CN109858310A - Vehicles and Traffic Signs detection method - Google Patents
Vehicles and Traffic Signs detection method Download PDFInfo
- Publication number
- CN109858310A CN109858310A CN201711239998.0A CN201711239998A CN109858310A CN 109858310 A CN109858310 A CN 109858310A CN 201711239998 A CN201711239998 A CN 201711239998A CN 109858310 A CN109858310 A CN 109858310A
- Authority
- CN
- China
- Prior art keywords
- traffic sign
- detection
- image
- interest
- area
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The invention discloses a kind of Vehicles and Traffic Signs detection methods, the described method comprises the following steps: obtaining the image of vehicle-mounted camera acquisition, and convert the image into grayscale image；Grayscale image is pre-processed to obtain area-of-interest, and multiple scaling processing is carried out to area-of-interest；HOG feature calculation and classification are carried out to the area-of-interest after each scaling processing, to obtain multiple target area rectangular arrays containing traffic sign；Degree of overlapping screening is carried out to all target area rectangular arrays according to the rectangular coordinates in each target area rectangular array, to obtain road traffic sign detection result.Detection method of the invention can carry out multistage scaling to input picture, and carry out HOG feature calculation and classification to every level-one zoomed image, to obtain road traffic sign detection as a result, to improve detection accuracy, reduce missing inspection probability.
Description
Technical field
The present invention relates to Vehicular intelligent driving technology field, in particular to a kind of method for traffic sign detection and a kind of vehicle
?.
Background technique
Road traffic sign detection and identification belong to the important content that Vehicular intelligent drives field, and process can be divided into detection-phase
And cognitive phase.
A kind of detection method is proposed in the related technology, combines a large amount of positive negative samples to be trained cascade classifier, with
Realize the detection positioning of traffic sign.But above-mentioned technology needs to carry out Haar feature, HOG (Histogram of
Oriented Gradient, histograms of oriented gradients) various features such as feature calculate, and centre is related to integrogram, gradient calculates,
Computational complexity is high, causes the real-time of algorithm poor, and Haar feature is more sensitive to edge, line segment etc., is not suitable for examining
Survey traffic sign.
It also proposed a kind of method for traffic sign detection based on color segmentation in the related technology, by carrying out to image
The color segmentation of rgb space obtains reddish yellow blue three-color segmented image, and carries out Morphological scale-space to the image after segmentation and obtain
The processing such as profile to traffic sign, to be detected to traffic sign.But above-mentioned technology is in the detection process, depends on
The color characteristic of traffic sign, in actual road conditions, due to weather, illumination, brightness or the conditions limitation such as fade lead to traffic mark
The cross-color of will accurately can not be partitioned into traffic sign profile according to color characteristic, so as to cause the failure of detection.
Summary of the invention
The present invention is directed at least solve one of the technical problem in above-mentioned technology to a certain extent.For this purpose, of the invention
First purpose be to propose a kind of method for traffic sign detection, multistage scaling can be carried out to input picture, and to every level-one
Zoomed image carries out HOG feature calculation and classification, to obtain road traffic sign detection as a result, to improve detection accuracy, reduces leakage
Examine probability.
Second object of the present invention is to propose a kind of non-transitorycomputer readable storage medium.
Third object of the present invention is to propose a kind of vehicle.
In order to achieve the above objectives, a kind of method for traffic sign detection that first aspect present invention embodiment proposes, including with
Lower step: the image of vehicle-mounted camera acquisition is obtained, and described image is converted into grayscale image；The grayscale image is located in advance
Reason carries out multiple scaling processing to the area-of-interest to obtain area-of-interest；It is emerging to the sense after each scaling processing
Interesting region carries out HOG feature calculation and classification, to obtain multiple target area rectangular arrays containing traffic sign；According to each
Rectangular coordinates in the rectangular array of target area carry out degree of overlapping screening to all target area rectangular arrays, to obtain traffic mark
Will testing result.
Method for traffic sign detection according to an embodiment of the present invention obtains the image of vehicle-mounted camera acquisition, and by image
Grayscale image is converted to, grayscale image is pre-processed to obtain area-of-interest, and area-of-interest is carried out at multiple scaling
Reason carries out HOG feature calculation and classification to the area-of-interest after each scaling processing, multiple containing traffic sign to obtain
Target area rectangular array, according to the rectangular coordinates in each target area rectangular array to all target area rectangular arrays into
Row degree of overlapping screening, to obtain road traffic sign detection result.This method can carry out multistage scaling to input picture as a result, and
HOG feature calculation and classification are carried out to every level-one zoomed image, to obtain road traffic sign detection as a result, to improve detection essence
Degree reduces missing inspection probability.
