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the invention being indicated by the following claims.
It will be understood that the invention is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the invention is limited only by the appended claims.
Claims (7)
1. A signboard information acquisition method is characterized by comprising the following steps:
carrying out Hough transformation on a target image to obtain a plurality of coordinate points in Hough parameter space, wherein the target image is an image of signboard information to be obtained;
selecting a target coordinate point from the plurality of coordinate points, and selecting a coordinate point located within a target circular area from the plurality of coordinate points, wherein the target circular area is a circular area with the target coordinate point as a circle center and a preset length as a radius;
determining coordinates of the target coordinate point in the Hough parameter space and coordinates of coordinate points located in the target circular area in the Hough parameter space;
calculating a mean shift vector by the following formula based on the coordinates of the target coordinate point in the Hough parameter space and the coordinates of the coordinate points located within the target circular region in the Hough parameter space;
wherein, in the above formula, mh,GFor the mean shift vector, q is a preset confidence coefficient, g (-) is a gradient probability density function, x is the coordinate of the target coordinate point in the Hough parameter space, xjThe coordinates of the j-th coordinate point in the target circular area in the Hough parameter space are obtained, h is the preset length, and n is the total number of coordinate points in the target circular area;
judging whether the vector value of the mean shift vector is smaller than a preset threshold value or not;
when the vector value of the mean shift vector is not smaller than a preset threshold value, determining the vector end point of the mean shift vector as the target coordinate point, and returning to the step of selecting the coordinate point located in the target circular area from the plurality of coordinate points until the mean shift vector with the vector value smaller than the preset threshold value is determined;
determining the coordinates of a target coordinate point of a mean shift vector with a vector value smaller than the preset threshold value as parameters in a parameter equation of a pixel point in the target image, and determining the position and the size of an image area of the traffic sign from the target image according to the parameter equation;
and identifying the image area of the traffic sign board to obtain the sign board information of the traffic sign board.
2. The method of claim 1, wherein the identifying the image area of the traffic sign to obtain the sign information of the traffic sign comprises:
and identifying the image area of the traffic sign board through a preset convolution network model to obtain sign board information of the traffic sign board, wherein the convolution kernel, the number of convolution layers and the number of full-connection layers of the preset convolution network model are all smaller than those of the specified convolution network model.
3. The method of claim 2, wherein the specified convolutional network model is an AlexNet network model.
4. A signboard information acquisition apparatus, comprising:
the conversion module is used for carrying out Hough conversion on a target image to obtain a plurality of coordinate points in Hough parameter space, wherein the target image is an image of signboard information to be acquired;
the determining module is used for determining an image area of the traffic sign from the target image through a mean shift algorithm based on the plurality of coordinate points;
the identification module is used for identifying the image area of the traffic sign board to obtain the sign board information of the traffic sign board;
wherein the determining module comprises:
the selection submodule is used for selecting a target coordinate point from the coordinate points and selecting a coordinate point located in a target circular area from the coordinate points, and the target circular area is a circular area which takes the target coordinate point as a circle center and takes a preset length as a radius;
a calculation submodule for calculating a mean shift vector based on the target coordinate point and coordinate points located within the target circular region;
the judgment submodule is used for judging whether the vector value of the mean shift vector is smaller than a preset threshold value or not;
a first determining sub-module, configured to determine, when a vector value of the mean shift vector is not less than a preset threshold, a vector end point of the mean shift vector as the target coordinate point, and return to the step of selecting a coordinate point located in a target circular area from the plurality of coordinate points until the mean shift vector having a vector value less than the preset threshold is determined;
the second determining submodule is used for determining the coordinates of a target coordinate point of the mean shift vector with the vector value smaller than the preset threshold value, which is obtained through calculation, as parameters in a parameter equation of a pixel point in the target image, and determining the position and the size of an image area of the traffic sign from the target image according to the parameter equation;
wherein the calculation submodule is specifically configured to:
determining coordinates of the target coordinate point in the Hough parameter space and coordinates of coordinate points located in the target circular area in the Hough parameter space;
calculating the mean shift vector by the following formula based on the coordinates of the target coordinate point in the Hough parameter space and the coordinates of the coordinate points located within the target circular region in the Hough parameter space;
wherein, in the above formula, mh,GFor the mean shift vector, q is a predetermined confidence, g (-) is a gradient probability density functionX is the coordinate of the target coordinate point in the Hough parameter space, xjThe coordinates of the j-th coordinate point located in the target circular area in the Hough parameter space are obtained, h is the preset length, and n is the total number of coordinate points located in the target circular area.
5. The apparatus of claim 4, wherein the identification module comprises:
the identification submodule is used for identifying the image area of the traffic sign board through a preset convolution network model to obtain the sign board information of the traffic sign board, and the convolution kernel, the number of the convolution layers and the number of the full-connection layers of the preset convolution network model are all smaller than those of the specified convolution network model.