In addition, the method for traffic sign detection proposed according to that above embodiment of the present invention can also have following additional skill
Art feature:
According to one embodiment of present invention, contain traffic sign in described image.
According to one embodiment of present invention, processing is successively zoomed in and out to the area-of-interest, and every time at scaling
The ratio of reason is identical.
According to one embodiment of present invention, to after each scaling processing area-of-interest carry out HOG feature calculation and
Classification, comprising: obtain detection window, detection block, detection unit and detection histogram port number；According to the detection window, institute
It states detection block, detection unit and detection histogram port number and HOG feature meter is carried out to the area-of-interest after each scaling processing
It calculates, and is classified using the traffic sign classifier that training obtains.
According to one embodiment of present invention, according to the following steps training traffic sign classifier: acquiring various roads
Video information under condition and weather condition, and interception has various shape traffic sign from the video frame of the video information
As positive sample, and from the video frame of video information, interception interferes image as negative sample to image；To the positive sample
The image of image and the negative sample carries out gray processing and normalized；According to the detection window, detection block, described
Detection unit and the detection histogram port number calculate the HOG Feature Descriptor number of each sample；According to each sample
HOG Feature Descriptor number is calculated and is saved to the HOG Feature Descriptor of all samples；By the HOG feature of all samples
Description son as input, using linear kernel function SVM (Support Vector Machine, support vector machines) classifier into
Row training, to obtain the traffic sign classifier.
According to one embodiment of present invention, the image of the image of the positive sample and the negative sample is square.
According to one embodiment of present invention, the corresponding shape of the positive sample is circular traffic sign, triangular traffic
Mark and rectangular traffic sign, and traffic sign classifier correspondence is divided into three kinds.
According to one embodiment of present invention, according to the rectangular coordinates in each target area rectangular array to all targets
Region rectangular array carries out degree of overlapping screening, comprising: is sat the rectangle in each target area rectangular array according to scaling
Mark reverts to the respective coordinates in original image；Degree of overlapping screening is carried out to the rectangular array after reduction.
In order to achieve the above objectives, second aspect of the present invention embodiment proposes a kind of non-transitory computer-readable storage medium
Matter is stored thereon with computer program, which realizes above-mentioned method for traffic sign detection when being executed by processor.
The non-transitorycomputer readable storage medium of the embodiment of the present invention, by executing above-mentioned road traffic sign detection side
Method can carry out multistage scaling to input picture, and carry out HOG feature calculation and classification to every level-one zoomed image, to obtain
Road traffic sign detection reduces missing inspection probability as a result, to improve detection accuracy.
In order to achieve the above objectives, third aspect present invention embodiment proposes a kind of vehicle, and the vehicle includes vehicle-mounted takes the photograph
As head, for acquiring road image, the vehicle further includes memory, processor and is stored in described deposit the vehicle-mounted camera
On reservoir and the road traffic sign detection program that can run on the processor, the road traffic sign detection program is by the processing
The step of device realizes above-mentioned method for traffic sign detection when executing.
The vehicle of the embodiment of the present invention can carry out multistage scaling to input picture, and carry out to every level-one zoomed image
HOG feature calculation and classification reduce missing inspection probability to obtain road traffic sign detection as a result, to improve detection accuracy.
Detailed description of the invention
Fig. 1 is the flow chart of method for traffic sign detection according to an embodiment of the present invention；
Fig. 2 is the division schematic diagram of grayscale image according to an embodiment of the invention；
Fig. 3 is the block diagram of area-of-interest according to an embodiment of the invention；
Fig. 4 is road traffic sign detection result schematic diagram according to an embodiment of the invention.
Specific embodiment
The embodiment of the present invention is described below in detail, examples of the embodiments are shown in the accompanying drawings, wherein from beginning to end
Same or similar label indicates same or similar element or element with the same or similar functions.Below with reference to attached
The embodiment of figure description is exemplary, it is intended to is used to explain the present invention, and is not considered as limiting the invention.
The method for traffic sign detection proposed according to embodiments of the present invention and vehicle described with reference to the accompanying drawing.
Fig. 1 is the flow chart of method for traffic sign detection according to an embodiment of the present invention.
As shown in Figure 1, the method for traffic sign detection of the embodiment of the present invention can comprise the following steps that
S1, obtains the image of vehicle-mounted camera acquisition, and converts the image into grayscale image.It wherein, include to hand in image
Logical mark, for example, including minimum stated speed sign and Maximum speed limit mark in image.
That is, the video image of the current road conditions acquired by vehicle-mounted camera, to each frame of video image into
Row processing, to obtain the image for including traffic sign, and is converted into gray level image (image indicated with gray scale).
S2 pre-processes grayscale image to obtain area-of-interest, and carry out multiple scaling processing to area-of-interest.
Wherein, grayscale image is pre-processed, specifically include that digitlization, geometric transformation, normalization, smooth, recovery and increased
Strong and etc..According to traffic sign position feature, area-of-interest is selected, to reduce calculation amount and improve efficiency, for example, Fig. 2
Bend part, area-of-interest are the 50% of original image area.
In an embodiment of the present invention, processing is successively zoomed in and out to area-of-interest, and the ratio of scaling processing every time
It is identical.For example, original image initial value is 1, setting scaling is 1.1, and scaling number is 8 times, then, every time after scaling
The length of image and the wide respectively length of original area-of-interest and 1/ wide (1.1)n, wherein n=1,2 ..., 8, it is, the
The length and width of image after primary scaling are originally 1/1.1, and the length and width of the image after scaling for the second time are original interested
The length in region and 1/ wide (1.1)2..., and so on, the grayscale image of the area-of-interest after scaling every time can be obtained.
S3 carries out HOG feature calculation and classification to the area-of-interest after each scaling processing, multiple containing friendship to obtain
The target area rectangular array of logical mark.
According to one embodiment of present invention, to after each scaling processing area-of-interest carry out HOG feature calculation and
Classification, comprising: obtain detection window, detection block, detection unit and detection histogram port number, according to detection window, detection block,
Detection unit and detection histogram port number carry out HOG feature calculation to the area-of-interest after each scaling processing, and utilize
The traffic sign classifier that training obtains is classified.
For example, as shown in figure 3, initial region of interest area image size is 640*360, selection detection window is 20*
20, choosing detection block size is 10*10, and detection unit size is 5*5, and detection histogram port number (bin value) is 9, that is,
The histogram of gradients in 9 directions of statistics in one detection unit, detection step-length are 5*5.It should be noted that detection window, inspection
Surveying block, detection unit and detection histogram port number is preset value (presetting in classifier training).
Detection window size is indicated with winSize (20,20), blockSize (10,10) indicates detection block size,
CellSize (5,5) indicates detection unit size, and blockStride (5,5) indicates detection step sizes, and nbins=9 indicates every
In a detection unit for the quantity of histogram of gradients.
So, a detection block includes A=(blockSize.width/cellSize.width) *
(blockSize.height/cellSize.height) a detection unit, above-mentioned parameter is brought into, A=(10/5) * (10/5)
=4, a detection block includes 9*A=9*4=36 histogram of gradients, that is, a detection block includes 36 feature vector dimensions
Number.One detection window include B=((winSize.width-blockSize.width)/(blockStride.width)+
1) * ((winSize.height-blockSize.height)/(blockStride.height)+1) a detection block, will be above-mentioned
Parameter is brought into, B=((20-10)/5+1) * ((20-10)/5+1)=9, and a detection window includes 9*A*B=9*4*9=ladder
324 histogram of gradients of histogram are spent, that is, a detection window includes 324 feature vector dimensions.
When region of interest area image size is 640*360, according to above-mentioned parameter, the feature vector dimension that can be taken is
324* (640/20) * (360/20)=186624 dimension.The feature vector dimension of the area-of-interest after scaling is according to scaling every time
Rear region of interest area image size calculates, and no longer calculates one by one here.
After the area-of-interest after scaling to first time carries out HOG feature calculation, trained obtained traffic sign is utilized
Classifier is classified, and can be obtained the target area rectangular array 1 containing traffic sign, above-mentioned steps be repeated, to every
Area-of-interest grayscale image after level-one scaling carries out HOG feature calculation and classification, the scaling number until reaching setting, can be with
It obtains multiple groups and contains the target area rectangular array 2,3 ... of traffic sign, 8.