6. The apparatus of claim 5, wherein the specified convolutional network model is an AlexNet network model.
7. A signboard information acquisition apparatus, comprising:
a processor;
a memory for storing processor-executable instructions;
wherein the processor is configured to:
carrying out Hough transformation on a target image to obtain a plurality of coordinate points in Hough parameter space, wherein the target image is an image of signboard information to be obtained;
selecting a target coordinate point from the plurality of coordinate points, and selecting a coordinate point located within a target circular area from the plurality of coordinate points, wherein the target circular area is a circular area with the target coordinate point as a circle center and a preset length as a radius;
determining coordinates of the target coordinate point in the Hough parameter space and coordinates of coordinate points located in the target circular area in the Hough parameter space;
calculating a mean shift vector by the following formula based on the coordinates of the target coordinate point in the Hough parameter space and the coordinates of the coordinate points located within the target circular region in the Hough parameter space;
wherein, in the above formula, mh,GFor the mean shift vector, q is a preset confidence coefficient, g (-) is a gradient probability density function, x is the coordinate of the target coordinate point in the Hough parameter space, xjThe coordinates of the j-th coordinate point in the target circular area in the Hough parameter space are obtained, h is the preset length, and n is the total number of coordinate points in the target circular area;
judging whether the vector value of the mean shift vector is smaller than a preset threshold value or not;
when the vector value of the mean shift vector is not smaller than a preset threshold value, determining the vector end point of the mean shift vector as the target coordinate point, and returning to the step of selecting the coordinate point located in the target circular area from the plurality of coordinate points until the mean shift vector with the vector value smaller than the preset threshold value is determined;
determining the coordinates of a target coordinate point of a mean shift vector with a vector value smaller than the preset threshold value as parameters in a parameter equation of a pixel point in the target image, and determining the position and the size of an image area of the traffic sign from the target image according to the parameter equation;
and identifying the image area of the traffic sign board to obtain the sign board information of the traffic sign board.
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 CN106557759A (en) | 2017-04-05 |
| CN106557759B true 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) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107122737B (en) * | 2017-04-26 | 2020-07-31 | 聊城大学 | Automatic detection and identification method for road traffic signs |
| GB2574372B (en) * | 2018-05-21 | 2021-08-11 | Imagination Tech Ltd | 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 |
| CN112036421B (en) * | 2019-05-16 | 2024-07-02 | 北京搜狗科技发展有限公司 | Image processing method and device and electronic equipment |
| CN110163153B (en) * | 2019-05-23 | 2022-01-04 | 阿波罗智能技术(北京)有限公司 | Method and device for recognizing traffic sign board boundary |
| CN112348876B (en) * | 2019-08-08 | 2024-06-11 | 北京地平线机器人技术研发有限公司 | Space coordinate acquisition method and device for signboards |
| 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 |
| 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 |
Citations (1)
| 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 |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8457373B2 (en) * | 2009-03-16 | 2013-06-04 | Siemens Aktiengesellschaft | 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 |
| US9053389B2 (en) * | 2012-12-03 | 2015-06-09 | Analog Devices, Inc. | Hough transform for circles |
| KR101737791B1 (en) * | 2014-10-23 | 2017-05-19 | 현대모비스 주식회사 | Traffic sign recognizing apparatus for vehicle and operating method threrof. |
| CN106022300B (en) * | 2016-06-02 | 2019-07-05 | 中国科学院信息工程研究所 | Traffic sign recognition method and system based on cascade deep study |
-
2016
- 2016-11-25 CN CN201611063821.5A patent/CN106557759B/en active Active
Patent Citations (1)
| 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 |
Non-Patent Citations (1)
| Title |
|---|
| In-vehicle camera traffic sign detection and recognition;Andrzej Ruta等;《Machine Vision and Applications》;20091203;第22卷(第2期);第359-375页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106557759A (en) | 2017-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106557759B (en) | Signpost information acquisition method and device | |
| CN106651955B (en) | Method and device for positioning target object in picture | |
| CN106557768B (en) | Method and device for recognizing characters in picture | |
| CN108121952B (en) | Face key point positioning method, device, equipment and storage medium | |
| RU2577188C1 (en) | Method, apparatus and device for image segmentation | |
| JP6392468B2 (en) | Region recognition method and apparatus | |
| CN108010060B (en) | Target detection method and device | |
| CN107944367B (en) | Face key point detection method and device | |
| CN107480665B (en) | Character detection method and device and computer readable storage medium | |
| CN107784279B (en) | Target tracking method and device | |
| CN106778773B (en) | Method and device for positioning target object in picture | |
| CN105631803B (en) | The method and apparatus of filter processing | |
| CN107958223B (en) | Face recognition method and device, mobile equipment and computer readable storage medium | |
| CN105095881A (en) | Method, apparatus and terminal for face identification | |
| CN109034150B (en) | Image processing method and device | |
| CN107025441B (en) | Skin color detection method and device | |
| CN104268864B (en) | Card edge extracting method and device | |
| CN108009563B (en) | Image processing method and device and terminal | |
| CN105678296B (en) | Method and device for determining character inclination angle | |
| CN110930351A (en) | Light spot detection method and device and electronic equipment | |
| CN114120034A (en) | Image classification method and device, electronic equipment and storage medium | |
| CN106372663B (en) | Construct the method and device of disaggregated model | |
| CN108596957B (en) | Object tracking method and device | |
| CN108154090B (en) | Face recognition method and device | |
| CN107480773B (en) | Method and device for training convolutional neural network model and storage medium |
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 |