It wherein, can be according to following steps training traffic sign classifier: acquiring the video under various road conditions and weather condition
Information, and image of the interception with various shape traffic sign is as positive sample from the video frame of video information, and from view
For interception interference image as negative sample, the image of image and negative sample to positive sample carries out gray processing in the video frame of frequency information
And normalized, the HOG of each sample is calculated according to detection window, detection block, detection unit and detection histogram port number
Feature Descriptor number is counted according to HOG Feature Descriptor of the HOG Feature Descriptor number of each sample to all samples
It calculates and saves, using the HOG Feature Descriptor of all samples as input, be trained using the SVM classifier of linear kernel function,
To obtain traffic sign classifier.Wherein, the image of positive sample and the image of negative sample are square.The corresponding shape of positive sample
Shape is circular traffic sign, triangle traffic sign and rectangular traffic sign, and traffic sign classifier correspondence is divided into three kinds.
Specifically, the video under various road conditions and weather condition is acquired, interception has friendship of various shapes from video frame
The image of logical mark is intercepted as positive sample easily by the interference image of erroneous detection as negative sample, and image interception is square, and every kind
Shape traffic sign acquisition positive sample 3000 opens figure, and negative sample 1000 opens figure.Then, the image to the positive sample of interception and negative sample
This image carries out gray processing and normalized, wherein normalization is exactly the image and negative sample various sizes of positive sample
This image be uniformly scaled to the size with detection window same size, and for different resize-windows, detection window is big
It is small be always it is identical, if detection window size is 20*20, and the image of positive sample is divided into three classes according to shape: round
Traffic sign, triangle traffic sign and rectangular traffic sign.
The HOG Feature Descriptor number for calculating each sample, specifically includes: choosing detection block size is 10*10, and detection is single
First size is 5*5, then the gradient information of each unit is counted with 9 histogram channels, and detection step-length is 5*5, each sample
The feature vector dimension of image is 9* (10/5) * (10/5) * ((20-10)/5+1) * ((20-10)/5+1)=324.Then, it counts
It calculates the HOG Feature Descriptor of all positive samples and negative sample and saves.It should be noted that detailed calculating process can refer to
Example is stated, which is not described herein again.
Using the HOG Feature Descriptor of all samples obtained above as input, and using the svm classifier of existing kernel function
Device is trained, and can obtain the three kinds of classifiers classified for traffic sign.
S4 is overlapped all target area rectangular arrays according to the rectangular coordinates in each target area rectangular array
Degree screening, to obtain road traffic sign detection result.
According to one embodiment of present invention, according to the rectangular coordinates in each target area rectangular array to all targets
Region rectangular array carries out degree of overlapping screening, comprising: is sat the rectangle in each target area rectangular array according to scaling
Mark reverts to the respective coordinates in original image, and carries out degree of overlapping screening to the rectangular array after reduction.
It specifically, can after the grayscale image to the area-of-interest after every level-one scaling carries out HOG feature calculation and classification
The target area rectangular array that multiple groups include traffic sign is obtained, for example, setting scaling number is 8, available 8 groups include
The target area rectangular array of traffic sign.Then, according to scaling, the coordinate of every group of obtained rectangular array is reverted to
Respective coordinates in original image (the area-of-interest grayscale image not zoomed in and out), for example, the rectangle that the first order scales
The coordinate of array is restored according to 1.1 ratio, and the coordinate for the rectangular array that the second level scales is according to (1.1)2Ratio also
Former ..., the 8th grade scales the coordinate of obtained rectangle data according to (1.1)8Ratio reduction.And to the rectangular array after reduction
The screening of (8 groups) progress degrees of overlapping, obtains final road traffic sign detection as a result, detection effect is as shown in Figure 4.
Wherein, the screening of degree of overlapping is the process of an image clustering, can be screened: be preset in the following manner
The position of the threshold value of cluster, the single rectangle in original image is determined by its upper left corner and lower right corner apex coordinate, when several
(rectangle herein is pros to the difference of the vertex abscissa of rectangle (rectangle after reduction) and ordinate with original rectangular side length
Shape) ratio be less than preset threshold value when, this several rectangle (rectangle after reduction) is merged, amalgamation result
Rectangle takes its vertex (upper left corner and the lower right corner) coordinate mean value.
As the above analysis, method for traffic sign detection of the invention, using grayscale image as input, can effectively avoid because
Weather, illumination, brightness or the conditions limitation such as fade, caused by traffic sign cross-color, the problem of can not accurately detecting,
Simultaneously by carrying out multistage scaling to original image, and various sizes of traffic sign is carried out with fixed size detection window
Detection, improves detection accuracy, reduces missing inspection probability.
In conclusion method for traffic sign detection according to an embodiment of the present invention, obtains the image of vehicle-mounted camera acquisition,
And grayscale image is converted the image into, grayscale image is pre-processed to obtain area-of-interest, and area-of-interest is carried out more
Secondary scaling processing carries out HOG feature calculation and classification to the area-of-interest after each scaling processing, multiple containing friendship to obtain
The target area rectangular array of logical mark, according to the rectangular coordinates in each target area rectangular array to all target area squares
Figurate number group carries out degree of overlapping screening, to obtain road traffic sign detection result.This method can carry out input picture multistage as a result,
Scaling, and HOG feature calculation and classification are carried out to every level-one zoomed image, to obtain road traffic sign detection as a result, to improve
Detection accuracy reduces missing inspection probability.
In addition, the embodiment of the present invention also proposed a kind of non-transitorycomputer readable storage medium, it is stored thereon with
Computer program, the program realize above-mentioned method for traffic sign detection when being executed by processor.
The non-transitorycomputer readable storage medium of the embodiment of the present invention, by executing above-mentioned road traffic sign detection side
Method can carry out multistage scaling to input picture, and carry out HOG feature calculation and classification to every level-one zoomed image, to obtain
Road traffic sign detection reduces missing inspection probability as a result, to improve detection accuracy.
In addition, the embodiment of the present invention also proposed a kind of vehicle, vehicle may include vehicle-mounted camera, and vehicle-mounted camera is used
In acquisition road image, vehicle further include memory, processor and storage on a memory and the friendship that can run on a processor
Logical Mark Detection program, road traffic sign detection program realize the step of above-mentioned method for traffic sign detection when being executed by processor
Suddenly.
The vehicle of the embodiment of the present invention can carry out multistage scaling to input picture, and carry out to every level-one zoomed image
HOG feature calculation and classification reduce missing inspection probability to obtain road traffic sign detection as a result, to improve detection accuracy.
In the description of this specification, reference term " one embodiment ", " some embodiments ", " example ", " specifically show
The description of example " or " some examples " etc. means specific features, structure, material or spy described in conjunction with this embodiment or example
Point is included at least one embodiment or example of the invention.In the present specification, schematic expression of the above terms are not
It must be directed to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described can be in office
It can be combined in any suitable manner in one or more embodiment or examples.In addition, without conflicting with each other, the skill of this field
Art personnel can tie the feature of different embodiments or examples described in this specification and different embodiments or examples
It closes and combines.
In addition, term " first ", " second " are used for descriptive purposes only and cannot be understood as indicating or suggesting relative importance
Or implicitly indicate the quantity of indicated technical characteristic.Define " first " as a result, the feature of " second " can be expressed or
Implicitly include at least one this feature.In the description of the present invention, the meaning of " plurality " is at least two, such as two, three
It is a etc., unless otherwise specifically defined.
Any process described otherwise above or method description are construed as in flow chart or herein, and expression includes
It is one or more for realizing custom logic function or process the step of executable instruction code module, segment or portion
Point, and the range of the preferred embodiment of the present invention includes other realization, wherein can not press shown or discussed suitable
Sequence, including according to related function by it is basic simultaneously in the way of or in the opposite order, Lai Zhihang function, this should be of the invention
Embodiment person of ordinary skill in the field understood.
Expression or logic and/or step described otherwise above herein in flow charts, for example, being considered use
In the order list for the executable instruction for realizing logic function, may be embodied in any computer-readable medium, for
Instruction execution system, device or equipment (such as computer based system, including the system of processor or other can be held from instruction
The instruction fetch of row system, device or equipment and the system executed instruction) it uses, or combine these instruction execution systems, device or set
It is standby and use.For the purpose of this specification, " computer-readable medium ", which can be, any may include, stores, communicates, propagates or pass
Defeated program is for instruction execution system, device or equipment or the dress used in conjunction with these instruction execution systems, device or equipment
It sets.The more specific example (non-exhaustive list) of computer-readable medium include the following: there is the electricity of one or more wirings
Interconnecting piece (electronic device), portable computer diskette box (magnetic device), random access memory (RAM), read-only memory
(ROM), erasable edit read-only storage (EPROM or flash memory), fiber device and portable optic disk is read-only deposits
Reservoir (CDROM).In addition, computer-readable medium can even is that the paper that can print described program on it or other are suitable
Medium, because can then be edited, be interpreted or when necessary with it for example by carrying out optical scanner to paper or other media
His suitable method is handled electronically to obtain described program, is then stored in computer storage.
It should be appreciated that each section of the invention can be realized with hardware, software, firmware or their combination.Above-mentioned
In embodiment, software that multiple steps or method can be executed in memory and by suitable instruction execution system with storage
Or firmware is realized.Such as, if realized with hardware in another embodiment, following skill well known in the art can be used
Any one of art or their combination are realized: have for data-signal is realized the logic gates of logic function from
Logic circuit is dissipated, the specific integrated circuit with suitable combinational logic gate circuit, programmable gate array (PGA), scene can compile
Journey gate array (FPGA) etc..
Those skilled in the art are understood that realize all or part of step that above-described embodiment method carries
It suddenly is that relevant hardware can be instructed to complete by program, the program can store in a kind of computer-readable storage medium
In matter, which when being executed, includes the steps that one or a combination set of embodiment of the method.
It, can also be in addition, each functional unit in each embodiment of the present invention can integrate in a processing module
It is that each unit physically exists alone, can also be integrated in two or more units in a module.Above-mentioned integrated mould
Block both can take the form of hardware realization, can also be realized in the form of software function module.The integrated module is such as
Fruit is realized and when sold or used as an independent product in the form of software function module, also can store in a computer
In read/write memory medium.
Storage medium mentioned above can be read-only memory, disk or CD etc..Although having been shown and retouching above
The embodiment of the present invention is stated, it is to be understood that above-described embodiment is exemplary, and should not be understood as to limit of the invention
System, those skilled in the art can be changed above-described embodiment, modify, replace and become within the scope of the invention
Type.
In the description of the present invention, it is to be understood that, term " center ", " longitudinal direction ", " transverse direction ", " length ", " width ",
" thickness ", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outside", " up time
The orientation or positional relationship of the instructions such as needle ", " counterclockwise ", " axial direction ", " radial direction ", " circumferential direction " be orientation based on the figure or
Positional relationship is merely for convenience of description of the present invention and simplification of the description, rather than the device or element of indication or suggestion meaning must
There must be specific orientation, be constructed and operated in a specific orientation, therefore be not considered as limiting the invention.
In the present invention unless specifically defined or limited otherwise, term " installation ", " connected ", " connection ", " fixation " etc.
Term shall be understood in a broad sense, for example, it may be being fixedly connected, may be a detachable connection, or integral；It can be mechanical connect
It connects, is also possible to be electrically connected；It can be directly connected, can also can be in two elements indirectly connected through an intermediary
The interaction relationship of the connection in portion or two elements.It for the ordinary skill in the art, can be according to specific feelings
Condition understands the concrete meaning of above-mentioned term in the present invention.
In the present invention unless specifically defined or limited otherwise, fisrt feature in the second feature " on " or " down " can be with
It is that the first and second features directly contact or the first and second features pass through intermediary mediate contact.Moreover, fisrt feature exists
Second feature " on ", " top " and " above " but fisrt feature be directly above or diagonally above the second feature, or be merely representative of
First feature horizontal height is higher than second feature.Fisrt feature can be under the second feature " below ", " below " and " below "
One feature is directly under or diagonally below the second feature, or is merely representative of first feature horizontal height less than second feature.
Although the embodiments of the present invention has been shown and described above, it is to be understood that above-described embodiment is example
Property, it is not considered as limiting the invention, those skilled in the art within the scope of the invention can be to above-mentioned
Embodiment is changed, modifies, replacement and variant.
Claims (10)
1. a kind of method for traffic sign detection, which comprises the following steps:
The image of vehicle-mounted camera acquisition is obtained, and described image is converted into grayscale image；
The grayscale image is pre-processed to obtain area-of-interest, and the area-of-interest is carried out at multiple scaling
Reason；
HOG feature calculation and classification are carried out to the area-of-interest after each scaling processing, it is multiple containing traffic sign to obtain
Target area rectangular array；
Degree of overlapping screening is carried out to all target area rectangular arrays according to the rectangular coordinates in each target area rectangular array,
To obtain road traffic sign detection result.
2. method for traffic sign detection as described in claim 1, which is characterized in that contain traffic sign in described image.
3. method for traffic sign detection as described in claim 1, which is characterized in that successively contract to the area-of-interest
Processing is put, and the ratio of each scaling processing is identical.
4. method for traffic sign detection as claimed in any one of claims 1-3, which is characterized in that after each scaling processing
Area-of-interest carry out HOG feature calculation and classification, comprising:
Obtain detection window, detection block, detection unit and detection histogram port number；
According to the detection window, the detection block, detection unit and detection histogram port number to each scaling processing after
Area-of-interest carries out HOG feature calculation, and is classified using the traffic sign classifier that training obtains.
5. method for traffic sign detection as claimed in claim 4, which is characterized in that according to the following steps training traffic mark
Will classifier:
The video information under various road conditions and weather condition is acquired, and is intercepted from the video frame of the video information with various
As positive sample, and from the video frame of video information, interception interferes image as negative sample to the image of shape traffic sign；
The image of image and the negative sample to the positive sample carries out gray processing and normalized；
Each sample is calculated according to the detection window, the detection block, the detection unit and the detection histogram port number
This HOG Feature Descriptor number；
It is calculated and is saved according to HOG Feature Descriptor of the HOG Feature Descriptor number of each sample to all samples；
Using the HOG Feature Descriptor of all samples as input, it is trained using the SVM classifier of linear kernel function, to obtain
Obtain the traffic sign classifier.
6. method for traffic sign detection as claimed in claim 5, which is characterized in that the image of the positive sample and the negative sample
This image is square.
7. method for traffic sign detection as claimed in claim 5, which is characterized in that the corresponding shape of the positive sample is circle
Traffic sign, triangle traffic sign and rectangular traffic sign, and traffic sign classifier correspondence is divided into three kinds.
8. such as method for traffic sign detection of any of claims 1-7, which is characterized in that according to each target area
Rectangular coordinates in rectangular array carry out degree of overlapping screening to all target area rectangular arrays, comprising:
The rectangular coordinates in each target area rectangular array are reverted into the respective coordinates in original image according to scaling；
Degree of overlapping screening is carried out to the rectangular array after reduction.
9. a kind of non-transitorycomputer readable storage medium, is stored thereon with computer program, which is characterized in that the program quilt
Such as method for traffic sign detection of any of claims 1-8 is realized when processor executes.
10. a kind of vehicle, which is characterized in that the vehicle includes vehicle-mounted camera, and the vehicle-mounted camera is for acquiring road
Image, the vehicle further include memory, processor and are stored on the memory and can run on the processor
Road traffic sign detection program is realized when the road traffic sign detection program is executed by the processor as appointed in claim 1-8
The step of method for traffic sign detection described in one.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711239998.0A CN109858310A (en) | 2017-11-30 | 2017-11-30 | Vehicles and Traffic Signs detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711239998.0A CN109858310A (en) | 2017-11-30 | 2017-11-30 | Vehicles and Traffic Signs detection method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109858310A true CN109858310A (en) | 2019-06-07 |
Family
ID=66888382
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711239998.0A Pending CN109858310A (en) | 2017-11-30 | 2017-11-30 | Vehicles and Traffic Signs detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109858310A (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110765224A (en) * | 2019-10-25 | 2020-02-07 | 驭势科技（北京）有限公司 | Processing method of electronic map, vehicle vision repositioning method and vehicle-mounted equipment |
| CN112508925A (en) * | 2020-12-16 | 2021-03-16 | 广州大学 | Electronic lock panel quality detection method, system, computer device and storage medium |
| CN112788260A (en) * | 2019-11-11 | 2021-05-11 | 纽伦百仕科技股份有限公司 | Intelligent identification single-chip system combined with image sensor |
| CN112966639A (en) * | 2021-03-22 | 2021-06-15 | 新疆爱华盈通信息技术有限公司 | Vehicle detection method and device, electronic equipment and storage medium |
| CN113298786A (en) * | 2021-05-26 | 2021-08-24 | 北京长木谷医疗科技有限公司 | Image recognition and model training method, and true mortar position recognition method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2028605A1 (en) * | 2007-08-20 | 2009-02-25 | Delphi Technologies, Inc. | Detection method for symmetric patterns |
| CN101561867A (en) * | 2009-05-19 | 2009-10-21 | 华中科技大学 | Human body detection method based on Gauss shape feature |
| CN102693427A (en) * | 2011-03-22 | 2012-09-26 | 日电（中国）有限公司 | Method and device for forming detector for detecting images |
| CN102930278A (en) * | 2012-10-16 | 2013-02-13 | 天津大学 | Human eye sight estimation method and device |
| CN105260712A (en) * | 2015-10-03 | 2016-01-20 | 上海大学 | Method and system for detecting pedestrian in front of vehicle |
-
2017
- 2017-11-30 CN CN201711239998.0A patent/CN109858310A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP2028605A1 (en) * | 2007-08-20 | 2009-02-25 | Delphi Technologies, Inc. | Detection method for symmetric patterns |
| CN101561867A (en) * | 2009-05-19 | 2009-10-21 | 华中科技大学 | Human body detection method based on Gauss shape feature |
| CN102693427A (en) * | 2011-03-22 | 2012-09-26 | 日电（中国）有限公司 | Method and device for forming detector for detecting images |
| CN102930278A (en) * | 2012-10-16 | 2013-02-13 | 天津大学 | Human eye sight estimation method and device |
| CN105260712A (en) * | 2015-10-03 | 2016-01-20 | 上海大学 | Method and system for detecting pedestrian in front of vehicle |
Non-Patent Citations (1)
| Title |
|---|
| 王斌 等: "基于MSER和SVM的快速交通标志检测", 《光电子·激光》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110765224A (en) * | 2019-10-25 | 2020-02-07 | 驭势科技（北京）有限公司 | Processing method of electronic map, vehicle vision repositioning method and vehicle-mounted equipment |
| CN112788260A (en) * | 2019-11-11 | 2021-05-11 | 纽伦百仕科技股份有限公司 | Intelligent identification single-chip system combined with image sensor |
| CN112508925A (en) * | 2020-12-16 | 2021-03-16 | 广州大学 | Electronic lock panel quality detection method, system, computer device and storage medium |
| CN112508925B (en) * | 2020-12-16 | 2023-07-04 | 广州大学 | Electronic lock panel quality detection method, system, computer device and storage medium |
| CN112966639A (en) * | 2021-03-22 | 2021-06-15 | 新疆爱华盈通信息技术有限公司 | Vehicle detection method and device, electronic equipment and storage medium |
| CN113298786A (en) * | 2021-05-26 | 2021-08-24 | 北京长木谷医疗科技有限公司 | Image recognition and model training method, and true mortar position recognition method and device |
| CN113298786B (en) * | 2021-05-26 | 2022-02-11 | 北京长木谷医疗科技有限公司 | Image recognition and model training method, and true mortar position recognition method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2017190574A1 (en) | Fast pedestrian detection method based on aggregation channel features | |
| CN109858310A (en) | Vehicles and Traffic Signs detection method | |
| Soilán et al. | Traffic sign detection in MLS acquired point clouds for geometric and image-based semantic inventory | |
| US10121245B2 (en) | Identification of inflammation in tissue images | |
| Parodi et al. | A feature-based recognition scheme for traffic scenes | |
| CN109214420A (en) | The high texture image classification method and system of view-based access control model conspicuousness detection | |
| CN108921089A (en) | Method for detecting lane lines, device and system and storage medium | |
| CN108090459B (en) | Traffic sign detection and identification method suitable for vehicle-mounted vision system | |
| CN111626277B (en) | Vehicle tracking method and device based on over-station inter-modulation index analysis | |
| Li et al. | A vehicle license plate recognition system based on analysis of maximally stable extremal regions | |
| CN112712057B (en) | Traffic signal identification method and device, electronic equipment and storage medium | |
| US20130342694A1 (en) | Method and system for use of intrinsic images in an automotive driver-vehicle-assistance device | |
| US9129156B2 (en) | Method for detecting and recognizing an object in an image, and an apparatus and a computer program therefor | |
| CN110288612B (en) | Nameplate positioning and correcting method and device | |
| CN110070545B (en) | Method for automatically extracting urban built-up area by urban texture feature density | |
| CN106709412A (en) | Traffic sign detection method and apparatus | |
| Mammeri et al. | North-American speed limit sign detection and recognition for smart cars | |
| CN106845458A (en) | A kind of rapid transit label detection method of the learning machine that transfinited based on core | |
| CN112927283A (en) | Distance measuring method and device, storage medium and electronic equipment | |
| CN104648397A (en) | System and method for warning lane departure | |
| CN105574490B (en) | Vehicle brand recognition methods and system based on headlight characteristics of image | |
| CN103136536A (en) | System and method for detecting target and method for exacting image features | |
| CN111666811A (en) | Method and system for extracting traffic sign area in traffic scene image | |
| CN110866435A (en) | Far infrared pedestrian training method with self-similarity gradient oriented histogram | |
| CN111695374A (en) | Method, system, medium, and apparatus for segmenting zebra crossing region in monitoring view |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190607 |